Back in the 90's when I was an on-site engineer for Silicon Graphics, I had a job to replace a failed power supply unit for an SGI workstation. After replacing the PSU, I proceeded to switch the workstation on to test it. All of a sudden, there was a Loud Bang emanating from the supposedly repaired machine, and the magic smoke was escaping from it. And all the electricity went off in the building. Everything just went dark. The office the machine was in went silent, and all eyes were on me. It was only when I - very puzzled - looked at the PSU, did I notice the voltage selector on it. Upon closer inspection, I further noticed that it was set to 120V, instead of the 240V it should have been - obviously this unit was manufactured outside of the UK and was set to 120V by default. Not something which was ever mentioned in any SGI training. That was a very awkward time in that particular customer's office for a few moments. I blamed the Loud Bang and the tripping of the building's circuit breakers on a "faulty replacement" - which, technically, was the truth ;)
Back ca. 1990 we were supplied with "Rockwell T50 Programming Terminals" £8K each - actually a rugged PC, and often the first such any of our techs would encounter. This was before auto supply ranging was common. For "safety" factory floor sockets were 110V (Round, blue). However technicians' workshop sockets were 240V (Round, Yellow). If we were lucky just the power supply went "poof" otherwise - a very dead box.
@@M3WDD Good heavens! If they'd changed the 110V sockets to Edison (US, Japan) style, it would probably been enough to make people think twice and make sure.
Indeed; We ran into that with some 'legacy' systems that used, IIRC, an S-100 style mainboard with add-in CPU and IO cards. The boxes had non-autoranging power supplies with that style of voltage switch. the vendor had to replace the power supply on one of them after someone failed to check the position of that switch before plugging it in, and that was entertaining, because not many people make that hardware anymore.. This is why any hardware I spec for [RedactedCo] states "power supply SHALL be auto-ranging between 120V and 208/220".
Great video as always. Could maybe have explained a bit more about how class D amps work for those who don't know, as they are a pretty fascinating technology. Those IC's are most likely re-branded IRS20957s or similar, these are the most common class D amp driver IC's. Frequently see them installed in amps with a OEM part code rather than their standard part name. One thing I disagree with however, is referring to class D amps as 'digital'. They are actually analogue in operation, and the term 'class D' is just coincidental. While they do use a square wave as the main carrier, it is just produced by a simple oscillator at a fixed rate, and is not linked to any clock or carries any bits of information. It is simply the fast switching that allows the output mosfets to be turned fully on and off as switches, and hence minimise power dissipation by not running them in their linear range. The audio is produced by modulating the pulse width (pwm) of the square wave in accordance with the incoming audio signal, which is analogue in nature. The output choke and capacitor on the output form a second order high pass filter, with a rolloff point high enough that it blocks most of the high frequency switching noise, but the pulse width modulation occurs at audio frequencies so these are passed through the the loudspeaker. In most designs, you still have a positive and negative half of the amp, much like a traditional class AB amp, where it differs is that the mosfets are being switched hard on and off, rather than directly being controlled by the audio waveform amplitude. And that line of resistors down the edge, is probably to indeed provide a local power supply as you suggested. But it is probably referenced to the negative rail, as the gate drivers need a +12v supply referenced to the negative rail, in order for the switching of the negative side mosfets as they usually use N channel mosfets for both halves of the amp, much like a quasi-complimentary class AB amp. Some amps I see use N and P channel fets, but usually they are all N channel, probably due to more even performance characteristics.
@@railgapThat doesn't make it a digital amp, any more than a pwm controller makes a Dyson motor a digital motor. A digital amp would be something like a TTL high-current logic buffer in an output to drive more TTL inputs. Most modern amplifiers could be referred to as "Digital" as they can be remotely controlled by rs232 or usb or an infrared remote control but that doesn't make them a "Digital" Amplifier. It's all marketing bullsh!t.
Precisely the kind of definition I was looking for. It struck me firstly that "digital" in reference to an essentially analogue device, was obviously bullshit designed to baffle brains and sell units. Thanks for taking the time time explain Class D amplification.
Thank you. As I have said before your channel introduces me to new concepts and circuit designs. I did not know about class D amplifiers, having grown up with traditional push pull amplifier designs. This has set me on a voyage of discovery. 75 and still learning.
I'm in love with the TPA3116D2 amplifier. It's a 50 watt Class D amplifier about the size of a pack of cigarettes and sounds amazing. I've got it running the sound from my computer to two 50 watt bookshelf speakers and it's doing the job VERY nicely. I've bought a bunch for when I need sound somewhere because they sound great and they're 5$ on the Chinese sites. Amplifiers have REALLY come a long way in the last few years.
Class D in general has come a long way. There's zero audible difference between a class A/B and a modern class D amp. Always love how much darn power a class D amp can produce despite being absolutely tiny. Also the efficiency is totally bonkers. Just beats my old heavy AB amp in every category.
If you want an Experience that is rather eye opening... take one of the new TPA3255 based amplifiers (A07, V3, etc) found on popular online vendors and cut it loose on something like a pair of Klipsh Heresies or JBL L-100s ... they're not just for bookshelf speakers anymore. In fact, the way these things get sold short is by always talking about them as though a tiny amp needs tiny speakers ... when that's just not true.
@@boydw1 the TPA3255 can produce upwards of 200 RMS watts per channel on 4 ohms, 100 on 8. And they are still a palm sized amplifier. Consider the power of a PM8006 for less than $100.
@@Douglas_Blake_579 Oh, for sure, the power these little class-D amps are putting out for the size/money is impressive. My point was more that the "large speakers needing a large amp" assumption is often backwards. Say you have a small bookshelf speaker with 87dB/1w/1m sensitivity, and a larger floorstander speaker with 93dB/1w/1m sensitivity (6dB delta). You'd need to give the bookshelf speaker 4x the power to reach the same listening level.
Back in the day playing in a band at various venues one would often find a very noisy wall AC outlet, and all the gear was plugged into this single outlet. But we did have a sophisticated earth lift that could fix the noise with some risk. Disconnect the earth at the wall plug.
From my perspective there are two kinds of ground - power ground (protective ground) and signal ground. So I'd like to have a dedicated signal ground post on the device. In most rack mounted solutions the air is taken in at the front and then the fan sucks out the air at the rear. That's why there's foam at the front - to capture the dust from the air sucked in at the front.
So, this is a FULL BRIDGE AMPLIFIER, haha! Looks nice. This is where switching mode power supplies and class D took us - a stage power amp in a single unit enclosure, so compact, clean and neat.
Not full bridge, half bridge, as the speaker has a connection to the mid point of the power supply, and a full bridge amplifier cannot be connected into a bridge configuration, as it is already running in that mode.
Some of the modern ultra high power 1HE amps are reliably pushing a total of 20kW nominal power into 4 chanels at 4 ohms. Those are slowly replacing the gold standard Labgruppen FP10000Q and FP140000 over here in europe. Although for low end professional use the high quality chinese FP series coppies are still unbeatable.
@@lukahierl9857 LABGRUPPEN has the PLM44 and others ( using LAKE processor) at least 4-6 yrs in the making , as of my knowledge these are 20000W (4ch) bridgeable and each can also be configured as 4 active p.amps with selectable slope configs... Also comes with full security locks & remote location networked control capabilities .
11:16 The fan is actually pulling cold air from the front venting sleeves over the circuitry out to the back of the amplifier. This design is very common in 1U devices like firewalls, routers, etc.....
Indeed. Case fans like this usually blow toward the side with the label, which would be outward in this unit. This is typical, as you don't want a fan blowing directly onto a PCB, as it will develop "cold spots," i.e., temperature gradients that will exert mechanical stresses on the fine electrical traces. Better to have the air flow in through a large inlet and get blown out through a small outlet, as this makes the cold air more diffuse.
5:23 Back in the mid 1980's, Bob Carver of the Carver stereo equipment manufacturer, started playing around with hybrid designs of digital power supplies and typical analog output stages for his amplifiers. He used a special high frequency transformer (that would massively saturate at the regular 60hz mains frequency) and then drove the transformer with a heavy duty Triac. The input circuitry would monitor the audio signal and when the circuitry would see a large musical spike coming it would very quickly ramp up the Triac for the brief instant and the transformer would momentarily output a higher voltage to cover that large musical signal. These amplifiers were called "rail switchers" (class H) as the transformer and Triac would be preset to 2 or 3 set levels of voltage that they would run the amplifiers power rails at, and switch between one of the 3 depending on the output demands of the amp. Later on, he then developed a power supply that would monitor the input signal, and run the rail voltages to the output transistors at exactly 4 volts above what the output was going to be at all times. That scheme worked well, as instead of ALWAYS having the entire power supply voltage across the transistors, and wasting all the excess power not being used as heat, the power going to the output transistors was always just slightly above what was going to be required at the output. These were known as class "G" amplifiers, and the benefits were that one could obtain high power out of a smaller case, because not as much heatsinking and not as large of a power supply transformer was needed to cover the excess heat and power demands as if the rail voltage was fixed at its maximum all the time. Another advantage was that the amps used conventional output stage designs, and sounded good.
That dounds similar to the Labguppen FP series. Those are somtimes described as Class T or TD. The construction is quite interesting mostly an normal AB output stage but with two buck converters per chanel that reduce the power supply voltage. The voltage drop on the AB transistors is in the region of 5-9V thereby reducing the heat dissapation.
@lukahierl9857 Those amps did not have a conventional class A/B output stage. They went a little further than Bob's design and used a similar tracking power supply, but then combined it with a class D output stage for even more efficiency.
Unfortunately, the commutation introduced new forms of distortion that most people of the time weren't measuring for. Carver said it measured good, everyone else said it sounded atrocious. I thought it sounded atrocious feeding mid-fi speakers and I don't consider myself to have golden ear at all. I've owned everything from high end to garbage, I am not on a high horse. But I'll grant his later stuff got better. Never justified the asking price, but he did pay attention after the very interesting amplifier challenge with Stereophile.
@railgap The rail switchers did create a spike in the output when the rail voltage stepped up or down. The later continuously variable power supply amps never had this issue. The perception of sound quality depended on who listened to them, as obviously everyone has their own tastes of what sounds "good or bad".
It would be great to see a multi-video series for this amplifier, eg: rail-voltages of the power-supply vs the IC, the push-pull nature of bridge-mode, and other fascinating aspects.
"They stop becoming an impedance, and start to become a resistance.." This is a GREAT explanation of what happens to a speaker voice coil when it's driven by the amplifier into clipping. Good one Clive... 🍻
Well, no, it isn’t really. There isn’t significant DC on a clipped output. It might be a squared-off waveform, but it’s not DC. For one thing, you have to be driving _pretty darned hard_ into clipping before you are actually generating a square wave. By then, nobody is going to want to listen to … whatever you call the result. It will be completely unintelligible. The only time you get DC on the output is when something has failed, and either a servo-balanced output has gone open loop, or an output transistor is conducting straight through to the rails. You’ll know when this happens because the speaker will release its emergency alert reserve of smoke.
Great topic and dare I say fairly new topic for the channel! I’ve built valve amps from old schematics, but around a year ago started to fix broken Cambridge Audio amps from the early 2000’s on to 2015. Plenty out there and are mostly through-hole construction with schematics easy to find. It’s taught me a lot and their stock op amp (5532) can be swapped for higher quality versions. Their boards had construction issues where a components pad would ‘periscope up’ if you grabbed a component whilst holding the board, but drop back almost invisibly and cause a fault once warm. Even continuity between pads would be fine until they weren’t. That was a pain to discover and add to the troubleshooting process.
Sherlock Holmes' favorite joke was to stand in front of the amp all day, pushing the power switch and saying "Watts off!, "Watts on." The Doctor hated him.
I have the exact same amp branded 'Citronic' for the New Zealand market. I'm running it in bridged mode as a bass amp for playing live gigs. It's effing loud and I've never run it loud enough to clip, you'd get a nosebleed before that ever happened. Nice and light, with a graphic eq and a compressor it's the best bass rig I've ever had.
This looks like quite a nice unit! It's basically a switched-mode power supply that can both source and sink current, delivering a stream of high frequency pulses with a variable duty cycle, and relying on the inertia of the moving parts to smooth it all back out into an audio waveform with a bit of help from the inductance of the speaker coil. The output transistors are (in theory) always either open-circuit (so carrying no current) or short-circuit (so dropping no voltage) and thus do not require massive heatsinks (here, the real world deviates somewhat from simplified ideals and _some_ sort of heatsink is required), unlike a traditional analogue amplifier where one or the other of the output transistors is dropping the voltage difference between the nearest supply rail and the output. The first generation digital amplifiers were a bit prone to high-frequency instability; blowing up tweeters -- and eventually themselves. But that could also happen with traditional amplifiers, if you replaced old 2N3055s (they don't like having the output short-circuited, and the fuse usually outlives them -- especially if someone has put the wrong value in) with modern ones, due to their higher turnover frequency ..... It's nice that they used dissipative regulators on the preamp supply, to avoid introducing digital noise into the small signals. I guess the real test will be how well they survive in the real world ..... Are they musician-proof? Are they roadie-proof? (And where are all the fibre-optic digital audio connections? All the relevant patents have long since expired by now!)
Ballast controllers are surprisingly common in these chinese 1U amps, I've come across them quite a few times. I guess they're cheaper than a dedicated on-line PSU controller? In any case, if you're ever fixing the PSU on one of these due to a blown fet, make sure you replace that IC as well because it's 100% dead every time and will instantly destroy the new fet. Not a mistake you want to make given the fets are almost always the most expensive components on the board apart from the transformer.
No way an IR2110 is cheaper than a China TL494. Impossible. Another weird thing: Big Clive says the IDC connector has power plus "6 control signals - power good etc". How is that possible with an IR2110? Something doesn't add up.
That is an extremely common solution if you want to make unregulated SMPS. It is commonly used in audio amplifiers (no mether of topology) because they do not need regulated imput voltage. I know (low) high-end (read high price, in over 10000$ range) class A amplifiers that use the same solution. It is cheaper to design it than regulated because the hardest part of SMPS design is to design compensation loop, and you do not need to spend any time testing stability of SMPS.
Comparators are used on amp board to drive the signal and overload leds, and also to enable the output stage as well with signal applied. The transistor and resistors are used to drive the fan, resistors to do most of the power dissipation, using cheap resistors in parallel, and in the path of cooling air, and then transistor to do the voltage regulation for the fan, so you do not need a heatsink on the fan drive, and also do not put the DC power noise from the fan on the opamp supply rails either. The mystery IC's will be a complete class D amplifier, complete with built in filtering, high side drive, and analogue input stage, plus some feedback from the output, to make a low part low cost amplifier, with power rating depending only on supply voltage and transistor ratings. One of hundreds of near identical IC's designed for that work.
In a class D amplifier, there will be a modulator that compares the input signal to a triangle wave and sends its output to drive the gate of the output transistors, but if the output transistors are large, this gate can become very capacitive so an additional "gate driver" IC is needed between the modulator output and the gate of the output transistors.
Those 40mm fans are sometimes annoyingly noisy with loud, specific frequency peaks. Unfortunately, not much space in 1U. Although one could put in a centrifugal blower that's 30mm or so thick but say 80mm "square" and still have only a small exhaust cross section. The blowers can develop much higher pressure gradient at lower noise. Airflow volume is typically less than for unrestricted axial fans, but the flow through axials drops sharply when there is a flow obstruction. Blower throughput is much more graceful. There's science and art (Engineering) in effective and efficient thermal management of such devices. Alas, electronic Engineers tend not to have the necessary background understanding... and will do what has worked for others in the past, instead of hiring a Mechanical Engineer for the project.
8:01 As an amateur radio operator, I got the kit and built the Hardrock 50 RF Power Amp, and the power MOSFETs were similarly mounted. They were on the back of the board, and during assembly, I would put the thermal paste on them, install the board, and then reach through a hole in the board with a screw and screwdriver, and screw the tabs on the transistors into the large, finned aluminum heat sink that formed the case to.
Nice teardown Clive. I love class D amps, so simple yet ingenious. I built one many years ago, from memory it generated a pure triangle wave using a pair of comparators, a flip flop and a pair of matching current sources/sinks, then fed that into a simple comparator as the A/D converter, and a full bridge output. It worked surprisingly well and sounded quite good though I never got round the issue of high quiescent current and ended up losing interest. I was going to build a car sub amp using the design one day but never got round to it. Think the schematics are still out there on diyaudio.
I believe the fan is an outtake so it's pulling air through the front panel so that foam acts as dust collection. Would love to see you take apart a D&B or L' Acoustics amplifier, the top of the line brands for live music.
It's important to note that Class-D amplifiers are not digital. They are fully analog. The D in the name was chosen simply because A,B and C had already been used. Although many class-D amps may have digital control systems and DSP the actual amplifier circuity operates as a continuous (non quantized) analog system. Any voltage ripple, jitter or noise will always have an effect on the output signal.
its kinda amazing to see this in comparison to something from the 70s or 80s, class b monsters that would regularly die from heat. Now we get several hundred watts from a single smd package
Got several Cerwin Vega Amps with much the same layout as this. They range from the CV900, CV2800 and CV5000. Thanks for posting this, reminded me to give them a clean out. 😂
I have a funky feeling that you got the direction of airflow through the amplifier wrong. The Fan does blow warm air out the back, pulling cool air in at the front and through the filters. Nobody in his right mind would push unfiltered air into an Amp across all components and filter the warm air on its way out.
Bridge mode potentially gives you twice the voltage, which is 4x the power, for a given load (assuming neither side amp current limits). Most of these designs can only achieve this on real musical sources, with a sensible peak to mean ratio. If you try it with sine waves, a limiter will kick in. Great for us ageing musicians. Together with a neodymium magnet speaker my bass rig has halved in weight and more than quadrupled in peak power. 😊😊😊
Using a Lamp driver does have a good thing in that they are fine with handling wildly varying loads, so can easily keep the regulation loop stable with audio, as the power draw can very wildly over a half second or so, so the power supply can still remain stable even with this kind of load.
@@fredfred2363 You can try IR2153 as a start (do it with soft start, overcurrent protection and IC+buffer [just 2+2 npn-pnp bjt conf]+MOS combination). You can also achieve kWs. 1.6kW is not a problem at all, but an 80x240mm circuit board (large input filter capacitance is the reason). It is better to make IR2156 largely SMD, finding good current sense etc. is a bit challenging as a first project. At home, 600+W feasible to make the size of a cigarette box, planar E38/8/25 transformer, windings from ultra-fine litz braided wire, switching frequency close to allowed maximum from the IC, very fast but light (infineon`s many are good to do this without a buffer) MOSFETs, shottky rectifiers at output and so on..
Good morning BC, I simply love your videos and care you take to share findings with us. Your approach reminds me of a famous movie quote “Life is like a box of chocolates. You never know what you’re gonna get.” This pretty much mirrors your channel. Good show BC! Carry on my friend I hope you and all who reads this have a wonderful day 😊
@@danmyers7827 You can do that by looking at the front panel. No need to open the case or plug it in. Hell, just pick it up and heft it. (but compare it to other Class D amps, duh, since they tend to be much lighter than their Class A/B cousins.
I recently switched from a class B power amp to a fully digital 2.1 amp in my office/home studio listening setup. I opted for a TPA3225 based design which is so much better than the old iron and heat sinks approach. Sure, it gets annoying when you crank it really loud, due to ringing, but I am normally under a quarter of its capacity, so it's fine. They managed to pack a 2x75W + 150W 2.1 amp in a box the size of two cigarette packs. It stays almost cold, even after hours of heavy use. It's lightning fast, you can hear it when modifying the transients of a recorded track. It draws 2W on idle. I won't go back to an analog amp anytime soon...
These amps are in fact analog, they do not have a code of zeros and ones to make the sound. They function with pulse width modulation and that is in essence still an analog principle.
@@ottonormalverbrauch3794 they use a specific digitising algorithm we know as delta-sigmoid modulation. It is equivalent to a digital signal in many ways as that's truly how simple DAC and ADC can be.
Ensure your TPA 3225 has sufficient voltage. Their distortion performance is far superior when they have the voltage headroom, both on the end stage amp and the signal stage with feedback.
@@AnnaVannieuwenhuysesorry, class D amps aren't digital. As explained above, they're purely analogue. A quick google will show the difference. It's a very common misconception that class D means digital, it doesn't. It just means switching.
@@AnnaVannieuwenhuysea digital signal needs to be quantised in both amplitude and duration. A class D amp has only one of these - amplitude. The PWM stream is infinitely variable in length, determined by the analogue input signal. If it was digital, those pulses would be discrete lengths.
Commonly with many digital amps the bridge switch only puts one signal to both inputs. One channel is already inverted to put a more symmetric load on the power supply.
Eek! as an Audiophile (have to be careful how you say that!') I am suprised at how little there is inside it. My Rotel RA971 MKII which is hooked up to, two RB971MKII and required some re-positioning of jumpers. And the board was, well full size and a ittle bit more populated. Hell's teeth how far tech has come.
clipping on opamps usually means that you are trying to drive above its max output voltage, wich you obviously cant so it flatten the "loudest" part to its max and distorts the sound so ill assume that i means basically the same here meaning that you are either hitting the power limit or the voltage limit of the output, both distort the sound but should not damage anything.
Digital amplifiers are honestly very simple. It is mainly just a comparator looking at the output voltage vs the input voltage. If it is "too high" it switches the output transistors low. If its "too low" it switches the output high. The LC network then smooths that square wave out into an analog signal. To aid the stability of the circuit one will also just sample the comparator's output at some fixed frequency. Preferably one that is N times higher than one's peak operational frequency, as to give room for adequate filtering. The two hard to identify chips are likely just a half H bridge driver IC meant to drive external transistors. It likely also has some clock input used to synchronize when it should update its output stage. It is honestly quite simple to build one of these amplifiers. All one needs is a comparator chip, a clocked latch and an half H bridge driver IC. One can go with a full H bridge driver too, then one effectively have a differential output, though one then needs two LC filters for said output, so pros and cons...
Noctua fans are quiet mostly because they spin at low RPMs. Low RPM also has an effect on the fan’s CFM rating - or, in other words, how much heat it can move away from things that get hot. Ergo, quiet fans tend to lead to _very_ quiet amplifiers, if they weren’t designed for low-n-slow airflow.
@@Okurka.tell that to some proffesional PAs. One I have serviced contained 4x 60mm Fans with 10W each. That thing was an freaking tornado on startup. But during operation it wasnt a problem because of the 15kW of output power put into 6x 18" subwoofers
Ive blanket banned Class D amps from any of my personal installations. When I first got into radio, I built my van up with a Class D amp for a subwoofer. Unfortunately it radiated a massive amount of noise on 145MHz, right on the 2m ham band. Its been a few years but I wanna say it was decreasing the sensitivity for the Kenwood TK-5720 that I was using for 2m by at least 10db. It didnt even have to be doing anything other than being 'On', just idling, to cause this. I located a Class A/B amp and replaced it, problem went away.
we amateurs must make many sacrifices... the XYL wants a dimmer in the dining room? Well, it's gonna cost a hundred-fifty dollars, not fifteen... *sigh*
@@jagmarc Im sure it could be done but I have too many projects to do everything I want to do as it is. Buying an A/B amp and replacing the Class D was the option I took. I have made some attempts to deal with RFI from my LED lightbar, which exceeds 30db, making my radio completely deaf. None of them were effective. I think the only way to fix it would be to remove the switching supply inside and either replace it with a better quality one or design a linear current regulator to take its place. As it is, I just deal with the fact I can't hear anything for those few times Im so far off road that I need the additional light.
@@clynesnowtail1257 wow, how amateur radio has changed, doesn't seem to be any design & development any more, now it's "not working right? - buy something else ready made". There was a time when a radio ham would design & build their own 0.25 uV sensitivity receiver that listened through the same antenna an active transmitter was using at same time (on a different band of course).
Loudspeakers are a linear motor connected to a diaphragm that moves to produce sound. They don't have any other power source or electronics built inside them, so in order to operate, they require an amplifier to send them a high voltage, high-current signal that can drive them. An amplifier is a big power supply that drives the linear motor inside the speaker to push and pull on the diaphragm. There is always a power supply section that rectifies the incoming AC power, into DC and there is always an output stage that acts like a valve for that power which modulates the DC voltage to an AC voltage in proportion to the incoming electrical signal, which is the musical signal. The output from the amplifier is just like the input waveform, only much higher voltage, and capable of much larger current. This amplifier makes it very easy to see the different parts of the unit that serve different functions; sometimes the circuit boards get very dense and it becomes difficult to see what is going on.
Looks ok for a lower cost unit assuming it sounds good, but I do not accept their rated output by standard testing practices. Let’s see how long it lasts under load too. Soon as that foam filter loads up with normal daily dust the high temperature cut-out will hopefully function as intended.
Isn't the output rating a peak measurement for most amplifiers? it's rarely going to sustain those loads for extended periods as it's amplifying a dynamic input. Usually it will had DSP to limit the input signal (or filter it) so you don't go over the peak, but the amp will clip if it tries to output too much, which you're supposed to address before the signal goes into the amp.
@@andygozzo72 I'm assuming it's implied that it's peak based off the nature of the device as rms would make no sense. They are designed to output only what they send in, so having its average power draw wouldn't be possible. I don't know what you mean by music power, but peak would make sense.
@@OscarSommerbo 180W suggests closer to 25V for each power rail, and as the ribbon, and the connectors, are rated for 1A each, 2A if you are being unrealistic, just using 6 in parallel will easily handle the power required, as it should be around 5A per power rail. With 24 pins, 4 being used for the opamp +-15V rails, 2 for fan and temperature sensing, and thus 6 for the +25V, 6 for -25V, and 6 for ground, will work out fine. Plus a quick almost foolproof connection, and reasonably robust.
An insight into the fact that Class D amplifiers are switched DC power supplies can be seen in the power ratings. It's usually something like "250 Wpc 8ohm, 500 Wpc 4ohm, 1000 Wpc 2 ohm". I love Class D amps because they are so light and efficient. In my outdoor theatre rig, I replaced two conventional A/B analog amps with a 4-channel Class D that weighs less than half as much and supplies more than twice the power. Yep, as far as ground-lifting goes, don't. Just don't. It's how you get the nice shiny bare aluminum Shure 58s going to line voltage, as dramatized in "Almost Famous".
The fan is not blowing into the enclosure, it is sucking air out from it. The intake is the front grille. They should have put a bigger fan in it, it can be done by a bracket, putting the bigger fan in a 45° angle and cutting out a bigger area on the rear panel for venting the air out, probably even a 80x80mm fan could have been installed into it this way, reducing the noise and increasing the cooling effect. This is small fan, and on top of that, the vents are covering most of the active area of it. Otherwise, it seems like a nice amplifier, if it is not too expensive. Of course all the noname electrolytics should be replaced if someone wants to use it in a heavy duty application, like in a bar or something. Hopefully the semiconductors are not counterfeit. Hopefully the secondary GND of the power supply is Earth grounded, so if the transformer has the usual Chinese-style dodgy insulation between the primary and secondary windings, it won't be lethal.
somewhere else on YT there isa review of a 1U "subwoofer amp", marketed into the live sound space. Dude opens it up and it's literally a kit type plate amp designed to be built into the side of a small subwoofer cabinet. 🤣
I am the proud owner of a couple Crown D-series amplifiers, to include a D-150A (Mk II), which would be a close contemporary of this amplifier in terms of power output to speakers. As a Class AB amplifier, it draws about twice the power for the same output, it's three times the size, and probably six times the weight. Astounding what can be done with high-speed switching FETs these days. I'd probably trust my D-150 more in a high reliability application, though, especially since I have schematics to repair it. (Edit: Not that it's terribly complicated...)
I love those old Crowns the lead sled power supplies they were bullet proof ! the only failure I had I pulled it to below 2 Ohms and burned out the mosfets But it ran like that for 6 hours before that happened driving 4 dual 18 band pass subs !! at a rave . I had extra amps in the rack to go to as a back up After we did the math it was running at almost a dead short !!
yup, I have that old D-300A with its 1kVA transformer in the back. No short circuit protection, no AC coupling, no speaker protection relays, no mercy :-)
I have a Crown CE4000, it is a very nice and robust class D amp. Not in any way related to the CE1000 and CE2000. Those are old school Class AB with iron core transformers. The CE4000 is not class D in the traditional sense as it dosent use an high frequency output stage and an LC filter. It can best be described as an class AB where the two transistors are replaced by two buck converters with an special control cirquit. The ripple current of the two buck converters cancels out and the difference of the high and low side currents is the output signal. And the service Manuals from crown are wonderfull.
If I may be so bold for a suggestion: a video explaining "push-pull". What it does, how it works, and why use it. You did a some "clive-splaining" videos already, and I suspect you'd do a good job about that as well.
@@CJDavies It happens, its just a REALLY bad idea as racks and everything else in them are designed to be fed cool air at the front and dump hot at the back. So youre either gonna need a dedicated rack or you'll be sucking hot air. It just makes sense to put your intake side where the user and aircon are.
@@CJDaviesi had an rack with two qsc amps. The befefit of reverse airflow is that one can mount an big fan with a proper paper wlement filter into the rear cover to keep the equipment from getting filthy.
There was a Sharp (!!) essay into high end audio in the 90s that included a $15K Class D "1-bit" amp which spit out PCM to the speakers, relying on the reactance of the speakers to filter the audio. The scheme sounded fraught with peril to me (speakers do vary quite a bit in their reactance curves, wot-wot) but Sterephile (where no one blinks at a $50K amp) gave it a gentle treatment. O_O Now they are everywhere, heh. We've come a long, long way from Bob Carver's ambitious but atrocious-sounding M-400.
A variable frequency drive and a inverter, is technically the same thing as a class-D audio amplifier. The working principle is basically the same. You are trying to make a sinewave, using PWM.
Digital amplifiers are surprisingly good, even the super cheap Chinese ones sound pretty good. I remember the days when a 360 watt amplifier would almost need a fork-lift truck to move it.
@@buddhistsympathizer1136 OK, I'll bite. How is an amplifier that's doing the amplification by a MOSFET and then using pulse width modulation to produce a pulse train output NOT a digital amplifier? This isn't a linear gain amplifier, it's a class D, digital amp.
@@PenryMMJ With respect, do research on Class D amplifiers and how they differ from the others. 'D' is an amplifier classification. It is not short for 'Digital'. The pulse train you speak of is a series of analogue pulses, not digital pulses. No part of the amplifier is 'digital' in nature.
@@buddhistsympathizer1136 this research perhaps? "Class-D amplifiers work by generating a train of rectangular pulses of fixed amplitude but varying width and separation. This modulation represents the amplitude variations of the analog audio input signal. In some implementations, the pulses are synchronized with an incoming digital audio signal removing the necessity to convert the signal to analog. The output of the modulator is then used to turn the output transistors on and off alternately. Since the transistors are either fully on or fully off, they dissipate very little power. A simple low-pass filter consisting of an inductor and a capacitor provides a path for the low frequencies of the audio signal, leaving the high-frequency pulses behind." The amp is amplifying the signal in it's digital state. It's not an analog device.
@@PenryMMJ Just because it uses square waves, does not mean it is "digital" ... PWM vaires the width of square waves in an analog fashion ... It's a modulation scheme, not a digital process.
They could have at least put one of those stacked double 60mm fans in this like they use in 1U servers. They sound like jet engines at full speed but they move a ton of air. That single fan doesn't seem nearly large enough to move air with all that heat sinking inside, particularly if this is stacked in a rack between other hot equipment running. That and the speaker posts you noted are my only complaints. Would love to see you break down more complex audio and industrial gear like this!
Here the heat spreader/ heatsink is just a plate. So the thermal design may be optimum for this fan. Such small fans as selected are mainly for " occasional" low airflow circulation to maintain a uniform spread of internal temperature in the amp. Excessive air flow from high speed fan..( especially those that run continuously) will always cause to pull in & have more dirt/ dust accumulation inside. and are more prone to fan failure..
@@analoghardwaretops3976 The surface area makes sense as you explain it. I'd get the dust argument too, but of course a dynamic speed fan would prevent that and be the best of both worlds. That said, I acknowledge your point and you are probably right.
@@hgbugalou yes there are many pieces of equipment that have 2-3 auto fan speed settings , (whilst some may auto variable) that are selected depending on temperature..and may also be normally off , but get turned on once a temp. threshold is crossed.
Some people call these "class D amplifiers" (or even "analog switching amplifiers") and leave the digital amplifier term to "power DAC" type amplifiers that take a digital input signal and only after the power switching stage use the L/C filter for analog conversion.
The locking system for power mosfet is "inspired" by QSC power module. Also the ics that are probably an irs20957 clone is used in QSC active speaker power amplifier. The TIP and the rail of resistor generate a +12V but referred to the V- rail. Usually when the rail voltage is high (and low considering it's dual) it is derived from an isolated secondary winding. Very strange the relay because the IRS20957 have a shutdown pin witch can be used to delay the turn on. Not a good choice the ribbon cable and the IDC connector to delivery power.
Well said... that was exactly my first thought... similar to an original K series Bucket of Power amp design, but not as refined. Since I live in a hot, humid country where plastics break down quickly and as far as I know QSC does not sell those plastic things (tensioners?), this "rigged" version idea may come in handy some day!!!
@@radioflyer2030 well if I remember correctly in the QSC version the plastic post push on the body of the MOSFET and so it doesn't align with the tab hole.
Those are not "super duper large" filter caps. I have an old (70:s or 80:s) Peavey powered mixer with rated output of just 2x100W and the filter caps are about the size of Red Bull cans, just a little shorter. Also, the power transformer looks like something out of a welder. I'm pretty sure that one actually outputs the rated power all day if necessary unlike most modern (at least the cheaper ones) amplifiers. But of course the Peavey also is the size of a small suitcase and weights something like 20 kilograms but that does not matter as it's just used at the training room and never transported anywhere. Works great despite it's age, just some of the pots are a little scratchy.
omg this amp is crazeh. i personally dont like the vacuum cleaner/dust collector style construction every dog and its breakfast is into these days, but the heatsinking just cannot be serious
It's so weird seeing inside an amplifier but not seeing a couple of huge masses of iron transformers...just still used to the old school style. I don't like how the ventilation fan has half its blade width covered by the case...the vent holes should have been much wider to allow proper airflow.
And given the long slot vents on the front panel ~ that fan looks small for the “ventilation architecture” and the heat-sink is dumping to the bottom case plate - I can imagine that easily overheating [particularly if stacked with its companions], would the long vents on the front be a bit deceiving with it cooling capacity? 😠
The problem with clipping is the waveform becomes distorted in the lower frequency (by the time you clip hf, the lf will have distorted long ago) but since a low frequency driver can't physically move as the amp is fighting its qms, it provides a lot of mechanical (and in turn electrical) resistance - that's what kills amplifiers and speakers.
Uh, no. Clipping doesn’t damage amplifiers. It’s a saturated output. The output signal has gone as close to the power rails as the output transistors can drive it, and so it can no longer follow the incoming signal’s waveform. That’s all. This doesn’t even hurt speakers, inherently. They can’t tell the difference between a clipping amplifier and intentional (recorded) distortion - partly because any naturally recorded distortion IS from a clipping amplifier. The only problem is that the average power of the output signal increases relative to its peak power - because the waveform is getting closer to a square wave, on account of the saturation at the peaks. This can drive a speaker into thermal overload, but notably, in exactly the same way as it would if you had turned the volume up to the same average power. Granted, the squaring of the waveform generates harmonics in increasing frequency, which will be sent by the speaker’s crossover to the tweeter. So the tweeter will be stressed to failure before any other driver. But it’s still just a case of too much power creating too much heat, and hitting one of the two natural limits of a driver. (The other being excursion.) There’s nothing inherently damaging about distortion. Not to the amp, not to the drivers, not to your soul. 🤘
@@nickwallette6201the nice thing about active crossovers is that when subwoofer amp goes clipping it can be easily herard and dosent damage the driver. Very usefull when someone turns up the base for mysterious reasons.
@@nickwallette6201 You're not inherently wrong with what you have said, but you're basing it on the sole premise that a given speaker can handle significantly more power than the amplifier - which is almost never the case. When you build any sort of audio system you always overspec the amp, because it's easier to do and leave a lot of headroom for speaker burn in, future upgrades, and to put it blunty: cheaper amps not being able to output full power at
letter D used to designate this amplifier class is simply the next letter after C and, although occasionally used as such, does not stand for digital. Class-D and class-E amplifiers are sometimes mistakenly described as "digital" because the output waveform superficially resembles a pulse-train of digital symbols, but a class-D amplifier merely converts an input waveform into a continuously pulse-width modulated analog signal. (A digital waveform would be pulse-code modulated.)
@11:20, I wouldn’t say the fan is stirring the air around the power supply. That isn’t how fans that pull or push air into/out of an enclosure work. By moving air in, some air has to go out somewhere, and generally speaking that should take heat out of the box, assuming the outside air is cooler!
The timing of this video :D Im currently in the process of repairing a QSC Pro Audio Amplifier.. With a blown Switch Mode Power Supply.... Picked it up off ebay for cheap.... Will be worth a bit of money when repaired, Being QSC... But im planning on adding it to my audio rack, to run it bridged for one of my woofers..
I'm currently servicing a Behringer amplifier that's like this inside. All tiny tiny surface mount except for the outputs. Big beefy mosfets strapped to the chassis. I'm lucky it's just a set of shorted output fets, that's easy, trying to repair it further would require one hell of a microscope we definitely don't have.
Hopefully not an NX6000. I tried reparing one of those with an blown output stage. I was lucky that i put the cover on before turning it on because one set of mosfets exploded very violently. Turns out there was more damage than just the output mosfets. I eventualy sold the power supply module because the amp wasent worth the effort to further diagnose. Then again I am now working on some Labgruppen FP10000Q and an FP14000 with an scary +-160V and +-200V supply with more capacitors thsn some 1000W AB amps
@@lukahierl9857LabGr. they're more likely 10000W & 14000W and not 1/10 of this as you mentioned.....exercise caution when live testing. protective eye gear , ear muffs etc.
When i design audio gear from scratch, i treat the earth connection like a mains neutral... think about it. (Most grid power installs tie the N and E together near the consumer unit) If you assume it is a neutral, and there are potentially leakage currents (and induced volts) on the earth line, there is a higher risk of mains hum. You have to assume that earth isn't really an earth for audio. Differential signal grounding is the solution. But few people bother! Nice amp.
The problem with clipping is that it is effectively an increase in average power, which is the only thing that will destroy a voice coil. It's not to do with waveform shape or resistivity in the load, it's simply an increase in average power. It's still an AC signal even if you clip the waveform to a proper square wave. It's also interesting to note that the reason why manufacturers get away with making claims that are double the RMS power of their amplifiers because if you clip a sine wave to a full square wave, you've doubled the power in that waveform. It's not real power or usable power but it has the heat load none the less.
“Clip” just means that a signal is exceeding a limit and the peak and/or trough of the signal is being cut off or “clipped”. It can be bad, it can be desirable. The guitar distortion effect would be impossible without clipping.
Not this model but just replaced one like this driving studio monitors for crackling in the speakers and a channel dropping out... reflowing/adding solder to the output jacks seems to have fixed it. Maybe when I redo that studio I will open the new one and see if it's a thing in small rack mount amplifiers. You would think considering what some of these go through on the road they would reinforce that part. I also wish more studio gear came with the ground lift switch. While I haven't tested it I suspect the output rating on many of these units is a bit... generous.
Wow...some of those circuit boards look as though they are Sim City maps 😅 Soooo many components clustered so close together!! I recently bought an old stereo amp with the intention of having a look inside...now that I see this, I know I'm in over my head. I'll still take it apart, but only to satisfy the scientist in me. Once I've had my fun it's going back together and being repurposed for my experiments.
The older style equipment tends to have a lot more dedicated components rather than the modern ones being full of chips of some sort. Depending on the era it's from, it's even more beautifully designed and looks even more like cities, especially before all the SMD stuff became a thing.
The mystery chips are possibly a clone of the obsolete International Rectifier /Infineon 98-1073pbf which is listed online for sale as 'ic audio driver digital'. No data sheet seems to be available but it's probably very similar to the common IRS2092. The power supply looks like it is a unregulated LLC design. The large caps close to the transformer are in series with the transformer primary so that is resonates at the drive frequency. Current in the primary is sinusoidal and switching losses and generated noise is minimized.
IR2156 is not even close to LLC, LLC is much different topology, yes it is in terms of interference noise, LLC is the best topology for powering an audio amplifier, but all these converters operate at a fairly high frequency, and when they stay in their modes, they usually do not have audible interference noise. The IR2156 is actually a ballast driver for a fluorescent lamp, but it can be ideally used as a half-bridge switching power supply controller. If it is done correctly, they are also quite reliable, they work even at frequencies 100kHz+ (quite extreme and must be a well-developed solution), it is possible to take 5-6kW of power from this kind of half bridge PSU to the powered device. Of course, this specific one in video does not correspond to these numbers (and don't have to acorrespond it, because amplifier is also much more modest)... LLC's standard example is ST microelectronics L6599, but it would make power supply much more detailed (dense) visually - the LLC power supply requires a very careful design and its (in L6599 case) power capacity is also quite limited (you cannot create a power supply unit of the same power as aforementioned numbers on basis, as is possible with these simple ballast drivers). With ballast drivers, it is also possible to use several protection functions - overcurrent, overvoltages, etc. that are present in full-blooded PWM drivers, but this has to be done externally, often this external design can also be better than on-chip solutions. I'm not afraid of them (ballast ICs), they've proven themselves quite well, and with add-ons, they can be made completely equivalent to special PSU drivers. And finally - an unregulated power supply is the best possible solution for an audio amplifier in 99% of cases. To understand this, dynamics of the audio signal over time must be studied...
@@inarinukka7729 I actually have a fixed frequency L6599 controlled power supply in front of me right now, powering my computer speakers amp. It's not very complicated. True careful design is required but complexity isn't necessary if regulation isn't required, just operating at the gain/load invariant point on the curve is OK. No reason this can' t be done with a IR2156. Just a hunch that the PS is LLC, might not be. Would depend on how much leakage inductance the transformer has.
@@stephendredge6077 but IR2156 is much-much-much simpler and reliable if properly constructed. Also, the IR2156 has a significantly simpler transformer design and you can use an IR2156+buffer+powerful MOS or IGBT+ colossal E55, E65 etc. transformer and rectify the output and take tons of power from it while keeping it simple and maintainable, with light weight. You dont need to calculate, wind it, to measure and recalculate and readjust - you simply just can calculate what power you need, what fq you have and you can make it once and it works at start, no gapped core and you get much higher power than same fq/volume LLC. Of course, manufacturers of all kinds of junk make things so that they work for a few years in a certain range set by the manufacturer, but if the magnetic materials get old, for example, then it's a kaboom and no one repairs it anymore, and you have to buy a new one again.
Something new for me intresting amp compared to the linear amps i still repair today fantastically explained as normal Clive you are brilliant certainly helped me over the years and using ur pinky with soldering lol epic
Few years back I needed a stage PA amp, so I bought some Class AB modules. Then ignored them, and bought some Class T modules instead. (Class T is like a clever sort of D) Because Class T and D are so much less fussy about their PSU, it was cheaper to re-buy the whole modules than it was to build a PSU for the AB I already had. The Class Ts were happy with a cheap switchmode and sounded just fine. A PSU for an AB would have needed an expensive toroidal transformer and all sorts of high-end bits.
I just tried going to the link in the description, mostly as I'm curious what this unit costs and received an "Access Denied" error. On the Pulse web site and others, it is listed for 157 Pounds. I know the description says you didn't test it for audio performance. As nice as it was to see the inside, I would sure like to know what it sounds like and does it meet the specs. Many decades ago I had an integrated pre-amp and power amp with 22 tubes and a massive transformer all to make 120 watts. In the winter you could heat the room with it. Then we went to solid state which was somewhat smaller, lighter, and less hear. But they still had the massive transformer and heat sinks for the output transistors. Now this amp is ridiculously small and light and almost nothing for heat sinking.
the tpa3255 is a tiny class-d chip that requires a few components to work, its rated to 480W at 1% thd in bridged mono into 2 ohms the only other chip that comes into my mind with such power levels is a TAS5630 its older and a bit bigger.
Thoughts on the unidentified chips: The comparators should be used in conjunction with a triangle or sawtooth wave generator which converts a voltage level into a PWM duty cycle. Are they connected in such a way that you are able to infer that kind of relationship?
Wouldn’t you expect the triangle/sawtooth wave generation and initial comparator circuit be on the input board? I would think the ribbon cable that goes between the input board and the amplifier board probably carries PWM signals and not linear analog signals.
@@timpontiusNot really. And it’s fairly standard practice to make a PWM sawtooth generator using a comparator too, but that implies that this isn’t switching at high frequency for efficiency. My guess too is the UICs are mosfet gate drivers.
Dedicated class D amplifier chip, which has all the parts needed to implement the circuit, from input analogue, to output drive with high side driver for the mosfets. Comparators are there to drive the front panel LED's, and to also enable the amplifier chips when they detect enough input signal, and also are there for DC offset protection as well, turning off the output relay if there is a DC offset at the speaker outputs, and providing the turn on anti thump delay as well. quad comparator, so 8 in total. 2 drive peak indicator for each channel, 2 do the signal detection, and thus amplifier enable, with a few second off delay for quiet input signals. Then the other 4 do DC offset protection on the output, and one does the power on delay, leaving one that is probably sitting being used as inverter to drive the mute control of those PWM driver chips, or is sensing heatsink temperature, to enable the fan to run.
@@tommihommi1 A ribbon cable can easily handle the frequency of digital audio signals. And it would be much less susceptible to noise. I don’t know if the ribbon cable here is carrying digital or analog, but if *I* were designing this thing I would convert to digital on the first board. And if I *HAD* to send the analog signal between boards I would probably use a shielded cable and not a ribbon cable.
I appreciate the plexiglass cover "protection" over the 110/220v, that's not a switch you want to move by mistake
Back in the 90's when I was an on-site engineer for Silicon Graphics, I had a job to replace a failed power supply unit for an SGI workstation.
After replacing the PSU, I proceeded to switch the workstation on to test it. All of a sudden, there was a Loud Bang emanating from the supposedly repaired machine, and the magic smoke was escaping from it. And all the electricity went off in the building. Everything just went dark. The office the machine was in went silent, and all eyes were on me.
It was only when I - very puzzled - looked at the PSU, did I notice the voltage selector on it. Upon closer inspection, I further noticed that it was set to 120V, instead of the 240V it should have been - obviously this unit was manufactured outside of the UK and was set to 120V by default. Not something which was ever mentioned in any SGI training.
That was a very awkward time in that particular customer's office for a few moments. I blamed the Loud Bang and the tripping of the building's circuit breakers on a "faulty replacement" - which, technically, was the truth ;)
@@FrickinLaserBeams Oops!
Back ca. 1990 we were supplied with "Rockwell T50 Programming Terminals" £8K each - actually a rugged PC, and often the first such any of our techs would encounter. This was before auto supply ranging was common. For "safety" factory floor sockets were 110V (Round, blue). However technicians' workshop sockets were 240V (Round, Yellow). If we were lucky just the power supply went "poof" otherwise - a very dead box.
@@M3WDD Good heavens! If they'd changed the 110V sockets to Edison (US, Japan) style, it would probably been enough to make people think twice and make sure.
Indeed; We ran into that with some 'legacy' systems that used, IIRC, an S-100 style mainboard with add-in CPU and IO cards. The boxes had non-autoranging power supplies with that style of voltage switch. the vendor had to replace the power supply on one of them after someone failed to check the position of that switch before plugging it in, and that was entertaining, because not many people make that hardware anymore..
This is why any hardware I spec for [RedactedCo] states "power supply SHALL be auto-ranging between 120V and 208/220".
Great video as always. Could maybe have explained a bit more about how class D amps work for those who don't know, as they are a pretty fascinating technology. Those IC's are most likely re-branded IRS20957s or similar, these are the most common class D amp driver IC's. Frequently see them installed in amps with a OEM part code rather than their standard part name. One thing I disagree with however, is referring to class D amps as 'digital'. They are actually analogue in operation, and the term 'class D' is just coincidental. While they do use a square wave as the main carrier, it is just produced by a simple oscillator at a fixed rate, and is not linked to any clock or carries any bits of information. It is simply the fast switching that allows the output mosfets to be turned fully on and off as switches, and hence minimise power dissipation by not running them in their linear range. The audio is produced by modulating the pulse width (pwm) of the square wave in accordance with the incoming audio signal, which is analogue in nature. The output choke and capacitor on the output form a second order high pass filter, with a rolloff point high enough that it blocks most of the high frequency switching noise, but the pulse width modulation occurs at audio frequencies so these are passed through the the loudspeaker. In most designs, you still have a positive and negative half of the amp, much like a traditional class AB amp, where it differs is that the mosfets are being switched hard on and off, rather than directly being controlled by the audio waveform amplitude. And that line of resistors down the edge, is probably to indeed provide a local power supply as you suggested. But it is probably referenced to the negative rail, as the gate drivers need a +12v supply referenced to the negative rail, in order for the switching of the negative side mosfets as they usually use N channel mosfets for both halves of the amp, much like a quasi-complimentary class AB amp. Some amps I see use N and P channel fets, but usually they are all N channel, probably due to more even performance characteristics.
Um, er, ACKCHOOULLY, Class D amps may be analog controlled, or digitally controlled, so there. :)
@@railgapThat doesn't make it a digital amp, any more than a pwm controller makes a Dyson motor a digital motor. A digital amp would be something like a TTL high-current logic buffer in an output to drive more TTL inputs. Most modern amplifiers could be referred to as "Digital" as they can be remotely controlled by rs232 or usb or an infrared remote control but that doesn't make them a "Digital" Amplifier. It's all marketing bullsh!t.
Agreed, PWM (Class D) is analogue. At no point is the audio signal quantised or sampled.
Precisely the kind of definition I was looking for. It struck me firstly that "digital" in reference to an essentially analogue device, was obviously bullshit designed to baffle brains and sell units. Thanks for taking the time time explain Class D amplification.
@@rarelycommentsMakes sense when you put it like that. 👍
Thank you. As I have said before your channel introduces me to new concepts and circuit designs.
I did not know about class D amplifiers, having grown up with traditional push pull amplifier designs.
This has set me on a voyage of discovery.
75 and still learning.
I'd love to see you cover more audio equipment! I love messing around with guitar effects pedals and replacing parts on them.
I'm in love with the TPA3116D2 amplifier. It's a 50 watt Class D amplifier about the size of a pack of cigarettes and sounds amazing. I've got it running the sound from my computer to two 50 watt bookshelf speakers and it's doing the job VERY nicely. I've bought a bunch for when I need sound somewhere because they sound great and they're 5$ on the Chinese sites. Amplifiers have REALLY come a long way in the last few years.
Class D in general has come a long way. There's zero audible difference between a class A/B and a modern class D amp. Always love how much darn power a class D amp can produce despite being absolutely tiny. Also the efficiency is totally bonkers. Just beats my old heavy AB amp in every category.
If you want an Experience that is rather eye opening... take one of the new TPA3255 based amplifiers (A07, V3, etc) found on popular online vendors and cut it loose on something like a pair of Klipsh Heresies or JBL L-100s ... they're not just for bookshelf speakers anymore.
In fact, the way these things get sold short is by always talking about them as though a tiny amp needs tiny speakers ... when that's just not true.
@@Douglas_Blake_579 Larger speakers can often have a higher sensitivity, in which case they can reach the same listening level with LESS power.
@@boydw1
the TPA3255 can produce upwards of 200 RMS watts per channel on 4 ohms, 100 on 8. And they are still a palm sized amplifier.
Consider the power of a PM8006 for less than $100.
@@Douglas_Blake_579 Oh, for sure, the power these little class-D amps are putting out for the size/money is impressive.
My point was more that the "large speakers needing a large amp" assumption is often backwards.
Say you have a small bookshelf speaker with 87dB/1w/1m sensitivity, and a larger floorstander speaker with 93dB/1w/1m sensitivity (6dB delta). You'd need to give the bookshelf speaker 4x the power to reach the same listening level.
Back in the day playing in a band at various venues one would often find a very noisy wall AC outlet, and all the gear was plugged into this single outlet. But we did have a sophisticated earth lift that could fix the noise with some risk. Disconnect the earth at the wall plug.
Shocking!!
Very common practice!
They sell something called an Isoline you put inline with your XLR which removes the ground loop noise without removing the ground.
You should never lift the earth...
@@AstrosElectronicsLab Exactly, only Atlas can lift the Earth...
I like the ribbon cable with the ruler pattern built into the insulation. That's a very neat idea.
I think it's mostly to help you keep track of each of the wires if you follow them
From my perspective there are two kinds of ground - power ground (protective ground) and signal ground.
So I'd like to have a dedicated signal ground post on the device.
In most rack mounted solutions the air is taken in at the front and then the fan sucks out the air at the rear. That's why there's foam at the front - to capture the dust from the air sucked in at the front.
XLR
So, this is a FULL BRIDGE AMPLIFIER, haha! Looks nice. This is where switching mode power supplies and class D took us - a stage power amp in a single unit enclosure, so compact, clean and neat.
Not full bridge, half bridge, as the speaker has a connection to the mid point of the power supply, and a full bridge amplifier cannot be connected into a bridge configuration, as it is already running in that mode.
@@SeanBZA right, it's hooking the load to two power amplifiers driven with inverted signals
Some of the modern ultra high power 1HE amps are reliably pushing a total of 20kW nominal power into 4 chanels at 4 ohms. Those are slowly replacing the gold standard Labgruppen FP10000Q and FP140000 over here in europe.
Although for low end professional use the high quality chinese FP series coppies are still unbeatable.
and sounds like flaming dog shite, but who cares about sound quality, right? It's only for live sound reinforcement!
@@lukahierl9857 LABGRUPPEN has the PLM44 and others ( using LAKE processor) at least 4-6 yrs in the making , as of my knowledge these are 20000W (4ch) bridgeable and each can also be configured as 4 active p.amps with selectable slope configs...
Also comes with full security locks & remote location networked control capabilities .
11:16 The fan is actually pulling cold air from the front venting sleeves over the circuitry out to the back of the amplifier.
This design is very common in 1U devices like firewalls, routers, etc.....
Indeed. Case fans like this usually blow toward the side with the label, which would be outward in this unit. This is typical, as you don't want a fan blowing directly onto a PCB, as it will develop "cold spots," i.e., temperature gradients that will exert mechanical stresses on the fine electrical traces. Better to have the air flow in through a large inlet and get blown out through a small outlet, as this makes the cold air more diffuse.
Spot On!
5:23 Back in the mid 1980's, Bob Carver of the Carver stereo equipment manufacturer, started playing around with hybrid designs of digital power supplies and typical analog output stages for his amplifiers. He used a special high frequency transformer (that would massively saturate at the regular 60hz mains frequency) and then drove the transformer with a heavy duty Triac. The input circuitry would monitor the audio signal and when the circuitry would see a large musical spike coming it would very quickly ramp up the Triac for the brief instant and the transformer would momentarily output a higher voltage to cover that large musical signal. These amplifiers were called "rail switchers" (class H) as the transformer and Triac would be preset to 2 or 3 set levels of voltage that they would run the amplifiers power rails at, and switch between one of the 3 depending on the output demands of the amp.
Later on, he then developed a power supply that would monitor the input signal, and run the rail voltages to the output transistors at exactly 4 volts above what the output was going to be at all times.
That scheme worked well, as instead of ALWAYS having the entire power supply voltage across the transistors, and wasting all the excess power not being used as heat, the power going to the output transistors was always just slightly above what was going to be required at the output. These were known as class "G" amplifiers, and the benefits were that one could obtain high power out of a smaller case, because not as much heatsinking and not as large of a power supply transformer was needed to cover the excess heat and power demands as if the rail voltage was fixed at its maximum all the time. Another advantage was that the amps used conventional output stage designs, and sounded good.
That dounds similar to the Labguppen FP series. Those are somtimes described as Class T or TD. The construction is quite interesting mostly an normal AB output stage but with two buck converters per chanel that reduce the power supply voltage. The voltage drop on the AB transistors is in the region of 5-9V thereby reducing the heat dissapation.
@lukahierl9857 Those amps did not have a conventional class A/B output stage.
They went a little further than Bob's design and used a similar tracking power supply, but then combined it with a class D output stage for even more efficiency.
Unfortunately, the commutation introduced new forms of distortion that most people of the time weren't measuring for. Carver said it measured good, everyone else said it sounded atrocious. I thought it sounded atrocious feeding mid-fi speakers and I don't consider myself to have golden ear at all. I've owned everything from high end to garbage, I am not on a high horse. But I'll grant his later stuff got better. Never justified the asking price, but he did pay attention after the very interesting amplifier challenge with Stereophile.
@railgap The rail switchers did create a spike in the output when the rail voltage stepped up or down. The later continuously variable power supply amps never had this issue. The perception of sound quality depended on who listened to them, as obviously everyone has their own tastes of what sounds "good or bad".
It would be great to see a multi-video series for this amplifier, eg: rail-voltages of the power-supply vs the IC, the push-pull nature of bridge-mode, and other fascinating aspects.
"They stop becoming an impedance, and start to become a resistance.."
This is a GREAT explanation of what happens to a speaker voice coil when it's driven by the amplifier into clipping.
Good one Clive... 🍻
Well, no, it isn’t really. There isn’t significant DC on a clipped output. It might be a squared-off waveform, but it’s not DC. For one thing, you have to be driving _pretty darned hard_ into clipping before you are actually generating a square wave. By then, nobody is going to want to listen to … whatever you call the result. It will be completely unintelligible.
The only time you get DC on the output is when something has failed, and either a servo-balanced output has gone open loop, or an output transistor is conducting straight through to the rails.
You’ll know when this happens because the speaker will release its emergency alert reserve of smoke.
@@nickwallette6201 WTF was that Garbage you expressed from your keyboard???
@@MrAudioBill Did you have a cogent argument, or just want to be rude to a stranger on the Internet?
Great topic and dare I say fairly new topic for the channel! I’ve built valve amps from old schematics, but around a year ago started to fix broken Cambridge Audio amps from the early 2000’s on to 2015. Plenty out there and are mostly through-hole construction with schematics easy to find. It’s taught me a lot and their stock op amp (5532) can be swapped for higher quality versions. Their boards had construction issues where a components pad would ‘periscope up’ if you grabbed a component whilst holding the board, but drop back almost invisibly and cause a fault once warm. Even continuity between pads would be fine until they weren’t. That was a pain to discover and add to the troubleshooting process.
Sherlock Holmes' favorite joke was to stand in front of the amp all day, pushing the power switch and saying "Watts off!, "Watts on." The Doctor hated him.
That's so bad it's good
I have the exact same amp branded 'Citronic' for the New Zealand market. I'm running it in bridged mode as a bass amp for playing live gigs. It's effing loud and I've never run it loud enough to clip, you'd get a nosebleed before that ever happened. Nice and light, with a graphic eq and a compressor it's the best bass rig I've ever had.
This looks like quite a nice unit! It's basically a switched-mode power supply that can both source and sink current, delivering a stream of high frequency pulses with a variable duty cycle, and relying on the inertia of the moving parts to smooth it all back out into an audio waveform with a bit of help from the inductance of the speaker coil. The output transistors are (in theory) always either open-circuit (so carrying no current) or short-circuit (so dropping no voltage) and thus do not require massive heatsinks (here, the real world deviates somewhat from simplified ideals and _some_ sort of heatsink is required), unlike a traditional analogue amplifier where one or the other of the output transistors is dropping the voltage difference between the nearest supply rail and the output.
The first generation digital amplifiers were a bit prone to high-frequency instability; blowing up tweeters -- and eventually themselves. But that could also happen with traditional amplifiers, if you replaced old 2N3055s (they don't like having the output short-circuited, and the fuse usually outlives them -- especially if someone has put the wrong value in) with modern ones, due to their higher turnover frequency .....
It's nice that they used dissipative regulators on the preamp supply, to avoid introducing digital noise into the small signals. I guess the real test will be how well they survive in the real world ..... Are they musician-proof? Are they roadie-proof?
(And where are all the fibre-optic digital audio connections? All the relevant patents have long since expired by now!)
Ballast controllers are surprisingly common in these chinese 1U amps, I've come across them quite a few times. I guess they're cheaper than a dedicated on-line PSU controller? In any case, if you're ever fixing the PSU on one of these due to a blown fet, make sure you replace that IC as well because it's 100% dead every time and will instantly destroy the new fet. Not a mistake you want to make given the fets are almost always the most expensive components on the board apart from the transformer.
Not only cheap amplifiers. I fixed an old Powersoft amplifier that used exactly the same ballast ic.
No way an IR2110 is cheaper than a China TL494. Impossible.
Another weird thing: Big Clive says the IDC connector has power plus "6 control signals - power good etc". How is that possible with an IR2110?
Something doesn't add up.
@@6AK5W-JAN IR2156, not 2110. You are right though its not cheap online, maybe 80c each compared to tl494 which is pennies.
That is an extremely common solution if you want to make unregulated SMPS. It is commonly used in audio amplifiers (no mether of topology) because they do not need regulated imput voltage. I know (low) high-end (read high price, in over 10000$ range) class A amplifiers that use the same solution.
It is cheaper to design it than regulated because the hardest part of SMPS design is to design compensation loop, and you do not need to spend any time testing stability of SMPS.
@@mrlazda A case of when "good enough" really is good enough?
Comparators are used on amp board to drive the signal and overload leds, and also to enable the output stage as well with signal applied. The transistor and resistors are used to drive the fan, resistors to do most of the power dissipation, using cheap resistors in parallel, and in the path of cooling air, and then transistor to do the voltage regulation for the fan, so you do not need a heatsink on the fan drive, and also do not put the DC power noise from the fan on the opamp supply rails either.
The mystery IC's will be a complete class D amplifier, complete with built in filtering, high side drive, and analogue input stage, plus some feedback from the output, to make a low part low cost amplifier, with power rating depending only on supply voltage and transistor ratings. One of hundreds of near identical IC's designed for that work.
In a class D amplifier, there will be a modulator that compares the input signal to a triangle wave and sends its output to drive the gate of the output transistors, but if the output transistors are large, this gate can become very capacitive so an additional "gate driver" IC is needed between the modulator output and the gate of the output transistors.
Those 40mm fans are sometimes annoyingly noisy with loud, specific frequency peaks.
Unfortunately, not much space in 1U. Although one could put in a centrifugal blower that's 30mm or so thick but say 80mm "square" and still have only a small exhaust cross section.
The blowers can develop much higher pressure gradient at lower noise. Airflow volume is typically less than for unrestricted axial fans, but the flow through axials drops sharply when there is a flow obstruction. Blower throughput is much more graceful.
There's science and art (Engineering) in effective and efficient thermal management of such devices. Alas, electronic Engineers tend not to have the necessary background understanding... and will do what has worked for others in the past, instead of hiring a Mechanical Engineer for the project.
If you can hear the fan the music is not loud enough 😉😂
@@MrJef06 exactly! 🤣
the Noctua 40mm fans are a fantastic drop in substitute that are whisper quiet in comparison
Not a problem when your amp rack is in an air conditioned closet. :D
Plenty of space in that unit to engineer a big fan.
8:01 As an amateur radio operator, I got the kit and built the Hardrock 50 RF Power Amp, and the power MOSFETs were similarly mounted. They were on the back of the board, and during assembly, I would put the thermal paste on them, install the board, and then reach through a hole in the board with a screw and screwdriver, and screw the tabs on the transistors into the large, finned aluminum heat sink that formed the case to.
Looks a tidy bit of kit thanks Clive
I'd love to see you running some test signals through the amp with an o-scope connected to those output transistors to see the PWN waveform.
Nice teardown Clive. I love class D amps, so simple yet ingenious. I built one many years ago, from memory it generated a pure triangle wave using a pair of comparators, a flip flop and a pair of matching current sources/sinks, then fed that into a simple comparator as the A/D converter, and a full bridge output. It worked surprisingly well and sounded quite good though I never got round the issue of high quiescent current and ended up losing interest. I was going to build a car sub amp using the design one day but never got round to it. Think the schematics are still out there on diyaudio.
I believe the fan is an outtake so it's pulling air through the front panel so that foam acts as dust collection. Would love to see you take apart a D&B or L' Acoustics amplifier, the top of the line brands for live music.
It's important to note that Class-D amplifiers are not digital. They are fully analog. The D in the name was chosen simply because A,B and C had already been used. Although many class-D amps may have digital control systems and DSP the actual amplifier circuity operates as a continuous (non quantized) analog system. Any voltage ripple, jitter or noise will always have an effect on the output signal.
Wikipedia disagrees with you: en.wikipedia.org/wiki/Class-D_amplifier
@@rivimey There is literally an entire section of that Wikipedia article (terminology) that confirms what I said.
its kinda amazing to see this in comparison to something from the 70s or 80s, class b monsters that would regularly die from heat. Now we get several hundred watts from a single smd package
Got several Cerwin Vega Amps with much the same layout as this. They range from the CV900, CV2800 and CV5000. Thanks for posting this, reminded me to give them a clean out. 😂
I have a funky feeling that you got the direction of airflow through the amplifier wrong. The Fan does blow warm air out the back, pulling cool air in at the front and through the filters. Nobody in his right mind would push unfiltered air into an Amp across all components and filter the warm air on its way out.
China would.
Bridge mode potentially gives you twice the voltage, which is 4x the power, for a given load (assuming neither side amp current limits). Most of these designs can only achieve this on real musical sources, with a sensible peak to mean ratio. If you try it with sine waves, a limiter will kick in. Great for us ageing musicians. Together with a neodymium magnet speaker my bass rig has halved in weight and more than quadrupled in peak power. 😊😊😊
Using a Lamp driver does have a good thing in that they are fine with handling wildly varying loads, so can easily keep the regulation loop stable with audio, as the power draw can very wildly over a half second or so, so the power supply can still remain stable even with this kind of load.
Using this IR2156 ballast driver, it is possible to achieve a power supplies of 5-6kW. Of course, construction gets more complex.
Genius idea. Like it a lot!
@@fredfred2363 You can try IR2153 as a start (do it with soft start, overcurrent protection and IC+buffer [just 2+2 npn-pnp bjt conf]+MOS combination). You can also achieve kWs.
1.6kW is not a problem at all, but an 80x240mm circuit board (large input filter capacitance is the reason). It is better to make IR2156 largely SMD, finding good current sense etc. is a bit challenging as a first project. At home, 600+W feasible to make the size of a cigarette box, planar E38/8/25 transformer, windings from ultra-fine litz braided wire, switching frequency close to allowed maximum from the IC, very fast but light (infineon`s many are good to do this without a buffer) MOSFETs, shottky rectifiers at output and so on..
Good morning BC, I simply love your videos and care you take to share findings with us. Your approach reminds me of a famous movie quote “Life is like a box of chocolates. You never know what you’re gonna get.” This pretty much mirrors your channel. Good show BC! Carry on my friend I hope you and all who reads this have a wonderful day 😊
Love to see more pro audio equipment! - And some repairs
It would be interesting to load the amplifier to 180 watts and look at the internals with an IR camera.
And then do a quick calculation as to the likely lifetimes of semiconductors and electrolytics.
Or measure distortion at full output.
@@danmyers7827 You can do that by looking at the front panel. No need to open the case or plug it in. Hell, just pick it up and heft it. (but compare it to other Class D amps, duh, since they tend to be much lighter than their Class A/B cousins.
@@railgap I use 20kw RMS per channel digital amps to drive large line array speaker systems, and they can be easily picked up in one hand.
@@railgap I'm thinking in terms of adequate ventilation for the components chosen.
That was REALLY interesting, I hadn't really considered how a "digital" amp worked.
I recently switched from a class B power amp to a fully digital 2.1 amp in my office/home studio listening setup. I opted for a TPA3225 based design which is so much better than the old iron and heat sinks approach. Sure, it gets annoying when you crank it really loud, due to ringing, but I am normally under a quarter of its capacity, so it's fine. They managed to pack a 2x75W + 150W 2.1 amp in a box the size of two cigarette packs. It stays almost cold, even after hours of heavy use. It's lightning fast, you can hear it when modifying the transients of a recorded track. It draws 2W on idle. I won't go back to an analog amp anytime soon...
These amps are in fact analog, they do not have a code of zeros and ones to make the sound. They function with pulse width modulation and that is in essence still an analog principle.
@@ottonormalverbrauch3794 they use a specific digitising algorithm we know as delta-sigmoid modulation. It is equivalent to a digital signal in many ways as that's truly how simple DAC and ADC can be.
Ensure your TPA 3225 has sufficient voltage. Their distortion performance is far superior when they have the voltage headroom, both on the end stage amp and the signal stage with feedback.
@@AnnaVannieuwenhuysesorry, class D amps aren't digital. As explained above, they're purely analogue. A quick google will show the difference. It's a very common misconception that class D means digital, it doesn't. It just means switching.
@@AnnaVannieuwenhuysea digital signal needs to be quantised in both amplitude and duration. A class D amp has only one of these - amplitude. The PWM stream is infinitely variable in length, determined by the analogue input signal. If it was digital, those pulses would be discrete lengths.
Commonly with many digital amps the bridge switch only puts one signal to both inputs. One channel is already inverted to put a more symmetric load on the power supply.
Clive, you and I can be Band-Aid (sticky plaster) buddies. I've got mine on the same finger. Thanks to sharp metal protruding from my fence.
Eek! as an Audiophile (have to be careful how you say that!') I am suprised at how little there is inside it. My Rotel RA971 MKII which is hooked up to, two RB971MKII and required some re-positioning of jumpers. And the board was, well full size and a ittle bit more populated. Hell's teeth how far tech has come.
clipping on opamps usually means that you are trying to drive above its max output voltage, wich you obviously cant so it flatten the "loudest" part to its max and distorts the sound so ill assume that i means basically the same here meaning that you are either hitting the power limit or the voltage limit of the output, both distort the sound but should not damage anything.
I think the word you were looking for is 'half bridge' (orrr push pull) which is combined into full bridge when you select bridge
Nice video to start the morning. That airflow is curiously small.
Thanks for another great video Clive 👍
Nice close look at the amplifier, I think that fan is a exhaust fan not intake fan.
Thumbs up 👍
Greetings
Digital amplifiers are honestly very simple.
It is mainly just a comparator looking at the output voltage vs the input voltage.
If it is "too high" it switches the output transistors low. If its "too low" it switches the output high. The LC network then smooths that square wave out into an analog signal.
To aid the stability of the circuit one will also just sample the comparator's output at some fixed frequency. Preferably one that is N times higher than one's peak operational frequency, as to give room for adequate filtering.
The two hard to identify chips are likely just a half H bridge driver IC meant to drive external transistors. It likely also has some clock input used to synchronize when it should update its output stage.
It is honestly quite simple to build one of these amplifiers.
All one needs is a comparator chip, a clocked latch and an half H bridge driver IC. One can go with a full H bridge driver too, then one effectively have a differential output, though one then needs two LC filters for said output, so pros and cons...
Thanks, Clive.
The fan looks to be an exhaust fan and I've replaced a lot of 1U fans with Noctua 40mm ones for noise
Noctua fans are quiet mostly because they spin at low RPMs. Low RPM also has an effect on the fan’s CFM rating - or, in other words, how much heat it can move away from things that get hot.
Ergo, quiet fans tend to lead to _very_ quiet amplifiers, if they weren’t designed for low-n-slow airflow.
@@Okurka.tell that to some proffesional PAs. One I have serviced contained 4x 60mm Fans with 10W each. That thing was an freaking tornado on startup. But during operation it wasnt a problem because of the 15kW of output power put into 6x 18" subwoofers
Ive blanket banned Class D amps from any of my personal installations. When I first got into radio, I built my van up with a Class D amp for a subwoofer. Unfortunately it radiated a massive amount of noise on 145MHz, right on the 2m ham band. Its been a few years but I wanna say it was decreasing the sensitivity for the Kenwood TK-5720 that I was using for 2m by at least 10db. It didnt even have to be doing anything other than being 'On', just idling, to cause this. I located a Class A/B amp and replaced it, problem went away.
we amateurs must make many sacrifices... the XYL wants a dimmer in the dining room? Well, it's gonna cost a hundred-fifty dollars, not fifteen... *sigh*
@@railgapUse a variac
Banned it? Thought effectively dealing with RFI to eliminate its effect while using it is the exact thing what ham radio all about
@@jagmarc Im sure it could be done but I have too many projects to do everything I want to do as it is. Buying an A/B amp and replacing the Class D was the option I took. I have made some attempts to deal with RFI from my LED lightbar, which exceeds 30db, making my radio completely deaf. None of them were effective. I think the only way to fix it would be to remove the switching supply inside and either replace it with a better quality one or design a linear current regulator to take its place. As it is, I just deal with the fact I can't hear anything for those few times Im so far off road that I need the additional light.
@@clynesnowtail1257 wow, how amateur radio has changed, doesn't seem to be any design & development any more, now it's "not working right? - buy something else ready made". There was a time when a radio ham would design & build their own 0.25 uV sensitivity receiver that listened through the same antenna an active transmitter was using at same time (on a different band of course).
I bought a velleman class D kit to pass an evening soldering a while ago 👍
Loudspeakers are a linear motor connected to a diaphragm that moves to produce sound. They don't have any other power source or electronics built inside them, so in order to operate, they require an amplifier to send them a high voltage, high-current signal that can drive them. An amplifier is a big power supply that drives the linear motor inside the speaker to push and pull on the diaphragm. There is always a power supply section that rectifies the incoming AC power, into DC and there is always an output stage that acts like a valve for that power which modulates the DC voltage to an AC voltage in proportion to the incoming electrical signal, which is the musical signal. The output from the amplifier is just like the input waveform, only much higher voltage, and capable of much larger current. This amplifier makes it very easy to see the different parts of the unit that serve different functions; sometimes the circuit boards get very dense and it becomes difficult to see what is going on.
Looks ok for a lower cost unit assuming it sounds good, but I do not accept their rated output by standard testing practices. Let’s see how long it lasts under load too. Soon as that foam filter loads up with normal daily dust the high temperature cut-out will hopefully function as intended.
Normal daily dust plus other more oily airborne by-products of the now legal recreational use of a popular music pleasure enhancer.😁
Isn't the output rating a peak measurement for most amplifiers? it's rarely going to sustain those loads for extended periods as it's amplifying a dynamic input. Usually it will had DSP to limit the input signal (or filter it) so you don't go over the peak, but the amp will clip if it tries to output too much, which you're supposed to address before the signal goes into the amp.
@@wk4958 wattage ratings should specify whether rms, peak, music power or peak music power ..
@@andygozzo72 I'm assuming it's implied that it's peak based off the nature of the device as rms would make no sense. They are designed to output only what they send in, so having its average power draw wouldn't be possible. I don't know what you mean by music power, but peak would make sense.
Although this does look like cheap Chinese shit
I have done it myself before, but somehow I find ribbon cable for power supply purposes a bit sketchy.
Since it is just +-12V, it should be fine. But an odd step given how well-designed the rest seems to be.
@@OscarSommerbo 180W suggests closer to 25V for each power rail, and as the ribbon, and the connectors, are rated for 1A each, 2A if you are being unrealistic, just using 6 in parallel will easily handle the power required, as it should be around 5A per power rail. With 24 pins, 4 being used for the opamp +-15V rails, 2 for fan and temperature sensing, and thus 6 for the +25V, 6 for -25V, and 6 for ground, will work out fine. Plus a quick almost foolproof connection, and reasonably robust.
An insight into the fact that Class D amplifiers are switched DC power supplies can be seen in the power ratings. It's usually something like "250 Wpc 8ohm, 500 Wpc 4ohm, 1000 Wpc 2 ohm". I love Class D amps because they are so light and efficient. In my outdoor theatre rig, I replaced two conventional A/B analog amps with a 4-channel Class D that weighs less than half as much and supplies more than twice the power.
Yep, as far as ground-lifting goes, don't. Just don't. It's how you get the nice shiny bare aluminum Shure 58s going to line voltage, as dramatized in "Almost Famous".
I’d expect input sources to feed a mixer with ground connected there, and the power amp input earths lifted if necessary.
The fan is not blowing into the enclosure, it is sucking air out from it. The intake is the front grille. They should have put a bigger fan in it, it can be done by a bracket, putting the bigger fan in a 45° angle and cutting out a bigger area on the rear panel for venting the air out, probably even a 80x80mm fan could have been installed into it this way, reducing the noise and increasing the cooling effect. This is small fan, and on top of that, the vents are covering most of the active area of it. Otherwise, it seems like a nice amplifier, if it is not too expensive. Of course all the noname electrolytics should be replaced if someone wants to use it in a heavy duty application, like in a bar or something.
Hopefully the semiconductors are not counterfeit.
Hopefully the secondary GND of the power supply is Earth grounded, so if the transformer has the usual Chinese-style dodgy insulation between the primary and secondary windings, it won't be lethal.
somewhere else on YT there isa review of a 1U "subwoofer amp", marketed into the live sound space. Dude opens it up and it's literally a kit type plate amp designed to be built into the side of a small subwoofer cabinet. 🤣
An exploration of a more classic class A/B amplifier would be interesting.
More amps please. Love em
I am the proud owner of a couple Crown D-series amplifiers, to include a D-150A (Mk II), which would be a close contemporary of this amplifier in terms of power output to speakers. As a Class AB amplifier, it draws about twice the power for the same output, it's three times the size, and probably six times the weight. Astounding what can be done with high-speed switching FETs these days. I'd probably trust my D-150 more in a high reliability application, though, especially since I have schematics to repair it. (Edit: Not that it's terribly complicated...)
I love those old Crowns the lead sled power supplies they were bullet proof ! the only failure I had I pulled it to below 2 Ohms and burned out the mosfets But it ran like that for 6 hours before that happened driving 4 dual 18 band pass subs !! at a rave . I had extra amps in the rack to go to as a back up After we did the math it was running at almost a dead short !!
yup, I have that old D-300A with its 1kVA transformer in the back. No short circuit protection, no AC coupling, no speaker protection relays, no mercy :-)
I have a Crown CE4000, it is a very nice and robust class D amp. Not in any way related to the CE1000 and CE2000. Those are old school Class AB with iron core transformers. The CE4000 is not class D in the traditional sense as it dosent use an high frequency output stage and an LC filter. It can best be described as an class AB where the two transistors are replaced by two buck converters with an special control cirquit. The ripple current of the two buck converters cancels out and the difference of the high and low side currents is the output signal.
And the service Manuals from crown are wonderfull.
If I may be so bold for a suggestion: a video explaining "push-pull". What it does, how it works, and why use it.
You did a some "clive-splaining" videos already, and I suspect you'd do a good job about that as well.
One obvious oversight - that fan is an exhaust - the fan frame is at the back, and you ALWAYS exhaust racks to the rear
QSC/JBL & C-Audio amps used to blow back to front (in the MPA1100/MPX1200 era).
@@CJDavies It happens, its just a REALLY bad idea as racks and everything else in them are designed to be fed cool air at the front and dump hot at the back. So youre either gonna need a dedicated rack or you'll be sucking hot air.
It just makes sense to put your intake side where the user and aircon are.
@@CJDaviesi had an rack with two qsc amps. The befefit of reverse airflow is that one can mount an big fan with a proper paper wlement filter into the rear cover to keep the equipment from getting filthy.
@@lukahierl9857 I hadn't even thought of that, but that's actually a pretty big benefit for some scenarios!
The foam behind the grill is likely to stop you from looking straight through the unit. Or as a dust filter of sorts since the fan is exhausting air.
There was a Sharp (!!) essay into high end audio in the 90s that included a $15K Class D "1-bit" amp which spit out PCM to the speakers, relying on the reactance of the speakers to filter the audio. The scheme sounded fraught with peril to me (speakers do vary quite a bit in their reactance curves, wot-wot) but Sterephile (where no one blinks at a $50K amp) gave it a gentle treatment. O_O Now they are everywhere, heh. We've come a long, long way from Bob Carver's ambitious but atrocious-sounding M-400.
oh poop, the M400 was not a Class D thing at all, I mis-remembered. It was it's own special hell.
Seems super minimalistic compared to standard amps that ì have explored.
very useful review, as I'm looking to buy one of these D-class type amplifiers... As if you read my mind. Thanks BigClive
A variable frequency drive and a inverter, is technically the same thing as a class-D audio amplifier. The working principle is basically the same. You are trying to make a sinewave, using PWM.
Digital amplifiers are surprisingly good, even the super cheap Chinese ones sound pretty good. I remember the days when a 360 watt amplifier would almost need a fork-lift truck to move it.
Except . . . it's not a Digital Amplifer
@@buddhistsympathizer1136 OK, I'll bite. How is an amplifier that's doing the amplification by a MOSFET and then using pulse width modulation to produce a pulse train output NOT a digital amplifier? This isn't a linear gain amplifier, it's a class D, digital amp.
@@PenryMMJ With respect, do research on Class D amplifiers and how they differ from the others. 'D' is an amplifier classification. It is not short for 'Digital'. The pulse train you speak of is a series of analogue pulses, not digital pulses. No part of the amplifier is 'digital' in nature.
@@buddhistsympathizer1136 this research perhaps? "Class-D amplifiers work by generating a train of rectangular pulses of fixed amplitude but varying width and separation. This modulation represents the amplitude variations of the analog audio input signal. In some implementations, the pulses are synchronized with an incoming digital audio signal removing the necessity to convert the signal to analog. The output of the modulator is then used to turn the output transistors on and off alternately. Since the transistors are either fully on or fully off, they dissipate very little power. A simple low-pass filter consisting of an inductor and a capacitor provides a path for the low frequencies of the audio signal, leaving the high-frequency pulses behind."
The amp is amplifying the signal in it's digital state. It's not an analog device.
@@PenryMMJ
Just because it uses square waves, does not mean it is "digital" ... PWM vaires the width of square waves in an analog fashion ... It's a modulation scheme, not a digital process.
They could have at least put one of those stacked double 60mm fans in this like they use in 1U servers. They sound like jet engines at full speed but they move a ton of air. That single fan doesn't seem nearly large enough to move air with all that heat sinking inside, particularly if this is stacked in a rack between other hot equipment running. That and the speaker posts you noted are my only complaints.
Would love to see you break down more complex audio and industrial gear like this!
Here the heat spreader/ heatsink is just a plate. So the thermal design may be optimum for this fan. Such small fans as selected are mainly for " occasional" low airflow circulation to maintain a uniform spread of internal temperature in the amp.
Excessive air flow from high speed fan..( especially those that run continuously) will always cause to pull in & have more dirt/ dust accumulation inside. and are more prone to fan failure..
@@analoghardwaretops3976 The surface area makes sense as you explain it. I'd get the dust argument too, but of course a dynamic speed fan would prevent that and be the best of both worlds. That said, I acknowledge your point and you are probably right.
@@hgbugalou yes there are many pieces of equipment that have 2-3 auto fan speed settings , (whilst some may auto variable) that are selected depending on temperature..and may also be normally off , but get turned on once a temp. threshold is crossed.
surprised to see a lampy explaining audio so well
jkjkjk
Some people call these "class D amplifiers" (or even "analog switching amplifiers") and leave the digital amplifier term to "power DAC" type amplifiers that take a digital input signal and only after the power switching stage use the L/C filter for analog conversion.
They're correct, as this is not a digital amplifier.
The locking system for power mosfet is "inspired" by QSC power module. Also the ics that are probably an irs20957 clone is used in QSC active speaker power amplifier. The TIP and the rail of resistor generate a +12V but referred to the V- rail. Usually when the rail voltage is high (and low considering it's dual) it is derived from an isolated secondary winding. Very strange the relay because the IRS20957 have a shutdown pin witch can be used to delay the turn on. Not a good choice the ribbon cable and the IDC connector to delivery power.
The very first comment that says only true and real info.
Well said... that was exactly my first thought... similar to an original K series Bucket of Power amp design, but not as refined. Since I live in a hot, humid country where plastics break down quickly and as far as I know QSC does not sell those plastic things (tensioners?), this "rigged" version idea may come in handy some day!!!
@@radioflyer2030 well if I remember correctly in the QSC version the plastic post push on the body of the MOSFET and so it doesn't align with the tab hole.
These class Ds are taking over the world esp with GaN getting cheap
Those are not "super duper large" filter caps. I have an old (70:s or 80:s) Peavey powered mixer with rated output of just 2x100W and the filter caps are about the size of Red Bull cans, just a little shorter. Also, the power transformer looks like something out of a welder. I'm pretty sure that one actually outputs the rated power all day if necessary unlike most modern (at least the cheaper ones) amplifiers. But of course the Peavey also is the size of a small suitcase and weights something like 20 kilograms but that does not matter as it's just used at the training room and never transported anywhere. Works great despite it's age, just some of the pots are a little scratchy.
omg this amp is crazeh.
i personally dont like the vacuum cleaner/dust collector style construction every dog and its breakfast is into these days, but the heatsinking just cannot be serious
It's so weird seeing inside an amplifier but not seeing a couple of huge masses of iron transformers...just still used to the old school style.
I don't like how the ventilation fan has half its blade width covered by the case...the vent holes should have been much wider to allow proper airflow.
And given the long slot vents on the front panel ~ that fan looks small for the “ventilation architecture” and the heat-sink is dumping to the bottom case plate - I can imagine that easily overheating [particularly if stacked with its companions], would the long vents on the front be a bit deceiving with it cooling capacity? 😠
The problem with clipping is that it generates a lot of harmonics that burns the high frequency (tweeter) speaker.
The problem with clipping is the waveform becomes distorted in the lower frequency (by the time you clip hf, the lf will have distorted long ago) but since a low frequency driver can't physically move as the amp is fighting its qms, it provides a lot of mechanical (and in turn electrical) resistance - that's what kills amplifiers and speakers.
@@cozmiumNAD went above and beyond to prevent that. Used some clever technologies.
Uh, no.
Clipping doesn’t damage amplifiers. It’s a saturated output. The output signal has gone as close to the power rails as the output transistors can drive it, and so it can no longer follow the incoming signal’s waveform. That’s all.
This doesn’t even hurt speakers, inherently. They can’t tell the difference between a clipping amplifier and intentional (recorded) distortion - partly because any naturally recorded distortion IS from a clipping amplifier.
The only problem is that the average power of the output signal increases relative to its peak power - because the waveform is getting closer to a square wave, on account of the saturation at the peaks. This can drive a speaker into thermal overload, but notably, in exactly the same way as it would if you had turned the volume up to the same average power.
Granted, the squaring of the waveform generates harmonics in increasing frequency, which will be sent by the speaker’s crossover to the tweeter. So the tweeter will be stressed to failure before any other driver. But it’s still just a case of too much power creating too much heat, and hitting one of the two natural limits of a driver. (The other being excursion.)
There’s nothing inherently damaging about distortion. Not to the amp, not to the drivers, not to your soul. 🤘
@@nickwallette6201the nice thing about active crossovers is that when subwoofer amp goes clipping it can be easily herard and dosent damage the driver. Very usefull when someone turns up the base for mysterious reasons.
@@nickwallette6201 You're not inherently wrong with what you have said, but you're basing it on the sole premise that a given speaker can handle significantly more power than the amplifier - which is almost never the case.
When you build any sort of audio system you always overspec the amp, because it's easier to do and leave a lot of headroom for speaker burn in, future upgrades, and to put it blunty: cheaper amps not being able to output full power at
letter D used to designate this amplifier class is simply the next letter after C and, although occasionally used as such, does not stand for digital. Class-D and class-E amplifiers are sometimes mistakenly described as "digital" because the output waveform superficially resembles a pulse-train of digital symbols, but a class-D amplifier merely converts an input waveform into a continuously pulse-width modulated analog signal. (A digital waveform would be pulse-code modulated.)
Awesome Video Big Clive
@11:20, I wouldn’t say the fan is stirring the air around the power supply. That isn’t how fans that pull or push air into/out of an enclosure work. By moving air in, some air has to go out somewhere, and generally speaking that should take heat out of the box, assuming the outside air is cooler!
The timing of this video :D
Im currently in the process of repairing a QSC Pro Audio Amplifier.. With a blown Switch Mode Power Supply.... Picked it up off ebay for cheap.... Will be worth a bit of money when repaired, Being QSC... But im planning on adding it to my audio rack, to run it bridged for one of my woofers..
4:54 it that a MOV or a NTC thermistor for inrush reduction? IME the latter are usually matt black like this one, MOVs being more shiny
I'm currently servicing a Behringer amplifier that's like this inside. All tiny tiny surface mount except for the outputs. Big beefy mosfets strapped to the chassis. I'm lucky it's just a set of shorted output fets, that's easy, trying to repair it further would require one hell of a microscope we definitely don't have.
Hopefully not an NX6000. I tried reparing one of those with an blown output stage. I was lucky that i put the cover on before turning it on because one set of mosfets exploded very violently. Turns out there was more damage than just the output mosfets. I eventualy sold the power supply module because the amp wasent worth the effort to further diagnose. Then again I am now working on some Labgruppen FP10000Q and an FP14000 with an scary +-160V and +-200V supply with more capacitors thsn some 1000W AB amps
@@lukahierl9857LabGr. they're more likely 10000W & 14000W
and not 1/10 of this as you mentioned.....exercise caution when live testing. protective eye gear , ear muffs etc.
@@lukahierl9857 Oh dear, that's exactly the one. With that in mind I'll slowly wind it up on the variac when the parts come in.
@@LunaphaseLasersOfficialat least it dosent have power factor correction to make the variac useless
@@lukahierl9857 Final update, it did not explode. It's acquired a second life it seems.
When i design audio gear from scratch, i treat the earth connection like a mains neutral... think about it.
(Most grid power installs tie the N and E together near the consumer unit)
If you assume it is a neutral, and there are potentially leakage currents (and induced volts) on the earth line, there is a higher risk of mains hum.
You have to assume that earth isn't really an earth for audio.
Differential signal grounding is the solution. But few people bother!
Nice amp.
The problem with clipping is that it is effectively an increase in average power, which is the only thing that will destroy a voice coil. It's not to do with waveform shape or resistivity in the load, it's simply an increase in average power. It's still an AC signal even if you clip the waveform to a proper square wave. It's also interesting to note that the reason why manufacturers get away with making claims that are double the RMS power of their amplifiers because if you clip a sine wave to a full square wave, you've doubled the power in that waveform.
It's not real power or usable power but it has the heat load none the less.
You should do an RMS test on these amplifier
“Clip” just means that a signal is exceeding a limit and the peak and/or trough of the signal is being cut off or “clipped”. It can be bad, it can be desirable. The guitar distortion effect would be impossible without clipping.
It's an exhaust fan... not intake.
The air comes in the front, flows over the board, and goes out the back via the 30mm fan.
Not this model but just replaced one like this driving studio monitors for crackling in the speakers and a channel dropping out... reflowing/adding solder to the output jacks seems to have fixed it. Maybe when I redo that studio I will open the new one and see if it's a thing in small rack mount amplifiers. You would think considering what some of these go through on the road they would reinforce that part. I also wish more studio gear came with the ground lift switch. While I haven't tested it I suspect the output rating on many of these units is a bit... generous.
Wow...some of those circuit boards look as though they are Sim City maps 😅 Soooo many components clustered so close together!!
I recently bought an old stereo amp with the intention of having a look inside...now that I see this, I know I'm in over my head. I'll still take it apart, but only to satisfy the scientist in me. Once I've had my fun it's going back together and being repurposed for my experiments.
The older style equipment tends to have a lot more dedicated components rather than the modern ones being full of chips of some sort. Depending on the era it's from, it's even more beautifully designed and looks even more like cities, especially before all the SMD stuff became a thing.
@@Dutch3DMaster I know, right? I absolutely love looking at well designed PCBs, especially the clean and shiny ones 😅
Nice, I love seeing what dj equipment would suck if I got it
Great video! The amp looks quite simple but it probably cost an arm and a leg. ;)
£125 according to the link in the description, so only an arm 💪 😉
@@JamieWhitehorn oh haha, sorry i missed that link. :O thanks!
super cheap .. just over 100.00
The mystery chips are possibly a clone of the obsolete International Rectifier /Infineon 98-1073pbf which is listed online for sale as 'ic audio driver digital'. No data sheet seems to be available but it's probably very similar to the common IRS2092.
The power supply looks like it is a unregulated LLC design. The large caps close to the transformer are in series with the transformer primary so that is resonates at the drive frequency. Current in the primary is sinusoidal and switching losses and generated noise is minimized.
IR2156 is not even close to LLC, LLC is much different topology, yes it is in terms of interference noise, LLC is the best topology for powering an audio amplifier, but all these converters operate at a fairly high frequency, and when they stay in their modes, they usually do not have audible interference noise.
The IR2156 is actually a ballast driver for a fluorescent lamp, but it can be ideally used as a half-bridge switching power supply controller. If it is done correctly, they are also quite reliable, they work even at frequencies 100kHz+ (quite extreme and must be a well-developed solution), it is possible to take 5-6kW of power from this kind of half bridge PSU to the powered device.
Of course, this specific one in video does not correspond to these numbers (and don't have to acorrespond it, because amplifier is also much more modest)...
LLC's standard example is ST microelectronics L6599, but it would make power supply much more detailed (dense) visually - the LLC power supply requires a very careful design and its (in L6599 case) power capacity is also quite limited (you cannot create a power supply unit of the same power as aforementioned numbers on basis, as is possible with these simple ballast drivers).
With ballast drivers, it is also possible to use several protection functions - overcurrent, overvoltages, etc. that are present in full-blooded PWM drivers, but this has to be done externally, often this external design can also be better than on-chip solutions.
I'm not afraid of them (ballast ICs), they've proven themselves quite well, and with add-ons, they can be made completely equivalent to special PSU drivers.
And finally - an unregulated power supply is the best possible solution for an audio amplifier in 99% of cases.
To understand this, dynamics of the audio signal over time must be studied...
@@inarinukka7729 I actually have a fixed frequency L6599 controlled power supply in front of me right now, powering my computer speakers amp. It's not very complicated. True careful design is required but complexity isn't necessary if regulation isn't required, just operating at the gain/load invariant point on the curve is OK. No reason this can' t be done with a IR2156. Just a hunch that the PS is LLC, might not be. Would depend on how much leakage inductance the transformer has.
@@stephendredge6077 but IR2156 is much-much-much simpler and reliable if properly constructed. Also, the IR2156 has a significantly simpler transformer design and you can use an IR2156+buffer+powerful MOS or IGBT+ colossal E55, E65 etc. transformer and rectify the output and take tons of power from it while keeping it simple and maintainable, with light weight. You dont need to calculate, wind it, to measure and recalculate and readjust - you simply just can calculate what power you need, what fq you have and you can make it once and it works at start, no gapped core and you get much higher power than same fq/volume LLC. Of course, manufacturers of all kinds of junk make things so that they work for a few years in a certain range set by the manufacturer, but if the magnetic materials get old, for example, then it's a kaboom and no one repairs it anymore, and you have to buy a new one again.
Something new for me intresting amp compared to the linear amps i still repair today fantastically explained as normal Clive you are brilliant certainly helped me over the years and using ur pinky with soldering lol epic
Few years back I needed a stage PA amp, so I bought some Class AB modules. Then ignored them, and bought some Class T modules instead. (Class T is like a clever sort of D)
Because Class T and D are so much less fussy about their PSU, it was cheaper to re-buy the whole modules than it was to build a PSU for the AB I already had. The Class Ts were happy with a cheap switchmode and sounded just fine. A PSU for an AB would have needed an expensive toroidal transformer and all sorts of high-end bits.
I just tried going to the link in the description, mostly as I'm curious what this unit costs and received an "Access Denied" error. On the Pulse web site and others, it is listed for 157 Pounds.
I know the description says you didn't test it for audio performance. As nice as it was to see the inside, I would sure like to know what it sounds like and does it meet the specs.
Many decades ago I had an integrated pre-amp and power amp with 22 tubes and a massive transformer all to make 120 watts. In the winter you could heat the room with it. Then we went to solid state which was somewhat smaller, lighter, and less hear. But they still had the massive transformer and heat sinks for the output transistors. Now this amp is ridiculously small and light and almost nothing for heat sinking.
Clive, it’s not digital. They use pulse width not pulse code modulation. A pulses width is infinitely variable, so analogue.
I saw the shorts and Here
Loved it
the tpa3255 is a tiny class-d chip that requires a few components to work, its rated to 480W at 1% thd in bridged mono into 2 ohms
the only other chip that comes into my mind with such power levels is a TAS5630 its older and a bit bigger.
Thanks a lot for the reverse engineering. It was quite interesting.
Thoughts on the unidentified chips: The comparators should be used in conjunction with a triangle or sawtooth wave generator which converts a voltage level into a PWM duty cycle. Are they connected in such a way that you are able to infer that kind of relationship?
Wouldn’t you expect the triangle/sawtooth wave generation and initial comparator circuit be on the input board? I would think the ribbon cable that goes between the input board and the amplifier board probably carries PWM signals and not linear analog signals.
@@timpontiusNot really. And it’s fairly standard practice to make a PWM sawtooth generator using a comparator too, but that implies that this isn’t switching at high frequency for efficiency. My guess too is the UICs are mosfet gate drivers.
@@timpontius I wouldn't want to push the signal that is relatively high frequency compared to everything else in the box over a ribbon cable like that
Dedicated class D amplifier chip, which has all the parts needed to implement the circuit, from input analogue, to output drive with high side driver for the mosfets. Comparators are there to drive the front panel LED's, and to also enable the amplifier chips when they detect enough input signal, and also are there for DC offset protection as well, turning off the output relay if there is a DC offset at the speaker outputs, and providing the turn on anti thump delay as well. quad comparator, so 8 in total. 2 drive peak indicator for each channel, 2 do the signal detection, and thus amplifier enable, with a few second off delay for quiet input signals. Then the other 4 do DC offset protection on the output, and one does the power on delay, leaving one that is probably sitting being used as inverter to drive the mute control of those PWM driver chips, or is sensing heatsink temperature, to enable the fan to run.
@@tommihommi1 A ribbon cable can easily handle the frequency of digital audio signals. And it would be much less susceptible to noise. I don’t know if the ribbon cable here is carrying digital or analog, but if *I* were designing this thing I would convert to digital on the first board. And if I *HAD* to send the analog signal between boards I would probably use a shielded cable and not a ribbon cable.