I just increased the µF on the filter capacitors of my stereo amplifier; it's a Nikko TRM-650. The unit initially had 2x6800µF in a single can, which i already replaced with two single capacitors of the same value from epcos, thinking they'd be "fresher" and would have a lower ESR. Your video here encouraged me to go a step higher on the capacity ladder, or two steps. I checked that the circuit would be able to handle the increased inrush current and after i found 4 thermistors at the right spots, i was fairly confident the could handle... 2x10000µF. Yeah, that might even have been 3 steps higher *lol* They're Vishay capacitors btw.. Fearing that I went too far, i turned the amplifier on, expecting the worst, but i was surprised how much the amplifier got held back by the 2x6800µF capacitors. Sound had more "authority" in general, sounded more rounded and real and high dynamic songs with lots of bass pretty much slapped you dizzy *lol* In my case increasing the value of the filter capacitors was totally worth it. Thank You for this video ^_^
Caps seem to have minimal effect on frequency in the sense that for this topic, they are acting as reserves of reactive power as we all know when the driver cones move forward they have to move back, if this occurs with a high signal ie like a loud part in the bass and the cone is in transit one way returning back... Then a new impulse comes in, driving it forward again. The inertia of the driver must be overcome. From the supply. If the supply has no head on it, the only reactive power will be in the transformer and the supply smoothing caps. Increasing the filter caps acts as a ballast or reserve. The true point of those caps is to filter out AC transients I think they can burn your amp up, or maybe the driver? Idk but since there's a relationship with frequency and capacity, larger capacities relate to longer waves lower f. But can result in a muddy dark tone, maybe at first there's a good extra kick to bass, but you might find that it can take away from the high end. I have thought of and tried somewhat, to parallel different caps so that they can hold more range. Hmm I hope that makes sense I'm not an engineer but getting some idea of how it works. More than anything good sound is about never having insufficient DC supply Voltage for your amp. When a peak of current demand goes thru, the supply needs to have extra volts left over because there's going to be more sound soon...if supply goes low you get inaudible clipping and stuff. This is a transistor amp right?
You can put a battery in parallel to your supply, if u know how. Then it can help the supply have tons of extra DC volts. It just depends on how many supply volts it runs Because I wouldn't use more than one battery. But an 18 or 20v tool battery has lots and lots of juice for an amp. ;) I would fuse it tho.
Thank you for explaining everything in a language that a hobbyist can understand! I do not have a Electronics engineering degree, I am a hobbyist. I love Electronics. You have explained everything perfectly! You have a new subscriber. I can not wait to check out your other videos as well!
That really was the *very first* time that I have seen the advantage/problem with power supply storage caps. Thank you very much, Tony! That was really an eye-opener and also very clear and precise explaination what was going on. BIG thumbs up!
I did this to my VTA ST70 amplifier and it provided much better bass impact and definition. In my opinion with the Amp I tried, I heard excellent results.
Old video, I know, but from the repair industry, at about 13 minutes in, a few things that are incorrect are mentioned. In regards to increasing the size of the main unregulated power filter, the increase of inrush current is an issue, but rarely for the transformer. Some transformers do have a thermal fuse buried in the winding's, and even though the inrush is too short of duration to make the transformer hot, it is often enough to damage the thermal fuse due to self heating internally. This is rare. Symptom is an open primary with no signs of overheating and burning. The big issue with many larger amps for consumer use is the pitting and burning of the power switch. It was quite common for vintage amplifiers to have a switch failure.
Great subject. Thank you for those tests. I'm one of those audiophiles with a cap passion. I bought a 24W class A power amp chi-fi that had 30kuF per channel of Elna caps. I didn't think they were genuine so I swapped them out with Supertech 2Ts. Clear improvement. Then I added an extra 10k per side so it's now 40k per side, I'm getting a more relaxed open bigger and more dynamic soundstage. So for me, not only add more but using better quality caps is the way. I'm sure Mundorfs will sound even better. More power to your channel.
Just as you were Tony, I was also surprised that the added capacitance in the power supply section made as much of a difference as it obviously did. Just goes to show, trying things in the "real world" versus relying solely on equations on paper, can have advantages also! Great job with the demonstration sir and I wish you and everyone watching (in the US) a Happy Memorial Day! God bless you all.
Nice experiment and very good idea to do a transient 5 cycle waveform test. It's not about getting more power from the amplifier but getting more reserve current in the capacitor banks to allow the output transistors to actually deliver what they need to when the signal reaches the crests of the waveform. And the results for sine wave testing, even 5 cycle bursts, is not a perfect reproduction of reality. In reality, the extra capacitors provide a much cleaner sound. Now regarding the endurance of bridge rectifier and transformer, it indeed can be a limiting factor. I don't condone what I did (40 years ago) but I did just that, and to a much higher extent. I had a Luxman L-3 lowly amplifier that was rated for 33WRMS. It came with two 3.300uF capacitors. I added 20.000uF per polarity - yes there was enough space inside - and the sound difference was huge. It really didn't deliver more power but what power was delivered was cleaner and the sound much sweeter even in average levels. I realized how much better the sound was when I had two output transistors blown after a high-school party. I had to get it serviced and of course I removed the extra caps before sending the unit in. When I received it fixed, I tried it out and it sounded horrible. Nothing like what I remembered. I quickly put the extra caps inside and the nice sound came back. The sound difference in the bass notes and high frequency clarity was enormous. I can still guess which amplifier has inadequate capacitors just by listening to it. It takes just a few minutes to identify that and I have not failed yet. My main amplifiers at the moment are a few Technics SE-9060 with 2x18.000uF per channel. They sound much more powerful than they are. And I need to repeat that it is very important to not strain the rectifier AND the transformer. In many amplifiers, they have cheaped out and either the transformer or the rectifier or both will not tolerate the inrush current.
First.. Thank you Tony for the inspiration. I've now repaired two integrated amps (Fisher CA-120 & Technics SU-8044) and one receiver (Technics SA-727). I'm far from any kind of an expert but I learn more from every video. The latest amp (Fisher CA-120) has two 6800uF 50V main caps. It's only pushing 30-35 watts. The 1000uF in that Kenwood seems mighty low in comparison. This amp only has 2 transistors and one amplifier IC (STK-465). That's it. I've never seen a unit with a transistor count that low. GOD Bless!
Take it from an electrical engineer and audio enthusiast, don't look at the time domain to try to determine if there's distortion. You will not be able to see it unless it's really really bad distortion. Look at the frequency domain on your Oscope - the FFT. Then you will really be able to tell how much distortion and harmonics pop up. You will not be able to tell from the time domain with the naked eye
Your videos are fine: As business owner who shares his limited time allowing us to look over your shoulder you do an excellent job, reading understanding schematics which is underlined by the ability to reverse engineer equipment. Troubleshooting and use of test equipment all come through to us your audience. Thanks for sharing and keep safe
I dont mean to be so offtopic but does anybody know a way to log back into an Instagram account..? I somehow lost the login password. I appreciate any tricks you can offer me
@Bodie Zeke Thanks for your reply. I found the site on google and I'm in the hacking process now. Seems to take quite some time so I will reply here later with my results.
Howdy. Very nice presentation. One comment though. Increasing the filter capacitance changes the transformer curret. The conduct time will decrease but the conduct current will increase. The current waveform will be more spiky. This in turn will increase the current RMS, althoug the DC consumption is the same. The current RMS times the mains voltage is the transformer VA power. It is possible, though unlikely, that the transformer VA power exceeds the rated VA value. I suggest being aware of this when increasing the capacitance. Regards.
The limiting factor on increases in power supply filter capacitance is sometimes the power switch on the front of the amp. Or the circuit breaker. Be careful out there!
Great video Tony. If you get the chance to test it here is something that I saw in an amplifier, the PSU capacitor bank was made out of 10 small capacitors 1500uF per branch 20 in total and the reasoning behind it was the time to recharge required for a small capacitor is a lot faster than a big one. This method seems to be a lot better than using big caps . Of course, things are much more complicated than that and it should be a very interesting test if you could do that.
Excellent analysis on power supply capacitance. Though the experiment contradicted your initial opinion, unless updates to supply is done, I would stay with about 20% max increase when replacing caps
i guess huge surge currents wont hurt transformers at all, they have a large thermal mass and its internal resistance limits the current anyways. it wont break if you dead short it for a second or two. rather the mains fuse or the rectifier might blow up
i accidentally shorted a 1 amp rated transformer for 30 seconds or some more , later i checked it gives 5.6 amps when shorted . and nothing happened to the transformer or its bridge . ppl just parrot what they heard and have no real idea about , to be able to say something negative .
I'm surprised at the 1000uf cap in the original. With the caps being expensive back then, I get it but wow. On a scope there must have been a nice amount of ac riding the rails. My guess would be 4700uf as a nice fit without cooking the diodes, or add a bridge. These videos are great!
In my experience, you would usually have to get pretty crazy with oversizing those capacitors to do any damage. Often, I oversize (and/or over voltage rate) them just so I can get capacitors that are physically large enough to fit the clamps. All other things being equal, new capacitors tend to be much smaller that what they are intended to replace.
Another alternative for improving the power supply is to use high powered zener diodes with sharp kneepoint, low noise and low dynamic impedance, They are available in normal or reverse polarity. Polarity is indicated by the position of the Zener symbol on the part. They are available in 46 voltages ranging from 6.8 V to 200 V. Devices' operating temperature range is -65 °C to +175 °C. Additional specifications include low forward voltage (VF max = 1.5 V) and 50 watts power derating (0.5 W/°C
This could be compensation for weak de-coupling caps in other parts of the circuit. Sometimes, changing a bad 22uf next to a transistor can make all the difference. Always check the other caps on that rail before deciding that the the power supply lacks main filter capacity
An interesting followup experiment would be to figure out why it improved the performance - was it the increase in capacitance, or was it the lowering of the effective ESR of the filter? As you pointed out, paralleling the caps lowers ESR. If you put in 4 caps total, each half the value of the previously installed pair of caps and repeated the experiment, we might learn something interesting. Perhaps ESR is the critical performance characteristic here more than capacitance.
A few years ago I picked up a Hafler d220 power amp for cheap money (known issues) and decided to replace the driver boards with better ones. The vFET outputs on these amps are pretty bullet proof so i was pretty confident they were ok. While I was at it I decided the original 25 year old 10,000uf caps should be replaced as well. I found I could get 22, 000uf 100v low ESR caps from Mouser for a good price so I went with them. Because of the inrush current I added inrush limiting to keep he beefy transformer happy, I also beefed up the wire sizes in the + / - 60v power supply using #14 wire. This amp wasn't working when I got it but now it sounds great and my inrush current limiting protects boththe transformer and ckt breaker from undue stress. The bias on this amp is set high so it also keeps that corner of the room a bit warmer.
With modern hi-fi amps bragging about total capacitance isn't unheard of, it absolutely makes a difference. Like many novice tweakers I'm currently in the process of running around doubling the capacitance of everything I have, haha. It's a little more complicated figuring out what's what inside my modern Yamaha receiver though. With mass produced products like this and many of the cheap Chinese amps I have, even though the components are relatively cheap this is still very much an area where corners are cut for cost.
Years ago I upgraded a Hafler d220 amp by replacing the driver board and decided to replace the main filter caps (10,000uf originally) because they were 20 years old and i wanted to bulk them up as well. This amp uses very fast FETs and was remarkable for it's time. I ended up using 22,000uf caps and i installed a 35a bridge as well. This stiffened up the supply and it increases the peak power capability, remember music is dynamic - we don't listen to test tones. When i did this i also added a couple of varistors in series with the power transformer so I would not fry diodes or the power switch. A class AB amp draws quite a bit of power so these varistors heat up pretty well and while they do limit the amps output they don't really interfere with the music peak power this can deliver. i also added a couple of high frequency 560uf caps along with a couple of 1uf film caps to improve the ESR at higher frequencies.
Adding parallel capacitors in this case lowers the effective ESR which can have beneficial effects such as the reduction of the oscillation originally noted in your analysis. Nice job!
The ages before the switching power supplies usage the term ESR was totally unknown because has to do with high frequency ripple currents flowing inside the circuit itself.In line transformer base power supplies there is no ESR because frequency is 50/60HZ.
It's true that most home audio equipment can be optimised. In this particular case adding power capacitors did make a difference. Adding a soft start circuit would be excellent . So let the 4700 MFD where it is . Thank you for your excellent demo and lecture. I am floored 😀 and subscribe d
Stumbled upon this video and found it worthy. I would be interested in also seeing the effects of added capacitance with a voltage doubler circuit. Anyways, i wonder what the waveforms looked like with the stock capacitance of only the 1000uF. You had already had installed a 200% increase over the stock designed value! I think much of that early to mid 70's consumer transistorized integrated stereo systems all pretty much had that same cookie cutter fizzy audio with flat under powered farty bass, scooped mids and ice pick upper tones that probably generated some sorts of ultra sonics too. As a junior high school kid I saved up for a dept store consumer rack component stereo (w/8track too!). Through out Jr High and High school I tried all kinds of speakers. Even the highly efficient EV drivers did little to present any greater sound. I was incredulous to observe that my 80 watt PEP rated stereo amplifier destroy my girlfriends dad's 250 watt RMS rated Klipsch loudspeakers and later realize that it also contributed to the death of my JBL PA cabinets for my band. For a long time I was perplexed as to such a great difference in perceived audio power from a 100 watt solid state stereo amplifier and say the simple single ended 15 watt tube driven Hi Fi amplifier. A pair of old fashioned dumpster dived 6v6 SE mono blocks would overpower any 100 watt and greater S/S stereo I could go against, even with 200% impedance mismatch of 8 ohms. Oh good times being a kid back then! Man I love that scope you have there! That is a nice piece of kit. Tektronix no less! I just gave away an old boat anchor digital storage scope! I want something like what you have but I have pretty much resolved to myself that I likely cant have a tektronix and will have to do with a rigol or hantek scope instead. That digital phosphor color screen is the cats meow! Oh yeah one last comment, couldnt a simple 555 timer and a simple transistor output to a relay make for a very simple kludged together soft start?
I think your video is a good demonstration that more capacitance in the power supply does make a slight difference, as you said, with diminishing returns. I have been experiencing the same thing lately in my vacuum tube amplifiers. My particular goal was reducing the spurs accompanying each of the harmonics in a spectral display watching it with a spectrum analyzer . I don't think the original designs are wrong or lacking. They are adequate for what the amplifier is supposed to do in its day and price category. If you stretch out your sine wave or use delayed sweep in your scope you will see that the oscillation (wiggle) on the bottom of your sine wave is either 60 Hz line frequency or 120 Hz full wave power supply ripple frequency and more capacitance helped reduce that and it you were looking at it with an audio spectrum analyzer you would see the 60 and 120 Hz spurs reduce quite a bit all along the spectrum with the extra PS capacitance. Also, you demonstrated the type of distortion typical of SS amplifiers vs vacuum tube amplifies. I thought that was really nice. A vacuum tube amplifier will distort different from the SS amplifier and not have the sharp transitions shown in your video. Our ears do not like sharp transitions in the waveform. Excellent demonstration, in my opinion. There is a little more stress on the rectifiers to keep the huge capacitors charged but it probably isn't hurting anything but it probably isn't helping much either. Thanks for posting.
It's an honor to hear from you! Love your channel! I wish I knew half as much as you do about this stuff. I agree about the oscillation being from line noise. Another viewer also pointed this out. Always something new to learn. Thanks!
THAT was something I've always suspected, from my multiple experiences adding capacitance to an amp's (or mixing board's) power supply. It always seemed to tighten up the bass. I've never gone nuts and overdone it, as I knew there was a limit. But boy can it make a difference. Now listen to the amp before and after and you'll hear it. I never suspected the *magnitude* of the difference - at 30Hz, you gained 50% again more power while lowering distortion! Now that's value, dawg. Thanks so much for doing this test- the *right* test. "Dynamic" test is a good name, but in audio a better name might be the "low-frequency transient" test. Fascinating to see what I've actually heard: taking little steps toward making a PS more like a battery really does help. Just don't overdo it. And by the way, you're not alone: my tech was surprised, too. When I asked him to add more capacitance to the power supply on a mixing console, he said it probably wouldn't make much difference. I told him I'd tried it on an older board, and it sure seemed to help tighten up the board's bass response. Well, he said, let's try it and see. Same results as you saw: sure enough, the original PS simply had been designed down to a price, and there was indeed plenty of room to work with in adding capacitance. Bass response and distortion improved pretty dramatically. I realize a console ain't a power amp, but mix consoles do have high power requirements, and they require very clean and responsive power, much like power amps.
The Marantz 1060 Integrated Amp is one of those classic amplifiers that people add larger filter caps for precisely this reason. The problem is this particular amp has no speaker relay so you get a bigger thump when the power is turned on, which can damage the speakers.
When selecting a replacement electrolytic cap you must know its original specifications such that ypu know from where you are starting. From there you can use a capacitor with higher ripple current rating and a lower ESR. This will deliver more current faster even with the same value cap.
I've always wondered about this. I always wondered if the rest of the power supply had to be up to the task of handling that extra capacity. Thanks for the video
You can calculate the ripple voltage at various capacitors fairly easy, q=C*dV and q=I*dt, so dV=(I*dt)/(C), Where C is capacitance in Farads, dV is ripple voltage in Volts and I is the maximum current pulled by the power amp. at 50Hz dt=100ms (full wave). As P=I^2 * R, I=sqrt(P/Zspeaker) (where Z speaker is 4R or 8R) Plug those values in and you can get an indication of average ripple.
I was recently working on a 1970s Rhodes Suitcase Piano with two 50W output channels. It uses +/-30V rails and has 11,000uF on each rail. But it also has a very large power transformer and a chassis-mounted high current rectifier bridge to handle the inrush current. With the Kenwood receiver, the designers probably calculated that amount of capacitance they installed would be adequate for how a typical user would run the amp. It also looks like there may have been size issues. Rhodes had plenty of room inside a large speaker enclosure.
I think what we see at 6:21 is not distortion, it is clipping, just your scope triks you because it is in AC mode, on DC mode you can clearly see the clipping. I noticed this phenomenon first when I tried to measure really low frequencies, my 100MHz scope just couldn't do it, but my 2MHz handheld scope did it well. Yesterday when I wanted to see how ugly the sine wave in the mains I noticed this again, the 100MHz scope measure higher voltage than a multimeter and I can see a distorted sinewave in AC mode, but when I put it in DC mode everything falls in place, it cuts the sine cap off and it measure the same voltage as the multimeter. The mains voltage is so bad here that the toroid transformer sometimes start to buzz loudly in my amplifier, the voltage is usually 230 instead of 240. Seeing this I'm considering to use switching power supply for amplifiers, over 4-5W load their switching freq goes over hearing freq so it shouldn't be a problem. I love my little DSO 068 scope, it is soo convinient to use and I can avoid my biggest fear of blowing up my scope trough ground loop.
thinking of changing the big filter capacitors on my '76 pioneer SX 750. Would I need bigger spec replacements in that Vintage?Can you even find exact matches for the originals? I guess going a little bigger is OK if the original spec arent available.
Be advised that if the amp has a tube rectifier, there are rules about maximum capacitance.... usually 60uF or less. If it's a SS rectifier, all bets are off..... you can go as big as you want, just know when an amp is turned on the inrush current to charge the larger caps could blow the mains fuse. An increase of lower frequencies and overall dynamics are the main benefits.
Seems to me that the designers of this amplifier were more comfortable with valve amp design and high SUPPLY rail voltages than transistors and their current demands. So they used a capacitance value more suited to a valve amp. Looks like it was designed in the "crossover era" (no pun intended).
I have an old Univox guitar/PA, amplifier, amplifier modules in each speaker cabinet, four speaker configuration in each cabinet. The filter capacitor is 47000 uf, it holds a charge alright! Two diode rectifier, simple but holds the extreme loudness of the amplifier, simple pre amp to transformer to a push pull output, a set of parallel darlington direct coupled output transistors. Electrolytic Capacitor output. Good non distorted output.
I picked up an already assembled kit amp which had issues turning on. And found that the amp had no inrush limiters and the entire power supply was really absurdly overkill for the power output of the amplifier. So in this case I reduced the capacitance size quite substantially. Perhaps smaller than what some designers would have done, but considering the amp only develops about 65 watts/side and has transformers rated at about 2400VA (!), and a really low volt swing, I only needed just enough to remove ripple. It is a kind of dual mono design so the transformers will be fast enough at that kind of output that performance wont matter at all. And the power on is not going to pop breakers or dim lights all over the house. And I say "kind of dual mono" because there are 2 transformers and sets of caps, but they are siamesed in parallel. So I could technically remove half of it and still be ok. And I may do that one day if I were to build another 2ch amp and want a football sized transformer for it.
Hi! This video is excellent, the way you have explained things is just excellent, I enjoy watching your video so much, keep up the good work... Have a great week..
There is nothing wrong with modifying and amplifier providing is done correctly with some good calculations if there is a small improvement is good. I'm not saying The Originals engineer didn't take this in consideration they already knew but it's all to do with cost at that time. I done the same thing many years ago and a JVC class A class B amplifier. I saw very similarity what you were showing on your silloscope and yes having the proper audio equipment then you can at least see it sometimes it's very difficult to hear it or you might not hear it at all all depends on how good your ears are and the type of music you're going to test it and you speakers. Excellent way of demonstrating and explaining. 👍👌
I remember back in the 80's walking by receivers at one of those warehouse sales and they had them playing and many of were showing power supply issues. Their display lights were flickering or dimming. I wondered if bigger caps would stop that. I have since gotten some very big computer grade caps to play with. But I don't remember if they worked.
Great video always wondered if this was worth doing. Tony any chance of ever doing a video showing how to fit Pre amp out or sub woofer out sockets. Lots of the older gear don't have these. I've had idea's where to take it from inside but I am sure your way would be best. Many Thanks.
I usually go up a size on filter caps. Lately I've been using film bypass caps as well. Most extreme change I've done is on a pioneer mj200 power amp. Went from 3400uf non bypass to 15,000uf with 0.1uf bypass. Its worth mentioning that the pcb had three rings printed on it with the standered caps being inside the smallest ring. The 15,000uf sits slightly outside the the largest ring. No problems to report yet 😆
@@johnsweda2999 sounds good. I use it for headphones and speakers. Only time I've noticed any difference with the bigger caps is with stuff like Bassotronics and DJ droppin or huge Japanese drums. Of more effect is the customised output filter. That really took the edge off things.
I listened with the volume off, and when you show the under side of the receiver, I was like "I saw that somewhere" I realised that I have almost the same receiver and worked on years ago. It's like my own trio kw-55. For mine, The biggest issue will likely be the rise of B+ and the subsequent voltage, in the case of a very lower rated transformer, it can also hurt the transformer. These high voltage transformers are rated in the 200-500mA and will likely burn in the blink of an eye if the in rush current is too high . For mine,The B+ rising too high will likely need some modifications because the valves will likely be ruined very quickly. If it's the receiver, I think it is. They are already pushed to their limits plates voltage with the correct capacitor in place.
Often the limiting factor is the power transformer and no amount of capacitance will fix that. I restored a Magnavox amp from a console in 1974 and it did 15 WPC into 8 ohms with the stock power transformer and I think it had a 1,000uF 63 volt can cap originally but I upgraded the rectifiers to higher current ones and used a 10,000uF cap in place of the 1,000uF cap. Just by replacing it with a transformer that is rated at 5 amps I was able to double the power to 30 WPC at 400 Hz with it being a little less at 20Hz. I noticed a weird distortion at 20Hz when I ran it to full output and I checked the power supply ripple and found it to be nearly 2Vrms. So I put in another 10,000uF cap and that completely took care of that issue and I now get full power at 20Hz and into four ohms I get around 49 watts. One indeed definitely MUST know exactly what they are doing and know the circuit well before even thinking of attempting mods like this and they must have the proper test equipment to verify they have not caused any issues and that the mod works as expected and actually does make the amp better. Physical capacitor size was definitely a determining factor back then. The 10,000uF 63 volt caps I used are about the same physical size as the stock 1,000uF can cap is. When one thinks about how large a cap of that value would have been and how expensive it would have been back then, they realize real quick why many amps use the bare minimum filter cap value they can get away with.
Not sure if it was covered but the effect of adding cap depends on the Class of the Amp. Class A would not see much change. It can however lower hum in some cases. It can be that the coupling caps are low. If you do add a filter cap make sure its WVDC significantly exceeds the supply voltage.
The improvement is almost certainly the result of lower impedance in power supply. Your transient response at 30hz has the caps getting recharged by the supply between each peak! A power supply transient response requires a faster transient. Even using higher frequency some multiple of mains would be better (full bridge@120 cycles).
Fitting extra capacitors does make a small difference but there can be issues with changing currents giving rise to buzzing , I built two identical amplifiers during lockdown one had two power supply’s mono block the second was single transformer with a higher secondary voltage than required , and I built a dual stabilised hi voltage regulator plus and minus 56v they are both very good but the one with stabilised supply has much better Bass and background noise/hum is absent 100W class AB but would be difficult to retrofit to an old amp and no supply line fuses as these are the weakest link when drawing large currents no ripple on supply line ! An external power stabilised supply could be built
I think what you called "oscillation" was power supply ripple which is only visible at clipping when the amplifier circuit power supply rejection doesn't work.
Is that right? I've often seen this but did always thought it was oscillation although not totally convinced. Your alternative explanation is interesting. Do you have time to expand on it here?
Outstanding, I have a NAD C352 amp, recapped it all as the originals were garbage. Just got the main filters to change, 22k uf at 63v. Really enjoyed your video! Subscribed.
Cool vid man! I like fixing (or at least TRYING to fix) audio equipment also, so this video was very interesting! Although you should've added a link to your other videos you mentioned in this one, especially the linear power supply one! (I'm gonna try to find that one in your overall list of videos, but a link to click on would make finding and watching that one SO much faster and easier!)
I found some Vishay 10000 µF 75V caps on sale at Newark. I'm re-capping two amps that have 8500 µF 75V caps, and another with 10000 µF 65V caps. All the original caps are 3.125" tall. The new caps are 4.125" tall. The tall caps will fit in both types of amp. The new caps have lower ESR and ~9A vs ~6A ripple current. The new caps are rated 3000 hours at 105°C. I want these 40+ year old amps to last another 40 years. I think I should add soft-start.
6:45 My assumption was that we were seeing a slowly moving phase difference between the FG and the 60Hz from the wall. The supply ripple will cause the power output to dip slightly when the peaks in the signal are at the low points in the ripple, right?
I have a Musical fidelity AS100, which uses just 6800uF per rail(35V dc) on each monoblock, which sounds pretty low to me (2.38 coulombs or 2.38A per second), so putting in some Naim 10,000uF additionally on each rail (4 in total), hopefully that will give it more powerful dynamics.
@@engjds That's crazy to hear, I would really want to know what they use. I just recapped my Sansui and it sounds fantastic compared to the original caps which were closing in on 50yo at this point. But I'm not sure if I could tell the difference between say Elna Silmics and Nichicon Fine Gold. I just try to fit from the same brand and line, and tend to choose what's made in Japan. I've had Elnas come from China...
@@maat8249 People swear Nichicon, but I have yet to find paper evidence, that they are actually better than the competition, they tend to hide their specs such as ESR unlike their competitors I think they might be riding on their reputation rather than golden eared reviewers.
If the capacitors are 2200uf and replace with 22000uf or more because you think the amplifier will sound better the fuse will blow due the high charging current.Or the internal fuse in the transformer will blow.In every case the circuit is designed for a certain maximum power with a maximum rated current and this does not change,but in the case the amplifier plays in low volume the original capacitors are more than enough because designed for the extrem conditions and passing the limits of the amp the caps will do nothing to protect from destroy.
I added an NTC to my mains input so that I can have ridiculously large reservoir capacitors for great dynamic ability and not pop the bridge or fuse. I don't want continuous high power.
I just beefed up my filter caps or power rail capacitors ;-) from 36,000 uF to 112,000 uF in a 100 Watt stereo amp. The new ones had the same physical size and voltage rating as the 40 year old ones. Seems to me a good practice to replace those old caps and have a bit of a bonus with the increased cap rating. Bridge is 600V and 50A, Xformer is 1000VA.
I think the oscillation on the power tests was probably not in fact oscillation but clipping ripple due to low valve smoothing caps that's why it vanished with higher value capacitors. I have seen this on many amplifiers in the past. I depends on how near to the rails the transistors go! Another way to get better voltage drop out is to use Schottky diodes. I never use any other diodes where power is relatively high. Most smallish transformer will deliver a fair current but will get quite hot if the drain is prolonged.
I changed the caps in a 35 year old Heathkit AA-1640 from the original two 7500uF to two 18,000uF. Seems to perform much better, especially on sub bass.
Of course. Thats simply because your 35 year old capactors would be degraded, and the ESR will have risen considerably. To get a valid comparison, you first need to compare old 7,500uF to new 7,500uF, and THEN compare the upgrade from new 7,500uF to new 18,000uF.
I have an amplifier that I had to recap and I replaced all the caps with exactly the same value and voltage, the only difference in the new caps is that they are 105 degree and not 85 degree so that's what I upgraded, only the main caps I also doubled with the same voltage value which was 63 volt but the old caps were 4700uF and the new caps are two 10000uF caps and also 105 degree. I haven't looked at the signals yet what it does now so I can't tell you if it's any better, to me it seems better because it sounds a lot better but that can as well be from replacing the other caps which were in series with the output to the speakers or those in the input part of the board, that's because at that point I didn't have an oscilloscope, now I have a new one and I can check the signals and it might be a good idea to do that and you also sparked my curiosity so I'm probably gonna do that soon.
That a great video. now if you put scope on power supply output you see a more stable voltage and less ripple . I like to use low ESR caps made in Japan for replacement. Even if you keep same value you see improvement. now in car audio 12V system with big bass amps a very bag cap in farads makes a improvement not just to the amp but the cars electrical system and less surges on alternator and battery. we call them super caps. even the newer class D amps improve. also super caps help with inverters that make AC from 12/24 or so to 120V ac. startup surges on appliance can be improved. this happens because of lower voltage higher amps. The voltage drop in wires becomes a factor. so placing large cap as close to load as possible in that case.
Be aware that you are reducing the conduction angle of the rectifiers, so in theory the transformer and rectifiers will have slightly greater losses. I have in the past gone to greater lengths with added capacitance, you might find it interesting to listen to the audio that gets superimposed on the reservoir capacitors. (Easy done using some dc blocking and headphones). Any high frequency content here points to high ESR rather than a lack of capacitance. I suspect the ringing you are seeing is related to ESR rather than the capacitance.
It's good for dynamics, kind of like what they do in car audio. I've increased cap size before in equipment, 9/10 no issue if done properly and it makes a difference in the bottom end of music. 30-50hz or lower is where you'll notice it. It can be a bridge/diode/fuse killer sometimes as people put too big a cap in and it's almost like a dead short before fully charged, just saying (for people out there who try it). Good video and you made me want to drink coffee! lol
I suspected this (Damage to the rectifier) would happen, especially on older equipment, have always stuck with the original size filter caps and had no problems, fresh caps really make an old amplifier roar like it should...
Interesting tests but I would like to see same tests at half power as that is how most people use their equipment, rarely do people run at full power. So would the caps make any difference testing at normal listening levels ?
I really should have mentioned this point. As far as you can see on test equipment, there is absolutely no evidence that anything changes, by adding a bigger capacitor. Some, however claim that you can hear a difference, even at low volumes. I think the bass may be a bit tighter sounding at mid volumes (because a transient can spike the power up quite high, even at mid volumes), but I absolutely can't hear any difference at lower listening levels. If you noticed, you could only see a change when the amp was close to its limits. Thanks for the comment!
Very good example I restore a lot of vintage receiver and amps and I up the filter caps by as much as 20% of the original Uf and the end result for me has been very good
I did not hear you mentioning anything about these effects of increased capacitor size. First, the increase causes higher peak currents (which you did mention), but what that higher peak can cause is more hum. Second, you send more distortion to the power cord, which can cause hum in other equipment attached to the same power branch. All that is also dependent on the transformer impedance, or let’s say its current limiting characteristics. Anyway, the overall consequences may be a little more wide spread than just the output power and sine wave distortion. And speaking of power, I have seen a statement that a normal home listening volume (with “normal” speakers) is less than 1 watt. In fact, that is also what my stereo indicator LEDs indicate. The strings have two ranges, 2 W full scale and 20 W full scale. I keep mine at the 2 W range and seldom see more than half of the string lit. Half way in the logarithmic scale represents about 1/2 W. Of course on a stereo, there are two channels, so I have two indicator LED strips, both below the half way, but of course being additive. And just a funny side note - you pointed out a change from germanium to silicon output transistors. I still have the germanium power on my first (and only) self designed and built 2 * 10 W stereo amplifier. I use it a few days per year, but it works fine. Never any capacitor changes or anything at all. Lucky me!
Nice video with useful information. All videos cannot contain everything, and there will be other opinions as usual. I do have one question though: how did you generate the quick 60hz burst of only 5 cycles? best regards
Hi thank you very much indeed for this very interesting test. I did a very trivial test ... i had a stereo amp originally with about 2x10mF total ... i replaced PS caps with 2x60mF huge caps by Mallory (refurbished ones) The solid state amp sounded more powerful and overall better to my ears ... when i switched it off the amp kept on playing for some 15-20 seconds running only on caps stored energy. Since then i believe that more capitance is a good thing. Maybe there is a point when adding more mF has no practical effect. I am looking for finding that point. Should i have stopped at 2X40mF maybe ? I can see in some high end amp huge caps banks ... scaring. And very expensive. thanks a lot again. Kind regards, gino
A little bit of ac ripple on a B+ supply is generally rejected within limits. By adding Capacitance narrows the Conduction angle. Can burn out supplies with big in rush current. I have come across lots of blown modules and often found the values of Caps resistors diodes could be substituted to increase reliability. Only way to find out is to try it. Thanks for taking the time to present.
This is very interesting. Thanks for the video. It seems like increasing power supply capacitance helps when the design under-specced capacitance. I'm curious if it would help in a more robust setup. Currently recapping a Yamaha CR-640 and didn't bother upsizing the PS capacitors because they were 10,000uf for a 45W amp. It might be interesting to run this again using a similar amp.
@xraytonyb, Thanks for putting this information out. It leads me to my question: I am recapping a Pioneer VSX D1S, it's a nice unit. 1990's...do you see any problems in increasing the main caps? Thanks for your help!!!! Cheers
@@chillidogkev Yeah, well I have amps that employ 2x 10.000 uf. So 4000 uf is quite small, but it really depends on the amps wattage delivery ratings and how big the power supply is. I mean more uf is always better, but the amps has to be cost effective too! Also upgrading with too large capacitors sizes can damage your diode bridge if it isn't rated for big filter capacitors, although this is mostly for concern in older equipment, but always look up your Full bridge rectifier specs for good measure!
Should be, but there wouldn't be much improvement, if any, to the performance of the amp. Going much larger in capacitance could cause issues with inrush current when you switch the amp on. This can damage the power switch over time, unless that amp has an inrush current limiting circuit (sometimes called a soft start circuit). That said, replacing the 15000 with a 22000 probably would work without any major issues in the SA9800.
I went up 20% on my sa 9500 and the heat sinks/transistors seemed to run alot cooler....I also had a failure in the power button on my unit and found a tower computer power switch that worked great!
Next step is the output capacitor. It should have at least 2000uF but I doubt it considering the small cap in the power supply. I think that is the reason why this amp delivers so much less power at 30Hz.
I just increased the µF on the filter capacitors of my stereo amplifier; it's a Nikko TRM-650. The unit initially had 2x6800µF in a single can, which i already replaced with two single capacitors of the same value from epcos, thinking they'd be "fresher" and would have a lower ESR. Your video here encouraged me to go a step higher on the capacity ladder, or two steps. I checked that the circuit would be able to handle the increased inrush current and after i found 4 thermistors at the right spots, i was fairly confident the could handle... 2x10000µF. Yeah, that might even have been 3 steps higher *lol* They're Vishay capacitors btw.. Fearing that I went too far, i turned the amplifier on, expecting the worst, but i was surprised how much the amplifier got held back by the 2x6800µF capacitors. Sound had more "authority" in general, sounded more rounded and real and high dynamic songs with lots of bass pretty much slapped you dizzy *lol* In my case increasing the value of the filter capacitors was totally worth it. Thank You for this video ^_^
aleluia ☺
Caps seem to have minimal effect on frequency in the sense that for this topic, they are acting as reserves of reactive power as we all know when the driver cones move forward they have to move back, if this occurs with a high signal ie like a loud part in the bass and the cone is in transit one way returning back... Then a new impulse comes in, driving it forward again. The inertia of the driver must be overcome. From the supply. If the supply has no head on it, the only reactive power will be in the transformer and the supply smoothing caps. Increasing the filter caps acts as a ballast or reserve. The true point of those caps is to filter out AC transients I think they can burn your amp up, or maybe the driver? Idk but since there's a relationship with frequency and capacity, larger capacities relate to longer waves lower f. But can result in a muddy dark tone, maybe at first there's a good extra kick to bass, but you might find that it can take away from the high end. I have thought of and tried somewhat, to parallel different caps so that they can hold more range. Hmm I hope that makes sense I'm not an engineer but getting some idea of how it works. More than anything good sound is about never having insufficient DC supply
Voltage for your amp. When a peak of current demand goes thru, the supply needs to have extra volts left over because there's going to be more sound soon...if supply goes low you get inaudible clipping and stuff. This is a transistor amp right?
You can put a battery in parallel to your supply, if u know how. Then it can help the supply have tons of extra DC volts. It just depends on how many supply volts it runs Because I wouldn't use more than one battery. But an 18 or 20v tool battery has lots and lots of juice for an amp. ;) I would fuse it tho.
Thank you for explaining everything in a language that a hobbyist can understand!
I do not have a Electronics engineering degree, I am a hobbyist. I love Electronics.
You have explained everything perfectly!
You have a new subscriber.
I can not wait to check out your other videos as well!
That really was the *very first* time that I have seen the advantage/problem with power supply storage caps. Thank you very much, Tony! That was really an eye-opener and also very clear and precise explaination what was going on. BIG thumbs up!
I did this to my VTA ST70 amplifier and it provided much better bass impact and definition. In my opinion with the Amp I tried, I heard excellent results.
Old video, I know, but from the repair industry, at about 13 minutes in, a few things that are incorrect are mentioned. In regards to increasing the size of the main unregulated power filter, the increase of inrush current is an issue, but rarely for the transformer. Some transformers do have a thermal fuse buried in the winding's, and even though the inrush is too short of duration to make the transformer hot, it is often enough to damage the thermal fuse due to self heating internally. This is rare. Symptom is an open primary with no signs of overheating and burning.
The big issue with many larger amps for consumer use is the pitting and burning of the power switch. It was quite common for vintage amplifiers to have a switch failure.
Great subject. Thank you for those tests.
I'm one of those audiophiles with a cap passion. I bought a 24W class A power amp chi-fi that had 30kuF per channel of Elna caps. I didn't think they were genuine so I swapped them out with Supertech 2Ts. Clear improvement.
Then I added an extra 10k per side so it's now 40k per side, I'm getting a more relaxed open bigger and more dynamic soundstage.
So for me, not only add more but using better quality caps is the way. I'm sure Mundorfs will sound even better.
More power to your channel.
Just as you were Tony, I was also surprised that the added capacitance in the power supply section made as much of a difference as it obviously did. Just goes to show, trying things in the "real world" versus relying solely on equations on paper, can have advantages also!
Great job with the demonstration sir and I wish you and everyone watching (in the US) a Happy Memorial Day! God bless you all.
Nice experiment and very good idea to do a transient 5 cycle waveform test.
It's not about getting more power from the amplifier but getting more reserve current in the capacitor banks to allow the output transistors to actually deliver what they need to when the signal reaches the crests of the waveform.
And the results for sine wave testing, even 5 cycle bursts, is not a perfect reproduction of reality.
In reality, the extra capacitors provide a much cleaner sound.
Now regarding the endurance of bridge rectifier and transformer, it indeed can be a limiting factor. I don't condone what I did (40 years ago) but I did just that, and to a much higher extent. I had a Luxman L-3 lowly amplifier that was rated for 33WRMS. It came with two 3.300uF capacitors. I added 20.000uF per polarity - yes there was enough space inside - and the sound difference was huge. It really didn't deliver more power but what power was delivered was cleaner and the sound much sweeter even in average levels.
I realized how much better the sound was when I had two output transistors blown after a high-school party. I had to get it serviced and of course I removed the extra caps before sending the unit in. When I received it fixed, I tried it out and it sounded horrible. Nothing like what I remembered. I quickly put the extra caps inside and the nice sound came back.
The sound difference in the bass notes and high frequency clarity was enormous.
I can still guess which amplifier has inadequate capacitors just by listening to it. It takes just a few minutes to identify that and I have not failed yet. My main amplifiers at the moment are a few Technics SE-9060 with 2x18.000uF per channel. They sound much more powerful than they are.
And I need to repeat that it is very important to not strain the rectifier AND the transformer. In many amplifiers, they have cheaped out and either the transformer or the rectifier or both will not tolerate the inrush current.
they tolerate short circuit for at least some 5-10 seconds . much longer than needed to fill those caps .
First.. Thank you Tony for the inspiration. I've now repaired two integrated amps (Fisher CA-120 & Technics SU-8044) and one receiver (Technics SA-727). I'm far from any kind of an expert but I learn more from every video. The latest amp (Fisher CA-120) has two 6800uF 50V main caps. It's only pushing 30-35 watts. The 1000uF in that Kenwood seems mighty low in comparison. This amp only has 2 transistors and one amplifier IC (STK-465). That's it. I've never seen a unit with a transistor count that low. GOD Bless!
Take it from an electrical engineer and audio enthusiast, don't look at the time domain to try to determine if there's distortion. You will not be able to see it unless it's really really bad distortion. Look at the frequency domain on your Oscope - the FFT. Then you will really be able to tell how much distortion and harmonics pop up. You will not be able to tell from the time domain with the naked eye
Your videos are fine:
As business owner who shares his limited time allowing us to look over your shoulder you do an excellent job, reading understanding schematics which is underlined by the ability to reverse engineer equipment. Troubleshooting and use of test equipment all come through to us your audience. Thanks for sharing and keep safe
I dont mean to be so offtopic but does anybody know a way to log back into an Instagram account..?
I somehow lost the login password. I appreciate any tricks you can offer me
@Onyx Deshawn instablaster ;)
@Bodie Zeke Thanks for your reply. I found the site on google and I'm in the hacking process now.
Seems to take quite some time so I will reply here later with my results.
@Bodie Zeke It worked and I now got access to my account again. I am so happy:D
Thanks so much you saved my ass !
@Onyx Deshawn Glad I could help =)
Howdy.
Very nice presentation.
One comment though. Increasing the filter capacitance changes the transformer curret. The conduct time will decrease but the conduct current will increase. The current waveform will be more spiky. This in turn will increase the current RMS, althoug the DC consumption is the same. The current RMS times the mains voltage is the transformer VA power. It is possible, though unlikely, that the transformer VA power exceeds the rated VA value. I suggest being aware of this when increasing the capacitance.
Regards.
It's more of a concern in tube amps with tube rectifiers. Too much capacitance will blow them up inside.
The limiting factor on increases in power supply filter capacitance is sometimes the power switch on the front of the amp. Or the circuit breaker. Be careful out there!
Great video Tony. If you get the chance to test it here is something
that I saw in an amplifier, the PSU capacitor bank was made out of 10 small capacitors 1500uF per branch 20 in total and the reasoning behind it was the time to recharge required for a small capacitor is a lot faster than a big one.
This method seems to be a lot better than using big caps .
Of course, things are much more complicated than that and it should be a very interesting test if you could do that.
Excellent analysis on power supply capacitance. Though the experiment contradicted your initial opinion, unless updates to supply is done, I would stay with about 20% max increase when replacing caps
i guess huge surge currents wont hurt transformers at all, they have a large thermal mass and its internal resistance limits the current anyways. it wont break if you dead short it for a second or two. rather the mains fuse or the rectifier might blow up
i accidentally shorted a 1 amp rated transformer for 30 seconds or some more , later i checked it gives 5.6 amps when shorted . and nothing happened to the transformer or its bridge . ppl just parrot what they heard and have no real idea about , to be able to say something negative .
I'm surprised at the 1000uf cap in the original. With the caps being expensive back then, I get it but wow. On a scope there must have been a nice amount of ac riding the rails. My guess would be 4700uf as a nice fit without cooking the diodes, or add a bridge. These videos are great!
In my experience, you would usually have to get pretty crazy with oversizing those capacitors to do any damage. Often, I oversize (and/or over voltage rate) them just so I can get capacitors that are physically large enough to fit the clamps. All other things being equal, new capacitors tend to be much smaller that what they are intended to replace.
Another alternative for improving the power supply is to use high powered zener diodes with sharp kneepoint, low noise and low dynamic impedance, They are available in normal or reverse polarity. Polarity is indicated by the position of the Zener symbol on the part. They are available in 46 voltages ranging from 6.8 V to 200 V. Devices' operating temperature range is -65 °C to +175 °C. Additional specifications include low forward voltage (VF max = 1.5 V) and 50 watts power derating (0.5 W/°C
Wow Tony. That was a marathon project for you. Good on you for persevering. Interesting piece on the capacitors. Thank you.
This could be compensation for weak de-coupling caps in other parts of the circuit. Sometimes, changing a bad 22uf next to a transistor can make all the difference. Always check the other caps on that rail before deciding that the the power supply lacks main filter capacity
An interesting followup experiment would be to figure out why it improved the performance - was it the increase in capacitance, or was it the lowering of the effective ESR of the filter? As you pointed out, paralleling the caps lowers ESR. If you put in 4 caps total, each half the value of the previously installed pair of caps and repeated the experiment, we might learn something interesting. Perhaps ESR is the critical performance characteristic here more than capacitance.
A few years ago I picked up a Hafler d220 power amp for cheap money (known issues) and decided to replace the driver boards with better ones. The vFET outputs on these amps are pretty bullet proof so i was pretty confident they were ok. While I was at it I decided the original 25 year old 10,000uf caps should be replaced as well. I found I could get 22, 000uf 100v low ESR caps from Mouser for a good price so I went with them. Because of the inrush current I added inrush limiting to keep he beefy transformer happy, I also beefed up the wire sizes in the + / - 60v power supply using #14 wire.
This amp wasn't working when I got it but now it sounds great and my inrush current limiting protects boththe transformer and ckt breaker from undue stress. The bias on this amp is set high so it also keeps that corner of the room a bit warmer.
It would be worth upgrading the rectifier diodes if you are increasing the smoothing capacitor. The peak current could be significantly increased.
Nice that you make the effort to broadcast this on RUclips, really appreciated.
With modern hi-fi amps bragging about total capacitance isn't unheard of, it absolutely makes a difference. Like many novice tweakers I'm currently in the process of running around doubling the capacitance of everything I have, haha. It's a little more complicated figuring out what's what inside my modern Yamaha receiver though. With mass produced products like this and many of the cheap Chinese amps I have, even though the components are relatively cheap this is still very much an area where corners are cut for cost.
Years ago I upgraded a Hafler d220 amp by replacing the driver board and decided to replace the main filter caps (10,000uf originally) because they were 20 years old and i wanted to bulk them up as well. This amp uses very fast FETs and was remarkable for it's time. I ended up using 22,000uf caps and i installed a 35a bridge as well. This stiffened up the supply and it increases the peak power capability, remember music is dynamic - we don't listen to test tones.
When i did this i also added a couple of varistors in series with the power transformer so I would not fry diodes or the power switch. A class AB amp draws quite a bit of power so these varistors heat up pretty well and while they do limit the amps output they don't really interfere with the music peak power this can deliver. i also added a couple of high frequency 560uf caps along with a couple of 1uf film caps to improve the ESR at higher frequencies.
Adding parallel capacitors in this case lowers the effective ESR which can have beneficial effects such as the reduction of the oscillation originally noted in your analysis. Nice job!
The ages before the switching power supplies usage the term ESR was totally unknown because has to do with high frequency ripple currents flowing inside the circuit itself.In line transformer base power supplies there is no ESR because frequency is 50/60HZ.
It's true that most home audio equipment can be optimised. In this particular case adding power capacitors did make a difference. Adding a soft start circuit would be excellent . So let the 4700 MFD where it is . Thank you for your excellent demo and lecture. I am floored 😀 and subscribe d
I wish you can do more of these videos because they are very educational.Thank you!
Stumbled upon this video and found it worthy. I would be interested in also seeing the effects of added capacitance with a voltage doubler circuit.
Anyways, i wonder what the waveforms looked like with the stock capacitance of only the 1000uF. You had already had installed a 200% increase over the stock designed value!
I think much of that early to mid 70's consumer transistorized integrated stereo systems all pretty much had that same cookie cutter fizzy audio with flat under powered farty bass, scooped mids and ice pick upper tones that probably generated some sorts of ultra sonics too. As a junior high school kid I saved up for a dept store consumer rack component stereo (w/8track too!).
Through out Jr High and High school I tried all kinds of speakers. Even the highly efficient EV drivers did little to present any greater sound. I was incredulous to observe that my 80 watt PEP rated stereo amplifier destroy my girlfriends dad's 250 watt RMS rated Klipsch loudspeakers and later realize that it also contributed to the death of my JBL PA cabinets for my band. For a long time I was perplexed as to such a great difference in perceived audio power from a 100 watt solid state stereo amplifier and say the simple single ended 15 watt tube driven Hi Fi amplifier.
A pair of old fashioned dumpster dived 6v6 SE mono blocks would overpower any 100 watt and greater S/S stereo I could go against, even with 200% impedance mismatch of 8 ohms.
Oh good times being a kid back then!
Man I love that scope you have there! That is a nice piece of kit. Tektronix no less! I just gave away an old boat anchor digital storage scope! I want something like what you have but I have pretty much resolved to myself that I likely cant have a tektronix and will have to do with a rigol or hantek scope instead. That digital phosphor color screen is the cats meow!
Oh yeah one last comment, couldnt a simple 555 timer and a simple transistor output to a relay make for a very simple kludged together soft start?
I think your video is a good demonstration that more capacitance in the power supply does make a slight difference, as you said, with diminishing returns. I have been experiencing the same thing lately in my vacuum tube amplifiers. My particular goal was reducing the spurs accompanying each of the harmonics in a spectral display watching it with a spectrum analyzer . I don't think the original designs are wrong or lacking. They are adequate for what the amplifier is supposed to do in its day and price category. If you stretch out your sine wave or use delayed sweep in your scope you will see that the oscillation (wiggle) on the bottom of your sine wave is either 60 Hz line frequency or 120 Hz full wave power supply ripple frequency and more capacitance helped reduce that and it you were looking at it with an audio spectrum analyzer you would see the 60 and 120 Hz spurs reduce quite a bit all along the spectrum with the extra PS capacitance. Also, you demonstrated the type of distortion typical of SS amplifiers vs vacuum tube amplifies. I thought that was really nice. A vacuum tube amplifier will distort different from the SS amplifier and not have the sharp transitions shown in your video. Our ears do not like sharp transitions in the waveform. Excellent demonstration, in my opinion. There is a little more stress on the rectifiers to keep the huge capacitors charged but it probably isn't hurting anything but it probably isn't helping much either. Thanks for posting.
It's an honor to hear from you! Love your channel! I wish I knew half as much as you do about this stuff. I agree about the oscillation being from line noise. Another viewer also pointed this out. Always something new to learn. Thanks!
THAT was something I've always suspected, from my multiple experiences adding capacitance to an amp's (or mixing board's) power supply. It always seemed to tighten up the bass. I've never gone nuts and overdone it, as I knew there was a limit. But boy can it make a difference.
Now listen to the amp before and after and you'll hear it. I never suspected the *magnitude* of the difference - at 30Hz, you gained 50% again more power while lowering distortion! Now that's value, dawg.
Thanks so much for doing this test- the *right* test. "Dynamic" test is a good name, but in audio a better name might be the "low-frequency transient" test. Fascinating to see what I've actually heard: taking little steps toward making a PS more like a battery really does help. Just don't overdo it. And by the way, you're not alone: my tech was surprised, too. When I asked him to add more capacitance to the power supply on a mixing console, he said it probably wouldn't make much difference. I told him I'd tried it on an older board, and it sure seemed to help tighten up the board's bass response. Well, he said, let's try it and see. Same results as you saw: sure enough, the original PS simply had been designed down to a price, and there was indeed plenty of room to work with in adding capacitance. Bass response and distortion improved pretty dramatically.
I realize a console ain't a power amp, but mix consoles do have high power requirements, and they require very clean and responsive power, much like power amps.
The Marantz 1060 Integrated Amp is one of those classic amplifiers that people add larger filter caps for precisely this reason. The problem is this particular amp has no speaker relay so you get a bigger thump when the power is turned on, which can damage the speakers.
When selecting a replacement electrolytic cap you must know its original specifications such that ypu know from where you are starting. From there you can use a capacitor with higher ripple current rating and a lower ESR. This will deliver more current faster even with the same value cap.
I'm learning something every time you drop a video. Thank you.
I've always wondered about this. I always wondered if the rest of the power supply had to be up to the task of handling that extra capacity. Thanks for the video
You can calculate the ripple voltage at various capacitors fairly easy, q=C*dV and q=I*dt, so dV=(I*dt)/(C),
Where C is capacitance in Farads, dV is ripple voltage in Volts and I is the maximum current pulled by the power amp.
at 50Hz dt=100ms (full wave).
As P=I^2 * R, I=sqrt(P/Zspeaker) (where Z speaker is 4R or 8R)
Plug those values in and you can get an indication of average ripple.
Hi Tony, great video, as always. I wish you measured the amplifier self noise with no input signal after beefing up the dc filter capacitance.
I was recently working on a 1970s Rhodes Suitcase Piano with two 50W output channels. It uses +/-30V rails and has 11,000uF on each rail. But it also has a very large power transformer and a chassis-mounted high current rectifier bridge to handle the inrush current. With the Kenwood receiver, the designers probably calculated that amount of capacitance they installed would be adequate for how a typical user would run the amp. It also looks like there may have been size issues. Rhodes had plenty of room inside a large speaker enclosure.
I know vintage Proton amps had dynamic power on demand which basically provided extra headroom via extra coil windings in the power transformer
Love your honesty and inuisitiveness Tony.
I think what we see at 6:21 is not distortion, it is clipping, just your scope triks you because it is in AC mode, on DC mode you can clearly see the clipping.
I noticed this phenomenon first when I tried to measure really low frequencies, my 100MHz scope just couldn't do it, but my 2MHz handheld scope did it well. Yesterday when I wanted to see how ugly the sine wave in the mains I noticed this again, the 100MHz scope measure higher voltage than a multimeter and I can see a distorted sinewave in AC mode, but when I put it in DC mode everything falls in place, it cuts the sine cap off and it measure the same voltage as the multimeter. The mains voltage is so bad here that the toroid transformer sometimes start to buzz loudly in my amplifier, the voltage is usually 230 instead of 240. Seeing this I'm considering to use switching power supply for amplifiers, over 4-5W load their switching freq goes over hearing freq so it shouldn't be a problem.
I love my little DSO 068 scope, it is soo convinient to use and I can avoid my biggest fear of blowing up my scope trough ground loop.
thinking of changing the big filter capacitors on my '76 pioneer SX 750. Would I need bigger spec replacements in that Vintage?Can you even find exact matches for the originals? I guess going a little bigger is OK if the original spec arent available.
Tony's amazing..any wonder his business involves Hospital measurement/monitor equipment repair ! ?
Be advised that if the amp has a tube rectifier, there are rules about maximum capacitance.... usually 60uF or less. If it's a SS rectifier, all bets are off..... you can go as big as you want, just know when an amp is turned on the inrush current to charge the larger caps could blow the mains fuse. An increase of lower frequencies and overall dynamics are the main benefits.
you r a n honest person . hard to find these days .
Seems to me that the designers of this amplifier were more comfortable with valve amp design and high SUPPLY rail voltages than transistors and their current demands. So they used a capacitance value more suited to a valve amp. Looks like it was designed in the "crossover era" (no pun intended).
I have an old Univox guitar/PA, amplifier, amplifier modules in each speaker cabinet, four speaker configuration in each cabinet. The filter capacitor is 47000 uf, it holds a charge alright! Two diode rectifier, simple but holds the extreme loudness of the amplifier, simple pre amp to transformer to a push pull output, a set of parallel darlington direct coupled output transistors. Electrolytic Capacitor output. Good non distorted output.
I picked up an already assembled kit amp which had issues turning on. And found that the amp had no inrush limiters and the entire power supply was really absurdly overkill for the power output of the amplifier. So in this case I reduced the capacitance size quite substantially. Perhaps smaller than what some designers would have done, but considering the amp only develops about 65 watts/side and has transformers rated at about 2400VA (!), and a really low volt swing, I only needed just enough to remove ripple. It is a kind of dual mono design so the transformers will be fast enough at that kind of output that performance wont matter at all. And the power on is not going to pop breakers or dim lights all over the house. And I say "kind of dual mono" because there are 2 transformers and sets of caps, but they are siamesed in parallel. So I could technically remove half of it and still be ok. And I may do that one day if I were to build another 2ch amp and want a football sized transformer for it.
Hi! This video is excellent, the way you have explained things is just excellent, I enjoy watching your video so much, keep up the good work... Have a great week..
There is nothing wrong with modifying and amplifier providing is done correctly with some good calculations if there is a small improvement is good. I'm not saying The Originals engineer didn't take this in consideration they already knew but it's all to do with cost at that time. I done the same thing many years ago and a JVC class A class B amplifier. I saw very similarity what you were showing on your silloscope and yes having the proper audio equipment then you can at least see it sometimes it's very difficult to hear it or you might not hear it at all all depends on how good your ears are and the type of music you're going to test it and you speakers. Excellent way of demonstrating and explaining. 👍👌
The negative feedback should cancel out the ac ripple only if the low part of the voltage dips to the output level and it clips.
I remember back in the 80's walking by receivers at one of those warehouse sales and they had them playing and many of were showing power supply issues. Their display lights were flickering or dimming. I wondered if bigger caps would stop that. I have since gotten some very big computer grade caps to play with. But I don't remember if they worked.
Great video always wondered if this was worth doing. Tony any chance of ever doing a video showing how to fit Pre amp out or sub woofer out sockets. Lots of the older gear don't have these. I've had idea's where to take it from inside but I am sure your way would be best. Many Thanks.
I always look at the AC ripple that shops up on the DC supply when drawing near full power.
Great job on troubleshooting and repairing this! I enjoy watching your videos!
I usually go up a size on filter caps. Lately I've been using film bypass caps as well.
Most extreme change I've done is on a pioneer mj200 power amp. Went from 3400uf non bypass to 15,000uf with 0.1uf bypass. Its worth mentioning that the pcb had three rings printed on it with the standered caps being inside the smallest ring. The 15,000uf sits slightly outside the the largest ring. No problems to report yet 😆
Yes using the 0.1 bypass cap how's it sounding that's where it counts
@@johnsweda2999 sounds good. I use it for headphones and speakers. Only time I've noticed any difference with the bigger caps is with stuff like Bassotronics and DJ droppin or huge Japanese drums. Of more effect is the customised output filter. That really took the edge off things.
I listened with the volume off, and when you show the under side of the receiver, I was like "I saw that somewhere"
I realised that I have almost the same receiver and worked on years ago. It's like my own trio kw-55. For mine,
The biggest issue will likely be the rise of B+ and the subsequent voltage, in the case of a very lower rated transformer, it can also hurt the transformer. These high voltage transformers are rated in the 200-500mA and will likely burn in the blink of an eye if the in rush current is too high . For mine,The B+ rising too high will likely need some modifications because the valves will likely be ruined very quickly. If it's the receiver, I think it is. They are already pushed to their limits plates voltage with the correct capacitor in place.
Often the limiting factor is the power transformer and no amount of capacitance will fix that.
I restored a Magnavox amp from a console in 1974 and it did 15 WPC into 8 ohms with the stock power transformer and I think it had a 1,000uF 63 volt can cap originally but I upgraded the rectifiers to higher current ones and used a 10,000uF cap in place of the 1,000uF cap.
Just by replacing it with a transformer that is rated at 5 amps I was able to double the power to 30 WPC at 400 Hz with it being a little less at 20Hz. I noticed a weird distortion at 20Hz when I ran it to full output and I checked the power supply ripple and found it to be nearly 2Vrms. So I put in another 10,000uF cap and that completely took care of that issue and I now get full power at 20Hz and into four ohms I get around 49 watts.
One indeed definitely MUST know exactly what they are doing and know the circuit well before even thinking of attempting mods like this and they must have the proper test equipment to verify they have not caused any issues and that the mod works as expected and actually does make the amp better.
Physical capacitor size was definitely a determining factor back then.
The 10,000uF 63 volt caps I used are about the same physical size as the stock 1,000uF can cap is. When one thinks about how large a cap of that value would have been and how expensive it would have been back then, they realize real quick why many amps use the bare minimum filter cap value they can get away with.
Not sure if it was covered but the effect of adding cap depends on the Class of the Amp. Class A would not see much change. It can however lower hum in some cases.
It can be that the coupling caps are low.
If you do add a filter cap make sure its WVDC significantly exceeds the supply voltage.
Fantastic explanation. Thank you
The improvement is almost certainly the result of lower impedance in power supply. Your transient response at 30hz has the caps getting recharged by the supply between each peak! A power supply transient response requires a faster transient. Even using higher frequency some multiple of mains would be better (full bridge@120 cycles).
Happy Memorial Day Tony, Family and friends !!! All the best to you all.
Fitting extra capacitors does make a small difference but there can be issues with changing currents giving rise to buzzing , I built two identical amplifiers during lockdown one had two power supply’s mono block the second was single transformer with a higher secondary voltage than required , and I built a dual stabilised hi voltage regulator plus and minus 56v they are both very good but the one with stabilised supply has much better Bass and background noise/hum is absent 100W class AB but would be difficult to retrofit to an old amp and no supply line fuses as these are the weakest link when drawing large currents no ripple on supply line ! An external power stabilised supply could be built
I think what you called "oscillation" was power supply ripple which is only visible at clipping when the amplifier circuit power supply rejection doesn't work.
Is that right? I've often seen this but did always thought it was oscillation although not totally convinced. Your alternative explanation is interesting. Do you have time to expand on it here?
I second that hypothesis :-)
Outstanding, I have a NAD C352 amp, recapped it all as the originals were garbage.
Just got the main filters to change, 22k uf at 63v.
Really enjoyed your video! Subscribed.
Cool vid man! I like fixing (or at least TRYING to fix) audio equipment also, so this video was very interesting!
Although you should've added a link to your other videos you mentioned in this one, especially the linear power supply one! (I'm gonna try to find that one in your overall list of videos, but a link to click on would make finding and watching that one SO much faster and easier!)
I found some Vishay 10000 µF 75V caps on sale at Newark. I'm re-capping two amps that have 8500 µF 75V caps, and another with 10000 µF 65V caps. All the original caps are 3.125" tall. The new caps are 4.125" tall. The tall caps will fit in both types of amp. The new caps have lower ESR and ~9A vs ~6A ripple current. The new caps are rated 3000 hours at 105°C. I want these 40+ year old amps to last another 40 years. I think I should add soft-start.
6:45 My assumption was that we were seeing a slowly moving phase difference between the FG and the 60Hz from the wall. The supply ripple will cause the power output to dip slightly when the peaks in the signal are at the low points in the ripple, right?
thank you Tony...I'm learning so much...your a good teacher!!!
I have a Musical fidelity AS100, which uses just 6800uF per rail(35V dc) on each monoblock, which sounds pretty low to me (2.38 coulombs or 2.38A per second), so putting in some Naim 10,000uF additionally on each rail (4 in total), hopefully that will give it more powerful dynamics.
Naim make their own capacitors?
@@maat8249 Not sure if they were in house or made by third party, but they are labelled Naim, I am guessing a 3rd party to Naims spec.
@@engjds That's crazy to hear, I would really want to know what they use. I just recapped my Sansui and it sounds fantastic compared to the original caps which were closing in on 50yo at this point. But I'm not sure if I could tell the difference between say Elna Silmics and Nichicon Fine Gold. I just try to fit from the same brand and line, and tend to choose what's made in Japan. I've had Elnas come from China...
@@maat8249 People swear Nichicon, but I have yet to find paper evidence, that they are actually better than the competition, they tend to hide their specs such as ESR unlike their competitors I think they might be riding on their reputation rather than golden eared reviewers.
If the capacitors are 2200uf and replace with 22000uf or more because you think the amplifier will sound better the fuse will blow due the high charging current.Or the internal fuse in the transformer will blow.In every case the circuit is designed for a certain maximum power with a maximum rated current and this does not change,but in the case the amplifier plays in low volume the original capacitors are more than enough because designed for the extrem conditions and passing the limits of the amp the caps will do nothing to protect from destroy.
I added an NTC to my mains input so that I can have ridiculously large reservoir capacitors for great dynamic ability and not pop the bridge or fuse.
I don't want continuous high power.
More capacity to a point doesn't affect the signal that much but will make the bottom end more balanced. But just a 10 or 15% value.
I just beefed up my filter caps or power rail capacitors ;-) from 36,000 uF to 112,000 uF in a 100 Watt stereo amp. The new ones had the same physical size and voltage rating as the 40 year old ones. Seems to me a good practice to replace those old caps and have a bit of a bonus with the increased cap rating. Bridge is 600V and 50A, Xformer is 1000VA.
More than entertaining, this video was very educational.
I think the oscillation on the power tests was probably not in fact oscillation but clipping ripple due to low valve smoothing caps that's why it vanished with higher value capacitors. I have seen this on many amplifiers in the past. I depends on how near to the rails the transistors go! Another way to get better voltage drop out is to use Schottky diodes. I never use any other diodes where power is relatively high. Most smallish transformer will deliver a fair current but will get quite hot if the drain is prolonged.
I changed the caps in a 35 year old Heathkit AA-1640 from the original two 7500uF to two 18,000uF. Seems to perform much better, especially on sub bass.
Of course. Thats simply because your 35 year old capactors would be degraded, and the ESR will have risen considerably. To get a valid comparison, you first need to compare old 7,500uF to new 7,500uF, and THEN compare the upgrade from new 7,500uF to new 18,000uF.
This guy is asking the real questions. Subscribed.
I really wish I had this kind of knowledge
It’s not that hard if I can build a amplifier from scratch anyone can
I have an amplifier that I had to recap and I replaced all the caps with exactly the same value and voltage, the only difference in the new caps is that they are 105 degree and not 85 degree so that's what I upgraded, only the main caps I also doubled with the same voltage value which was 63 volt but the old caps were 4700uF and the new caps are two 10000uF caps and also 105 degree.
I haven't looked at the signals yet what it does now so I can't tell you if it's any better, to me it seems better because it sounds a lot better but that can as well be from replacing the other caps which were in series with the output to the speakers or those in the input part of the board, that's because at that point I didn't have an oscilloscope, now I have a new one and I can check the signals and it might be a good idea to do that and you also sparked my curiosity so I'm probably gonna do that soon.
That a great video. now if you put scope on power supply output you see a more stable voltage and less ripple . I like to use low ESR caps made in Japan for replacement. Even if you keep same value you see improvement. now in car audio 12V system with big bass amps a very bag cap in farads makes a improvement not just to the amp but the cars electrical system and less surges on alternator and battery. we call them super caps. even the newer class D amps improve. also super caps help with inverters that make AC from 12/24 or so to 120V ac. startup surges on appliance can be improved. this happens because of lower voltage higher amps. The voltage drop in wires becomes a factor. so placing large cap as close to load as possible in that case.
Be aware that you are reducing the conduction angle of the rectifiers, so in theory the transformer and rectifiers will have slightly greater losses. I have in the past gone to greater lengths with added capacitance, you might find it interesting to listen to the audio that gets superimposed on the reservoir capacitors. (Easy done using some dc blocking and headphones). Any high frequency content here points to high ESR rather than a lack of capacitance. I suspect the ringing you are seeing is related to ESR rather than the capacitance.
It's good for dynamics, kind of like what they do in car audio. I've increased cap size before in equipment, 9/10 no issue if done properly and it makes a difference in the bottom end of music. 30-50hz or lower is where you'll notice it. It can be a bridge/diode/fuse killer sometimes as people put too big a cap in and it's almost like a dead short before fully charged, just saying (for people out there who try it). Good video and you made me want to drink coffee! lol
I suspected this (Damage to the rectifier) would happen, especially on older equipment, have always stuck with the original size filter caps and had no problems, fresh caps really make an old amplifier roar like it should...
Interesting tests but I would like to see same tests at half power as that is how most people use their equipment, rarely do people run at full power. So would the caps make any difference testing at normal listening levels ?
I really should have mentioned this point. As far as you can see on test equipment, there is absolutely no evidence that anything changes, by adding a bigger capacitor. Some, however claim that you can hear a difference, even at low volumes. I think the bass may be a bit tighter sounding at mid volumes (because a transient can spike the power up quite high, even at mid volumes), but I absolutely can't hear any difference at lower listening levels. If you noticed, you could only see a change when the amp was close to its limits. Thanks for the comment!
Very good example I restore a lot of vintage receiver and amps and I up the filter caps by as much as 20% of the original Uf and the end result for me has been very good
I did not hear you mentioning anything about these effects of increased capacitor size. First, the increase causes higher peak currents (which you did mention), but what that higher peak can cause is more hum. Second, you send more distortion to the power cord, which can cause hum in other equipment attached to the same power branch. All that is also dependent on the transformer impedance, or let’s say its current limiting characteristics. Anyway, the overall consequences may be a little more wide spread than just the output power and sine wave distortion. And speaking of power, I have seen a statement that a normal home listening volume (with “normal” speakers) is less than 1 watt. In fact, that is also what my stereo indicator LEDs indicate. The strings have two ranges, 2 W full scale and 20 W full scale. I keep mine at the 2 W range and seldom see more than half of the string lit. Half way in the logarithmic scale represents about 1/2 W. Of course on a stereo, there are two channels, so I have two indicator LED strips, both below the half way, but of course being additive. And just a funny side note - you pointed out a change from germanium to silicon output transistors. I still have the germanium power on my first (and only) self designed and built 2 * 10 W stereo amplifier. I use it a few days per year, but it works fine. Never any capacitor changes or anything at all. Lucky me!
Linear power supplies will just waste more as heat. Narrows the conduction angle higher peak current .
Well you said diminishing returns, not that there wouldn’t be any returns. So you were right. It didn’t increase total power, just cleaned it up more.
Nice video, Tony. I learned a little here.
Excellent stuff Tony really enjoyed this video
Nice video with useful information. All videos cannot contain everything, and there will be other opinions as usual. I do have one question though: how did you generate the quick 60hz burst of only 5 cycles? best regards
I was the opposite of what you thought, I figured it would improve the signal. Cool video.
Hi thank you very much indeed for this very interesting test. I did a very trivial test ... i had a stereo amp originally with about 2x10mF total ... i replaced PS caps with 2x60mF huge caps by Mallory (refurbished ones) The solid state amp sounded more powerful and overall better to my ears ... when i switched it off the amp kept on playing for some 15-20 seconds running only on caps stored energy. Since then i believe that more capitance is a good thing. Maybe there is a point when adding more mF has no practical effect. I am looking for finding that point. Should i have stopped at 2X40mF maybe ?
I can see in some high end amp huge caps banks ... scaring. And very expensive.
thanks a lot again. Kind regards, gino
A little bit of ac ripple on a B+ supply is generally rejected within limits. By adding Capacitance narrows the Conduction angle. Can burn out supplies with big in rush current. I have come across lots of blown modules and often found the values of Caps resistors diodes could be substituted to increase reliability. Only way to find out is to try it. Thanks for taking the time to present.
This is very interesting. Thanks for the video. It seems like increasing power supply capacitance helps when the design under-specced capacitance. I'm curious if it would help in a more robust setup. Currently recapping a Yamaha CR-640 and didn't bother upsizing the PS capacitors because they were 10,000uf for a 45W amp. It might be interesting to run this again using a similar amp.
@xraytonyb, Thanks for putting this information out. It leads me to my question: I am recapping a Pioneer VSX D1S, it's a nice unit. 1990's...do you see any problems in increasing the main caps? Thanks for your help!!!! Cheers
4000 µF is shockingly little capacitaince for a power supply :o
I thought that was quite a lot...although when I see a lot of my decent old Pioneer stuff is using 8000 uf I can see it isn't.
@@chillidogkev Yeah, well I have amps that employ 2x 10.000 uf. So 4000 uf is quite small, but it really depends on the amps wattage delivery ratings and how big the power supply is. I mean more uf is always better, but the amps has to be cost effective too! Also upgrading with too large capacitors sizes can damage your diode bridge if it isn't rated for big filter capacitors, although this is mostly for concern in older equipment, but always look up your Full bridge rectifier specs for good measure!
does the new generation professional amplifiers need an extra capacitors? thanks
Would it be o.k. for a SA-9800 to replace the 4x 15000µF 80V with 22000µF 80V same dimensions ?
Should be, but there wouldn't be much improvement, if any, to the performance of the amp. Going much larger in capacitance could cause issues with inrush current when you switch the amp on. This can damage the power switch over time, unless that amp has an inrush current limiting circuit (sometimes called a soft start circuit). That said, replacing the 15000 with a 22000 probably would work without any major issues in the SA9800.
I went up 20% on my sa 9500 and the heat sinks/transistors seemed to run alot cooler....I also had a failure in the power button on my unit and found a tower computer power switch that worked great!
Next step is the output capacitor. It should have at least 2000uF but I doubt it considering the small cap in the power supply. I think that is the reason why this amp delivers so much less power at 30Hz.