The speaker doesn't like DC through the coil. That is constant heat dissipation that gets overlaid on whatever AC power from reproducing music. It needs to be coupled through a capacitor (like 1000uF, larger is better) and the quiescent current has to be at least Vcc/speaker impedance 😉
Yeah, no decoupling capacitor is the first thing I noticed! Depending on exact impedance of the loudspeaker at low frequency (assuming it is near 8 ohm) that tiny speaker isn't going to be able to produce much on the low frequency side so a 200 microfarad capacitor should be plenty for a roughly -3 dB roll off of around 100 Hz. Still results aren't too bad considering the DC. What value resistor would you use on the cathode?
@@AllThingsOnePlace I wouldn't use a resistor, active load would be better. After thinking again about it, you would need somewhat less resistance than speaker impedance. A better way would be to use an inductor (but a very large value, maybe a classic fluorescent lamp balast?) on the plate and setup the bias point with negative voltage on the grid. outputs Would still be coupled through a capacitor. Should allow for 4 times more output power with only twice higher quiescent current?
@@akkudakkupl Yeah, giant inductor would be best for performance, wouldn't really fit the project though. It isn't going to be efficient ever since its a class A amp but yeah squeezing as much of that AC into the loudspeaker as possible is one way to go. I say resistor simply to match the project or make it work within the confines of the current project. Swap the 100k for a much lower value (has to meet power requirements of course), add capacitor to the loudspeaker (possibly hide it inside the already made enclosure)? I guess could put a lamp ballast under the whole box and route the wires around it too. Either way, I am very new to tubes and slowly learn a little on each video I watch.
I actually did try to put a few different caps in series when testing the design to remove any DC through the coil, but no matter what size cap I tried, it always decimated the sound. Like removed greater than 90% of volume. It could be that I just didn't get a large enough capacitor, but I think it mostly had to do with the insanely low voltages I'm running at. The 12B4A was moving around 40mA through each tube, which at 48V should mean an internal resistance of about 1200 ohms per tube. With six in parallel that should be something like 200 ohms of series resistance with the speaker. I think the impedance is just so ridiculously mismatched that adding a series capacitor just wrecked whatever weird balance I had going on here. Or I could be totally off the mark, this is all uncharted territory for me!
@@UsagiElectric it was because the tubes have to run current at least equal to VCC/speaker impedance anyway, so it can swing up or down. This is why class A is ineficient.
What a cool journey through non-conventional tube amp design! Just today I was studying cathode - followers for guitar amps and it is so cool to see them drive a speaker directly! Thank you for such a wonderful content!
Thank you so much! It's interesting how many different ways there are to tackle tube amplifier design! OTL amps really appeal to me, so this definitely won't be the last one I build. There's a ton of place I can improve on the design and get even better sound out of it. Plus, I have a handful of 6146 Power tubes that when triode strapped should be able to move about three times as much current as the 12B4A, which should be pretty exciting!
Your construction techniques are one of my favorite parts of your videos, always so clever and resourceful. Mounting the pots that way with the header pins is a great idea to take the stress off the joints.
Thank you so much! I actually really struggled with a good way to mount the potentiometers using a single sided board. I discovered that If I give the header pins a slight bend, they slip right into the pot connection holes on the back side making it super easy to solder up, but also adding some strength to the mounting. I'm really happy with how this one turned out visually, it's a bit dense and runs way too hot, but it looks awesome!
Back around 2014 while in college I built my own 4 channel OTL surround sound amp and used +/- 60V rails for the output (needed -120V output bias and something like 350V for the input gain stages). With 4 6AS7 outputs per channel (8 in stereo mode) it's plenty loud for a moderate sized room on quadraphonic mode, and rattle the windows and walls loud in stereo. It's clearer than conventional tube and SS amps too...It's really interesting to hear a tube amp deliver razor sharp synth from prog rock while at the same time being able to render crisp punchy DEEP bass from say an action film with gun shots and explosions, and get things like piano and vibraphone right to a surprising level of realism. You should try collecting vacuum tube TVs from the 60's. Once you get used to working with and surviving regularly getting shocked by 25KV, voltages over 60V stop being scary, and become pains that you learn to avoid touching.
You are right, would be a better test with a high efficiency full range hi-fi speaker. Bigger speaker with the extra bass content takes a ton more current to drive. You should use a capacitor to protect the speaker from any DC, I can't remember if there are any caps in the pre speaker circuit? No there aren't. 🙂 Also restrict the bass frequencies to the audible range by trying different (smaller) cathode bypass cap values in the preamp stages, 100uF seems flabby. Saves wasting energy driving sub audible bass. Try losing or a bigger resistor in the neg feedback loop to the first preamp? Might drive a real speaker?? Exciting stuff.
Yes, actually pro audio speakers like the brand Faital Pro have 96, 98db that will work great on low power amps that needs high efficency. I've tried that with my 1w diy tube amp, also compression drivers have 105db minimum in most of the cases
Thank you! I really wanted to stuff it all in that tiny footprint, but in the future, I'm definitely going to spread things out and run more speaker for some proper full on sound! I happen to have a handful of 6146 power pentodes, that when triode strapped can move well more than double what the 12B4A can at a given voltage, so some of those mixed with a bigger speaker and maybe some proper design to eliminate DC bias and I think we can get something really wild going on!
Thank you so much! That depends a bit on what you define as metal. I always listen to music when I work, but as a translator, I can't listen to music with lyrics or my translations end up... strange, haha. So, over the past few years I've gotten big into different types of instrumental metal. Mick Gordon's work on the Doom and Doom Eternal soundtracks are perfect, and Andrew Hulshult and David Levy both knocked it out of the park with The Ancient Gods soundtrack. I listen to all three of those on the regular. Hulshult's work in general is really fantastic (both the Dusk and Prodeus soundtracks get a lot of playtime too). If I need something a little less... violent, I tend to fall back on Animals as Leaders, the Algorithm, or some instrumental Djent (which is actually what's playing in the background right now). Instrumental metal is a bit tough to find, but on the occasions some gems show up, like some of Meshuggah's stuff!
Wow that is sooo cool. I have a tube amplifier that I started restoring a few years ago and I kind of lost my forward motion on it. watching your vids really gets me psyched to work on it again.
Thank you so much! And that's awesome to hear that you're getting fired up to work on it again! If you get going, let me know, I'd love to see some pictures or a video of it running!
@@UsagiElectric Absolutely. I have a question can I use a 1b3 (13b? rectifier valve) to perform a full wave bridge or do I need four of them? I believe I need to rectify 2000 volts or 5000 volts at a high frequency such as 200khz. what will turn that sine wave into a unipolar pulse train?
@@johnkoury1116 That's a lot of voltage! The 1B3 is just a half-wave rectifier, so you would need two of them to do full-wave rectification if you have a center tapped transformer. If your transformer only has two taps, then you would need four to get full-wave rectification. Though, with such a high frequency, half-wave rectification and a good HV cap should be totally fine. The HP 150A actually uses a 1X2B as a half-wave rectifier with a 1500pF smoothing capacitor for it's HV generation. Definitely check out some old HP and Tektronix oscilloscope manuals for ideas on how they did HV rectification! Page 63 in this one: hparchive.com/Manuals/HP-150A-AR-Manual.pdf Page 114 in this one: w140.com/tekwiki/images/3/3d/070-198.pdf
@@UsagiElectric I am only going to be rectifying 3,000 volts but it is at a high frequency. I want to capture at a nanosecond level the energy from a spark gap. I have silicon rectifiers that will capture at 100nanoseconds but it takes some time to conduct from what I understand. I have been reading and it seems that the 1b3 because of the grid heater would be ready to conduct almost instantaneously. I have seen setups where they have used four of these valves together but I have to delve into it much further. And yes I want to fill a capacitor.
absolutely one, two, victor seven. let them squirm in the floor and rithe in anger, you be you keep up the amazing work. I am mad RUclips didn't recommend your channel sooner.
Output transformers definitely make way more sense! But, there was just something super appealing about the idea of brute forcing sound through a speaker. I think it's because my simple brain just says "more toob is good," haha. The tung oil made such a massive difference with the wood, I was blown away with how nice it came out. I never thought of using gun oil though, that's a brilliant idea!
Did make a mosfet class A amplifier a few years ago, used a few parallel mosfets, and an iron core transformer to provide the DC removal for the speaker. Could adjust the bias to go from full class A, and a really hot set of mosfets to something like AB, where at least the current was under 1A per transistor. Heatsinks were old Pentium CPU coolers, one per transistor, and an old mains transformer with 12-0-12VAC secondary, with primary disconnected. Loud on 12V.
I have two push pull amplifiers built with fets and power transformers that are literaly 2x115V to 36V. Fun to build, guital sounds great through them.
That's awesome! I love the idea of using old CPU coolers are heat sinks for the mosfets too, that's brilliant because those can often be found for cheap or free, and if you get really desperate, they already have provisions for cooling fans on them!
@@UsagiElectric I have used them a lot for power supplies, as you can easily mount a few TO220 or TO247 power devices onto them with no problem, just drill and tap the M2 holes for the hardware to mount them. Some with no insulation, because the heatsink was easier to mount on isolating bushes, and I have done a number of LED lamps using the round CPU heatsinks, they are really good for that to mount a star PCB on. For that easiest is to use 2 part metal loaded epoxy, you can use the thermal stuff, but for me just the regular Pratley quickset steel worked well.
You built a V8 racket machine. Congrats! Keep up the cool engineering work :) And I agree that this amp could use ceramic tube bases, possibly on a chassis rather than PCB.
Thank you Keri! It means a ton to hear encouraging words from you after seeing your phenomenal amp projects! Ceramic bases and a bit more space between the tubes would have made a massive difference. For the next OTL I build, I think we'll throw space restraints out and just let the amp get big!
Great! I love OTLs as a concept - they are the peak of inefficiency in tube audio (of course), but with a simplistic charm of their own. Not critical here, but it's generally considered bad practice to bias a tube from the wiper on a pot. The reasoning is that the connection could become noisy or intermittent over time which would clearly upset the DC bias conditions and act as a noise source. A simple solution is to add a single resistor from the wiper / grid to ground. This could be 1M, or even as much as 4.7M or so. The grid input current is extremely low, so a high resistor value will not affect the DC conditions, and will also allow a high value pot, permitting a high input impedance to the stage if needed.
Thank you! It's a completely improper use of vacuum tubes for audio, but I think that's why I find it so fascinating! Thank you for the information on the additional resistor from the grid to ground. I never thought about the potentiometer potentially causing biasing issues, I'll definitely slip that in on future audio builds!
If you're really interested in building 2 amps getting RCA receiving in the back part there's charts but tell you at which voltage to use which resistor value on which pin to get max output without distortion. You can adjust these values to get more or less gain ad needed.
You could use a series synthetic inductor to set the DC bias and tap off the audio with a large capacitor. A suitable "gyrator" circuit can be made from a single p-type FET.. ..which suggests that it could be done with a (suitably low impedance) tube as well.
That's definitely a bit beyond my skill level! But, I do have some 6146s that are quite a bit lower impedance than the 12B4A, so I think for the next OTL journey will involve those big boys!
@@UsagiElectric Consider using an inverted triode arrangement - plate at large negative bias becomes a control electrode with very little input capacitance, grid at positive bias becomes the output, cathode is handled as usual. The output impedance is much lower (nothing in the way of the electrons other than weak repulsion from the other side of the grid), at the expense of inverted voltage gain. However, the input still draws almost no current, so a tiny step-up signal transformer with a ratio similar to the "normal" gain of the tube can correct for this.
The proper output transformer is going to match your impedance where you will get a very flat response and the warmth and bottom end will increase greatly. But that litle speaker is not going to demonstrate that as well as a bigger driver will.
You are into cool stuff. I had a ton of 12B4’s 10 years ago to build a HV OTL amp but never did. If going for tough stuff you can use the 12B4 in a simple amp that is powered by a HV SMPS supplying some 170-200V. Or if going minimalistic, use a 6AS7G (or better still a 5998A) which is a dual power triode and a 12AT7 or similar for a 2-tube stereo amp. Requires beefier psu and output transformer but is an absolute killer.
I wouldn't put DC thru the speaker at the cathode, at the very least in this topology put a large capacitor in series with the speaker. The DC conditions at the cathode will be maintained and only the AC will pass thru to the speaker.
@@fabriziobrutti1205 You have no idea what you are talking about. Maybe *YOU* should go lookup what it means to bypass a cathode at AC. The worst thing about your comment is total lack understanding of what would happen to a low impedance voice call having DC running through it.
@@vincei4252 actually i didn't get your answer about the DC through the voice coil(?)... I know that there mustn't be any, or otherwise the speaker will get some damage, so a capacitor must be put in series, but by what I know the 100k resistance should be replaced with an 8 Ohm resistance... Anyway I'll check the part about bypassing a chatode at AC, since I don't know much about vacuum tubes (btw I'm just 16)😅
@@fabriziobrutti1205 That's fine. In a nutshell at AC, provided the capacitor is big enough, it looks like a short and the impedance seen at the cathode to AC is the impedance of the speaker. If you look at transistor circuits you will see something called emitter degeneration where a capacitor is placed across the emitter resistor or portion of the total emitter resistance. Doing this has the magical effect of increasing the gain of a common emitter amplifier because the AC impedance seen at the emitter is lower but the DC conditions at the emitter are maintained for biasing purposes. Look it up it's interesting stuff. Good luck in your journey.
As much as I enjoy OTLs, output transformers are not the villains that you suppose them to be. Being made up of inductor sections, transformers can and do increase signals dynamic behavior. In fact, they are a fundamental tool in that they allow us to shape the final timbre the amp is going to present. People wrongly believe the more straightforward a circuit is, the better. But this is not always true.Sometimes a more complex circuit carrying transformers means more harmonic richness, which is a good thing in musical terms. And let us not forget that harmonic richness is that which musicians are looking for whenever they overdrive their guitar tube amps!
First digits are the heater voltage, the letters are a factory designation and the last number is the number of useful elements in the tube, for tubes that use that numbering convention. But that doesn't take into account ECC34 KT88 and other tube naming conventions. I don't get upset by it though. I know that things are read a certain way over the radio. I often use phonetic alphabet when reading things out
As many pointed out is the DC you get through the speaker in this design an issue that will affect the sound and generate a fair amount of heat in the voice coil. But it is a fun project anyway. I would probably gotten for a big capacitor in series with the speaker and then match the resistor in parallell with that for decoupling but I know some gonna say a choke is better but that depends how much bias you wanna have the tube at. many would say just omit the resistor and put the choke there instead but you also gonna remove any bias of the tube and this design is cathode biased which gonna need that resistor to work, the same thing is actually happening with that low resistance speaker across it anyway so there will be barely no bias at all which might as well be a good things at a low voltage. but in a way you still wanna have the parallell capacitor on the resistor to increase gain and if you put the speaker in series with this you actually gonna get sound out of it as signal passes by but no DC. however I could go on forever what pros and cons you gonna have with each solution but the golden rule is still to test them and see how they really do. you already play the tubes at lower voltage than they want and it gonna change everything in the end so one solution may still not be correct even if it should be. experiments is what gives the progress of things.
Thank you for all the excellent information! And you're totally right, because I'm operating in such a weird region for the tubes, experimentation is really the only way to figure out what still sounds good. For the next OTL I build, I definitely want o work on getting the DC bias out of the speaker, so I'll definitely give these ideas a shot! I'm planning on using a handful of 6146 tubes triode strapped to move about twice as much current as the 12B4A here per tube. Granted, those need more than twice the filament current, so I suppose that's not unexpected, haha. It'll be interesting to see how they sound through a properly sized speaker, even if we're still at a really low voltage.
@@UsagiElectric no problem. I am happy to help and it is very fun to watch someone learning tubes like you do too. It is a great kind of learning. If you are interested you could get my discord if you want some help with anything :)
Great idea! I am also a fan of OTL amplifiers. Here are a few suggestions: (1) That tiny speaker has two er, three fatal flaws regarding OTL: (1) It is probably very low efficiency and of course (2) Very poor bass response. You might want to try something like an old 1960's large speaker, those were many times more efficient and therefore louder and less distorty at reasonable volume levels. (2) Put the speaker in a sealed box, that will greatly improve the bass. (3) Put an 8 ohm resistor or a 1 Henry choke from cathodes to ground and couple the speaker through a 470uF capacitor, that will allow the speaker to run centered and perhaps less distorty. (2) There is a huge impedance mismatch between the tubes and the speaker. Maybe look around for a 40 ohm voice coil speaker? Or put a transformer between cathodes and ground? Adequate choices might be a wall-wart 120V to 6V transformer. You will likely get like TEN TIMES the power to the speaker. I know this shivs the OTL principle, but it really isn't an "output transformer" per se and you will get TONS more output!!!! (3) Another work-around is to use a PA speaker, those come with included "70.7V" transformers riveted on, and that will be a much better impedance match to your tubes. True, there still is an "OT", but you didn't design it in! Fudgy, but so much louder. For comparison I made an OTL amplifier using two 6080 tubes. Without an OT I only got about 300 milliwatts, very weak. With an old Tektronix power transformer as OT I got 4 watts, quite enough for an efficient speaker to blast out adequate sound. Regards and great respects! -- George
Thank you! And thank you for the excellent suggestions! The speaker is definitely ill-suited for this application, but it was what I had on hand. The wood box made a huge difference to base response though and both volume and range nearly doubled when I put the speaker in place. I also did try to put a few different coupling caps in series when testing the design to remove any DC through the coil, but no matter what size cap I tried, it always decimated the sound. Like removed greater than 90% of volume. It could be that I just didn't get a large enough capacitor, but I think it mostly had to do with weird impedance mismatching and the insanely low voltages I'm running at. Which is what you mentioned next! The speaker is an 8 ohm speaker, and I think the 6x 12B4As in parallel have an impedance of around 200 ohms, which is just way too much. Keeping with the concept of pure OTL, I have a few ideas. Another commenter suggested using the White Cathode Follower design, which looks really promising, but I have to do a lot more reading to figure out what the output impedance would be using that design. Also, I happen to have 14x 6146s hanging about that I'm not doing anything with. Even as a standard cathode follower, the 14 of them in parallel would bring output impedance down to like 50 ohms, which with some speakers in series perhaps could get the impedance very close indeed! I know an output transformer is the right way to do it, and if I were building an amplifier to actually have a nice sounding amp to listen to, that is what I would do. But, I already happen to have a Magnavox Concert Grand from 1962 that has two 40W amplifiers using a cadre of 6V6s and sounds absolutely sublime! So, this OTL shenanigans is purely a learning exercise to see just what all I can accomplish with some brute force ideas!
@@UsagiElectric The choke to ground and 470uF capacitor to speaker *should* have worked. Maybe you used too small a capacitor? The 6146 tube, if you look at its plate curves, is not very peppy at low plate voltages and has a high plate resistance, so it's not going to be a great match to an 8 ohm speaker. You could use 16 of them in parallel, then the heaters could be run off 120 VAC, and the plates off a half-wave or voltage doubler, so you could eliminate the power transformer too!!. Be careful though to keep the cathodes going to the white wire (neutral) side of the AC power. A better choice would be some tubes designed for lower plate voltage, like the classic 6AS7 or its military brother the 6080 or the pepped-up 6336. Somewhere in my attic I have a case of 6080's and a brass chassis, for a long-delayed 20-tube OTL. Keep up the good work!
Oh another little detail-- it looks like the tubes are running at zero or just grid-leak bias-- that's kinda iffy for audio. It might sound better with like -9 volts bias on the grids from a 9-volt battery through a 1Meg resistor. That will center the tubes a bit better into their operating range. Grid-leak bias is more for Class C operation or for tubes designed for zero bias. Regards!
Did a little digging last night as I thought I recognized one of the Schematics, one of them was the Dickie Jr. , comes from an article I think from 1955. No iron at all on that one. fun stuff.
Good eye! That is indeed the Dickie Jr which I've had saved on my PC for a while now as inspiration. I would love to build a proper OTL copy of that design someday!
@@UsagiElectric Yeah a fairly simple design as it also eliminates the mains transformer as well, talk about a widower maker ;) 100% hot chassis. Not something you would like as you shy away from HV ;) Was thinking recreating this one myself but with four 6U10s.
You should try the 6С19П triode. It can pass a lot of current for it's heating. I tried to do a OTL with two of these in a SRPP configuration, a coupling capacitor and serial connected speakers. It kind of worked
Whoa, I hadn't heard of the 6С19П triode before, but you're right, that thing can move some serious current! My next plan was to actually use a bunch of 6146s triode strapped because even at 50V they can move something like 70mA (more than double the 12B4A). OTL is super interesting because it definitely requires a shift in thinking just to even get some sound out of it!
I'm curious of the output transformers in the original amplifier were a large part of the issues you encountered. They remind me of the tiny transformers used in early transistor applications (while the engineers versed in vacuum tubes were making the transitions to transistors). Relatively small number of windings. Small core size. Small flux density. This worked well for the voltages involved in transistor applications. The limitations might explain why the bass falls flat, etc. Your extremely low plate voltages seem to change several variables though. I have a bunch of 48v telecom rectifiers on-hand... I might have to give this all a try. 🙂
You're probably absolutely right! They were mystery transformers I bought on eBay years ago, so I have no idea what the winding ratio or even what they were originally designed for. Though, I'm fairly confident they weren't designed to take low voltage vacuum tube output, haha. "Low voltage vacuum tube audio amplifier output transformer" is an interesting field that I think might need a bit more research!
Thank you! I'm curious how long it'll actually go for. I've got a few more amplifier projects I want to build, so once I get something much better going, I may stress test this little guy to see just how long it'll survive for.
The speaker is shunting the Cathode resistor, and needs decoupling with a large capacitor. You can also increase the value of that resistor to limit Anode current whilst giving a higher AC voltage across the speaker, and therefore more volume whilst improving output tube Cathode life.
Thank you! I really love experimenting with tubes in way that not many others have before. I learn a ton about how the tubes react in weird situations, which gives me a lot of fun ideas about what to try next!
@@UsagiElectric Howdy. Yeah. I'm with You. The DC through the speaker is problematic. One idea is to use a really large inductor at the cathode to pass the DC and divert the AC through a large el. cap. to the speaker. With an 8 ohm speaker the inductor needs to be about 30 H for a transition frequency of 40 Hz. An enormous chunk of iron. If solid state components are tolerated one may use a constant current cathode load instead. That would be feasable. Or one could design a SRPP (Series Regulated Push Pull) output stage. Regards.
A wonderful hack. Try the other direction. Rather than decreasing output impedance with more tubes, try several speakers in series for more load impedance! Class B or A/B to get rid of the DC is also good.
I wonder if you could get good results from 3 12bh7 tubes, or even 2 6360 tubes. I'd love to see what you come up with when trying to scale the size of this down. I've been dreaming of making an opt guitar amp and this design seems like it could be a great start. Hell, I bet you could easily turn this into a great headphone amp too.
Different tubes have different naming conventions. Depends if it's named under RETMA, RMA, JAN, Mullard/Phillips, Eimac, EIA, etc... 4cx250 is an Eimac tube. In their convention 4 = Tetrode (3 for triode, 5 for pentode, etc). C = ceramic construction. X = external anode. 250 = 250w anode dissipation. Next letter indicates the variant (filament voltage, special use case, etc). A next number indicates amplification factor in the case of triodes. They also use a 4 digit EIA part number, and other manufacturers (RCA, etc) produced tubes identical to Eimac with this number.
I don't think you should have to explain yourself for the tube type readouts. Anyone who knows tubes surely understand what you're saying and who cares, say it as you like!
Thank you! The truth is that it did bother a lot of people, and so I figured I would just let them know that there is a logical reason behind why I said it the way I did, but I also figured I would have a bit of fun with the filming of it too. Filming that section was actually a blast!
Look at headphone amplifiers for ideas on how to construct an OTL amp. This design will eventually fry the speaker, since you have the full output tubes (idle) current going through 8 ohms (100k across 8 ohms is 8 ohms) speaker, and speakers do not like DC across them. Pluss, the output tubes are cathode biased, but with only 8 ohms from the cathode to ground, they have near zero bias, so the tube are basically turned fully on.
I forgot to mention that you could also try using two amps or at least two output stages in antiphase, wires across the speaker to double the voltage swing (which will double the currant), this should give you an extra 6db or so. And you forgot to ADD the toast rack 🙂
I actually did initially try building a push pull type amplifier with driver triodes before and after the speaker to get a full range voltage swing out of it, but it absolutely refused to work. It made some sound, but it was extremely faint and distorted beyond belief. That was when I said "Just go as simple as you can idiot" and fell back on the design here, which worked surprisingly well! Aside form the heat that is, then again, not many toasters also play music!
Funny you are going full circle. Starting in aviation and ending up in aviation. OTL was one way they cut down on size on early light aircraft radios. One schematic I have handy uses two triodes for the output stage. Of course you need a set of crystal earphones to use with them but the same effect. Might I suggest a by-pass cap on our output to cut down on the tinny sound. .001 uf. As well you might want to use self bias vs cathode for the first stage you will get better gain. Pardon the sacrilege the 'Champ' was not a good design to start with ;)
That's actually really awesome, I had no idea that they would use OTL in aircraft radios to cut down on size. That makes a ton of sense though because space is at an absolute premium in a cockpit. I actually did try a couple by-pass caps on the driver tube output into the speaker coil to cut down on DC bias across the speaker, but the impedance mismatching is so ferociously bad, it didn't work at all. I'll have to give the whole design a proper rethink. And while the Champ Amp may not be a very good design, boy it sure is simple, haha!
@@UsagiElectric Yeah impedance matching is a real key to any out-put stage. That is why they put 'Transconduntance' on data sheets. Of course that is only matching to an output transformer but I wonder if you could use in in a formula for a direct couple to a speaker. Have a nice time trying to find a speaker other than 4 or 8. There are 3.2 out there but few an far between. You might come up with something like 6.68 or alike ;)
Oooh, the 4-125 is a beefy tube! I love the idea of using transmitter tubes for completely different jobs. I actually have some 6146 beam power tubes that should dump out about 100mA at 50V when triode strapped, which is like triple what the 12B4A is rated for, so I would love to build another OTL with six or eight of those pushing enough power through a big speaker to rattle the windows! Thanks again and hello to Austria from Texas!
@@UsagiElectric I've commented a while ago that you could use a led voltage drop for negative grid bias. I've tested this in a circuit with a pl805 tube @170V. A power led between the tube and ground is enough to raise the cathode ~ 10v above ground level. This should work with direct heated tubes too. Just make sure the led get enough cooling :D
Vacuum Tube amplifier is one thing but important is to have a high quality DAC, suitable speakers and not to mention a quality sound source, ie some lossless format
If you're going for some proper, Hi-FI audio quality, you're absolutely right! Here, with a simple cell phone playing a compressed MP3 out of a tiny speaker, the audio quality was never going to be that great. But, if I'm being totally honest, I'm deaf in one ear so proper audio quality is often lost on me, so the main purpose of the build was to see what could be done at low voltage and as simplistic as possible!
Very nice project, although I would stick to the design with an output transformer, it's not too hard to the tubes. Instead of trying to find a suitable output transformer, you could use a speaker designed for 70 Volt or 100 Volt wiring, the kind of speaker used for playing background music and making announcements in public places. Such speakers have high-to-low impedance conversion transformer builtin, so it should be possible to connect them to the tube amplifier directly.
Thank you! And for sure, using an output transformer makes way more sense, but the idea of just brute forcing power into a speaker by running a bunch of tubes in parallel was just way too appealing to me for some reason. I think it's because my simple mind just says "More toob is good," haha!
philips came up with a transformerless output circuit using either 2 EL86/6CW5s, or a EL86(or UL84) and a EL84/6BQ5 , with a special 800ohm impedance speaker , capacitor coupled, unfortunately, as said in the video, it needed quite a high HT/B+ for 'good' output (one record player that used the circuit had a roughly 400v HT line!)
I've always wanted to build an OTL amp. My goal would be use big octal tubes as the outputs, and have enough that my paralleled plate impedance would match my speaker impedance. The trick then is to build 32+ ohm speakers.
Buy commercial grade speakers that have a 70 volt transformer and various wattage taps built in. JBL, Klipsch, Tannoy, QSC, Community, Atlas, and many more brands all make 70v speakers in all sizes and form factors for commercial use...
I just learnt about how tubes work 2 days ago and now I'm hooked. I want to make a tube preamp for a gainclone lm3886 amp. What do you think, I go with the 12AX7?
I'm not sure how good it would be to test that as it would be limited by the speaker and microphone capabilities wouldn't it? As it is, I had to crank the recording volume on my Tascam recorder way down to prevent the microphone from clipping. I did check the distortion of the two amplifier stages on the scope though. This is a 1V P-P signal being input, and the output signal that goes to the grids of the driver tubes is 12V P-P and has surprisingly little distortion. It gets just a slight bit of distortion at the bottom of the sine wave, but other than that, it was way cleaner than I was expecting! i.postimg.cc/rw6pH9R3/IMAG0463.jpg
It helps to more adequate judgement of voice quality, if spoken through also the qualities of loudspeakers, a resonance frequency meaning related to them and how this affects the imaging of soundtrack played. Mind, with original amp, with transformer output, You used really teeny-tiny speakers, which likely cut the frequencies heavily already from 500Hz downwards. The OTL however, I'd guess, had a speaker load with a resonance frequency nearer to 200 Hz and therefore also allowing better image of base. Try them both amps with a very simple card cone 6 to 8 inch full range speaker and the improvement will be immediate, despite the lowest price of the speaker. Also, you are handy (respect ) on woodwork, like it shows. This could lead to another exploration, related to word couple - Open Baffle. Keep up. Love your thinking and deeds. ;)
You're definitely right, it wasn't really a fair comparison of the two because the speakers were pretty dramatically different between them. I do really want to build another amp with a proper output transformer and speaker to see how good of sound I can get out of something compact. I have an idea of building a tube amp that sits on top of a wooden cylindrical speaker box with the speaker mounted on the bottom facing downwards. I think that could result in a really beautiful design with some work and it sounds like the perfect candidate for a proper output transformer amp!
@@UsagiElectric Oh... I am far from being ultimately right here. Didn't touch the issue with tiny output trannies. Small as these, they do limit the frequency delivery also. Neverthless, especially in guitar amp making, is not so critical always. Also, the whole side kick science, not recognized/accepted with true proud audiophools, goes about using power supply transformers for output. And these go for fraction of price compared to specialized output transformers. Can have lots of fun cutting/bending the corners in this area. ;)
Absolutely! I happen to have about 14x 6146 power pentodes, which when triode strapped move well more than double the current as the 12B4A at a given voltage. I'm thinking a bunch of those in parallel would be quite quite exciting!
If you put enough of them in parallel, I guess you could actually impedance match with a 16 ohm speaker. 4 16 ohm in series with 15 12B4 in parallel should match up pretty close.
Yup, even two of the 8 ohm speakers in series would have made a massive difference, but I only had one of the speakers unfortunately. For the next OTL I build, I think I'm going to use big 6146 tubes triode strapped. At 50V, they should move about 60 to 70mA. I have 15 of those in my collection, so if I run 14 of them in parallel, that should get me to about 50 ohms of impedance, which with 6x 8 ohm speakers in series is just about perfect! (Granted, I'll need 100+W of power just for the filaments, haha)
@@UsagiElectric Match all your filament currents and put them all in series. If you get your tube count right, you many end up with enough that you can run the heaters off line voltage and you won't need an A supply at all.
Yes, any output noise from the supply goes straight through, just like it does in a solid state unit. Use a high quality S.M.P.S. As far as RFI/EMI interference goes, keep the converter away from any transformers, as any magnetic interference WILL couple thru the transformers. Other than that, there's nothing different than there would be with a solid state design.
A higher voltage would definitely give better performance and volume, but that defeats the purpose here, which is to see what we can accomplish at ridiculously low voltages. By that same logic, I should have built the vacuum tube computer at higher voltages, but I think what makes it (and this OTL amplifier) really unique and fun is that it's at super low voltages! Nothing new on the 555 kit yet, Chinese New Year slowed some things down and I've had a nightmare of a time figuring out a reliable way to produce the legs. I have some more ideas to try, but I gotta recover some money first.
@@UsagiElectric true true in that case could you try building a otl with pl95 tubes I have like 10-30 of those around and I kind of wanna use them for something
Recently a famous audio engineer playing with some rather expensive speakers found that letting a 100 ma or so dc curent actualy lowered the speaker distortion by up to 50%. !
Hmmmm, I have a decent handful of pentode Mullard EL34 (that's 6CA7 to you fellas over the pond). I wonder how well they'd do as a parallel cathode follower final stage OTL 48V. The heaters on them are 1.5 Amp if my (aging) memory is correct so that'd be 56 Watts of heaters for a '6-pack'
The DC coupling definitely helps with the metal rock. The DC bias it puts on the speaker may or may not be making much of a difference. It depends on the capability of the speaker. Too much bias and it will bottom or top out the cone at the extremes of its travel well before it would do so without the bias. Using a choke in the vicinity of 1 henry (with the smallest resistance practical) as a direct load and connecting the speaker via a large capacitor in the vicinity of 1000 uF would eliminate the bias. The high value resistor from the cathodes to ground shouldn't make a difference at all given the low impedance of everything else. With powers this small "red plating" is not likely. Overstressing the tiny cathodes would be more likely.
Thank you for the excellent info! I actually did try to put a few different caps in series when testing the design to remove any DC through the coil, but no matter what size cap I tried, it always decimated the sound. Like removed greater than 90% of volume. It could be that I just didn't get a large enough capacitor, but I think it mostly had to do with the insanely low voltages I'm running at. The 12B4A was moving around 40mA through each tube, which at 48V should mean an internal resistance of about 1200 ohms per tube. With six in parallel that should be something like 200 ohms of series resistance with the speaker. I think the impedance is just so ridiculously mismatched that adding a series capacitor just wrecked whatever weird balance I had going on here. Or I could be totally off the mark, this is all uncharted territory for me!
@@UsagiElectric You'll need a sufficient DC load for the cathodes to get any beef from this amplifier from a capacitor isolated speaker if the speaker isn't the DC load (that 100K is a joke) That's why I suggested the choke. It's possible to get low resistance toroidal chokes today for ridiculous cheap prices. The choke will block AC, letting it be able to pass on all the better to the speaker. Mismatch is part of why you don't have the very best output possible. But it still should be possible to spare the speaker a DC bias. Remember the cathodes can pull (positive) but cannot push (negative). That's what a DC load is for.
I think you should try white cathode follower (WCF) next time. And I recommend you not to quit from output transformers, they can be funny part of designing tube amp, espesialy when you use tetrode or beam-tetrode/pentode where you cen make many diffrent circuit like ultra-linear or super pentode or even unity coupled like in McIntosh apms. 😉
Ooh, the white cathode follower is very, very cool! I'll definitely have to experiment with that one next time! I wouldn't say that I have given up on output transformers, they're definitely the right way to do it, but I've got the OTL bug now, I totally want to build a properly wicked OTL amplifier!
I have no clue how to go about checking something like that, but I did actually test input vs. output of the amplifier stages on the scope! This is a 1V P-P signal being input, and the output signal that goes to the grids of the driver tubes is 12V P-P and has surprisingly little distortion. It gets just a slight bit of distortion at the bottom of the sine wave, but other than that, it was way cleaner than I was expecting! i.postimg.cc/rw6pH9R3/IMAG0463.jpg
With that circuit, the only grid bias you're getting is from the current through the speaker! So the operating point is in a very nonlinear portion of the curve. If you were to put a DC voltmeter across any of those 1K resistors, I'm sure you'll find that the tubes are drawing grid current. You DON'T want THAT, in a Class A amp!
I've spent the last 40 years designing at 5v. 8 years ago I renovated 1932 Philco Cathedral, been doing tubes since. Jumping to 3, 4, 500 volts is a little scary but just listen to Alanis and keep your Hand in Your Pocket and you'll be safe. Thanks for the video.
Thanks for checking the video out! I actually started learning electronics in earnest with BEAM robots, which are even lower voltage than 5v since they're usually solar powered. The transition from that to tubes was an interesting one because BEAM stuff is all about being as absolutely efficient as possible and the average tubes heater consumes more power in a minute than a BEAM bot will in its entire life! I am slowly getting more comfortable working with higher voltages, particularly after learning and working on various HP equipment, but there's still a massively fun allure to working tubes at low voltages!
@@UsagiElectric yeah, I can see low voltage tube design being a real challenge. The important thing is having fun and learning, which you look like your having.
Numb(?) question - will tubes like 6bj6 and 6ak5 work “acceptably” at 48 volts, or perhaps a little more - perhaps as much as 72 volts? I got *tossed* by ~300 volts in my teens, (1 tube Amateur Radio transmitter) and I’m not fond of “high” voltages.
Thank you! It was a little disheartening at first to see all the negative comments about something as trivial as pronunciation, but in the end, I realized they weren't trying to be malicious, most were just trying to educate me on how things are conventionally done. And it's really easy for me to forget that very few people actually know my background or know what kind of weird muscle memory kicks in when talking to a camera all by yourself. But also, it was just a lot of fun to film that segment, haha.
Oh I don't know about getting rid of the output transformers.... Just speaking aesthetically.... you need output transformers to give it a "proper yesteryear feel". But I suppose 6 driver tubes is more tubes, so that's "old-timey" too. And it's certainly got some "go juice" in it! The whole thing with the filament voltage having to be a "correct number"... I wonder if that's why some people call logic ICs "seventy four L. S. twenty" where I'd always say "seven four L. S. two zero" (I'm dyslexic... that's kinda the same as talking to aircraft on a radio)
All the tubes on the OTL amplifier are 12B4A single triodes. For the older amplifier that we built a long time ago, I use 12BH7 dual triodes for amplification and 12B4As to drive the output transformer.
In German, the hundreds, ones and then the tens are usually mentioned. But when I'm alone, I say the numbers number by number: hundreds, tens, ones. In English, only the numbers up to nine-teen are twisted. So teen-nine, than twenty-one.
I'm not sure I'm following you here? Thermistors are usually used for limiting inrush current, right? But, the "red-plate" potentiometer is just controlling the voltage level on the grids of the output tubes, there shouldn't be any appreciable current flowing through there at all. As for powering the tube filaments from the power source, in this case it didn't really seem to be a big issue, the little 48V supply I'm using seemed to handle the inrush with no problems!
I enjoy your computer stuff but this low voltage no transformer audio amplifier is super cool. Why would anybody build such a good looking vacuum tube low voltage amplifier? Because you can. Stan Paddock
I don't think the output tubes are getting much negative bias with only an 8 ohm speaker in the cathode circuit. What is the voltage from cathode to grid?
The cathode followers should essentially just put the cathode at the same level as the grid voltage. However, with just an 8-ohm speaker, there's all sorts of horrific impedance mismatching going on. It's actually pretty surprising it sounds as good as it does.
@@UsagiElectric There is an impedance mismatch like most OTL amplifiers. Looking at the curves, the plate current is not too far off from proper class A operation. The curves are hard to read at only 48 volts. The negative feedback is good to reduce distortion but some audiophiles would say the sound is not as clear. I would like to know the cathode to grid voltage to study this further. Also could you measure the resistance of the speaker with an ohmmeter to help calculate the DC operating condition.
Usagi, I don’t want to open a can of worms for you, but have you ever considered re-winding your speaker voice coils? Since there’s an impedance mid-match from the tube output to an 8 ohm speaker, then change the speaker to high impedance. Say, 1K? If I remember correctly, there was an ancient article in “Hi-Fi Magazine” about doing this, and they reported amazing response from the speakers when not running through an output transformer.
Always follow the rules. I before E, except when your foreign neighbor Keith receives eight counterfeit beige sleighs from feisty caffeinated weightlifters.
Oooh, I don't know if I have any directly heated triodes hanging around to test that out with. Most directly tubes run at like 1V or 2V though, right? I suppose with big batteries on them you can get all the hum out, but that feels a bit like cheating, haha.
In my initial testing, I ran it as low as 24V with just three output tubes and it still put sound through the speaker. It was very loud, but it had enough volume to be able to hear what was going on. With more output tubes, I reckon it'll put some decent sound through the speaker. But, the margins on distortion from the amplifier stages get a lot tighter. Still, could totally be done!
Hi, wonderfull and I would like to build a replica myself. However I am a complete beginner. Can you please give me hint how to connect the tubes. What I see is that they have 9 pins but in your schematics there are only 3 pins connected? Shurely I missed some background on tubes ..... Can you please give me some support.
The speaker doesn't like DC through the coil. That is constant heat dissipation that gets overlaid on whatever AC power from reproducing music. It needs to be coupled through a capacitor (like 1000uF, larger is better) and the quiescent current has to be at least Vcc/speaker impedance 😉
Yeah, no decoupling capacitor is the first thing I noticed! Depending on exact impedance of the loudspeaker at low frequency (assuming it is near 8 ohm) that tiny speaker isn't going to be able to produce much on the low frequency side so a 200 microfarad capacitor should be plenty for a roughly -3 dB roll off of around 100 Hz. Still results aren't too bad considering the DC. What value resistor would you use on the cathode?
@@AllThingsOnePlace I wouldn't use a resistor, active load would be better. After thinking again about it, you would need somewhat less resistance than speaker impedance.
A better way would be to use an inductor (but a very large value, maybe a classic fluorescent lamp balast?) on the plate and setup the bias point with negative voltage on the grid. outputs Would still be coupled through a capacitor. Should allow for 4 times more output power with only twice higher quiescent current?
@@akkudakkupl Yeah, giant inductor would be best for performance, wouldn't really fit the project though. It isn't going to be efficient ever since its a class A amp but yeah squeezing as much of that AC into the loudspeaker as possible is one way to go.
I say resistor simply to match the project or make it work within the confines of the current project. Swap the 100k for a much lower value (has to meet power requirements of course), add capacitor to the loudspeaker (possibly hide it inside the already made enclosure)? I guess could put a lamp ballast under the whole box and route the wires around it too.
Either way, I am very new to tubes and slowly learn a little on each video I watch.
I actually did try to put a few different caps in series when testing the design to remove any DC through the coil, but no matter what size cap I tried, it always decimated the sound. Like removed greater than 90% of volume. It could be that I just didn't get a large enough capacitor, but I think it mostly had to do with the insanely low voltages I'm running at.
The 12B4A was moving around 40mA through each tube, which at 48V should mean an internal resistance of about 1200 ohms per tube. With six in parallel that should be something like 200 ohms of series resistance with the speaker.
I think the impedance is just so ridiculously mismatched that adding a series capacitor just wrecked whatever weird balance I had going on here.
Or I could be totally off the mark, this is all uncharted territory for me!
@@UsagiElectric it was because the tubes have to run current at least equal to VCC/speaker impedance anyway, so it can swing up or down. This is why class A is ineficient.
What a cool journey through non-conventional tube amp design! Just today I was studying cathode - followers for guitar amps and it is so cool to see them drive a speaker directly!
Thank you for such a wonderful content!
Thank you so much!
It's interesting how many different ways there are to tackle tube amplifier design! OTL amps really appeal to me, so this definitely won't be the last one I build. There's a ton of place I can improve on the design and get even better sound out of it. Plus, I have a handful of 6146 Power tubes that when triode strapped should be able to move about three times as much current as the 12B4A, which should be pretty exciting!
Your construction techniques are one of my favorite parts of your videos, always so clever and resourceful. Mounting the pots that way with the header pins is a great idea to take the stress off the joints.
Thank you so much!
I actually really struggled with a good way to mount the potentiometers using a single sided board. I discovered that If I give the header pins a slight bend, they slip right into the pot connection holes on the back side making it super easy to solder up, but also adding some strength to the mounting.
I'm really happy with how this one turned out visually, it's a bit dense and runs way too hot, but it looks awesome!
Really enjoyed watching this.
Thank you so much!
Your absolutely right when naming vacuum tube numbers.
Back around 2014 while in college I built my own 4 channel OTL surround sound amp and used +/- 60V rails for the output (needed -120V output bias and something like 350V for the input gain stages). With 4 6AS7 outputs per channel (8 in stereo mode) it's plenty loud for a moderate sized room on quadraphonic mode, and rattle the windows and walls loud in stereo. It's clearer than conventional tube and SS amps too...It's really interesting to hear a tube amp deliver razor sharp synth from prog rock while at the same time being able to render crisp punchy DEEP bass from say an action film with gun shots and explosions, and get things like piano and vibraphone right to a surprising level of realism.
You should try collecting vacuum tube TVs from the 60's. Once you get used to working with and surviving regularly getting shocked by 25KV, voltages over 60V stop being scary, and become pains that you learn to avoid touching.
I'm sure that if you connect the amp to a 10" speaker it will sound even better! And of course with a two way, nice project as always!
And don't forget the water cooling
You are right, would be a better test with a high efficiency full range hi-fi speaker. Bigger speaker with the extra bass content takes a ton more current to drive. You should use a capacitor to protect the speaker from any DC, I can't remember if there are any caps in the pre speaker circuit? No there aren't. 🙂 Also restrict the bass frequencies to the audible range by trying different (smaller) cathode bypass cap values in the preamp stages, 100uF seems flabby. Saves wasting energy driving sub audible bass.
Try losing or a bigger resistor in the neg feedback loop to the first preamp? Might drive a real speaker?? Exciting stuff.
Yes, actually pro audio speakers like the brand Faital Pro have 96, 98db that will work great on low power amps that needs high efficency. I've tried that with my 1w diy tube amp, also compression drivers have 105db minimum in most of the cases
But maybe is better the idea of a full range as you mentioned, so you don't have to worry about pairing with a driver, crossover..
Thank you!
I really wanted to stuff it all in that tiny footprint, but in the future, I'm definitely going to spread things out and run more speaker for some proper full on sound!
I happen to have a handful of 6146 power pentodes, that when triode strapped can move well more than double what the 12B4A can at a given voltage, so some of those mixed with a bigger speaker and maybe some proper design to eliminate DC bias and I think we can get something really wild going on!
This is quickly becoming my favourite channel!
Thank you so much!
Very cool as always 😊
Thank you very much!
bnuuuuy!
I'm glad you gave the explainer rant because I had no idea what the numbers meant. So I learned a thing today!
I can watching your vacuum tube videos for hours.
Are you a metalhead?
Man you just keep getting cooler every video.
Amazing work as always!
Thank you so much!
That depends a bit on what you define as metal. I always listen to music when I work, but as a translator, I can't listen to music with lyrics or my translations end up... strange, haha. So, over the past few years I've gotten big into different types of instrumental metal. Mick Gordon's work on the Doom and Doom Eternal soundtracks are perfect, and Andrew Hulshult and David Levy both knocked it out of the park with The Ancient Gods soundtrack. I listen to all three of those on the regular. Hulshult's work in general is really fantastic (both the Dusk and Prodeus soundtracks get a lot of playtime too). If I need something a little less... violent, I tend to fall back on Animals as Leaders, the Algorithm, or some instrumental Djent (which is actually what's playing in the background right now).
Instrumental metal is a bit tough to find, but on the occasions some gems show up, like some of Meshuggah's stuff!
@@UsagiElectric
Maybe you want to take a look at Master boot records on youtube, it has some Doom vibes to it and is purely instrumental :)
man i love this channel!
Thank you so much!
Just stumbled on your channel great job thanks for the video
I love this so much :) ha! makes me smile! especially when the volume went up :D
Thank you so much!
It turned out better than I could have hoped!
It puts out a ton of volume and still sounds excellent!
Wow that is sooo cool. I have a tube amplifier that I started restoring a few years ago and I kind of lost my forward motion on it. watching your vids really gets me psyched to work on it again.
Thank you so much!
And that's awesome to hear that you're getting fired up to work on it again! If you get going, let me know, I'd love to see some pictures or a video of it running!
@@UsagiElectric Absolutely. I have a question can I use a 1b3 (13b? rectifier valve) to perform a full wave bridge or do I need four of them? I believe I need to rectify 2000 volts or 5000 volts at a high frequency such as 200khz. what will turn that sine wave into a unipolar pulse train?
@@johnkoury1116 That's a lot of voltage!
The 1B3 is just a half-wave rectifier, so you would need two of them to do full-wave rectification if you have a center tapped transformer. If your transformer only has two taps, then you would need four to get full-wave rectification. Though, with such a high frequency, half-wave rectification and a good HV cap should be totally fine. The HP 150A actually uses a 1X2B as a half-wave rectifier with a 1500pF smoothing capacitor for it's HV generation.
Definitely check out some old HP and Tektronix oscilloscope manuals for ideas on how they did HV rectification!
Page 63 in this one: hparchive.com/Manuals/HP-150A-AR-Manual.pdf
Page 114 in this one: w140.com/tekwiki/images/3/3d/070-198.pdf
@@UsagiElectric I am only going to be rectifying 3,000 volts but it is at a high frequency. I want to capture at a nanosecond level the energy from a spark gap. I have silicon rectifiers that will capture at 100nanoseconds but it takes some time to conduct from what I understand. I have been reading and it seems that the 1b3 because of the grid heater would be ready to conduct almost instantaneously. I have seen setups where they have used four of these valves together but I have to delve into it much further. And yes I want to fill a capacitor.
@@UsagiElectric Thank you so much. This is awesome information!! I believeI will have to study more before i can even understand it...
absolutely one, two, victor seven. let them squirm in the floor and rithe in anger, you be you keep up the amazing work. I am mad RUclips didn't recommend your channel sooner.
Pretty cool
Think I am happy with output transformers.
Wood looks beautiful, tung oil is awesome. Gun stock oil is nice too.
Output transformers definitely make way more sense!
But, there was just something super appealing about the idea of brute forcing sound through a speaker. I think it's because my simple brain just says "more toob is good," haha.
The tung oil made such a massive difference with the wood, I was blown away with how nice it came out. I never thought of using gun oil though, that's a brilliant idea!
I love it. At 48 volt what a surprise.
Thank you!
Did make a mosfet class A amplifier a few years ago, used a few parallel mosfets, and an iron core transformer to provide the DC removal for the speaker. Could adjust the bias to go from full class A, and a really hot set of mosfets to something like AB, where at least the current was under 1A per transistor. Heatsinks were old Pentium CPU coolers, one per transistor, and an old mains transformer with 12-0-12VAC secondary, with primary disconnected. Loud on 12V.
I have two push pull amplifiers built with fets and power transformers that are literaly 2x115V to 36V. Fun to build, guital sounds great through them.
That's awesome!
I love the idea of using old CPU coolers are heat sinks for the mosfets too, that's brilliant because those can often be found for cheap or free, and if you get really desperate, they already have provisions for cooling fans on them!
@@UsagiElectric I did use the fans, just connected in series so they would be quieter, running all of 12VDC.
@@UsagiElectric I have used them a lot for power supplies, as you can easily mount a few TO220 or TO247 power devices onto them with no problem, just drill and tap the M2 holes for the hardware to mount them. Some with no insulation, because the heatsink was easier to mount on isolating bushes, and I have done a number of LED lamps using the round CPU heatsinks, they are really good for that to mount a star PCB on. For that easiest is to use 2 part metal loaded epoxy, you can use the thermal stuff, but for me just the regular Pratley quickset steel worked well.
You built a V8 racket machine. Congrats! Keep up the cool engineering work :)
And I agree that this amp could use ceramic tube bases, possibly on a chassis rather than PCB.
Thank you Keri!
It means a ton to hear encouraging words from you after seeing your phenomenal amp projects!
Ceramic bases and a bit more space between the tubes would have made a massive difference. For the next OTL I build, I think we'll throw space restraints out and just let the amp get big!
Great! I love OTLs as a concept - they are the peak of inefficiency in tube audio (of course), but with a simplistic charm of their own.
Not critical here, but it's generally considered bad practice to bias a tube from the wiper on a pot. The reasoning is that the connection could become noisy or intermittent over time which would clearly upset the DC bias conditions and act as a noise source. A simple solution is to add a single resistor from the wiper / grid to ground. This could be 1M, or even as much as 4.7M or so. The grid input current is extremely low, so a high resistor value will not affect the DC conditions, and will also allow a high value pot, permitting a high input impedance to the stage if needed.
Thank you!
It's a completely improper use of vacuum tubes for audio, but I think that's why I find it so fascinating!
Thank you for the information on the additional resistor from the grid to ground. I never thought about the potentiometer potentially causing biasing issues, I'll definitely slip that in on future audio builds!
If you're really interested in building 2 amps getting RCA receiving in the back part there's charts but tell you at which voltage to use which resistor value on which pin to get max output without distortion. You can adjust these values to get more or less gain ad needed.
You do you! Read it off however you'd like!
Thank you!
Mostly though, filming that section was a lot of fun, haha.
You could use a series synthetic inductor to set the DC bias and tap off the audio with a large capacitor. A suitable "gyrator" circuit can be made from a single p-type FET.. ..which suggests that it could be done with a (suitably low impedance) tube as well.
That's definitely a bit beyond my skill level!
But, I do have some 6146s that are quite a bit lower impedance than the 12B4A, so I think for the next OTL journey will involve those big boys!
@@UsagiElectric Consider using an inverted triode arrangement - plate at large negative bias becomes a control electrode with very little input capacitance, grid at positive bias becomes the output, cathode is handled as usual. The output impedance is much lower (nothing in the way of the electrons other than weak repulsion from the other side of the grid), at the expense of inverted voltage gain. However, the input still draws almost no current, so a tiny step-up signal transformer with a ratio similar to the "normal" gain of the tube can correct for this.
The proper output transformer is going to match your impedance where you will get a very flat response and the warmth and bottom end will increase greatly. But that litle speaker is not going to demonstrate that as well as a bigger driver will.
Yep, you said it.
You are into cool stuff. I had a ton of 12B4’s 10 years ago to build a HV OTL amp but never did.
If going for tough stuff you can use the 12B4 in a simple amp that is powered by a HV SMPS supplying some 170-200V. Or if going minimalistic, use a 6AS7G (or better still a 5998A) which is a dual power triode and a 12AT7 or similar for a 2-tube stereo amp.
Requires beefier psu and output transformer but is an absolute killer.
ngl, this is pretty bussin.
Thank you!
I wouldn't put DC thru the speaker at the cathode, at the very least in this topology put a large capacitor in series with the speaker. The DC conditions at the cathode will be maintained and only the AC will pass thru to the speaker.
Nope, you would have a 100k output impedence, check out class A amplifiers
@@fabriziobrutti1205 You have no idea what you are talking about. Maybe *YOU* should go lookup what it means to bypass a cathode at AC. The worst thing about your comment is total lack understanding of what would happen to a low impedance voice call having DC running through it.
@@vincei4252 actually i didn't get your answer about the DC through the voice coil(?)... I know that there mustn't be any, or otherwise the speaker will get some damage, so a capacitor must be put in series, but by what I know the 100k resistance should be replaced with an 8 Ohm resistance... Anyway I'll check the part about bypassing a chatode at AC, since I don't know much about vacuum tubes (btw I'm just 16)😅
Yes! It is very possible to split the DC and AC paths of the cathode signal.
@@fabriziobrutti1205 That's fine. In a nutshell at AC, provided the capacitor is big enough, it looks like a short and the impedance seen at the cathode to AC is the impedance of the speaker. If you look at transistor circuits you will see something called emitter degeneration where a capacitor is placed across the emitter resistor or portion of the total emitter resistance. Doing this has the magical effect of increasing the gain of a common emitter amplifier because the AC impedance seen at the emitter is lower but the DC conditions at the emitter are maintained for biasing purposes. Look it up it's interesting stuff. Good luck in your journey.
Unless I'm having a stroke, putting 4 speakers in series should give you four times the volume.
I use proteus to simulate circuits before making my tube amp. and it works in first attempt.
As much as I enjoy OTLs, output transformers are not the villains that you suppose them to be. Being made up of inductor sections, transformers can and do increase signals dynamic behavior. In fact, they are a fundamental tool in that they allow us to shape the final timbre the amp is going to present. People wrongly believe the more straightforward a circuit is, the better. But this is not always true.Sometimes a more complex circuit carrying transformers means more harmonic richness, which is a good thing in musical terms. And let us not forget that harmonic richness is that which musicians are looking for whenever they overdrive their guitar tube amps!
الله ينور عليك يا هندسة
Thank you!
First digits are the heater voltage, the letters are a factory designation and the last number is the number of useful elements in the tube, for tubes that use that numbering convention.
But that doesn't take into account ECC34 KT88 and other tube naming conventions.
I don't get upset by it though. I know that things are read a certain way over the radio. I often use phonetic alphabet when reading things out
As many pointed out is the DC you get through the speaker in this design an issue that will affect the sound and generate a fair amount of heat in the voice coil. But it is a fun project anyway. I would probably gotten for a big capacitor in series with the speaker and then match the resistor in parallell with that for decoupling but I know some gonna say a choke is better but that depends how much bias you wanna have the tube at. many would say just omit the resistor and put the choke there instead but you also gonna remove any bias of the tube and this design is cathode biased which gonna need that resistor to work, the same thing is actually happening with that low resistance speaker across it anyway so there will be barely no bias at all which might as well be a good things at a low voltage. but in a way you still wanna have the parallell capacitor on the resistor to increase gain and if you put the speaker in series with this you actually gonna get sound out of it as signal passes by but no DC. however I could go on forever what pros and cons you gonna have with each solution but the golden rule is still to test them and see how they really do. you already play the tubes at lower voltage than they want and it gonna change everything in the end so one solution may still not be correct even if it should be. experiments is what gives the progress of things.
Thank you for all the excellent information!
And you're totally right, because I'm operating in such a weird region for the tubes, experimentation is really the only way to figure out what still sounds good.
For the next OTL I build, I definitely want o work on getting the DC bias out of the speaker, so I'll definitely give these ideas a shot! I'm planning on using a handful of 6146 tubes triode strapped to move about twice as much current as the 12B4A here per tube. Granted, those need more than twice the filament current, so I suppose that's not unexpected, haha. It'll be interesting to see how they sound through a properly sized speaker, even if we're still at a really low voltage.
@@UsagiElectric no problem. I am happy to help and it is very fun to watch someone learning tubes like you do too. It is a great kind of learning. If you are interested you could get my discord if you want some help with anything :)
Great idea! I am also a fan of OTL amplifiers. Here are a few suggestions:
(1) That tiny speaker has two er, three fatal flaws regarding OTL: (1) It is probably very low efficiency and of course (2) Very poor bass response. You might want to try something like an old 1960's large speaker, those were many times more efficient and therefore louder and less distorty at reasonable volume levels. (2) Put the speaker in a sealed box, that will greatly improve the bass. (3) Put an 8 ohm resistor or a 1 Henry choke from cathodes to ground and couple the speaker through a 470uF capacitor, that will allow the speaker to run centered and perhaps less distorty.
(2) There is a huge impedance mismatch between the tubes and the speaker. Maybe look around for a 40 ohm voice coil speaker? Or put a transformer between cathodes and ground? Adequate choices might be a wall-wart 120V to 6V transformer. You will likely get like TEN TIMES the power to the speaker. I know this shivs the OTL principle, but it really isn't an "output transformer" per se and you will get TONS more output!!!!
(3) Another work-around is to use a PA speaker, those come with included "70.7V" transformers riveted on, and that will be a much better impedance match to your tubes. True, there still is an "OT", but you didn't design it in! Fudgy, but so much louder.
For comparison I made an OTL amplifier using two 6080 tubes. Without an OT I only got about 300 milliwatts, very weak. With an old Tektronix power transformer as OT I got 4 watts, quite enough for an efficient speaker to blast out adequate sound.
Regards and great respects!
-- George
Thank you!
And thank you for the excellent suggestions!
The speaker is definitely ill-suited for this application, but it was what I had on hand. The wood box made a huge difference to base response though and both volume and range nearly doubled when I put the speaker in place.
I also did try to put a few different coupling caps in series when testing the design to remove any DC through the coil, but no matter what size cap I tried, it always decimated the sound. Like removed greater than 90% of volume. It could be that I just didn't get a large enough capacitor, but I think it mostly had to do with weird impedance mismatching and the insanely low voltages I'm running at.
Which is what you mentioned next! The speaker is an 8 ohm speaker, and I think the 6x 12B4As in parallel have an impedance of around 200 ohms, which is just way too much. Keeping with the concept of pure OTL, I have a few ideas. Another commenter suggested using the White Cathode Follower design, which looks really promising, but I have to do a lot more reading to figure out what the output impedance would be using that design. Also, I happen to have 14x 6146s hanging about that I'm not doing anything with. Even as a standard cathode follower, the 14 of them in parallel would bring output impedance down to like 50 ohms, which with some speakers in series perhaps could get the impedance very close indeed!
I know an output transformer is the right way to do it, and if I were building an amplifier to actually have a nice sounding amp to listen to, that is what I would do. But, I already happen to have a Magnavox Concert Grand from 1962 that has two 40W amplifiers using a cadre of 6V6s and sounds absolutely sublime! So, this OTL shenanigans is purely a learning exercise to see just what all I can accomplish with some brute force ideas!
@@UsagiElectric The choke to ground and 470uF capacitor to speaker *should* have worked. Maybe you used too small a capacitor? The 6146 tube, if you look at its plate curves, is not very peppy at low plate voltages and has a high plate resistance, so it's not going to be a great match to an 8 ohm speaker. You could use 16 of them in parallel, then the heaters could be run off 120 VAC, and the plates off a half-wave or voltage doubler, so you could eliminate the power transformer too!!. Be careful though to keep the cathodes going to the white wire (neutral) side of the AC power. A better choice would be some tubes designed for lower plate voltage, like the classic 6AS7 or its military brother the 6080 or the pepped-up 6336. Somewhere in my attic I have a case of 6080's and a brass chassis, for a long-delayed 20-tube OTL. Keep up the good work!
Oh another little detail-- it looks like the tubes are running at zero or just grid-leak bias-- that's kinda iffy for audio. It might sound better with like -9 volts bias on the grids from a 9-volt battery through a 1Meg resistor. That will center the tubes a bit better into their operating range. Grid-leak bias is more for Class C operation or for tubes designed for zero bias. Regards!
Did a little digging last night as I thought I recognized one of the Schematics, one of them was the Dickie Jr. , comes from an article I think from 1955. No iron at all on that one. fun stuff.
Good eye!
That is indeed the Dickie Jr which I've had saved on my PC for a while now as inspiration. I would love to build a proper OTL copy of that design someday!
@@UsagiElectric Yeah a fairly simple design as it also eliminates the mains transformer as well, talk about a widower maker ;) 100% hot chassis. Not something you would like as you shy away from HV ;) Was thinking recreating this one myself but with four 6U10s.
You should try the 6С19П triode. It can pass a lot of current for it's heating. I tried to do a OTL with two of these in a SRPP configuration, a coupling capacitor and serial connected speakers. It kind of worked
Whoa, I hadn't heard of the 6С19П triode before, but you're right, that thing can move some serious current! My next plan was to actually use a bunch of 6146s triode strapped because even at 50V they can move something like 70mA (more than double the 12B4A).
OTL is super interesting because it definitely requires a shift in thinking just to even get some sound out of it!
Whoa! That makes me wanna try using that little bugger with a 100R speaker I got recently...
I'm curious of the output transformers in the original amplifier were a large part of the issues you encountered. They remind me of the tiny transformers used in early transistor applications (while the engineers versed in vacuum tubes were making the transitions to transistors). Relatively small number of windings. Small core size. Small flux density. This worked well for the voltages involved in transistor applications. The limitations might explain why the bass falls flat, etc. Your extremely low plate voltages seem to change several variables though. I have a bunch of 48v telecom rectifiers on-hand... I might have to give this all a try. 🙂
You're probably absolutely right! They were mystery transformers I bought on eBay years ago, so I have no idea what the winding ratio or even what they were originally designed for. Though, I'm fairly confident they weren't designed to take low voltage vacuum tube output, haha.
"Low voltage vacuum tube audio amplifier output transformer" is an interesting field that I think might need a bit more research!
That's awesome. It'll be short lived, but AWESOME!!! :)
Thank you!
I'm curious how long it'll actually go for. I've got a few more amplifier projects I want to build, so once I get something much better going, I may stress test this little guy to see just how long it'll survive for.
@@UsagiElectric That should be fun. Ha ha.
The speaker is shunting the Cathode resistor, and needs decoupling with a large capacitor. You can also increase the value of that resistor to limit Anode current whilst giving a higher AC voltage across the speaker, and therefore more volume whilst improving output tube Cathode life.
Howdy. Nice.
I like Your persistence of vision and thinking outside the box. Makes me want to try using the beasts KT88, 6V6 or 6L6.
Regards.
Thank you!
I really love experimenting with tubes in way that not many others have before. I learn a ton about how the tubes react in weird situations, which gives me a lot of fun ideas about what to try next!
@@UsagiElectric Howdy. Yeah.
I'm with You. The DC through the speaker is problematic.
One idea is to use a really large inductor at the cathode to pass the DC and divert the AC through a large el. cap. to the speaker. With an 8 ohm speaker the inductor needs to be about 30 H for a transition frequency of 40 Hz. An enormous chunk of iron.
If solid state components are tolerated one may use a constant current cathode load instead. That would be feasable.
Or one could design a SRPP (Series Regulated Push Pull) output stage.
Regards.
A wonderful hack. Try the other direction. Rather than decreasing output impedance with more tubes, try several speakers in series for more load impedance! Class B or A/B to get rid of the DC is also good.
the first letter is filament voltage category
An interesting hack amp, would be good to see two banks of output tubes, for push-pull, without running DC through the speaker.
I wonder if you could get good results from 3 12bh7 tubes, or even 2 6360 tubes. I'd love to see what you come up with when trying to scale the size of this down. I've been dreaming of making an opt guitar amp and this design seems like it could be a great start. Hell, I bet you could easily turn this into a great headphone amp too.
Different tubes have different naming conventions. Depends if it's named under RETMA, RMA, JAN, Mullard/Phillips, Eimac, EIA, etc... 4cx250 is an Eimac tube. In their convention 4 = Tetrode (3 for triode, 5 for pentode, etc). C = ceramic construction. X = external anode. 250 = 250w anode dissipation. Next letter indicates the variant (filament voltage, special use case, etc). A next number indicates amplification factor in the case of triodes. They also use a 4 digit EIA part number, and other manufacturers (RCA, etc) produced tubes identical to Eimac with this number.
In theory, having a few loudspeakers hooked up in series can pull up the impedance to better match the output impedance of the valve output stage...
I'll have to try that. Thanks!
I don't think you should have to explain yourself for the tube type readouts. Anyone who knows tubes surely understand what you're saying and who cares, say it as you like!
Thank you!
The truth is that it did bother a lot of people, and so I figured I would just let them know that there is a logical reason behind why I said it the way I did, but I also figured I would have a bit of fun with the filming of it too. Filming that section was actually a blast!
Look at headphone amplifiers for ideas on how to construct an OTL amp.
This design will eventually fry the speaker, since you have the full output tubes (idle) current going through 8 ohms (100k across 8 ohms is 8 ohms) speaker, and speakers do not like DC across them.
Pluss, the output tubes are cathode biased, but with only 8 ohms from the cathode to ground, they have near zero bias, so the tube are basically turned fully on.
I forgot to mention that you could also try using two amps or at least two output stages in antiphase, wires across the speaker to double the voltage swing (which will double the currant), this should give you an extra 6db or so.
And you forgot to ADD the toast rack 🙂
I actually did initially try building a push pull type amplifier with driver triodes before and after the speaker to get a full range voltage swing out of it, but it absolutely refused to work. It made some sound, but it was extremely faint and distorted beyond belief. That was when I said "Just go as simple as you can idiot" and fell back on the design here, which worked surprisingly well!
Aside form the heat that is, then again, not many toasters also play music!
Funny you are going full circle. Starting in aviation and ending up in aviation. OTL was one way they cut down on size on early light aircraft radios. One schematic I have handy uses two triodes for the output stage. Of course you need a set of crystal earphones to use with them but the same effect. Might I suggest a by-pass cap on our output to cut down on the tinny sound. .001 uf. As well you might want to use self bias vs cathode for the first stage you will get better gain. Pardon the sacrilege the 'Champ' was not a good design to start with ;)
That's actually really awesome, I had no idea that they would use OTL in aircraft radios to cut down on size. That makes a ton of sense though because space is at an absolute premium in a cockpit.
I actually did try a couple by-pass caps on the driver tube output into the speaker coil to cut down on DC bias across the speaker, but the impedance mismatching is so ferociously bad, it didn't work at all. I'll have to give the whole design a proper rethink.
And while the Champ Amp may not be a very good design, boy it sure is simple, haha!
@@UsagiElectric Yeah impedance matching is a real key to any out-put stage. That is why they put 'Transconduntance' on data sheets. Of course that is only matching to an output transformer but I wonder if you could use in in a formula for a direct couple to a speaker. Have a nice time trying to find a speaker other than 4 or 8. There are 3.2 out there but few an far between. You might come up with something like 6.68 or alike ;)
Ah no one two be for a is just fine. 👌🏻 we’ll understand.
Thank you!
Honestly, it's kind of become a running joke among friends now, we find it quite funny!
4-125 tubes are rated for ~300 ma. You should try one of those. ;) I like your work! Great project! Greetings from Austria.
Oooh, the 4-125 is a beefy tube! I love the idea of using transmitter tubes for completely different jobs.
I actually have some 6146 beam power tubes that should dump out about 100mA at 50V when triode strapped, which is like triple what the 12B4A is rated for, so I would love to build another OTL with six or eight of those pushing enough power through a big speaker to rattle the windows!
Thanks again and hello to Austria from Texas!
@@UsagiElectric I've commented a while ago that you could use a led voltage drop for negative grid bias. I've tested this in a circuit with a pl805 tube @170V. A power led between the tube and ground is enough to raise the cathode ~ 10v above ground level. This should work with direct heated tubes too. Just make sure the led get enough cooling :D
Vacuum Tube amplifier is one thing but important is to have a high quality DAC, suitable speakers and not to mention a quality sound source, ie some lossless format
If you're going for some proper, Hi-FI audio quality, you're absolutely right!
Here, with a simple cell phone playing a compressed MP3 out of a tiny speaker, the audio quality was never going to be that great. But, if I'm being totally honest, I'm deaf in one ear so proper audio quality is often lost on me, so the main purpose of the build was to see what could be done at low voltage and as simplistic as possible!
@@UsagiElectric In that case, you did very well and I understand the reasons that led you to do so.
Thanks and good luck ;-)
Very nice project, although I would stick to the design with an output transformer, it's not too hard to the tubes.
Instead of trying to find a suitable output transformer, you could use a speaker designed for 70 Volt or 100 Volt wiring, the kind of speaker used for playing background music and making announcements in public places. Such speakers have high-to-low impedance conversion transformer builtin, so it should be possible to connect them to the tube amplifier directly.
Thank you!
And for sure, using an output transformer makes way more sense, but the idea of just brute forcing power into a speaker by running a bunch of tubes in parallel was just way too appealing to me for some reason. I think it's because my simple mind just says "More toob is good," haha!
Agreed! OTLs are more of a seductive theoretical concept than a worthwhile audible result in practical terms.
philips came up with a transformerless output circuit using either 2 EL86/6CW5s, or a EL86(or UL84) and a EL84/6BQ5 , with a special 800ohm impedance speaker , capacitor coupled, unfortunately, as said in the video, it needed quite a high HT/B+ for 'good' output (one record player that used the circuit had a roughly 400v HT line!)
I've always wanted to build an OTL amp. My goal would be use big octal tubes as the outputs, and have enough that my paralleled plate impedance would match my speaker impedance. The trick then is to build 32+ ohm speakers.
...Or use headphones with an OTL amp. Virtually all headphones are 32 ohms or more.
@@dmark2639 headphones are never my goal.
Buy commercial grade speakers that have a 70 volt transformer and various wattage taps built in. JBL, Klipsch, Tannoy, QSC, Community, Atlas, and many more brands all make 70v speakers in all sizes and form factors for commercial use...
I just learnt about how tubes work 2 days ago and now I'm hooked. I want to make a tube preamp for a gainclone lm3886 amp. What do you think, I go with the 12AX7?
Run, little bunny! Remember what they did to the bunny in DOOM!
Very nice can you upload pcb design somewhere?
Could you use the RightMark Audio Analyzer 6.4.5 to measure frequency response and distortion?
I'm not sure how good it would be to test that as it would be limited by the speaker and microphone capabilities wouldn't it? As it is, I had to crank the recording volume on my Tascam recorder way down to prevent the microphone from clipping. I did check the distortion of the two amplifier stages on the scope though. This is a 1V P-P signal being input, and the output signal that goes to the grids of the driver tubes is 12V P-P and has surprisingly little distortion. It gets just a slight bit of distortion at the bottom of the sine wave, but other than that, it was way cleaner than I was expecting!
i.postimg.cc/rw6pH9R3/IMAG0463.jpg
It helps to more adequate judgement of voice quality, if spoken through also the qualities of loudspeakers, a resonance frequency meaning related to them and how this affects the imaging of soundtrack played.
Mind, with original amp, with transformer output, You used really teeny-tiny speakers, which likely cut the frequencies heavily already from 500Hz downwards. The OTL however, I'd guess, had a speaker load with a resonance frequency nearer to 200 Hz and therefore also allowing better image of base.
Try them both amps with a very simple card cone 6 to 8 inch full range speaker and the improvement will be immediate, despite the lowest price of the speaker.
Also, you are handy (respect ) on woodwork, like it shows. This could lead to another exploration, related to word couple - Open Baffle.
Keep up. Love your thinking and deeds. ;)
You're definitely right, it wasn't really a fair comparison of the two because the speakers were pretty dramatically different between them. I do really want to build another amp with a proper output transformer and speaker to see how good of sound I can get out of something compact. I have an idea of building a tube amp that sits on top of a wooden cylindrical speaker box with the speaker mounted on the bottom facing downwards. I think that could result in a really beautiful design with some work and it sounds like the perfect candidate for a proper output transformer amp!
@@UsagiElectric Oh... I am far from being ultimately right here. Didn't touch the issue with tiny output trannies. Small as these, they do limit the frequency delivery also. Neverthless, especially in guitar amp making, is not so critical always. Also, the whole side kick science, not recognized/accepted with true proud audiophools, goes about using power supply transformers for output. And these go for fraction of price compared to specialized output transformers. Can have lots of fun cutting/bending the corners in this area. ;)
Need a bigger triode, perhaps a beam power tetrode running in triode mode, or one of the bugger transmitter tubes instead.
Absolutely!
I happen to have about 14x 6146 power pentodes, which when triode strapped move well more than double the current as the 12B4A at a given voltage. I'm thinking a bunch of those in parallel would be quite quite exciting!
If you put enough of them in parallel, I guess you could actually impedance match with a 16 ohm speaker. 4 16 ohm in series with 15 12B4 in parallel should match up pretty close.
Yup, even two of the 8 ohm speakers in series would have made a massive difference, but I only had one of the speakers unfortunately. For the next OTL I build, I think I'm going to use big 6146 tubes triode strapped. At 50V, they should move about 60 to 70mA. I have 15 of those in my collection, so if I run 14 of them in parallel, that should get me to about 50 ohms of impedance, which with 6x 8 ohm speakers in series is just about perfect!
(Granted, I'll need 100+W of power just for the filaments, haha)
@@UsagiElectric Match all your filament currents and put them all in series. If you get your tube count right, you many end up with enough that you can run the heaters off line voltage and you won't need an A supply at all.
The 0A2 is without a doubt a zero-volt hater... because it's a cold-cathode tube, it doesn't have a heater. :)
Is there any effect from SMPS's or Boost/Buck converters on tube systems?
Yes, any output noise from the supply goes straight through, just like it does in a solid state unit. Use a high quality S.M.P.S.
As far as RFI/EMI interference goes, keep the converter away from any transformers, as any magnetic interference WILL couple thru the transformers.
Other than that, there's nothing different than there would be with a solid state design.
I wish you were to use a bit of a higher voltage
because I think it would preform even better and louder
any new yet on the 555 kit?
A higher voltage would definitely give better performance and volume, but that defeats the purpose here, which is to see what we can accomplish at ridiculously low voltages. By that same logic, I should have built the vacuum tube computer at higher voltages, but I think what makes it (and this OTL amplifier) really unique and fun is that it's at super low voltages!
Nothing new on the 555 kit yet, Chinese New Year slowed some things down and I've had a nightmare of a time figuring out a reliable way to produce the legs. I have some more ideas to try, but I gotta recover some money first.
@@UsagiElectric true true
in that case
could you try building a otl with pl95 tubes
I have like 10-30 of those around and I kind of wanna use them for something
Recently a famous audio engineer playing with some rather expensive speakers found that letting a 100 ma or so dc curent actualy lowered the speaker distortion by up to 50%. !
Hmmmm, I have a decent handful of pentode Mullard EL34 (that's 6CA7 to you fellas over the pond).
I wonder how well they'd do as a parallel cathode follower final stage OTL 48V.
The heaters on them are 1.5 Amp if my (aging) memory is correct so that'd be 56 Watts of heaters for a '6-pack'
The DC coupling definitely helps with the metal rock.
The DC bias it puts on the speaker may or may not be making much of a difference. It depends on the capability of the speaker. Too much bias and it will bottom or top out the cone at the extremes of its travel well before it would do so without the bias. Using a choke in the vicinity of 1 henry (with the smallest resistance practical) as a direct load and connecting the speaker via a large capacitor in the vicinity of 1000 uF would eliminate the bias. The high value resistor from the cathodes to ground shouldn't make a difference at all given the low impedance of everything else.
With powers this small "red plating" is not likely. Overstressing the tiny cathodes would be more likely.
Thank you for the excellent info!
I actually did try to put a few different caps in series when testing the design to remove any DC through the coil, but no matter what size cap I tried, it always decimated the sound. Like removed greater than 90% of volume. It could be that I just didn't get a large enough capacitor, but I think it mostly had to do with the insanely low voltages I'm running at.
The 12B4A was moving around 40mA through each tube, which at 48V should mean an internal resistance of about 1200 ohms per tube. With six in parallel that should be something like 200 ohms of series resistance with the speaker.
I think the impedance is just so ridiculously mismatched that adding a series capacitor just wrecked whatever weird balance I had going on here.
Or I could be totally off the mark, this is all uncharted territory for me!
@@UsagiElectric You'll need a sufficient DC load for the cathodes to get any beef from this amplifier from a capacitor isolated speaker if the speaker isn't the DC load (that 100K is a joke) That's why I suggested the choke. It's possible to get low resistance toroidal chokes today for ridiculous cheap prices. The choke will block AC, letting it be able to pass on all the better to the speaker.
Mismatch is part of why you don't have the very best output possible. But it still should be possible to spare the speaker a DC bias.
Remember the cathodes can pull (positive) but cannot push (negative). That's what a DC load is for.
I think you should try white cathode follower (WCF) next time. And I recommend you not to quit from output transformers, they can be funny part of designing tube amp, espesialy when you use tetrode or beam-tetrode/pentode where you cen make many diffrent circuit like ultra-linear or super pentode or even unity coupled like in McIntosh apms. 😉
Ooh, the white cathode follower is very, very cool! I'll definitely have to experiment with that one next time!
I wouldn't say that I have given up on output transformers, they're definitely the right way to do it, but I've got the OTL bug now, I totally want to build a properly wicked OTL amplifier!
Cool, I would still like to see a frequency response and have a look at THD just for a laugh !...cheers.
I have no clue how to go about checking something like that, but I did actually test input vs. output of the amplifier stages on the scope!
This is a 1V P-P signal being input, and the output signal that goes to the grids of the driver tubes is 12V P-P and has surprisingly little distortion. It gets just a slight bit of distortion at the bottom of the sine wave, but other than that, it was way cleaner than I was expecting!
i.postimg.cc/rw6pH9R3/IMAG0463.jpg
@@UsagiElectric Thanks !
With that circuit, the only grid bias you're getting is from the current through the speaker! So the operating point is in a very nonlinear portion of the curve. If you were to put a DC voltmeter across any of those 1K resistors, I'm sure you'll find that the tubes are drawing grid current. You DON'T want THAT, in a Class A amp!
How many watts is this? And how many ohm speakers do I need?
I've spent the last 40 years designing at 5v. 8 years ago I renovated 1932 Philco Cathedral, been doing tubes since. Jumping to 3, 4, 500 volts is a little scary but just listen to Alanis and keep your Hand in Your Pocket and you'll be safe. Thanks for the video.
Thanks for checking the video out!
I actually started learning electronics in earnest with BEAM robots, which are even lower voltage than 5v since they're usually solar powered. The transition from that to tubes was an interesting one because BEAM stuff is all about being as absolutely efficient as possible and the average tubes heater consumes more power in a minute than a BEAM bot will in its entire life!
I am slowly getting more comfortable working with higher voltages, particularly after learning and working on various HP equipment, but there's still a massively fun allure to working tubes at low voltages!
@@UsagiElectric yeah, I can see low voltage tube design being a real challenge. The important thing is having fun and learning, which you look like your having.
Would you consider a cooling solution to get some longevity out of it?
its just a guess, if its coming just prepare cash (and internet), since my be3 purchase took 2 months of stock waiting and lasted like 3 weeks or smth
Numb(?) question - will tubes like 6bj6 and 6ak5 work “acceptably” at 48 volts, or perhaps a little more - perhaps as much as 72 volts?
I got *tossed* by ~300 volts in my teens, (1 tube Amateur Radio transmitter) and I’m not fond of “high” voltages.
don't let petty souls wear you down
Thank you!
It was a little disheartening at first to see all the negative comments about something as trivial as pronunciation, but in the end, I realized they weren't trying to be malicious, most were just trying to educate me on how things are conventionally done. And it's really easy for me to forget that very few people actually know my background or know what kind of weird muscle memory kicks in when talking to a camera all by yourself. But also, it was just a lot of fun to film that segment, haha.
Oh I don't know about getting rid of the output transformers.... Just speaking aesthetically.... you need output transformers to give it a "proper yesteryear feel". But I suppose 6 driver tubes is more tubes, so that's "old-timey" too. And it's certainly got some "go juice" in it!
The whole thing with the filament voltage having to be a "correct number"... I wonder if that's why some people call logic ICs "seventy four L. S. twenty" where I'd always say "seven four L. S. two zero" (I'm dyslexic... that's kinda the same as talking to aircraft on a radio)
Which tubes are used on the 2nd amp?
All the tubes on the OTL amplifier are 12B4A single triodes.
For the older amplifier that we built a long time ago, I use 12BH7 dual triodes for amplification and 12B4As to drive the output transformer.
In German, the hundreds, ones and then the tens are usually mentioned.
But when I'm alone, I say the numbers number by number: hundreds, tens, ones.
In English, only the numbers up to nine-teen are twisted. So teen-nine, than twenty-one.
lol why do your tubes in your schematics look like a sad robocop with a handlebar mustache dang it what is seen can never be unseen XD
Hahaha, I never noticed that before! That's just how the models in TINA look, but man, now that's all I'm ever going to see!
you could put a thermistor betwwen the power source and the red-plate potentiometer
I'm not sure I'm following you here?
Thermistors are usually used for limiting inrush current, right? But, the "red-plate" potentiometer is just controlling the voltage level on the grids of the output tubes, there shouldn't be any appreciable current flowing through there at all.
As for powering the tube filaments from the power source, in this case it didn't really seem to be a big issue, the little 48V supply I'm using seemed to handle the inrush with no problems!
@@UsagiElectric after the amp has been on a while it is still apt to redplate if the pot isn't turned to about 2/3
I enjoy your computer stuff but this low voltage no transformer audio amplifier is super cool.
Why would anybody build such a good looking vacuum tube low voltage amplifier?
Because you can.
Stan Paddock
why you did not share the PCB ?
Try this with a 6080wa or 6as7 for otl amplifier
I don't think the output tubes are getting much negative bias with only an 8 ohm speaker in the cathode circuit. What is the voltage from cathode to grid?
The cathode followers should essentially just put the cathode at the same level as the grid voltage.
However, with just an 8-ohm speaker, there's all sorts of horrific impedance mismatching going on. It's actually pretty surprising it sounds as good as it does.
@@UsagiElectric There is an impedance mismatch like most OTL amplifiers. Looking at the curves, the plate current is not too far off from proper class A operation. The curves are hard to read at only 48 volts. The negative feedback is good to reduce distortion but some audiophiles would say the sound is not as clear. I would like to know the cathode to grid voltage to study this further. Also could you measure the resistance of the speaker with an ohmmeter to help calculate the DC operating condition.
how many watts does this put out?
try a push pull tube amp!
Usagi, I don’t want to open a can of worms for you, but have you ever considered re-winding your speaker voice coils? Since there’s an impedance mid-match from the tube output to an 8 ohm speaker, then change the speaker to high impedance. Say, 1K?
If I remember correctly, there was an ancient article in “Hi-Fi Magazine” about doing this, and they reported amazing response from the speakers when not running through an output transformer.
génial
Merci!
Always follow the rules. I before E, except when your foreign neighbor Keith receives eight counterfeit beige sleighs from feisty caffeinated weightlifters.
Brilliant!
And the fact that you retained the 100K resistor, in parallel with the speaker?
BTW the first two digits are the filament voltage BTW
Except for when they're not 😄
the real challenge is building a pure dht amp.(driver and output tubes are DHT.) if you can do that without hum you will be the best.
Oooh, I don't know if I have any directly heated triodes hanging around to test that out with.
Most directly tubes run at like 1V or 2V though, right? I suppose with big batteries on them you can get all the hum out, but that feels a bit like cheating, haha.
You've managed to use 3 more tubes than you need for a radio.
I wonder how much lower you could go on the OTL plate voltage 🤔
In my initial testing, I ran it as low as 24V with just three output tubes and it still put sound through the speaker. It was very loud, but it had enough volume to be able to hear what was going on. With more output tubes, I reckon it'll put some decent sound through the speaker. But, the margins on distortion from the amplifier stages get a lot tighter. Still, could totally be done!
@@UsagiElectric interesting, I’m gonna have to try this out when I get up and running; just ordered worktop today so progress is being made!
Hi, wonderfull and I would like to build a replica myself. However I am a complete beginner. Can you please give me hint how to connect the tubes. What I see is that they have 9 pins but in your schematics there are only 3 pins connected? Shurely I missed some background on tubes ..... Can you please give me some support.