I have decided to build a very low power tube amp with a single 6J5 as output tube with a custom made output transformer. Expected output power is 0,5-1W. Since this tube is not normally used for power output I can't find any data on operating points. Should I first decide on +B-voltage and cathode current before I can calculate a suitable load resistance?
Would it be possible to drive it directly from a low level signal from a CD-player or do I need a preamp-stage? Can it be done with an interstage transformer?
Would it be possible to drive it directly from a low level signal from a CD-player or do I need a preamp-stage? Can it be done with an interstage transformer?
6j5 is slightly different from 1/2 6sn7. It's not designed for output stage. You can parallel some to make more power. SE 6SN7 amp....
I think it's not a good idea. Two 6sn7/channel can give 0.5-1w se class A. So you'll need 4 6j5/channel and 5k opt (much more easier to find than 10k or 20k). If you want to make a spud low power amp, I suggest 6c45pi/we437/we417/5842. With 5k opt you could get 1.5w se amp.
For e.g:
I think it's not a good idea. Two 6sn7/channel can give 0.5-1w se class A. So you'll need 4 6j5/channel and 5k opt (much more easier to find than 10k or 20k). If you want to make a spud low power amp, I suggest 6c45pi/we437/we417/5842. With 5k opt you could get 1.5w se amp.
For e.g:
290~300 volts supply after filtering, a 10k load, and ~6 volts across the cathode or so would be a good start, maybe use an adjustable cathode resistor, and set it so that you are at 4~4.5 watts plate dissipation, and then play with the bias point to get you the most symmetrical output waveform.
It might work if your CD player has a hot enough output, but another 6J5 (or a single 6SN7, split between channels) with an unbypassed cathode resistor, and a moderate to higher resistance load would be a pretty good match as a voltage amplifier to drive it. should have enough headroom to implement some global feedback too.
.5W or so is about what I would expect, at clip.
Personally, I would do push-pull with a pair per channel, using some circuit like the Flea amplifier project in my signature. 1-1.5 watts is easy and simple.
It might work if your CD player has a hot enough output, but another 6J5 (or a single 6SN7, split between channels) with an unbypassed cathode resistor, and a moderate to higher resistance load would be a pretty good match as a voltage amplifier to drive it. should have enough headroom to implement some global feedback too.
.5W or so is about what I would expect, at clip.
Personally, I would do push-pull with a pair per channel, using some circuit like the Flea amplifier project in my signature. 1-1.5 watts is easy and simple.
Last edited:
IMHO. having used 6J5 in a 2A3 amp.
I found that high quality 6J5 equivalents were much cheaper than collectable 6SN7.
A very good valve to use is the L63 in bottle shaped glass.
As with everything you need to try them, some are quite microphonic.
I found that high quality 6J5 equivalents were much cheaper than collectable 6SN7.
A very good valve to use is the L63 in bottle shaped glass.
As with everything you need to try them, some are quite microphonic.
6J5 has half the dissipation rating of a 6sn7 with only one triode used. I've used 2 6sn7 tubes to get 2 triodes by wiring opposite triodes in each bottle so I could swap them and use the opposite triode. That would leave you the option of paralleling the sections in the future so the amp would work with lower primary Z transformers.
Triode OPT needs to be at least 2X Rp. In the case of the 6J5 that would be 15K. Better still needs to be 3X Rp to keep distortion down.
Thank you all for very good ideas and input. I now realize that the 6J5 may not be the most suitable tube for what I would like to achieve. At least not in a single tube configuration. I will look into the other tubes suggested as well as the recommendations about output transformer and load resistance with regards to plate resistance.
About the difficulty finding an output transformer with a very high primary (10-20k) I am not really worried yet because I think I will have to order one specifically made for my requirements. And the additonal cost for a custom interstage transformer probably wouldn't set me back too much (I hope) so if that is what I have to do to make it work I may very well do it.
The reason I looked at the 6J5 was partly the attractive price and availability.
About the difficulty finding an output transformer with a very high primary (10-20k) I am not really worried yet because I think I will have to order one specifically made for my requirements. And the additonal cost for a custom interstage transformer probably wouldn't set me back too much (I hope) so if that is what I have to do to make it work I may very well do it.
The reason I looked at the 6J5 was partly the attractive price and availability.
Here is a list of double triodes that would work for you. All with the exception of the two at the top are suitable for a small two stage amp capable of 2W or more, depending on the power supply used.
MU1 column refers to the voltage amplifier section while MU2 covers the power section. Yes, I've built one using a 6EM7, it worked very well!🙂
MU1 column refers to the voltage amplifier section while MU2 covers the power section. Yes, I've built one using a 6EM7, it worked very well!🙂
Mikael, if you are going to have an OPT made, do you really want to go cheap-@$$ and use a $5 tube because it's cheap? 😉 I'm the dutch guy, not you 🙂))
Last edited:
The way I see it, the OPT is one of the, if not THE, most important component in making a a tube amp sound good and the "flea power" amps have intrigued me for along time. A good output transformer will cost a bit and I figure a custom made will probably not cost a ton more than a readily availabe so I might as well go all the way. Also, depending on the specs I decide upon, it may work well with other low power tubes as well.
The price of the 6J5 is one of the things I like about it, but availability is probably even more appealing to me.
I liked the specs of the 6GL7 tube but it will make the amp more "advanced" and also that tube was not as easy to find as the 6J5.
Perhaps I should come and visit you the next time I am in the Netherlands and we can talk about tubes over a nice beer 😉
Assuming I would like to use the 6J5 despite its shortcomings, primarily the output power, is there anyway I can "fake" the OPT to make some initial experiments and measurements?
I suspect an OPT with a 35k primary will have a rather larger DC resistance. I have found some schematics with the 6J5 as a pre tube and the value of the plate resistor is close to 100k. I guess that in order to make any useful measurements I need to get an estimate of the DC resistance of such a transformer, right? The signal part of the output should be correct using a cap and a ~35k load resistor I think. And as suggested by Lingwendil a variable resistor at the cathode should make it possible to decide a resistor value that produces a symmetrical output.
I suspect an OPT with a 35k primary will have a rather larger DC resistance. I have found some schematics with the 6J5 as a pre tube and the value of the plate resistor is close to 100k. I guess that in order to make any useful measurements I need to get an estimate of the DC resistance of such a transformer, right? The signal part of the output should be correct using a cap and a ~35k load resistor I think. And as suggested by Lingwendil a variable resistor at the cathode should make it possible to decide a resistor value that produces a symmetrical output.
The theoretical efficiency of an SE Class A triode is at best 25%. That happens when the load is 2x the plate resistance. So if the plate volts X the plate current computes up to One Watt we can expect no more than 250 mW at the output.
Good design would probably go for a load of 3X the plate resistance to reduce distortion.
The output transformer will reduce that further by 10-20% depending on its efficiency. So perhaps we would finally have 200 mWatts of audio for the One Watt of input power at the plate.😱
Good design would probably go for a load of 3X the plate resistance to reduce distortion.
The output transformer will reduce that further by 10-20% depending on its efficiency. So perhaps we would finally have 200 mWatts of audio for the One Watt of input power at the plate.😱
This is only roughly true and only for the "matched" condition (which does NOT optimize power output). You can raise efficiency toward 50% by raising the load impedance. You also raise the B+ to take advantage of the improved loading. What always happens is you exceed the tube plate voltage rating. You also can exceed the 5% THD we normally work for. But you can easily take one of the big transmitter triodes and get over 30% eff at not much over 5% THD.
But the 6J5 is no transmitter tube. Aside from low ratings, it is not linear over a w-i-d-e range. It was tuned for good gain at moderate output into medium-high impedance loads. Actually transformers, but plate-to-grid where the actual load takes "no" power but transformer losses must be covered. And at these impedances, transformers can not be wide-band (hi-fi).
Two plots.
One respects tube ratings and would work near 0.5W out _if_ you could wind a good 35K output transformer (an incredibly difficult thing). Efficiency, even though off-matched for better output, is only 20%.
You are hooked on 6J5 because you have a heap of them? Then foo on ratings. If one goes weak you can replace many times. I'm sure a 6J5 will not curl up and die at 3W 400V. Second plot takes this to 1W. That requires a very asymmetric grid swing, and will be high THD unless massive overall NFB is used.
In any case your grid swing will be over 10V peak, and even more on the down-swing. Yes, a 600r:10K transformer will get this toward the 2.8V peak out of a good DAC. However tubes have been cheaper than good audio transformers since my father was a boy, over 70 years. The dead-obvious thing to do, with a heap of 6J5, is use one as the gain-stage in front of your "power" tube.
Attachments
Thank you all for very informative input and good suggestions. My plan was to drive an array of small 8 Ohm speakers. I aim for background listening levels which seem to be possible.
I have a price quote from Sowter on a 35k transformer and I believe that ~30k primary, 20mA current rating and ~2W power output specs can be used with a 6J5 as well as quite a few other small signal triodes, correct?
The "speaker" I have looked at is this one
I have a price quote from Sowter on a 35k transformer and I believe that ~30k primary, 20mA current rating and ~2W power output specs can be used with a 6J5 as well as quite a few other small signal triodes, correct?
The "speaker" I have looked at is this one
Just a thought... Since this is a low power project that is really a one-off experiment, why not get a conventional, reusable, transformer like a 8K or 10K to 4R speaker and load it with some 8 ohms or 16 ohms in series to get the high primary Z and then if the project is a letdown you can use the tranny for a common amp topology to recover the amp to a normal output tube.
As already pointed out with rather high plate resistance tubes like 6j5 one can expect 20% anode efficiency at best. However I doubt this can be reached with a 35K transformer. Having 18K, with 280V anode voltage at 8 mA anode current one can expect about 0.45W at about 5-6% THD. 18K is a bit more than 2X the plate resistance which is about 8K.
Similarly, as pointed out in its original Philips datasheet, the 1/2 ECC40 running at 250V/6mA will achieve 0.28W @8.5% THD with just 15K load. However the ECC40 in those conditions has 11K plate resistance hence higher distortion.
Similarly, as pointed out in its original Philips datasheet, the 1/2 ECC40 running at 250V/6mA will achieve 0.28W @8.5% THD with just 15K load. However the ECC40 in those conditions has 11K plate resistance hence higher distortion.
Excellent point that I should have considered. What are the downsides of using a 4 ohm secondary with 8 or 16 (or even) 32 ohm loads to get a higher reflected primary load?
One other thing that I have been thinking about is the input. I may need to use a small input transformer to get the desired voltage swing from a solid state source, but won't the secondary of that transformer see the negative DC of the grid and does that mean that is have to be gapped for the DC? Or do I still need an input capacitor?
- Status
- Not open for further replies.