I built both the EL84 and EL34 versions of the amp from tubecad above - they are excellent sounding, but not the simplest amp due to the number of valves involved, a bit more demanding on power transformer etc. But they sound really great. I put a CCS on the cathodes of the output valves, and an LED on the cathodes of the input side from memory.
For those experienced audiers, much of the loop stability disadvantages surrounding the Williamson design can be mitigated and optimized if one uses the Y Sweep sawtooth of a storage scope injected at the amp input at sub audio frequencies, i.e a few Hz. Any input capacitor to the first stage will have to be bypassed. Not all scopes have an external sweep accessory or long enough screen persistence usless a digital version is used. The trick is use very low power levels to avoid o/p tranny sat issues and study the loop or effectively the "Q" factor with scope input cannected the the ouput via dummy load. One can then select the right coupling capacitors, usually the first stage to phasesplitter. This is only a brief analysis, but I find this method given that 20dB GNFB is used at 1Khz, and after the treatment one can increase GNFB by 10-15dB without any signs of LF instability. This is assuming proper value LF decoupling caps are in the B+ lines.
I know this is a tough nut for many to crack, but given the already superlative sectional qualities of the Williamson or Partridge output transformers, the resultant bass attack that many digital sources can provide will result is excellent dynamic sound.
Keep at it,
rJ
HollowState said:
Probably avoid the Williamson because it requires an exceptional output transformer do to two interstage capacitors instead of one.
Years after building the Williamson's, I came across Talbot Wrights modification for the Williamson and applied it. Measured performance was improved but audibly, I cannot tell much difference. I do however believe the modification did help with any "possible" low frequency instability that may arise under the right conditions. I have re-chassised the Williamson now and I am using Edcor output transformers. I can't tell any difference in performance, but the feedback network did have to change to compliment the output transformers. I am using approxinately 18.5dB of feedback.
If more people actually built an amp instead of just simulating it, some positive conversations might actually come of it. Yes, simulation is great, but it is nothing compared to actually building a circuit and fine tuning it the old fashioned way.
Probably avoid the Williamson because it requires an exceptional output transformer do to two interstage capacitors instead of one.
I built a pair of Williamson mono amps 26 years ago. My initial build used a solid state power supply, not a tube as in the original. I used Hammond 1650P output transformers and never had any issues with low frequency instability. I had to tweak the feedback network to work with the 1650P's, but it was not that difficult. I suspect a lot of the original complaints about low frequency instability were due to the original tube based power supply.
Years after building the Williamson's, I came across Talbot Wrights modification for the Williamson and applied it. Measured performance was improved but audibly, I cannot tell much difference. I do however believe the modification did help with any "possible" low frequency instability that may arise under the right conditions. I have re-chassised the Williamson now and I am using Edcor output transformers. I can't tell any difference in performance, but the feedback network did have to change to compliment the output transformers. I am using approxinately 18.5dB of feedback.
If more people actually built an amp instead of just simulating it, some positive conversations might actually come of it. Yes, simulation is great, but it is nothing compared to actually building a circuit and fine tuning it the old fashioned way.
What is it that you want to know? I haven't even logged in for years. The old address associated with the account hasn't even been active for a year or more. Imagine my surprise this morning to see this thread revived and active. These are extremey expensive projects to undertake, and I have a career and family to think about. I am merely a hobbyist. Not a pro. I asked the question just as much with the intention of learning how to design a circuit as actually building one in practice. At this point, I am practicing winding my own audio and power transformers using microwave oven iron. Despite much discouragement from people saying it wasn't worth the effort, I have had extremely good success thus far. Granted, the experiment was for a simple SE 6AS7G gain stage, but I am seeing good bass performance down to 30hz with it. Very low distortion overall. Better than anything else that I have ever worked on, in fact. One thing I am blessed with is patience, and patience is an absolute must when hand-winding transformers. My dream is to build an amp from the ground up, using as many DIY parts as is possible. Suitable iron for a PP EL34 amp is wildly expensive. How nice it would be to be able to skip that expense altogether by utilizing free iron, free magnet wire and free patience. I am also practicing shaping sheet metal so that I can build my own enclosure. All this stuff takes time. In my mind, "soon" may be five years down the road. For what it's worth, I highly appreciate anyone that decides to take a moment out of their day to help me understand more about the hobby. I have been fascinated with vacuum tubes since at 4-years old, I looked in the vent holes on the back of the TV and saw a bunch of little glowing "light bulbs" in there. It was like bottled magic. Now, I just want to have a little fun.
One reason I don't post here often is the atmosphere. Always naysayers and "experts" who give the worst advice most times. You'll know the real experts when they post, not the armchair ones. Too much simulation sometimes and not enough building.
I think it is great you are experimenting and learning on your own. Keep it up.
Good push pull output transformers are not too expensive, take for example Edcor. You can get good ones from about $120 a piece which is not too bad. Not as cheap as DIY though. Depends on power output too.
read my text.
do you find it aggressive or negative?
I'm just making an observation.
I don't ask for an explanation.
do you find it aggressive or negative?
I'm just making an observation.
I don't ask for an explanation.
... but you can understand the guy repudiating your observation.
It is nice to be 'part of the journey', so hearing about these recycled microwave traffos would be interesting! I think the buzzword these days is 'upcycling'.
A word of caution for those new to electronics: I feel obligated to mention that microwave transformers by default put out 2500+ volts, potentially at more than an amp. If you touch that, you'll be dead before you hit the floor. Just cut those original windings off of it right off the bat. These things can be positively LETHAL, and they have killed many hobbyists. There are people of all skill levels here, so I just wanted to mention that.
@ THD+N
I'm a big fan of Edcor, and have ordered several of their transformers in the past. They're certainly cheaper than an equivalent Hammond or similar model, but like you said, it's hard to beat DIY in terms of cheap. I do love experimenting too.
@ OldHector
I have my test amp disassembled at the moment, but I'll probably start a new thread with more details when I have more to share. I'll feed it a square wave and post pics of the output on a scope. I 3d-printed new bobbins out of nylon, and tried two different methods of winding them. The original iron is already in an E-I configuration, so it has something of an air gap already in it. I have also ground one down to increase the gap, but didn't see any appreciable improvement. Laying the secondary on top of the primary (isolated with Kapton tape) works, but the corners of the square wave were a tad rounded off. That produced a softer tone. I read that interleaving the primary and secondary produces better results, and I was able to confirm that on my setup. Effectively, 4-layers of primary, 1-layer of secondary, 4-layers primary, etc. All isolated with more Kapton tape. That config knocked my socks off. I have a comparable Edcor OPT, and bass performance really dropped off at about 60hz. The DIY OPT went all the way down to 30hz before it produced a similar-looking waveform. Now, the iron volume on the DIY OPT is probably 2x the Edcor OPT, so that may well have been why. I don't know for sure. There was a bit of slant at low frequencies, but little to no overshoot. I was using it with a 6AS7G (dual power triode) in single-ended stereo. Preamp end was a 6DJ8 in cascode for each channel. No negative feedback was used. The 6AS7G apparently has extremely low output impedance. Possibly as low as a few hundred ohms. I'm still having trouble coming to terms with that, but either way, I have the OPT set up for roughly 2500ohms:24ohms (for headphones). That just worked best for me. If I remember correctly, that comes out to about 1500:150 turns. I'm not saying that these are "just as good" as any particular OPT. I don't have the expertise to make any kind of claims like that. I'm just saying that they are very surprisingly good, all things considered.
@ THD+N
I'm a big fan of Edcor, and have ordered several of their transformers in the past. They're certainly cheaper than an equivalent Hammond or similar model, but like you said, it's hard to beat DIY in terms of cheap. I do love experimenting too.
@ OldHector
I have my test amp disassembled at the moment, but I'll probably start a new thread with more details when I have more to share. I'll feed it a square wave and post pics of the output on a scope. I 3d-printed new bobbins out of nylon, and tried two different methods of winding them. The original iron is already in an E-I configuration, so it has something of an air gap already in it. I have also ground one down to increase the gap, but didn't see any appreciable improvement. Laying the secondary on top of the primary (isolated with Kapton tape) works, but the corners of the square wave were a tad rounded off. That produced a softer tone. I read that interleaving the primary and secondary produces better results, and I was able to confirm that on my setup. Effectively, 4-layers of primary, 1-layer of secondary, 4-layers primary, etc. All isolated with more Kapton tape. That config knocked my socks off. I have a comparable Edcor OPT, and bass performance really dropped off at about 60hz. The DIY OPT went all the way down to 30hz before it produced a similar-looking waveform. Now, the iron volume on the DIY OPT is probably 2x the Edcor OPT, so that may well have been why. I don't know for sure. There was a bit of slant at low frequencies, but little to no overshoot. I was using it with a 6AS7G (dual power triode) in single-ended stereo. Preamp end was a 6DJ8 in cascode for each channel. No negative feedback was used. The 6AS7G apparently has extremely low output impedance. Possibly as low as a few hundred ohms. I'm still having trouble coming to terms with that, but either way, I have the OPT set up for roughly 2500ohms:24ohms (for headphones). That just worked best for me. If I remember correctly, that comes out to about 1500:150 turns. I'm not saying that these are "just as good" as any particular OPT. I don't have the expertise to make any kind of claims like that. I'm just saying that they are very surprisingly good, all things considered.
Here is the thread about my microwave oven transformer project if anyone is interested:
https://www.diyaudio.com/community/...transformers-into-output-transformers.391040/
https://www.diyaudio.com/community/...transformers-into-output-transformers.391040/