Recently I was given a pair of SOWTER model U082 output transformers.
SOWTER TYPE U082
They are good for EL84 type tubes but I am keen to know if these can be adapted for 807 tubes. I have seen the STC datasheet and the mention of running 807's into 8k primaries. These are pretty generously proportioned heavy products.
Tubes at hand include small signal triodes of 12AX7/U7/T7 family. Several 6SN7/SL7's. 68J8 family tubes. And several pairs of 807 NOS tubes.
Some might suggest to follow the Williamson design. If the Sowter type U082 ironware is suitable then I would surely go ahead. Unfortunately I don't have any T&M setup to verify. I do have a few digital multimeters and a surplus 5A variac. At pinch I can buy and learn to use a cheap oscilloscope.
I have experience building basic tube stuff to a print. Such as SE EL34 type amps, linestages etc,. I am an industrial worker thus aware of the hazards including those arising out of 807 having a top connection.
Before committing myself to this project I wish to absorb as much as possible. Would strictly try not to use any complicated CCS and other fancy stuff.
I will still be required to procure the power transformer, chokes, casings and other bits. Planning to make it a weekends only slow paced build in a monoblock configuration.
Looking forward to ideas and help.
SOWTER TYPE U082
They are good for EL84 type tubes but I am keen to know if these can be adapted for 807 tubes. I have seen the STC datasheet and the mention of running 807's into 8k primaries. These are pretty generously proportioned heavy products.
Tubes at hand include small signal triodes of 12AX7/U7/T7 family. Several 6SN7/SL7's. 68J8 family tubes. And several pairs of 807 NOS tubes.
Some might suggest to follow the Williamson design. If the Sowter type U082 ironware is suitable then I would surely go ahead. Unfortunately I don't have any T&M setup to verify. I do have a few digital multimeters and a surplus 5A variac. At pinch I can buy and learn to use a cheap oscilloscope.
I have experience building basic tube stuff to a print. Such as SE EL34 type amps, linestages etc,. I am an industrial worker thus aware of the hazards including those arising out of 807 having a top connection.
Before committing myself to this project I wish to absorb as much as possible. Would strictly try not to use any complicated CCS and other fancy stuff.
I will still be required to procure the power transformer, chokes, casings and other bits. Planning to make it a weekends only slow paced build in a monoblock configuration.
Looking forward to ideas and help.
I'd say there is no specific primary impedance for each tube. The 807 can handle higher plate voltage that a 6L6, you can make a pretty powerful PP amp with 807s and a U082.
If you go for 807 with higher B+, stay away from UL setup, g2 does not enjoy high voltages at all.
EL84's remain an option. But only if they the Sowters are totally unsuitable with 807 to reproduce the full audio range with sufficient wattage and without getting whacked.
I wish to employ the 807's with me but the EL84's I will need to buy.
jcalvarez:
I will prefer triode or UL whilst remaining mindful of the screen limitations. Seen several Williamson triode strapped 807 designs being fed about 425v .
I am also open to pure tetrode operation with a Zener regulates screen supply. But I really do not need too many watts to feed a my of vintage 12" Philips fullrangers.
don't take into account what I said, I answered too quickly.
on the other hand, even if it means buying EL84s, buying 6P14P-ev in large quantities and selecting them, they are really good tubes
on the other hand, even if it means buying EL84s, buying 6P14P-ev in large quantities and selecting them, they are really good tubes
I fully understand the soundness of your advice. But at the moment I would prefer to save some money if possible.
What are the unique requirements of a Williamson that a modern wideband OPT such as our Sowtar can not match? If that can be settled, then I am veering towards the 1st schematic presented in the attached scan.
Any obvious pitfalls to be mindful of with this?
I looked up some data published by MJ some 30 yrs ago.
Triode connected 807's : Vb= 400v
Rk= 330 ohms (common)
Zo= 8k
Ia = 116ma (2 tubes)
Po= 10,8w
Triode connected 807's : Vb= 400v
Rk= 330 ohms (common)
Zo= 8k
Ia = 116ma (2 tubes)
Po= 10,8w
Sowter U802 OPT:
https://www.sowter.co.uk/specs/9015.htm
Nice OPTs! Nice gift!!
Did you say you'd prefer triode output tubes? And you have a set of good 807 tubes?
Would you be happy with perhaps 10W of (mostly) class A output?
If yes, I think you're good with what you have.
Those OPTs are rated for 40W. Yes, they'd be a bit overkill for push-pull 807-triode. So what? They'd work great. Less possibility of core saturation at low frequencies, more tolerant of mismatched output tubes, etc.
A triode output stage with minimal NFB is going to be easier to design cookbook-style than a pentode output stage requiring more NFB applied.
Did you say how much output power you want? If 10W per channel is OK, you might want to use your 807s triode wired with about a 400V B+. 8k ohm primary would be just about perfect.
https://www.sowter.co.uk/specs/9015.htm
Nice OPTs! Nice gift!!
Did you say you'd prefer triode output tubes? And you have a set of good 807 tubes?
Would you be happy with perhaps 10W of (mostly) class A output?
If yes, I think you're good with what you have.
Those OPTs are rated for 40W. Yes, they'd be a bit overkill for push-pull 807-triode. So what? They'd work great. Less possibility of core saturation at low frequencies, more tolerant of mismatched output tubes, etc.
A triode output stage with minimal NFB is going to be easier to design cookbook-style than a pentode output stage requiring more NFB applied.
Did you say how much output power you want? If 10W per channel is OK, you might want to use your 807s triode wired with about a 400V B+. 8k ohm primary would be just about perfect.
Go for the 807 tubes.
. . . They Look Beautiful
Can be used in Triode wired (400V max B+); less output power, but no negative feedback needed, so no adjustments needed.
Can be used in Beam Power mode (300V max Screen voltage). Will definitely need global NFB or Schade NFB (negative feedback)
Can be used in Ultra Linear mode (with a 40% UL tap, you should be able to use 350V B+) might not need Schade or global negative feedback, much depends on your speakers.
Have fun, try all 3 modes
Do not use a magnetic steel chassis, or you are asking for hum . . . hmmm.
Just my opinions
Happy designing and happy building!
. . . They Look Beautiful
Can be used in Triode wired (400V max B+); less output power, but no negative feedback needed, so no adjustments needed.
Can be used in Beam Power mode (300V max Screen voltage). Will definitely need global NFB or Schade NFB (negative feedback)
Can be used in Ultra Linear mode (with a 40% UL tap, you should be able to use 350V B+) might not need Schade or global negative feedback, much depends on your speakers.
Have fun, try all 3 modes
Do not use a magnetic steel chassis, or you are asking for hum . . . hmmm.
Just my opinions
Happy designing and happy building!
I just noticed something about that Sowter U802 output transformer... It has only one secondary, rated for an 8 ohm load.
That means that if your speakers present a 4 ohm load (which almost all contemporary speakers do) then the OPT primary will present a 4k ohm load to the output tubes. Many contemporary speakers say they are '6 ohms' nominal, but significant portions of their impedance curves go down to or below 4 ohms.
You can't design the amplifier without taking into account what it will be used to drive. So...
What speakers are you planning to use with this amplifier?
That means that if your speakers present a 4 ohm load (which almost all contemporary speakers do) then the OPT primary will present a 4k ohm load to the output tubes. Many contemporary speakers say they are '6 ohms' nominal, but significant portions of their impedance curves go down to or below 4 ohms.
You can't design the amplifier without taking into account what it will be used to drive. So...
What speakers are you planning to use with this amplifier?
I am a big fan of the Williamson and have built quite a few versions with various output transformers. It's such a beautiful, natural sound. Assuming the Sowters have the requisite bandwidth to tolerate 16-20dB of feedback, you'll still want a 'scope handy to check for high-frequency stability. 807's will tolerate 400 volts on the screen without a problem, either in triode mode or UL. With a B+ of around 420-430 VDC and 36 volts cathode bias, you'll get 12 wpc in triode and 24 wpc in UL, maybe even more with the lower plate load.
I always use the Sarser & Sprinkle "Musician's Amplifier" as a basis for my amps:
http://oestex.com/tubes/mus_amp .pdf
with the following changes:
1) You only need one choke in the power supply. 7-8 henries is sufficient.
2) I up the B+ to about 430VDC. Subbing a 5V4 for the 5U4 accomplishes this handily.
3) You need a low-pass filter across the first stage's plate resistor. This is pretty much necessary to avoid HF oscillation. You can start with the standard Williamson values, 4.7K ohms + 200pF in series, and see if that keeps the amp stable. If not, I find that an "all-purpose" filter of 680 ohms + 1200pF works very well without significantly decreasing the HF response in the audible range. It's about a 40kHz roll-off and works well with "unknown" transformers. Or if you have the skills (I don't) you can experiment with different values to nail down the OPT's specific HF resonances.
4) You'll need a phase-lead cap across the feedback resistor to further tame any HF anomalies. 100-300pF is a typical range, maybe even less, depending on the Sowter's characteristics.
5) You'll probably need to adjust the feedback resistor upward a bit to compensate for the lower plate load.
I also typically use the UL version, or the "Gilded Lily":
https://dalmura.com.au/static/Gilding the Lily - July 1952 - Audio.pdf
In line with the authors, I find that UL gives up very little, if anything, in terms of naturalness of tone, and the extra power is nice. The only drawback is that UL can be less stable than triode mode, and tougher to tame, so consider that. Low-frequency stability can be an issues as well.
An good primer for building any Williamson is Dave Gillespie's excellent thread at Audiokarma:
https://audiokarma.org/forums/index.php?threads/regilding-the-gilded-lily-heaths-w-2m.767851/
Dave lays out the basic issues of stability and build quality. Of particular note are the changes he makes in the coupling cap values to improve LF stability. I always use this arrangement and it works very well. His low-pass filter for the W2 is quite steep, and I find that with a Peerless OPT you needn't go that far. But it makes for a rock-solid amp.
Sorry to go on (!) but this is a pet project of mine. ;-)
Regarding a scope, you can buy them used for very little. For basic amp testing, the rudiments are not hard to learn. I did it in a couple of weeks when I realized that I would need to if I wanted to continue experimenting with Williamsons. Anyone who wants to build a quality amp should learn how to use one, especially if feedback is involved. I use a Picoscope 2206B because the built-in signal generator goes up to 1MHz, which is important for checking ultra-high-frequency stability, and I like the software a lot, but that may be more than you want to spend.
Anyway, hope this helps! Feel free to PM me if I can be of any assistance. I've walked a number of folks through Williamson builds. It's a challenge but it's fun.
I always use the Sarser & Sprinkle "Musician's Amplifier" as a basis for my amps:
http://oestex.com/tubes/mus_amp .pdf
with the following changes:
1) You only need one choke in the power supply. 7-8 henries is sufficient.
2) I up the B+ to about 430VDC. Subbing a 5V4 for the 5U4 accomplishes this handily.
3) You need a low-pass filter across the first stage's plate resistor. This is pretty much necessary to avoid HF oscillation. You can start with the standard Williamson values, 4.7K ohms + 200pF in series, and see if that keeps the amp stable. If not, I find that an "all-purpose" filter of 680 ohms + 1200pF works very well without significantly decreasing the HF response in the audible range. It's about a 40kHz roll-off and works well with "unknown" transformers. Or if you have the skills (I don't) you can experiment with different values to nail down the OPT's specific HF resonances.
4) You'll need a phase-lead cap across the feedback resistor to further tame any HF anomalies. 100-300pF is a typical range, maybe even less, depending on the Sowter's characteristics.
5) You'll probably need to adjust the feedback resistor upward a bit to compensate for the lower plate load.
I also typically use the UL version, or the "Gilded Lily":
https://dalmura.com.au/static/Gilding the Lily - July 1952 - Audio.pdf
In line with the authors, I find that UL gives up very little, if anything, in terms of naturalness of tone, and the extra power is nice. The only drawback is that UL can be less stable than triode mode, and tougher to tame, so consider that. Low-frequency stability can be an issues as well.
An good primer for building any Williamson is Dave Gillespie's excellent thread at Audiokarma:
https://audiokarma.org/forums/index.php?threads/regilding-the-gilded-lily-heaths-w-2m.767851/
Dave lays out the basic issues of stability and build quality. Of particular note are the changes he makes in the coupling cap values to improve LF stability. I always use this arrangement and it works very well. His low-pass filter for the W2 is quite steep, and I find that with a Peerless OPT you needn't go that far. But it makes for a rock-solid amp.
Sorry to go on (!) but this is a pet project of mine. ;-)
Regarding a scope, you can buy them used for very little. For basic amp testing, the rudiments are not hard to learn. I did it in a couple of weeks when I realized that I would need to if I wanted to continue experimenting with Williamsons. Anyone who wants to build a quality amp should learn how to use one, especially if feedback is involved. I use a Picoscope 2206B because the built-in signal generator goes up to 1MHz, which is important for checking ultra-high-frequency stability, and I like the software a lot, but that may be more than you want to spend.
Anyway, hope this helps! Feel free to PM me if I can be of any assistance. I've walked a number of folks through Williamson builds. It's a challenge but it's fun.
VOB,
Use a good DMM to measure the DCR of your Vintage Philips 12 inch full range loudspeakers.
If the full range are truly a single driver, then the DCR is the minimum impedance of the loudspeaker.
Let us know what that DCR is, please.
And if your loudspeakers are really efficient, you can use only a small portion (percentage) of the maximum output power of your 807 amplifiers.
That will help to keep the distortion low.
Use a good DMM to measure the DCR of your Vintage Philips 12 inch full range loudspeakers.
If the full range are truly a single driver, then the DCR is the minimum impedance of the loudspeaker.
Let us know what that DCR is, please.
And if your loudspeakers are really efficient, you can use only a small portion (percentage) of the maximum output power of your 807 amplifiers.
That will help to keep the distortion low.
The spec says "6CA7 Etc. Rl = 8000 Ohms". 6CA7 is second cousin to 6L6 which is 807 in different bottle. Which makes sense because you won't get 40 Watts out of EL84 at any voltage they will live with, but 40W@8K is in-reach for any of the 400+ volt tubes 807 6L6 6CA7 EL34 etc.
I don't see an issue with an 8k Ra-Ra OPT. I've seen schematics using 10k OPTs with 807s, and load lines I've played with seem to find a higher value is better than a lower with 807s/1625s.