I built an 4X el36 triode amp using the Williamson circuit, 6GU7-5687. 5K P-P out put transformer.
I was wondering if the EL36 as a triode could be driven by just a single gain stage and a Cathodyne phase splitter/ Recently a circuit came out in AudiXpress using the ECC99 to drive El34's.
Not sure how much drive the EL36 needs as a triode, would be no problem driving them as pentodes or ultra linear..
Phil
I was wondering if the EL36 as a triode could be driven by just a single gain stage and a Cathodyne phase splitter/ Recently a circuit came out in AudiXpress using the ECC99 to drive El34's.
Not sure how much drive the EL36 needs as a triode, would be no problem driving them as pentodes or ultra linear..
Phil
Hi Phil, great to hear about your Williamson amp. The EL36/6CM5/PL36 family are very linear triode strapped - Tom Schlangen from this forum and others have done a lot of work with this valve as an output in a SET amp. In addition, the mu of this valve as a triode makes it possible to be driven easily by just one triode as driver - I am working on a SET amp at the moment using triode strapped 6CM5's driven by just one triode and hope to post full details here soon
Last edited:
Patrick Turner published a push-pull design using these tubes in his website...2323-triode-integrated-6cm5
i tend to favor the mullard 5-20 topology over concertina types though....
i tend to favor the mullard 5-20 topology over concertina types though....
Yes, they are: Le Renard Main Schemo
PP 6BQ6 Loadline
I designed this project based on the 6BQ6. I came across a Portuguese (Brazilian?) ham mag from the late 1950s that described an AM plate modulator project that used 6BQ6s as audio finals. That looked very odd, as you'd expect to see 6L6s used here. So I took a closer look, and the loadline looked very promising. Besides, 6BQ6s were going very cheap, about a buck a pop, at the time (though the price has gone up -- sorry 'bout that).
As an audio final, the 6BQ6 is every bit as good as the 6V6, but can handle a good deal more power. Though rated like a 6V6, the 6BQ6 can be biased a bit hotter (Pd= 17W Increase Ipq to 50mA) without red plating, and that moves the Q-Point farther towards Class A with a definite improvement in sonic performance. The RCA gray plate 6BQ6s won't red plate pulling 50mA. At 70mA, these show a bit of color in the dark, whereas the Sylvania black plate 6BQ6s aren't glowing even then (though this would stress the PTX a bit). As far as distortion, the 6BQ6 makes mainly H3 with very little H5 and higher. Even run open loop, they sound a bit "edgy", but not "pentode nasty", which is not the case with 6L6-oids which need the extra help of local NFB to tame that higher order harmonic nastiness. All you need is just enough gNFB to take the excessive edge off. I suppose that's why the Portuguese (Brazilian?) hams decided i favor of the 6BQ6 -- less nastiness going out over the ham bands,
I guess the reason there is no audio spec is since these can't be run as Class A audio amps. The best part of the plate characteristic for Class A is well into red plate territory. (George confirmed this.)
6BQ6GTBs and 6BQ6GAs both work. Avoid the 6BQ6GTAs, however, these aren't quite so tough, and red plate badly with these Q-Points. (Internally, these seem to have smaller, thinner cathodes.)
Other TV HD power finals also look promising as audio finals, some of which can make some really bigwatts with the right OPTs.
Even those 6BQ7s, which have no audio applications mentioned in the spec sheet, are excellent performers, at least when used in the cascode LTP. As for more conventional audio use, I dunnow how they'd perform. The plate characteristic looks odd (undocumented variable-u characteristic?) and you need to get that Vpk up to hit the most linear part of the characteristic.
One thing I did discover is that you want to avoid 6BQ7s with series connected heaters: they ring like bells. The ones with parallel heaters aren't noticeably microphonic.
Just wanted to update anyone planning to build my 6CM5 amp that I have had some feedback from constructors that idle current can be too high with some 6CM5's even at max setting of the bias pot. The solution is to increase the value of C7 from 220nf up to 470nf or higher. It seems some valves need in excess of -70V bias
I did the write up over at the Hawthorne Audio forums under a different handle: Part I: VT Amp Design Walk Through
thanks, i took the liberty of linking that to our local website, hope you don't mind....
Hi All, I built an amp (two channels) using 6CB5A TV set tubes, Push Pull UL, in class AB. I wound two OP transformers of the Arcosound incorporated with a Williamson designs with the idea so I could use the low required 200 VDC on the screen grid circuits and still have 750 VDC on anode circuits without destroying the tubes. Lowest distortion was found at 25 % UL taps on the tertiary windings. ( oscilloscope waveform analyzing) I drive it with a low distortion Williamson circuit and when I finished testing and setting it up, what an incredible sounding amplifier . When it was out in my workshop, I carried out the loudspeakers , hooked them up in place of the dummy loads , played music instead of sine waves , and said to my family you had better come and listen to this . One has the impression you are in front of performers. The set up is now in the living room giving us hours of listening pleasure. By the way it will do 100 W RMS output before clipping or top of waveform distortion, into the 8 ohm wire wound resistor network sitting in a cup of water for cooling. It will start to boil the water in about 5 minutes at 100 Watts in to it.
Nice to hear you had a successful project.
Did you wind the OT's with a few different primary taps in order to substantiate the 25%, and were you 'eyeballing' the distortion levels from scope waveforms as the comparison ?
I am part way through a 6CM5 UL amp, but it is very much non-hifi (as compared to Grant's 15W). The default screen taps available from the cheap (in Australia) Altronics M1115 line transformer with 8k PP are 42% taps.
Did you wind the OT's with a few different primary taps in order to substantiate the 25%, and were you 'eyeballing' the distortion levels from scope waveforms as the comparison ?
I am part way through a 6CM5 UL amp, but it is very much non-hifi (as compared to Grant's 15W). The default screen taps available from the cheap (in Australia) Altronics M1115 line transformer with 8k PP are 42% taps.
Hi there,
Yes I did put taps in the tertiary windings to experiment with, starting at 50% then 10 % taps down to 25% I put in a 43 % tap as well . They were a mission to wind but I got there in the end. A good turns counter is a must and also a 1000 volt insulation tester and variac. The primary is untapped apart from HT CT, but can quite easily be tapped as each layer has to be brought out to the side of the bobbin to connect the next layer as there is a section of secondary and tertiary in between them so there are a porcupine of connections down the sides. 8 k ohms to 12 ohms or roughly 27 to 1 is the turns ratio, actually 3600 turns anode to anode and each secondary section 70 turns times two series, or 875 volts AC to 34 volts AC. This seems to match the 8 ohm Philips speakers ok. The bass will rattle the china cabinet well, in fact other things like some walls. Its loud when that's happening but still clean audio. The core size is what you would use for a 50 or 60 Hz power transformer of around 750 to 1000 VA in size so the amp is heavy with two of these bolted onto the chassis plus the power transformer made of the same bobbin and laminations. I live in Tonga so one has to make do with what you can find for projects like this. There are some good articles on the net about distributed loading for ultra linear output circuits. I did eye ball waveforms as I have no distortion meter for audio use . I do have a Fluke power analyzer but I found that no good at frequencies other than 50 or 60 Hz or mains voltages. I have a dual channel Hitachi V422 scope and fed direct waveforms from audio generator into one channel the the output of amp into the other. I could see by lining up the two waveforms and even inverting and adding of one channel to see distortions and phase differences. I ended up with around 6 dB global negative feedback, this seems to keep the amp stable and sounding vibrant. It was a little trial and error to achieve good high freq stability , the whole thing was wanting to oscillate at around 100 kHz in the beginning. This was caused by phase shift of the output transformers at high frequencies as the interwinding capacitances are a little high (around 5000 pF from memory prim to sec) due to the interlayering of the windings to obtain good coupling. The fix was to put a a low pass filter RC network on the anode of the first input amplifier tube, to roll off the high freq end as others I read about had had similar experiences and done the same thing. My bias is set at 22 mA cathode current each tube, I think I saw crossover distortion down around 10 or12 mA current on each of the tubes. ( I should write everything down at the time of testing) 750 VDC HT anode supply The amp was going to be for experimental purposes with all the tertiary taps on the OPT's to try a few different output tubes, which I have, but now it sounds so good , I have left it as it is. I have begun building another amp like this for a friend , I have nearly finished the two output transformers but I have run short of the 0.8 mm secondary winding wire . Hope to visit the motor rewinding shop next week. I have a set of 807s to try and also a set of used 6DQ6's, which I used in an amp I built around 15 years ago now. My new design now is a much better amp than what I built back then. One lives and learns along the way.
Good luck with your project.
Best Regards
Murray.
Yes I did put taps in the tertiary windings to experiment with, starting at 50% then 10 % taps down to 25% I put in a 43 % tap as well . They were a mission to wind but I got there in the end. A good turns counter is a must and also a 1000 volt insulation tester and variac. The primary is untapped apart from HT CT, but can quite easily be tapped as each layer has to be brought out to the side of the bobbin to connect the next layer as there is a section of secondary and tertiary in between them so there are a porcupine of connections down the sides. 8 k ohms to 12 ohms or roughly 27 to 1 is the turns ratio, actually 3600 turns anode to anode and each secondary section 70 turns times two series, or 875 volts AC to 34 volts AC. This seems to match the 8 ohm Philips speakers ok. The bass will rattle the china cabinet well, in fact other things like some walls. Its loud when that's happening but still clean audio. The core size is what you would use for a 50 or 60 Hz power transformer of around 750 to 1000 VA in size so the amp is heavy with two of these bolted onto the chassis plus the power transformer made of the same bobbin and laminations. I live in Tonga so one has to make do with what you can find for projects like this. There are some good articles on the net about distributed loading for ultra linear output circuits. I did eye ball waveforms as I have no distortion meter for audio use . I do have a Fluke power analyzer but I found that no good at frequencies other than 50 or 60 Hz or mains voltages. I have a dual channel Hitachi V422 scope and fed direct waveforms from audio generator into one channel the the output of amp into the other. I could see by lining up the two waveforms and even inverting and adding of one channel to see distortions and phase differences. I ended up with around 6 dB global negative feedback, this seems to keep the amp stable and sounding vibrant. It was a little trial and error to achieve good high freq stability , the whole thing was wanting to oscillate at around 100 kHz in the beginning. This was caused by phase shift of the output transformers at high frequencies as the interwinding capacitances are a little high (around 5000 pF from memory prim to sec) due to the interlayering of the windings to obtain good coupling. The fix was to put a a low pass filter RC network on the anode of the first input amplifier tube, to roll off the high freq end as others I read about had had similar experiences and done the same thing. My bias is set at 22 mA cathode current each tube, I think I saw crossover distortion down around 10 or12 mA current on each of the tubes. ( I should write everything down at the time of testing) 750 VDC HT anode supply The amp was going to be for experimental purposes with all the tertiary taps on the OPT's to try a few different output tubes, which I have, but now it sounds so good , I have left it as it is. I have begun building another amp like this for a friend , I have nearly finished the two output transformers but I have run short of the 0.8 mm secondary winding wire . Hope to visit the motor rewinding shop next week. I have a set of 807s to try and also a set of used 6DQ6's, which I used in an amp I built around 15 years ago now. My new design now is a much better amp than what I built back then. One lives and learns along the way.
Good luck with your project.
Best Regards
Murray.
Hi Murray, many are using a PC soundcard for audio spectrum analysis nowadays, including distortion assessment. Even better if you can get a 2nd hand USB audio interface like a Focusrite. It makes the distortion assessment a breeze with free software like REW but requires a little effort for interfacing to and from the DUT amp.
One good aspect of your scope is the ability to do X-Y phase and gain margin testing at low and high frequencies if you have an oscillator with suitable range. Aiming for Williamson style bandwidth and feedback levels pretty much requires a good awareness of the raw gain and phase response. Although I can see that the experimental use of tertiary windings would be a significant impost to leakage inductance and shunt capacitance, as 5nF is almost an order of magnitude higher than other PP OT benchmarks used for Williamson amp.
I can appreciate the 'have to do with what I've got, or can easily get' situation you are in. I should have waved to you as I flew to Niue a few years ago.
Ciao, Tim
One good aspect of your scope is the ability to do X-Y phase and gain margin testing at low and high frequencies if you have an oscillator with suitable range. Aiming for Williamson style bandwidth and feedback levels pretty much requires a good awareness of the raw gain and phase response. Although I can see that the experimental use of tertiary windings would be a significant impost to leakage inductance and shunt capacitance, as 5nF is almost an order of magnitude higher than other PP OT benchmarks used for Williamson amp.
I can appreciate the 'have to do with what I've got, or can easily get' situation you are in. I should have waved to you as I flew to Niue a few years ago.
Ciao, Tim
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.