DCAudioDIY.com

I'll also add that there aren't too many other push-pull designs. Somehow, somewhere, you've got to split the signal into two opposite phases. Apart from the Williamson, which uses a "cathodyne" splitter, you can use a transformer to split the phase, or you can use a Mullard circuit, which directly couples the voltage amp stage to a common-cathode "long-tailed" phase splitter. There's also a boot-strapped phase splitter, the "paraphase" and the (my memory fails me here) Van Voycik or something like that, all of which are more complicated.

Thanks for the explanation. It seems that it is harder to design an excellent pp amp than an excellent se amp because of the phase splitter and nfb issues.

ray

Yes, exactly, Ray. The SE amp craze has been a boon for beginners because it's quite easy to contruct a stable, nice-sounding 2-10 wpc SE amplifier. Triode push-pull amplifiers can be more challenging but still with the grasp of a novice. Once you get into feedback, it's a daunting challenege, even for experienced builders. I've learned enough oscilloscope skills to make rough/cursory adjustmetnts on my Williamson projects, but many of the technical aspects of feedback are still way beyond me. On the heels of his excellent Heathkit W-2 thread, Dave Gillespie produced another fascinating pair of threads about the Heathkit W-3, which employed the ultra-bandwidth Acro TO-300 transformer, with near-disastrous results. The complicated feedback network Dave came up with to stabilize this amplifier is admirable, and beyond the skills of many of us:

https://audiokarma.org/forums/index.php ... 3m.769029/

https://www.audiokarma.org/forums/index ... am.772780/

Also, if you read these threads, you understand the significance of the square wave shots I posted for the Acro Williamson project.

Grover,

What do you think about using UTC LS-55 for a classic triode connected 6L6 Williamson (i.e. the original circuit). I have a pair sitting on the shelf unused, and this might be a good project if it has the bandwidth (the catalog info says 7-50kHz +/- 1 dB, though that is usually debatable). Output power is limited (20W).

David

Well, as chance would have it, here's a video from a guy who built a triode-wired EL34 Williamson with the LS-57, which is the same transformer without the line output taps. He does a lot of impressive measurements. As the caption says, he's using 16dB of feedback. This is a very recent vid, so you could probably reach out to the guy and find out what feedback resistor he's using to get the 16dB. At one point he mentions Acrosound transformers, but I think he misspoke because he's talking about the UTCs.

https://m.youtube.com/watch?v=-dVyePOA7-c

I've used EL34s and they make a great-sounding Williamson. But you could probably adjust the feedback for 6L6s.

Grover,

Thanks for the link to the video. I think that just might be my next amp project! Got to clean up my bench first though (and what am I going to do about all my other amplifiers, I have 4 or 5 stereo pairs just sitting around gathering dust....).

I used those UTC transformers in a PP 2A3/300B amp you heard years ago. I replaced them with a pair of nice 60W 5K Hashimoto transformers I bought in Japan which are probably overkill for the amp. I had thought about putting the UTCs back in to free up the Hashimotos for my current PP amp. The Hashimotos also would allow a bit more output (and better LF) than the UTCs. So have to think about which way to go.

Will have to check on what he was using for feedback, so perhaps I will drop him a line.

David

Heck, use the UTCs! Look at the measurements he got with a simple feedback arrangement. 20wpc at .5% THD! That's one powerful, stable Williamson! Ask him for a schematic, if he's willing, or at the least what shelf network he employed, if any, and what feedback resistor. Otherwise it looks like he prefers the stock Heathkit circuit, though he does use that balance pot on the driver stage plate resistors, probably dials it in for lowest distortion. He says in the video he used 150pF for the phase lead capacitance.

Grover,

I think I will use the UTCs with a PP EL34 triode Williamson. Yes, I did note the 150pF value on the phase lead capacitor. I will have to dig up the Heathkit W4 schematic and look at it and contact him for his schematic (or changes in his feedback network). I might just build it as a stereo amp as well, I have a studio down in the basement that I am currently using my Aria 5Rs with my stereo Mosfet Hybrid (25W/ch), so a nice (and compact) stereo tube amp like this would work out great, and would not be bothered by the mediocre LF response of the UTCs. This could be the perfect amp for a pair of bookshelf speakers of reasonable efficiency with a biamped subwoofer. Hmm.

Oh, and I have a quad of Genelex GL EL34 sitting on a shelf.

David

That sounds great. I've never gotten ahold of a pair of LS-63s or 61s to compare them to the Peerless. Sometimes they go for next to nothing and sometimes people want a fortune for them.

Grover,

Had a nice discussion with David Beard at ElPassTubeAmps via YouTube comments. I think I have the feedback and snubber arrangements that he used after viewing the video again.

I believe he is taking the feedback off the 8 ohm tap on the UTC transformer (there really is no way to do otherwise due to the way the transformer winding are interconnected. The video was clear enough for me to read the color codes off the resistor, and he is using a 6.5K resistor, with a 150pF parallel capacitor to get the excellent 10kHz squarewave response. He is also using a input snubber on the grid of the phase splitter referenced to ground, 220pF in series with 4700 ohms. I assume he is using the Heath W4 coupling cap values (0.22uF) between the sections. So I have a pretty good starting point.

I also have a Chicago P-6315 power transformer, 370-0-370 @ 275mADC, with a 5V/5A winding and 6.3V/7A winding that should be perfect for a stereo build. I suspect it was from your stash when you moved out to Oregon.

David