Yes, this is quite a bit more encouraging. ;-)
Williamson power and distortion measurements
- Thread starter zackthedog
- Start date
Yes, this is quite a bit more encouraging. ;-)
I think triode mode is probably going to produce less distortion than UL. Also, I think the S-265-Q, whether the original Peerless or the Heyboer, is not going to produce as good results as a Partridge or similar OPT that. In triode mode, the best I was able to produce was 12 watts in the midband before the onset of clipping, which is completely in line with the original "Musician's Amplifier."
A while back, I restored a vintage "Musician's Amplifier" with a Partridge output that some well-heeled DIYer had ordered from England back in the day. Unfortunately it must have been in bad condition, because it produced a *terrible* 10kHz square wave. Maybe humidity or some other factor had taken its toll on the windings. I can hardly believe it was performing as intended. I had to roll it off at 18K to get any sort of stability out of it, so it was impossible to judge what it was originally supposed to sound like.
A friend of mine with deep pockets wants me to build him a pair of Williamson monoblocks with Finemet OPTs. Oddly, they make a 10K p-p 60 watt model with 50% taps. For what, I'm not sure, but it's a perfect match for a "Gilded Lily." Someday I would also like to try the Sowter "Williamson" OPTs, which, relatively speaking, are not completely outrageous in price. ;-)
Maybe next weekend I'll pop in some 5881s, Russian 807s, Chinese KT66s and some other tubes I've got and compare the THD figures for each.
I meant to add that, yes, the amp likes them, and I think NOS tubes might well deliver lower distortion. The RCA 807s are quite "romantic" with a bit of a halo around the music. Not a bad thing if that's what you want. The Gold Lions have a slightly drier, more "modern" sound, if that makes any sense. ;-)
For comparison, my Williamson mono's (B+ 435VDC, Edcor CXPP7.6k/25W OPT) :
1 Watt / 8 Ohm
Full Power @ 2% THD+N, 1kHz / 8 Ohm
Full Power @ 2% THD+N, 20kHz / 8 Ohm
FFT / 1 Watt / 8 Ohm
Power Response / Frequency Response @ Rated Power (Power level set to 20 watts, 2% THD, 8 Ohm)
1 Watt / 8 Ohm
Full Power @ 2% THD+N, 1kHz / 8 Ohm
Full Power @ 2% THD+N, 20kHz / 8 Ohm
FFT / 1 Watt / 8 Ohm
Power Response / Frequency Response @ Rated Power (Power level set to 20 watts, 2% THD, 8 Ohm)
Last edited:
You should not have to lower anything for 807's. I'm running my 1625's at "normal" voltage levels as found in the original Williamson circuit.
@zackthedog
Have you read DaveG's article about his update for the W5's? He solved the distortion issue at higher frequencies at higher power by ditching the 50% screen tap connections, adding a fixed bias supply and using the secondary as a cathode feedback winding. I just finished doing this for a pair of W5's owned by a friend. Subjectively they have very high detail resolution, the treble at all volumes is possibly the cleanest I have ever heard and they maintain composure right up to the higher power end of things.
If the Heyboer clones have a well balanced secondary you should be able to experiment with the cathode feedback on your amps.
John
edit: for spelling
Have you read DaveG's article about his update for the W5's? He solved the distortion issue at higher frequencies at higher power by ditching the 50% screen tap connections, adding a fixed bias supply and using the secondary as a cathode feedback winding. I just finished doing this for a pair of W5's owned by a friend. Subjectively they have very high detail resolution, the treble at all volumes is possibly the cleanest I have ever heard and they maintain composure right up to the higher power end of things.
If the Heyboer clones have a well balanced secondary you should be able to experiment with the cathode feedback on your amps.
John
edit: for spelling
Last edited:
I guess I was just being a little cautious. ;-) I'll try them at the regular voltage.
No, the Heyboers are built like the original 265-Qs, with two center-tapped windings in series or parallel rather than taps off a full winding. I do have two sets of W5 OPTs though and plan to try Dave's build.
Does anyone know if Heyboer has ever been asked to reverse engineer the Peerless 16309 or 16458 outputs from the W5's? Just curious. Seems like there should be some failed 309's out there that could be offered up if someone wants to pay the fees. I might be willing.
I don't believe so. Pacific Transformer made a copy of the 16309 for a DIYer years ago. I have a pair here, I actually breadboarded them and they work, but that was well before I was able to use a scope. I could take another look at them. There are two issues with duplicating the 16309, as I see it. 1) The nickel stripe in the lams that extends the low-frequency response of such a small OPT (the Pacific copy lacks it), and 2) the built-in cap across one primary (which the Pacific copy also lacks). The cap could be added externally provided the value can be determined. I just saw a pair of bad 16309s on eBay, don't know if they're still there. But we'd have to ask Heyboer if they can source the nickel. The 16458 would probably be easier to replicate, especially since Dave speculates that it is probably a 16277 with reconfigured secondaries. The problem there is that they rarely, if ever, failed. ;-) So a good one would need to be sacrificed. I have a pair here, and you could probably find an orphan single on eBay, they do turn up. Then the question is, would Heyboer get the secondaries exactly right to allow for the balance needed for the cathode winding. Hmmmmm. They did a very nice job with the 265-Q, but...? It could cost $700-800 and be a bust.
There have been many articles over the years that have validated using 807 or 1625's in a Williamson circuit even though the screen voltage is exceeded in the Williamson design.
I know. It was just a rectifier swap. I can try them at full tilt. ;-)
I emailed Arlyn at Heyboer asking about the availability of nickel lams. I have a single 16309 I'd be willing to sacrifice.
It seems Lundahl, Intact and others are able to source nickel lams, so it ought to be possible. Price is another discussion.
trainbuftony
Electron Herder
I'll play!
These aren't technically 100% Williamson, but basically are the Williamson design using an Acrosound TO-300.
Making 25w, just over 1% distortion midband.
At 20 hz we are still "flat" same at 20k.
Distortion at 20 hz is 1.5% at full power.
Distortion at 20khz 1.8%.
These aren't technically 100% Williamson, but basically are the Williamson design using an Acrosound TO-300.
Making 25w, just over 1% distortion midband.
At 20 hz we are still "flat" same at 20k.
Distortion at 20 hz is 1.5% at full power.
Distortion at 20khz 1.8%.
@trainbuftony, what is your B+ going into the OPT primary?
...and can we see some square waves? ;-) Please play along, by all means!
^^^^^!!!!!