Using a UTC OPT for "ultralinear" operation

zackthedog

zackthedog

AK Subscriber
Subscriber
I was reading this interesting article at Tim Robbins' Williaosn site:


I've cropped out the schematic for easy reference:

IMG_0417.jpeg

The authors have taken a UTC LS-57, which has 5K and 3K primary (60%) taps, and reversed the P and B leads to provide a 40% tap for the screens. My question is this. A UTC LS-63 has taps for 10K and 6K operation. Is it safe to say that this represents the same 60/40 relationship, such that one could equally use the 6K taps as 40% screen taps by reversing the P and B leads? I assume so but knowing as little as I know about OPT winding rations and the loads they represent, I thought I'd check.
 
Register to hide this ad
Yes, that's the same impedance ratio, although I can't tell you which tubes it would be appropriate for. That schematic is certainly drawn in an odd way. Never seen one where the primary windings didn't line up with the tubes driving them.

Jack
 
LOL. It does take a little staring to figure it out. Interesting use of plate-to-plate feedback, too. No reason you couldn't use this arrangement in a KT66 Williamson, or any other UL desing as well. I have some UTC OPTs and was resigned to using them only in triode or pentode mode, but this shows you can use them as UL trannies as well.
 
I actually thought i was looking at a unity coupled setup at first before i realized where all of the leads actually went.
 
It's interesting that the UTC catalogs don't make any mention of separate B+ taps for each primary half winding, so that was an interesting observation of the article's author, given it was 1953.

That author appears to have had some 'issues' with instability, and perhaps principally due to the modified OPT primary winding configuration, given the need to include not only local feedback from output stage to driver stage, but also additional roll-off 16kHz RC on the driver common cathode, although no first stage step filter or feedback comp capacitor were used. The author also makes some remarks about the choice of first stage valve, which were not substantiated with any measurement data, and I note the Editors took the auther to task on some other comments made in that article.

That re-use of a standard PP output transformer for UL was also done in Oz for the Ferguson OPM7A, with 7k and 10k PP taps in a 6GW8 guitar amp magazine project, where the B+ connection was made with separate taps to allow 'reverse' P and B+ connections. Of a range of about 20 models in the OPM series, only 2 models had that style of taps, so it would appear to have been done to increase the opportunity for increased sales.
 
trobbins said:
It's interesting that the UTC catalogs don't make any mention of separate B+ taps for each primary half winding, so that was an interesting observation of the article's author, given it was 1953.

That author appears to have had some 'issues' with instability, and perhaps principally due to the modified OPT primary winding configuration, given the need to include not only local feedback from output stage to driver stage, but also additional roll-off 16kHz RC on the driver common cathode, although no first stage step filter or feedback comp capacitor were used. The author also makes some remarks about the choice of first stage valve, which were not substantiated with any measurement data, and I note the Editors took the auther to task on some other comments made in that article.

That re-use of a standard PP output transformer for UL was also done in Oz for the Ferguson OPM7A, with 7k and 10k PP taps in a 6GW8 guitar amp magazine project, where the B+ connection was made with separate taps to allow 'reverse' P and B+ connections. Of a range of about 20 models in the OPM series, only 2 models had that style of taps, so it would appear to have been done to increase the opportunity for increased sales.
Click to expand...

So Tim, in your opinion, is this an invitation to trouble? ;-) Based on my experienced, even in triode mode the LS-57 is a bit tricky to stabilize, but not impossible. ;-)
 
Where's the fun if you don't have a go! I'd anticipate that UL mode should be as mundane as triode mode, but I haven't worked through any Williamson's with UL, or where the primary half-windings could be reconfigured. So the outcome could well be benign and no cause for hesitation, but similarly one could find that oscillation kicks in for some extreme load and then you may or may not want to do the hard yards to provide a stable outcome whilst considering the consequences that may have on audible response.

As background, I felt like I was flying blind before I started using FRA4Picoscope and could confidently relate gain and phase changes above 96kHz with methods to suppresses resonances at frequencies where I could see a looming gain peak (especially when trying to avoid oscillation for known adverse loading conditions like capacitance only), and could relate such a gain peak's frequency with the oscillation frequency. Sometimes it seemed like squarewave responses were telling me what I needed to know, but there was sometimes doubt with interpretation. As anticipated, I've noticed that vintage OPT's designed for Williamson triode-mode amps are inherently easier to make unconditionally stable, and may only require a whiff of compensation, and so the 'fun' I've been having is to work through what can be done for those OPT's that are not so credentialed. I managed to get unconditional stability on one Williamson operating pentode mode - perhaps assisted by the OPT being a WWFB - but that required some effort.
 
Remember not to confuse impedance ratios with turns ratios. If you're looking to use 6L6 class tubes as the authors did, the target screen tap location for the Ultralinear specification is 43% of the plate winding -- not 43% of the plate impedance. In the article then, relative to the UL specification, the screen tap location is actually at 22.5% of the winding using their configuration, or about half of that needed. If you use a higher impedance transformer in this configuration with the same percentage of impedance spread between the primary taps, you will still get the same percent of tap placement relative to turns ratio, or 22.5%.

With 6L6 class tubes, this configuration is theoretically no doubt a better way to use the 3KΩ taps for tapped screen operation than to use it with the P and B+ terminals conventionally connected, as that configuration would really cause performance to suffer. However --

This "backwards" UL configuration using a 10KΩ transformer with 6KΩ taps should be well suited for 6V6 or 6BQ5 class tubes, since they operate best with an Ultralinear screen tap placement at 24-25% of the winding.

Bring on the instability challenge!

Dave
 
^^^^Okay, THIS is what I mean! This is hard for me to wrap my head around. Thanks for clearing that up, Dave! Brewer, if I googled the right guy, was an amateur ham operator and not an EE at all. Odd that this wasn't "peer reviewed"!
 
If you look through some of those old magazines there are circuits and designs that we'd now consider to be less than optimal ideas. Its the benefit of hindsight :)
 
Yes, I agree, though I find it odd that the editor didn't catch the impedance vs. turns ratio thing. Though perhaps in the next issue, someone did...
 
Last edited:
You must log in or register to reply here.
Back
Top Bottom