I neglected to mention previously I'd added variable attack and release to it which required a good bit of switch rewiring. In stereo mode the right side attack and release control everything. Some pics from 2011.
Serviced this one again last week. The output tube sockets had gotten in such bad shape they would hardly hold the tubes, leading to some red plating in one from lack of grid connection. Replaced the sockets, also burnt cathode resistance on one side, went ahead and replaced all cathode resistances on both output sides as all were significantly out of spec.
Discovered it would not align for gain reduction that matched in both dual mono and stereo. It was down to the 2N1307 shunt transistors not matching for HFE, along with associated resistances having drifted significantly. Germanium not matching, imagine that! Surprisingly there was a relatively inexpensive source of NOS pieces, so I bought 10 to select 4. 3-4 were borderline DOA, the remainder just barely got me acceptable matches.
R501 (39K) on the transistor daughter boards is the threshold setting, this is not documented anywhere. This unit had been modified to differing thresholds in each channel, both lower than stock. Going back to 39K pushes the threshold much closer to the headroom of the amp, so transients from slower attack times get distortion/limiting from hitting max output. From listening 28K or so seems like a reasonable threshold for use with an attack modification such as I did, keeping transient overshoot further away from max level.
While the iron looks small compared to a BA-6, it's all pretty good. The 1:1 interstage measures 256H on each side at 120Hz with a handheld meter, which is about double the inductance found in the Sowter BA-6 repro transformer with the same meter. With the SS daughter boards in there we have 4 transformers in the signal path, and response is just about +/-1dB 20-20K.
I once noticed a unit being sold with sales copy stating 'improved with deluxe UTC A-18 interstage transformer'. I think that was just all someone could find to do a replacement with after one died. For one, the ratio is wrong, and assuming it was connected as a step-down like many limiters, would lead to a different threshold point.
Comparing manuals and units is a bit confusing, as the BA-6 style Bal 1 / Bal 2 controls are no longer critical components since this one is a fixed gain tube amp with a SS shunt limiter front end. The SS front end seems most particularly to have increased attack time to a point that called for critically matched 6K7's in the vari-mu version. I imagine this was the #1 reason Gates made the change. I added a few new docs in the Gates thread of the Technical Documents area.
The client (Strange Weather) liked the sound and did not feel a need to explore changing it to vari-mu operation. I would agree, not having heard the vari-mu version either. I have been told it sounds good too.