Scott AW23 Volume Expander Info

[J.N. Shetler]

I have an optional volume expander for the AW23. I have the expander schematic dated November 30^th, 1936 (from the Scott Info Archive). I have found some huge changes in my own unit and I’d like to know if they were a later revision or an earlier version that should be changed to match the schematic. The parts look period correct and the solder joints look undisturbed, so I believe the changes are not a "hack". Although I know the expanders are uncommon, I'm hoping a few members of the site have rebuilt one and found the same deviations. First, the bias supply on mine differs from the print in that the 400ohm (later noted as 750 ohm) resistor off of the power cable is omitted completely. The bias from the power supply connects directly to the junction of the cathode and grid resistors of the 76 tube.  The 2500 ohm and 800 ohm series connected resistors from the negative bias to ground are gone and one 3300 ohm resistor is in their place. The 2 Meg resistor from the grid of the 6C6 should go to the junction of the 2500 ohm and 800 ohm in the schematic. It is now grounded. That covers the bias supply. The other revision is in the screen supply for the 6A7s. On the print there are a 10K and a 2K connected in series off of the B+.  The junction between them supplies grid 4 of the 6A7s and grid 3 of the 6C6. These two resistances are actually a huge ceramic candohm resistor. The 10K is correct but the 2K is actually 8K. This dual-resistor is marked as such. Also at this junction is a .25mfd cap. The schematic alternately lists a .5 mfd in parentheses. Which is preferred? I assuming this adjusts the return time constant. Any thoughts on these changes?  Something that should be left as a later factory version, or something that should be repaired to match the schematic?  Thank you for any help or advice you may be able to provide!

[J.N. Shetler]

I just took another look at the print. That change from a 2K resistor (on the schematic) to an actual 8K resistor (in the expander) is connected to the high side of the rheostat for the cathodes of the 6A7s. I think the 8K value would lower the range of operator adjustment of the 6A7 cathode voltage. Max setting would have the cathodes grounded. Minimum setting would place a higher potential on the cathodes and minimize tube current. I can only assume this resistor change was done for a reason.

[J.N. Shetler]

I had a few minutes to get back to the workbench this morning. The situation is even more puzzling. I decided to draw up a new amended schematic to match the chassis and once again double check all of the parts and resistances. In addition to the above deviations, there is one other notable change involving that 2 section candohm resistor that supplies screen and suppressor potentials. On the schematic, Pins 4 and 5 of the 6C6 are jumpered and grounded. In the unit, they are jumpered but do not go to ground. They go to the high side of the cathode rheostat for the 6A7 tubes. This is the junction of the 300 ohm rheostat and the 8K (2K on the print) section of the candohm.  This would elevate the 6C6 cathode and suppressor grid to positive potential and cut the gain. I also noticed I mistakenly identified the common screen grid decoupling cap in my first post. It's the cap that has either .25 or .5 listed on the print. Still curious if anyone has tackled this before.... or if there are any schematics of the AW27 or Quaranta expander circuits out there to compare. It would be good to know if the 11/30/1936 schematic was a workable unit. In any other chassis it'd be easy to try a few different arrangements and pick the best performer, but this is such a difficult unit to get apart and service (you have to desolder all the cables and half the components on the lower phenolic board), it's a one shot repair.

[J.N. Shetler]

With no other options/replies, I configured and tested the expander three ways: in the "as found" format (likely early production), in the version of the November 36' revised schematic, and also as was done in the AW27 (I was lucky enough to find high quality photos of the underside of the upper/lower chassis to give me parts values).  There were very distinct operating differences. I examined: the "expansion level" of 6A7 grid 1 (pin5) bias relative to input signal, the operating characteristics of the 6A7 and the 6C6 in each version, and the overall effectiveness of the feature. Scott seems to have been playing with values to "dial in" the circuit during production (the gain of the tubes, biasing, idle current, time constants, etc.). Some of their changes work far better than others. Ultimately, I used some circuit segments of all three versions and did a little math of my own with receiving tube spec sheets. I will say that when in optimal working order, it performs quite well and doesn't deserve the bad reputation the expanders are sometimes given. I think the earliest version used a time constant that was too slow for both ramp up and return, also the "expansion" relative to the input signal was so high it sounds unnatural. In the later versions, this was addressed and it sounds/functions much better. There is a threshold signal level under which the 76 diode effects no change on grid 1 of the 6A7. This I admit I modified to my own liking. From the factory, the expansion would only work at higher volume (they concede this in the operating instructions). Now it is active at any level when the expansion control is rotated. Case closed on this one, maybe the observations above may be helpful to someone in the future who stumbles onto this thread.