The Fine Things are Always Hand Made
In the process of restoring the Warsaw custom Philharmonic I learned that the record scratch suppressor circuit was not working. With the help of Bill Liers I was able to fix the circuit but the problem is an unusual but likely common one that is easy to fix once the problem is understood. The same circuit was probably used in the Phantom and other models, therefore the solution may also be applicable there.
During the late 1930s, some radio manufacturers were employing tubes designed for RF use in audio circuits. The EH Scott Radio Laboratories was one of these companies with the use of the 6B8G tube for the record scratch suppressor "diode". These tubes were designed to meet RF specifications therefore their performance in audio circuits may vary widely depending on the tube and the circuit. Later production 6B8G tubes were designed with greater transconductance for improved RF performance but not necessarily so for audio applications. Credit for this tube history information is due a presently unnamed acquaintance of Bill Liers who has been in the electronic service business for many years.
The symptom in the Warsaw chassis was "overaction" of the record scratch suppressor circuit when the circuit was engaged (record scratch suppressor turned on). Overaction consisted of a substantial reduction in the mid to high frequency audio notes and some reduction in volume. The entire broadcast or reproduced program became very bassy with the record scratch suppressor turned on. Scott service data suggests addressing this condition by replacing the coupling capacitor between the grid and plate of the 6J7G scratch suppressor control tube with a lower value. The original capacitor is a 35-pf mica capacitor (C55, not the erroneously noted C35 0.01-mfd, 200-volt capacitor identified on the schematic). Reducing the value of that capacitor did not affect the problem. Further investigation yielded low plate voltage and near normal screen grid and control grid voltages on the 6B8G scratch suppressor "diode". This was the case for a variety of 6B8G tubes tried in the circuit although weaker tubes yielded higher but still insufficient plate voltages. All supplies, resistors, and capacitors were tested and found to be well within acceptable limits. Ultimately it was substantially determined that the reduced difference between plate voltage and screen voltage was causing the tube to run "hot" (but not to the touch) and keep the record scratch suppressor full on for all input signal levels. The proper function of the record scratch suppressor circuit was restored by shunting the screen grid to ground with a 560k resistor (1m will be sufficient with a weak tube).
It is likely that the EH Scott Radio Laboratories was aware of this problem. During the run of Philharmonic receivers the laboratories increased the top section of the B+ divider resistor from 1200-ohms to 1700-ohms. The voltage charts in Scott service data were not revised with this change. Therefore the later divider resistor was most likely specified to correct voltages in the set because the original divider did not produce voltages expected by the circuit designer (plate voltages 10- to 15-volts low but other taps substantially correct). This change also serves to increase the difference between the screen voltage and plate voltage on the 6B8G record scratch suppressor "diode" and reduce circuit sensitivity to the 6B8G tube installed.
In addition to this discussion, I'd like to suggest to those restoring one of these Scotts, that there are often either "ruggedized" or low microphonic audio specific versions of many pentodes. These include:
using 1612 for a 6L7, or using a 1620 for a 6J7. For 6K7s and 6B8s you might be better off with the mil version if you can find them; VT-86 or VT-93 respectively. Of course, many std grade tubes work fine, but often you will find poorer noise performance and microphonics in the later pentodes.I especially like to use the 1612 tubes in the expander circuit. Typically very quiet and easy to match. As far as the differences in RF pentodes made after it was fashionable to use them in audio circuits ('37-'38), I cannot say for sure if the tubes were designed differently, possibly many of the test constraints that applied to audio were then simply overlooked.
This is good information guys... Thanks for posting.