This is a later production example which uses a tertiary winding IF transformer similar to model 38-89.

Some problem in the radio causes excessive AVC voltage about -20 volts at the RF and mixer end of the AVC buss with no signal; even with their grid caps disconnected. All voltages except the AVC and osc. grid (about -40 V) are normal. The radio is re-capped, re-resistored, and aligned. Tubes are OK and tube swaps make no difference. All coil resistances of normal. Osc. grid leak parts are OK. Occasionally the radio operates normally (but with some "staticy" noise) making this a somewhat intermittent problem. I suspect some sort of parasitic oscillation that causes excessive grid current (and excessive bias) that somehow works into the AVC buss. I can shunt the AVC buss with a resistor and make the radio work. But I'm stumped as regards the cause and real cure for the problem. Does anyone have suggestions? Thank you.

It is best not to replace all resistors unless they are faulty, as resistors don't often fail, and so if in good condition now, they will likely perform for years. However, if you did replace all resistors with their correct values, and didn't make any mistakes, then the set should perform properly.

You say that resistor 15 is of correct value. Be sure that it is so that the oscillator can't run away with oscillation.

Did you replace and/or test resistor 50? This is a multi-tap resistor in the negative portion of the power supply. If it is failing or opening up, especially on the chassis side of things, this can drive the AVC line, which is connected to this circuit, heavily negative. Also, other grids biased at points on this resistor below the chassis (more negative than) will exhibit more negative voltage than specified.

The problem you describe also sounds like it could be generated by spurious oscillation. Make sure that all filter capacitors big and small are connected and are operating properly. Capacitors 27, 17, and 47 (and any others I missed) filter RF and high frequency AF out of the B supply. Capacitor 6 filters the AVC line. If this capacitor is open or of too small a value, feedback can occur between the tubes the AVC circuit serves. Same is true of the B supply filter capacitors.

Be sure that all shielding that is supposed to be installed is installed. Be sure that grid and plate and antenna wires are kept away from each other.

Oscillation can take place above human hearing range, and so though you can't hear it, it can disrupt the set's operation.

:It is best not to replace all resistors unless they are faulty, as resistors don't often fail, and so if in good condition now, they will likely perform for years. However, if you did replace all resistors with their correct values, and didn't make any mistakes, then the set should perform properly.
:
:You say that resistor 15 is of correct value. Be sure that it is so that the oscillator can't run away with oscillation.
:
:Did you replace and/or test resistor 50? This is a multi-tap resistor in the negative portion of the power supply. If it is failing or opening up, especially on the chassis side of things, this can drive the AVC line, which is connected to this circuit, heavily negative. Also, other grids biased at points on this resistor below the chassis (more negative than) will exhibit more negative voltage than specified.
:
:The problem you describe also sounds like it could be generated by spurious oscillation. Make sure that all filter capacitors big and small are connected and are operating properly. Capacitors 27, 17, and 47 (and any others I missed) filter RF and high frequency AF out of the B supply. Capacitor 6 filters the AVC line. If this capacitor is open or of too small a value, feedback can occur between the tubes the AVC circuit serves. Same is true of the B supply filter capacitors.
:
:Be sure that all shielding that is supposed to be installed is installed. Be sure that grid and plate and antenna wires are kept away from each other.
:
:Oscillation can take place above human hearing range, and so though you can't hear it, it can disrupt the set's operation.
:
Thanks for your reply and I apologize in advance for this long winded follow up. It seems better to try and say everything in one shot.

I agree that oscillation is the problem, probably 6A8 grid 1/grid 2 caused, and have tried to address this. This must have been a problem for some time as the AVC buss was grounded (by a past serviceman ?) through a 5 k resistor from the rf trans. 8 term. 3 to ground when I got the radio. This reduces the avc bias but kills avc action. Let me address the points you mentioned first.

All resistors were more than 10% off value except 5, 23, and 50. Resistor 15 was originally 70 k not 32 k (as mentioned before this radio uses a design similar the the 38-89,with osc grid leak of 120k, but with component arrangement of the 37-89). Substitution with values from 12 to 120 k don't kill the problem. The 6K7 IF is excluded from the AVC buss and its suppressor is grounded through a single turn tertiary IF transformer winding. The IF secondary is grounded. I replaced the tert. winding as the rubber insulation had disintegrated. The radio oscillates nicely if this is poled wrong (50/50 chance). I will check through this work for mistakes. All tube shields were in place. As far as I know there is no under or above chassis shielding.

As an "amateur engineer" I made 6 design changes: (1) increased decoupling cap. 38 to 8 uf, (2) shunted the 35 ohm portion of res. 50 with a 68 ohm unit to bring the bias of the 6Q7 up to the right value ( I see now that the excessive bias might be related to the AVC problem), (3) relocated res. 34 and cap. 35 to the end of the term. board to ground a shielded wire to the 6Q7 grid, (4) jettisoned the death cap. 54 in favor of a 3 wire cord, (5) The ONLY ground for the tuning components seems to be through a wire to the antenna grounding terminal so I connected a substantial bonding jumper between these two chassis, (6) installed a 100 uf cap. across res. 50 to improve the audio. I ran a wire from res. 29 to res. 16 routing around the front of the chassis. I think the original went across the middle. All other replaced (bad insulation) wiring was installed as originally found. Given the old 5k expedient mentioned above I don't think any of this bears on the problem.

I substituted a different cap. for 14 (which tested OK on a bridge) and have tried simple neutralization tricks on the 6A8. Also tried screen and avc bypass caps closer to the sockets. The next steps will be to re-pole the IF winding, increase res. 18 value to drop the voltage to about 100, insert parasitic suppressors in the leads for osc. tube terminals, fabricate under chassis shielding, then . . . .

Thanks again for your help, and any other thoughts you have. I believed these old radios were simple machines and that simple machines have simple problems. Now I'm not so sure.

Well, there may be a miswire or bad connection somewhere within the radio causing this trouble. The trouble probably wasn't there when the radio was first produced, though sometimes sets can have design flaws.

You might consider grounding the control grid of each RF stage and see if that has any affect on the AVC voltage. Also measure the voltage on resistor 50 from the point where the AVC connects to the chassis. If this voltage is rather negative, consider disconnecting the AVC biasing resistor from this point and connecting it directly to the chassis, or making appropriate repairs to fix the excessively negative bias voltage. There could also be some part within the set that is drawing too much current, or was connected to the wrong part of the bias resistor that would cause excessive current to flow through the resistor and cause an excessive voltage drop across the resistor.

Make sure that your electrolytics are fresh (regarding parasitic oscillations), and try paralleling with a .05 to .1 MFD cap to see if this improves anything (if parasitic oscillations are in fact present).

Try pulling RF tubes one by one to see if this affects AVC voltage.

The AVC voltage normally contains little current, and is difficult to read by an ordinary 100,000 ohm per volt meter. If the AVC circuit really pulls a hard negative 40 volts, I strongly suspect that something is wrong with either resistor 50 or something drawing too much current on it, or a mis-connection. Of course if the voltage on the R50 side of the AVC bias resistor isn't -40, then the trouble does lie on the AVC side of the AVC bias resistor. Incidentally, the AVC bias resistor establishes an initial negative bias for the RF tubes to operate by when no actual AVC voltage is present.

The 3-wire cord should have little affect on the set's internal function, though it does make the set safer from a shock point-of-view, and also offers a possibly better RF ground. Aesthetically, if the thick black cord wears on you, you can obtain attractive 3-wire cloth cords from www.sundialwire.com, and they, or maybe someone else, may have a vintage looking 3-prong plug.

Typically the problem with the 'death cap' is encountered when musicians use multiple amplified instruments and microphones with exposed metal, where there can be a shock potential between two instruments and/or microphones. Since the chassis of a radio is rarely contacted by the user, there is little risk for shock. Still, when one goes to make an antenna or ground connection with the set plugged in, there can be a little surprise, which will usually be entirely eliminated by a 3-wire cord that properly grounds the chassis.

T.

Also consider that the AVC resistor may be connected to the wrong tap on R50.

T.

...Though you say that the problem is only there about 50% of the time, so maybe not.

T.

Also, be sure that C6 is of correct value and is connected properly (I did not see this mentioned in your check list). It filters the AVC voltage and prevents parasitic oscillation.

T.