Beltone AP-50: Last of the 60Hz Hum

[garytoosweet]

Hey all,

I'm resurrecting a Beltone AP-50.

New electrolytics, all new coupling caps (except 1), new rectifier tube, 2x new NOS 6AU6's in the preamp, all new plate resistors.

It's now sounding great but there is a bit too much 60Hz hum remaining.

Please excuse the crudely drawn schematic below. I was intending to re-draw on the computer, and will delete and re-upload a new one once that is done.

Grounding

The grounding scheme was terrible (massive 60Hz hum) and was redone. The PT HT CT is connected to the negative on the 16+16 can cap (replacing the old 20+20), and then to a grounded tag board next to the cap. The OT secondary ground is also connected to this tagboard. The power tube grounds (cathode, grid leak) are attached to a buss bar with a wire running to the grounded terminal strip. The PI grounds are attached to a separate terminal strip with a wire running to the power tube buss bar. The two 6AU6 preamp tubes are grounded to a their own terminal strip (a long with the 10uf preamp supply cap) with a wire running to the input jack ground (new switchcraft jacks, lock washers, chassis stripped of paint and cleaned under the lock washers. (I tried modifying this slightly but nothing I could think of was more quiet than this configuration).


The Hum:

10mV 60HZ hum at idle (ok not too bad).
15mV AC hum at bedroom (knobs at 9’oclock) (kind of ok).
46-78mV at full (depending on tone control setting) (anything above bedroom level is unusable).

* All these measurements here and below are with input jack shorted.

Clues on the source of the hum:

AC ground Pin 1 (Grid) of V1 makes the hum shoot way up (to ~150mV)
AC grounding Pin 5 (Plate) of V1 drops the hum down to 20mV.
Pulling V1 drops the hum down to 15mV, not changing with volume control.

AC grounding V2 Pin1 Grid brings it down tfrom 40-10mV (at MAX volume zero tone). Stays at 10mV regardless of volume knob position.
AC grounding other side of that .01 cap at volume pot wiper brings it down to 25mV. (This is the only original coupling cap in the amp, left it original as it seems to be doing it's job and should not have much voltage on it).
AC grounding V2 Pin5 (Plate) does not change hum at all.
Pulling V2 does some as pulling V1.

AC grounding Pin 1 (Grid) of 6SN7 PI lowers the hum to 5-25mV depending on volume control).
AC grounding Pin 2 (Plate) of PI does nothing.

Also: There is 50VDC on Pin 1 Grid of V2 (no DC on the tone stack side of the original .01uf coupling capacitor)

Preamp tube voltages:

V1 (6AU6)
Pin 1 (G1): -1.0VDC
Pin 5 (P): 104VDC
Pin 6: (G2): 36.25VDC

V2 (6AU6)
Pin 1 (G1): 49.67VDC
Pin 2 (K): 79VDC
Pin 5 (P): 155VDC
Pin 6: (G2): 155VDC

V3 (6SN7)
Pin 1: 0VDC
Pin 2: 96.7VDC
Pin 3: 3.5VDC
Pin 4: 0VDC
Pin 5: 97.2VDC
Pin 6: 3.5VDC

Things I have tried:

- Humdinger (no change, currently heaters are referenced to 10VDC at power tube cathode through a pair of 100 ohm resistors - this was quieter than referenced to ground or using humdinger)
- Running heaters offf of DC bench supply (no change)
- Swapped PI tube (no change - tried 2 other tubes)
- Swapped both 6AU6's (no change)
- New EL84's (no change - the original tubes in there are matched)
- Disconnected pilot light as it is above OT secondaries (no change)
- Shielding with metal plate on bench (no change)
- Tried amp back in cabinet (bit of shielding) in different room, different breaker (no change)

Questions:

- Why is V2 connected to ground through a 100k resistor? I have seen this set up on Hammond AO44 which also uses 2 6AU6's.
- Why is there 50V on the grid of V2? I'm assuming this is right and that's why there is a second coupling cap before the grid but would love to know what's going on to learn more about it.
- What else can I try to get rid of the hum?

Thanks in advance to those who got through this novel.

[Helmholtz]

Is the amp safety grounded?

Does hum lower when you shield the chassis by covering it with a grounded metal plate or foil?

Is the 5Y3 good? When not symmetrical it will introduce 60Hz hum on B+.

What happens if you pull the PI?

What happens if you short the two PI grids?

The V2 wiring is unusual indeed. The 100k resistor to ground lifts cathode and grid voltages

[garytoosweet]

Is the amp safety grounded? Yes

Does hum lower when you shield the chassis by covering it with a grounded metal plate or foil? No

Is the 5Y3 good? When not symmetrical it will introduce 60Hz hum on B+. Seems ok. New JJ in there now. Tried swapping with an NOS GE, no change.

What happens if you pull the PI? No hum with the 6SN7 removed.

What happens if you short the two PI grids? No hum

The V2 wiring is unusual indeed. The 100k resistor to ground lifts cathode and grid voltages Any idea what they were going for there? I'm attaching a Hammond AO-44 schematic where I've seen the same idea.

Some additional thoughts:

1. I tried swapping V1 with a Westinghouse pull I had. The other tubes I were trying were all NOS JAN from the same stash. With the Westinghouse the 60Hz hum tops out at around 28mV Instead of 78mV like the JAN tubes. The Westinghouse is quite a bit more microphonic than the JAN's (believe it was pulled from a previous amp for being too microphonic), and the hum still isn't 100% solved, but seems like a big clue. The amp is more or less usable with the Westinghouse, I may be expecting too much from an old Teisco.

V1 is grid leak biased. Attached is an excerpt from Radiation Designers Handbook stating that grid leak biased tube stages can cause hum. Found in this thread over at TDPRI. I'm not sure why they are susceptible to hum, but I wonder if converting V1 to cathode bias may help. I'm assuming the grid leak bias was an economical decision more than anything else.

2. PI and Preamp plate are fed from the same node. Seems unusual, but probably not related.

[Helmholtz]

Ok, the hum is not generated in the power amp.

The AO-44 schematic shows a typical split-load or Concertina PI. Here the lower 100k resistor makes sense, as it's used for the non-inverting output.
With your circuit it doesn't seem to make sense, especially as it drops the gain of the stage to just below 1.

Sure, the wiring is original?

[garytoosweet]

Ok, the hum is not generated in the power amp.

Sure, the wiring is original?

Yes, I had this thought as well so I found photos of a couple other AP-50's online and they are all like this.

EDIT:

Based on my result in swapping Westinghouse tube for V1 in last post I think some tube rolling is in order. I only have NOS JAN 6AU6's so will find some more of various brands.

I have also tried:
- Converting V1 to cathode bias (quick and dirty, 1M resistor in parallel with the 5M grid leak, 22uf and 2.2k cathode to ground). No change in hum (returned to stock).
- Shorting the 100k resistor to ground on V2: Out of control gain 750mV of noise coming through with volume max.

[g1]

This is probably not related but I don't think your PI (6SN7) circuit is drawn right?
Pin 4 should be getting signal from somewhere, maybe the other end of that .01 connected to pin 4 goes somewhere other than ground?

[Helmholtz]

This is probably not related but I don't think your PI (6SN7) circuit is drawn right?
Pin 4 should be getting signal from somewhere, maybe the other end of that .01 connected to pin 4 goes somewhere other than ground?

Oh, you're right. I thought this was a LTPI, but it is not, as the cathodes are AC grounded.

[garytoosweet]

This is probably not related but I don't think your PI (6SN7) circuit is drawn right?
Pin 4 should be getting signal from somewhere, maybe the other end of that .01 connected to pin 4 goes somewhere other than ground?

The schematic is correct. However it's possible that the 6SN7 is not actually the PI, and is actually a differential amplifier. The previous 6AU6 is set up like a PI except only signal from the plate is taken (??).

Here is a site that details a bit of this set up that apparently is found in a Danelectro 1335: https://www.ampbooks.com/mobile/amp-.../balanced-amp/

[Helmholtz]

, and is actually a differential amplifier.

A LTPI is a differential amplifier. Your circuit is not. The 2 triodes are not coupled via the cathodes, thus working independently from each other.

[g1]

As drawn, that half of the 6SN7 would do nothing, and the power tube connected to point '3' would not get any signal.

[garytoosweet]

As drawn, that half of the 6SN7 would do nothing, and the power tube connected to point '3' would not get any signal.

I just checked quickly and you are correct there is no signal getting to grid of that power tube.

The amp is only putting out about 8W and only starts to become asymmetrical once it is really clipping.

Tomorrow I will have to figure out what is supposed to be going on here.

[garytoosweet]

The .01uf cap on Pin 4 of the 6SN7 should actually go to junction of 100k and 3k resistor off the cathode of V2, making it a typical Concertina PI like in the Hammond.

I must (very) embarrassingly admit this appears to be my doing when I re-did the grounds.

Now that the amp is operating at full power (about 13.5W) the 60Hz hum is much louder, 18mV at idle and up to 190mV with volume maxed (input shorted).

Tomorrow I will continue troubleshooting the hum.

Apologies for my error and thank you both for the help.

[g1]

For sure there is no NFB loop in this circuit?

[garytoosweet]

For sure there is no NFB loop in this circuit?

For sure no NFB.

[Helmholtz]

The .01uf cap on Pin 4 of the 6SN7 should actually go to junction of 100k and 3k resistor off the cathode of V2, making it a typical Concertina PI like in the Hammond.

So you have a Cathodyne/split-load/Concertina phase inverter (V2) followed by separate gain stages (V3a/b) for inverted and non-inverted signals.
As V3 is not wired as differential amp (no tail resistor), your amp differs from the Danelectro circuit.