Hammond Chassis to TC15 Hummmmmm and questions

[lord preset]

This a build I set aside a while back and have come back to. It's a Hammond AO-39 chassis gutted, modded and built as the 12ax7 channel of a Trinity TC15, which is in turn pretty close to a DC-30. When I first built it had very little headroom and not much volume. More recently I've been using my new-to-me oscilloscope to try to see why it was not performing as expected. I found that V1 was not clipping until very late on the dial, but that the V2 cathode follower (specifically V2a) was clipping almost immediately. I tried removing the 25uf cathode bypass cap on V2a and the signal cleaned up considerably - and sounded good even, but volume was still way too low. Looking at the PI, I wondered if it did not have enough gain and compared it to a AC30 PI to see if there were values I might change. What I noticed is that there is a 100K resistor (R11 on the schematic) going to ground just before the PI that is not there in an AC30. I couldn't see why it should be there, and on the theory that anything I don't understand must not be important I removed the 100K resistor and voila, volume is up to what I know AO-39 iron can do.

I'm glad it sounds good now but I don't quite understand why I had to mess with a known good design to make it so.

So now that the amp sounds good I have turned my attention to it's other big problem - way too much hum.

I have tried:
1. Elevating the heater CT to ~10vdc off the power tube cathodes, and then to ~50vdc with a voltage divider off the screen supply.
2. Disconnecting the CT ground and installing a 2x100r artificial CT.
3. Chopsticking wires all over the amp.
4. Separating the ground for the preamp filter caps from the plate and screen supply ground.

I would be willing to redo the heater wiring and/or try a humdinger if I thought it would help but I'm not convinced it's heater related as nothing I've tried has reduced the hum in a noticeable way. Suggestions appreciated.

In looking for grounding issues I realized that when I omitted the EF86 channel I left one capacitor - C8 - that would have been the PI entrance for that channel unconnected. The PI works without it but I thought this might be a source of the hum as diagrams of phase inverters seem to show this cap connected to ground when it doesn't serve as an entrance for a 2nd channel or for negative feedback. But when it's grounded, the amp produces a high frequency squeal my dog objects to. So my question here is should this cap be connected in some other way or just omitted?

Note: The pics were taken before I removed the 25uf cap and 100k resistor.

[Stevem]

First off tape a nice sheet of aluminum foil to a big enought scrap of card board to cover the bottom of the amp and ground it with a clip lead or something and sit it foil side down on the bottom of the amp, does this greatly tame the hum issue?
Since you have a scope now is it 60hz heater hum you are chasing down, or 120 hz power supply ripple?
120hz sounds close to a B played on your A string just for reference.

Do you read any dc voltage on the down stream side of any of your coupling caps off of any of the plates,if so you likely have a leaky cap.

If you yank out V1 thru V3 one at a time does the hum issue stop, and if so with what tube, if not do you have the output tube heaters wired up pin 4 to 4 and pin 5 to 5?
Have you tryed different known good preamp tubes?

If yanking the tubes changes nothing than take a voltmeter set for AC and measure the ripple voltage at the input to the plate load resistor and report back as to when that MV is.

[lord preset]

So it's 120hz hum as best as I can tell from the scope. Definitely not 60hz.

Pulling v1 and/or v2 does not affect the hum but pulling the PI does. I covered the amp bottom with a grounded aluminum plate - no effect.

Now I guess I hunt for where the 120hz signal 1st appears

[lord preset]

I discovered that touching the DMM probe to pin 2 of the PI made the hum go away. What might that mean?

[lord preset]

I never did figure out the hum on this one and set it aside for a while. Getting back to it I'm still stumped. The only clues I have are that the hum goes away when touching the DMM probe to pin 2 or 1 of the PI. Volume is reduced when doing this but not dramatically so. What is the DMM doing? Is its internal resistance low enough to provide a significant ground?

I'm also still not sure why I had to remove the 100K resistor (R11) to get reasonable volume and wonder if that has anything to do with the hum.

Halppp!!

[Phil_S]

I see you didn't get any follow up before, so I'll take a whack at it. Remember, I'm an amateur, so take what I say with a grain of salt.

The 100K you removed looks to me like a simple voltage divider, though it also connects to the PI tail resistor. My first reaction is that it's there to attenuate the signal on the top channel and maybe without the EF86 channel it is overkill. I don't see how it affects the PI much. I'm thinking the ground end is just ground scheme design. If removing it works, let it be.

Without the EF86 channel, C8 needs to be grounded. I am thinking probing the grid with the meter is providing a missing ground reference. Looking at some of the Fender designs, there is a low value (47Ω or 100Ω) resistor between the cap and ground, but that's with NFB. The Spitfire as an example without NFB, it goes straight to ground.

If it squeals, that suggests your amp might benefit from a reversal of phase. I recognize this problem typically occurs with a NFB loop and you have none, so maybe this is a fool's errand. Nevertheless, you have little to lose by reversing phase and it is something you can do quickly. The simplest thing to do is reverse the OT wires at the EL84's or you can make any equivalent swap between the PI and the EL84's.

Since there's no NFB, before looking at phase reversal (predicting no change there), I think I'd look at how you routed the OT secondary to the output jack. I see the OT is at the opposite end of the chassis from the PT. I can't see all of it, but am thinking you've got the OT secondary (black and green?) running under the board. This is a ripe opportunity to allow oscillation to develop. What I can see is the green/black twist laying on the yellow cathode wire (p3) and crossing over top the red grid wires. Maybe it's nothing, but that's worth moving around. If that fails, I think I'd reroute the OT wires around the outside of the chassis. Since you've put the heater wires at the fold, you may want to use a few clips to hold the OT wires away from the heater wires, following a similar route and coming to the output jack from above instead of below.

[lord preset]

I mis-poke a bit when I said that the amp squeals when C8 is connected to ground. It does, but only at high volumes. At lower volumes it introduces ratty, buzzy distortion that goes into squealing feedback as you crank it up. I tried a 470 ohm resistor between C8 and ground, still buzzy and humming. A 1 meg resistor yielded little buzzing but no real improvement in hum. For whatever reason, this amp does not seem to want C8 to be in the circuit.

I have tried replicating what the DMM may have been doing re. high resistance a ground reference by adding a 20 Meg resistor from PI pin 2 to ground. I get about the same results - significant hum reduction but also gain/volume reduction that's enough to suck some of the mojo out of the amp. I will try moving OT wires as suggested but I'm skeptical that this will help - it seems to me there is still something amiss in the circuit itself.

[Phil_S]

Really, I'm must guessing now. I can't go over the whole build -- too much of a challenge for me to do from pictures. I continue to think the main problem is oscillation, but I think the hum constitutes a 2nd problem. If that is true (who knows?) then I'd try to lick the problems one at a time. I'd go for the oscillation first. I'll go back to the OT wires. If moving them a bit as they are seen in the pic doesn't help, I'd reroute them.

Meanwhile, I'd connect C8 to the ground side of 47K tail resistor and leave that alone for the time being.

Then, it's back to basics. 1st pull V3 (PI) and see if it quiets down. Replace V3 and pull V2. Replace V2 and pull V1. This should help you isolate the location of the problem. If it is between two of the tubes and you can find out which one. I'm assuming post PI is OK, but if it's not, that has to get looked at, too.

Maybe you get lucky and the noise goes away when you pull V2. In that case, jumper the 1st grid of V2 directly to the PI input (C7). If you get good results, the problem is in the CF. That's the sort of thing that needs doing. It is time consuming work.

The hum is possibly filtering or maybe a ground loop. I'm not clear about the ground scheme. I can't decipher it from the pictures. I'm going to guess you have under-board wires. I did that one time and learned my lesson. Put as much as you possibly can topside, just for situations like this.

I'm hoping someone else will come along and say a few words. I'm not seeing the real problem. Good luck.

[frankdrebin]

i would put back the 100k (or 220k) in front of PI,that cuts gain and chance to oscillate.
1,5k on power tubes grids is always encouraging it,i'd go for at least 10k.
Also i never liked that 1,2k for PI biasing.
Many things can lead to oscillation and hum,lead dress,grounding.
My rule of thumb is to ground all the preamp by itself and power tubes near the PT.

[lord preset]

Back this this crazy-uncle in the attic build. I took Phil's suggestion of rerouting the OT wires to it's logical conclusion and ran them outside the chassis altogether. That took care of the oscillation issue which made it possible to ground C8 and put the 100K resistor back that I had removed. Spent some more time swapping 12ax7s in and out of my extensive inventory of Questionable Tubes (TM) to get ones that don't sizzle and pop. The net is that the amp is stable at all volumes and sounds pretty good. Hum is less than it was but still higher than I would like , but I think I am going to declare this done, slap it in a head cabinet and attempt to foist it off on an unsuspecting victim.

[tictac]

Is the ground lug of the input jack actually grounded? I can only see the shield connected....

Need a ground connection there....

TT

[lord preset]

Is the ground lug of the input jack actually grounded? I can only see the shield connected....

Need a ground connection there....

TT

Yes it is, by a black wire in shadows.