Hum and buzz

[guittube]

I have recently been given a half-built tube amp that was badly assembled and had really bad buzz and hum. When I got it it was in pieces, although the basic circuit boards were still intact.

It's two channel, with an fx loop and reverb. The caps are doghouse style (side by side in a line), every single one. However, the filter caps "split" into two strings after the power section caps. One string is the preamp channel filter caps and the other is the fx and reverb filter caps.

To try and reduce the hum i thought I would do the following (can someone check my theory, please)...

1) Connect power valve cathodes (2 x 6l6 fixed bias) and the CT at the power cap negative end and create a star ground nearby
2) Connect all the filter cap negatives on both strings using a bus bar and connect the bus bars to the star ground
3) Connect the ground points from each preamp to their respective filter cap negative point on the bus bar

Here is a jpg of the schematic (I can't post a picture of the layout as there's currently nothing to show except bits)... can someone check that I'm connecting the grounds of the right parts to the right places? I've been told that sometimes where in a circuit you ground MV pots etc can make a big difference wrt hum.

[loudthud]

You need to supply more details about the amp, a schematic would be a big help. Pulling one or more preamp tubes should kill the hum. Which ones? Which of the tone or volume controls affect the hum? Do any of them kill the hum completely? Does the amp have a standby switch? When you turn standby switch from "play mode" to "standby mode", does the hum die immediately, or does it fade out over over a period of a couple of seconds? Where are the input jacks grounded with respect to the buss bars?

[guittube]

The schematic is in my first post. It's hard to see it because of the colour scheme of these pages but it's an attachment at the end of my post.

In answer to your other hints and tips, all of this is secondhand knowledge from the previous owner as the amp is currently in bits. The circuit boards are intact but the actual wiring isn't.

Pulling one or more preamp tubes should kill the hum. Which ones?
I can't test this as the amp is in bits. Apparently, the main power section of the amp was buzz and hum free.

Which of the tone or volume controls affect the hum?
The preamp gain controls (the higher gain channel was more noticeable than the lower gain one)

Do any of them kill the hum completely?
There was no buzz/hum with the gain pots at zero or at maximum. Between 9 o'clock and 3 o'clock was where the buzz was.

Does the amp have a standby switch? When you turn standby switch from "play mode" to "standby mode", does the hum die immediately, or does it fade out over over a period of a couple of seconds?
I don't know and as it's in bits I can't test this.

Where are the input jacks grounded with respect to the buss bars?
The original owner had wired each and every point that needed to be grounded to a very large star ground (about 30 wires in all!) I'm guessing this was the main problem, but I want to wire it up as good as possible and not have to redo any of it. I thought a bus bar for each of the channels would make for a simple neat job.

However, I'm loathed to redesign the circuit boards and so the filter caps are kinda stuck in a row nowhere near the main circuit board.

I hope this helps you to help me!

Thanks.

[loudthud]

Ok, that link looked like a smudge on my screen. What is that component H11F1? Relay contacts?

The ground side of the output transformer, does/did that go to the star?

Since there is no detail for the power supply, I need to ask: does the heater supply (6.3VAC) have a center tap or the virtual center tap consisting of a pair of 100 ohm resistors? The absence of a center tap will make the amp hum no matter how well the ground is done. It can also cause strange things like you described the hum only at the center of pot rotation.

Star grounding is not my favorite, I prefer a local star for each stage and a buss connecting the stars basically left to right on the schematic. The decoupling/bypass cap for each stage should be connected to the local star point. The ground for the secondary of the OT should be connected to the phase inverter ground point to avoid injecting noise into the feedback loop.

[guittube]

Thanks for the swift reply. OK, I'll try to answer your questions...

What is that component H11F1?
Optofets

The ground side of the output transformer, does/did that go to the star?
It did. I'll connect it to the PI ground when i rebuild

does the heater supply (6.3VAC) have a center tap
Yes, and it's accurate (no need for a humdinger or 100 ohm resistors)

Star grounding is not my favorite, I prefer a local star for each stage and a buss connecting the stars basically left to right on the schematic.
Me too, but the bypass caps for each stage are all in a line, not near their respective stages. Instead, what I thought I'd do is (to my knowledge) the next best thing... make a bus from all the negatives of the bypass caps and then connect each stages (or group of stages) star ground to their respective bypass cap (as in the schematic I posted).

[loudthud]

The comment about gain controls humming from 9 o'clock to 3 o'clock makes me wonder about the general layout. Was the layout done by someone with experence or by someone trying to cram too much in too little space and no experience? Has this circuit been built before with these optofets?

Are/were there long unshielded lines running next to heater wires? When you say "circuit boards" are you speaking of eyelet or turret board construction or PCBs? How are the heater wires routed to the preamp tubes? Are the optofets close to where the signals need to be switched? How does the power supply for the optofets look? Is it something that was improvised using the heater winding?

[guittube]

Was the layout done by someone with experence
A small amount, but not a lot

Has this circuit been built before with these optofets?
Yes, it's a copy of some Randall MTS preamp modules

Are/were there long unshielded lines running next to heater wires?
No. All long wires were shielded (ground connection at one end only) and all signal wires were at 90 degrees to the heater wires and raised aboout 3/4" above them

When you say "circuit boards" are you speaking of eyelet or turret board construction or PCBs?
Tagboard

How are the heater wires routed to the preamp tubes?
Twisted pair, 6.3V AC, running along the chassis floor. PT -> Power tubes -> Preamp tubes

Are the optofets close to where the signals need to be switched?
Right next to the signals, on the tagboard

How does the power supply for the optofets look? Is it something that was improvised using the heater winding?
It's built on a separate breadboard/veroboard circuit (yes, I know... not ideal). It's based on a known circuit, but I forget by who. It's fed from a separate 6.3V winding (not the heater winding). The switching circuit itself is as far from the tubes as possible with a 5v DC supply going from the switching circuit to the optos. The circuit was tried with the switching circuit turned off and the optos bypassed and it still buzzed the same as before. The switching circuit wasn't grounded to the chassis though. Perhaps it should have been.

The mystery continues...

[loudthud]

Depending on how the power transformer is constructed, there can be a huge voltage (line voltage or higher) riding on a low voltage winding due to inter-winding capacitance. This is one reason why a heater winding needs to be grounded, usually with a center tap. You can measure this AC voltage, simply apply power to the transformer primary and with a DVM measure the AC voltage between a transformer bolt or the chassis and the un-grounded center tap of the heater winding.

If there was no ground reference to the switching circuit power supply, this "common mode" voltage could be riding on the control side of the optofets, in close proximity to all the sensitive circuits where switching takes place. It's not clear if this possibility was eliminated when experiments were done bypassing the switching elements.

Your grounding scheme, although not perfect, shouldn't be causing as much hum as you describe.

[guittube]

If there was no ground reference to the switching circuit power supply, this "common mode" voltage could be riding on the control side of the optofets
When the previous owner tested to see if the optofet power supply was causing the hum, he completely disconnected the AC supply to the switching circuit and hardwired a bypass loop over the optofets. It still buzzed, so although the switching circuit has no ground reference, that doesn't seem to be the culprit.

Your grounding scheme, although not perfect, shouldn't be causing as much hum as you describe.
Well, I hope it's better than the original ground wire scheme which consisted of "everything goes straight to the star ground". There were a lot of very long ground wires all bundled together. I'm also going to include a hum blocking resistor / capacitor network between an isolated main star and the chassis / heater CT connection. Fingers crossed that should sort out any remaining hum.

Thank you for your suggestions

[loudthud]

When the previous owner tested to see if the optofet power supply was causing the hum, he completely disconnected the AC supply to the switching circuit and hardwired a bypass loop over the optofets. It still buzzed, so although the switching circuit has no ground reference, that doesn't seem to be the culprit.

That's exactly what you would have to do to eliminate the switching circuit as a possible cause of the hum. At this point I'm stumped. If you can get access to an oscilloscope to look at the hum, that would give us another clue. 60Hz hum usually comes from the heater supply or the ungrounded opto supply. It will fade out when you turn off the B+ with the standby switch. 120Hz hum or buzz comes from switching noise in the power supply, usually associated with poor grounding location of the HV CT. It will cut off immediately when the standby switch cuts off the first filter cap from the rectifiers.

[guittube]

I'll start re-assembling the amp using my revised grounding schematic. Once rebuilt (within a few weeks hopefully) I'll report back with how it went. If it's still got some hum / buzz I'll get my hands on an oscilloscope and, if necessary, a signal generator.

Once again, thanks for the help.

[guittube]

Ok... progress...

I've rebuilt part of the amp (only part) and I've located the specific point where the buzz enters the amp. It's on the input to the tube where the gain pot is.

This is what I've found out...

With all circuitry BEFORE the tube removed and a signal source applied to pin 7 it's buzz free and sounds GREAT.
With a resistor attached between pin 7 and the signal source it buzzes. The original resistor wasn't 1k (as on the diagram). It was, infact, 470k. I've tried both values. I haven't tried it with the resistor attached directly to the tube socket. Would this be the wisest next move?

I've highlighted the affected area on a diagram (see attached).

Does this help isolate the problem?

Thanks again!

***UPDATE*** - placed the grid resistor directly on pin 7. This improved things slightly. Changing the resistor from 470k to 220k improved things a bit more. However, with the gain pot in play (which is grounded right where the cathode cap is) the buzz / hum is back with vengance. Let me know if it's time to dig out the oscilloscope. Not that I can remember how to use one mind you...

[loudthud]

The wiring between the pot, resistor and tube will be very vulnerable to hum pickup. And, the higher the value of the resistor, the more vulnerable it is. The wiring should be shielded and kept as far as possible from any heater wiring. The best place for the resistor is right at the tube socket (IMHO). Use a physically small resistor like a 1/4W. You can try leaving the resistor out altogether as an experiment. The purpose of the resistor is to improve the tone at the low end of the pot's rotation which can be too bright and sterile. Make sure the bushing and back of the pot are well grounded to a quiet ground.

Options to try: If you locate the resistor away from the tube socket, that may be an advantage because there will be less un-shielded signal exposed at the tube socket. If you can find some braded shield, you could actually shield the resistor.

[guittube]

I've got some nice shielded cable... I'll try to embedd a heatshrunk resistor under the braid.

With regards to the grounding of the pot, all the pots are currently just attached to the chassis by a shakeproof washer on the inside and a normal washer and nut on the outside. The back of all the pots are just "as is". The ground lug of the pot is connected to the ground point of the following tubes cathode resistor/capacitor network. Is this ok?

[guittube]

Ok... I've shielded the resistor and connected as follows...

Resistor directly to the tube socket pin 7
Shielded cable directly to the gain pot (braid to the ground lug and signal wire directly to the middle lug)
Gain pot ground lug connected to the following tube cathode ground point (the ground end of the resistor / capacitor network)
Gain pot "hot" lug connected with shielded cable to the preceeding resistor/capacitor network (shield also connected to the following tube cathode ground point)

There's a lot of hum. 60Hz, 120Hz, a mix of both... I can't easily tell. I'll bring the CRO home from work tomorrow night just in case (be warned, it's a basic one!)