Transformer hum?

[txstrat]

Here're some pics with descriptions of the actual grounding right now.
Hope it's clear enough.

[txstrat]

Will also check if it's still 100Hz buzz.

Wow, Tom. How'd you know.
It's indeed two sorts of hum.
A deep frequency (50Hz) hum and a 100Hz buzz. The Hum (50Hz) increases with the volume pot. The buzz is constant.
When I pull the first tube, the buzz (100Hz) remains. If I pull both 6SL7 all hum and buzz are gone.
OK so the 50Hz seems to be heater hum from the first tube, while the 100Hz seems to be buzz from the PI tube (the buzz level doesn't change when I pull the first tube).
Regarding the hum I'm gonna focus on the first tube, regarding the buzz I'm gonna stick with the second tube. Hope this makes sense.
I've read somewhere that the 6SL7's first triode is more prone to heater hum. That's why I'm sending the input signal to pin 4 (second grid on the 6SL7).
Gonna try it the other way round to see if the hum changes.
Besides I'm gonna try insulated jacks for the speakers and different grounding spots for the PI and OT ground.

[R.G.]

Power line frequency (50Hz) always comes from the AC mains, as leakage, magnetic induction, or capacitive pickup. Can come from power wires, heater wires, anywhere there is AC power line frequency.

Double power line frequency comes from the full wave rectification of the power supply. That's the only plausible place for it to be. Could be filter cap ripple, rectifier turn-off buzz (squark of RF when solid state rectifiers turn off), fluorescent light fixtures in the house you're playing in, bad connection/wiring choices in the transformer ground return, routing power supply return currents across the chassis, etc.

Pulling a tube to see what stays and what goes is a useful test. So is turning off the power and seeing what vanishes before the power supply runs down. If it stops instantly, it's related to what the AC power is doing, either in the AC power coming in, heater wiring, or ripple. If it keeps on, it's pickup from the environment outside the amp, perhaps house ground voltages, fluorescent lights, etc.

One interesting one: a tube amp can have hum in the speakers with the output tubes pulled. This is induction from the power transformer to the output transformer. Try running THAT one down without already knowing the trick.

[txstrat]

OK here's what I did.
Pulled the power tubes (only). No hum no buzz, just a sweet silence - just a barely noticeable hum from the PT I wouldn't bother.
Pulled the PI tube only. Same thing, sweet silence.

Pulled the first tube. Hum gone, buzz still there.

When I throw the std by switch with the first tube pulled (either on or off stand by), there is a 50Hz hum coming up for a couple seconds which fades out and keeps away.

When all the tubes are in their sockets and I throw the stb by switch from play to std by mode the buzz (100Hz) disappears at once, while the hum (50Hz) fades out slowly.

Generally the hum increases with the volume pot.

The buzz is there from volume at zero until the volume pot is on 4 - then it doesn't disappear completely but gets veeery quiet - until the pot is on 7 from where it increases until full open.

Have taken the amp from my basement to the living room (just to the delight of my wife ) but the amps behavior stays the same.

The electrical installation in the house is quite new. I don't expect it to be the culprit. All my other amps are dead quiet.

Could try the change the filter caps even though they are quite new, maybe a year.

[R.G.]

OK here's what I did.
Pulled the power tubes (only). No hum no buzz, just a sweet silence - just a barely noticeable hum from the PT I wouldn't bother.
Pulled the PI tube only. Same thing, sweet silence.

OK. That eliminates the magnetic induction from PT to OT.

Pulled the first tube. Hum gone, buzz still there.

That says that most of the hum is a first/second stage issue. Good. When it's isolated to one tube, it's much easier to pin down and kill.

When I throw the std by switch with the first tube pulled (either on or off stand by), there is a 50Hz hum coming up for a couple seconds which fades out and keeps away.

I'll have to think about that for a while. Nothing leaps out from it.

When all the tubes are in their sockets and I throw the stb by switch from play to std by mode the buzz (100Hz) disappears at once, while the hum (50Hz) fades out slowly.

That says to me that the buzz is generated from the power supply/rectifiers. With no AC power in, the rectifiers don't do anything. We'll look at that in depth below.

Generally the hum increases with the volume pot.

Another vote for hum happening mostly in the section before the volume pot.

The buzz is there from volume at zero until the volume pot is on 4 - then it doesn't disappear completely but gets veeery quiet - until the pot is on 7 from where it increases until full open.

Another one to think about for a while for me.

Have taken the amp from my basement to the living room (just to the delight of my wife ) but the amps behavior stays the same.

The electrical installation in the house is quite new. I don't expect it to be the culprit. All my other amps are dead quiet.

That's a good start at eliminating the house wiring.

Could try the change the filter caps even though they are quite new, maybe a year.

I wouldn't start there.

Let's tackle buzz first.
1. The wiring of the return from the rectifiers to PT is a major and unrecognized source of buzz. It simply has to be done correctly, or the buzz will never disappear. Here's how to think about this.

The PT secondary making high voltage, the rectifiers, and the first filter cap form a current loop. The current comes (conceptually!) from the two halves of the PT HV winding in the case of a CT power circuit, or from each side of the PT in a full wave bridge circuit. The current goes through the rectifiers, being made into pulsating DC. The DC part goes out into the amp and the filter caps shunts as much of the residual AC into the rectifier/PT return as it can. From the cap negative, the rectified pulses return to the PT HV secondary.

The rectifier currents are short, sharp pulses of DC, maybe 10X the average DC current going out to the circuit. They're big enough to cause a noticeable fraction of a volt of change just in the wiring resistances from the filter cap negative back to the PT. It is CRITICAL for low noise that the negative terminal of the first filter cap be the only place where "ground" currents from the amp return to. If you happen to ground the transformer CT to the chassis or to a wiring point, return the circuit signal grounds to that, and run a wire to the filter cap negative, it *will* buzz.

The right way to do this is to connect the return wires from the PT CT or from the negative side of the rectifiers if you have a full wave bridge directly to the first filter cap negative. The circuit ground returns also go to that capacitor negative. Even an inch or two of wire "shared" by the circuit signal grounds and the rectifier return wires will cause buzz.

The first filter cap negative lead is a somewhat magical point. This is the single place where the power tube cathode circuits need to come through a dedicated wire, shared by nothing else. Likewise, this is a great place to run the single wire from signal ground to the metal chassis, although not the only one which works. Ideally, almost all of the other ground wires should come here individually, giving a pure star ground. that's impractical, but you can easily enough be sure the rectifier currents start and end here.

The next thing to suspect for 100Hz buzz is solid state rectifier snap-off oscillations. Tube rectifiers are much more forgiving about this. When the voltage reverses on a solid state diode, the junctions are full of conductors, and they stay conducting until the reversed voltage sweeps the charge carriers out of the junction. When that happens, the junction snaps off very sharply, causing a square edge of current that makes any wiring inductance ring like a bell. A pulse of RF often gets radiated, and that happens at twice the power line frequency. The cure is to put a ceramic disk capacitor rated at high enough voltage to take the rectifier reverse voltage across the rectifier. Generally 0.001uF / 1kV will work. In hard cases, you may need to add a resistance of 10R to 100R in series to the cap to damp the ringing as well.

Let me know where I've made it more confusing rather than less.

[txstrat]

First off, I'm learning more with troubleshooting this amp than I've ever thought I would. Thank you all.

The right way to do this is to connect the... ...full wave bridge directly to the first filter cap negative.

That's what I've done. Bridge rectifier and first filter cap are side by side on the terminal board connected by a solid wire. See pic 2 in my post #76.

The circuit ground returns also go to that capacitor negative.

The second filter (for the screen grids) is also connected to this "line" of a solid wire from the rectifier ground. Next are the power tubes 1ohms resistors for which I used a stranded wire. Another wire comes from the bias supply ground directly to the 1ohms resistors.

I've added a layout scheme of the actual grounding. The solid wires are "pink", the stranded wires are "green/yellow". I left out most other wires not to make the layout scheme indecipherable since the grounding is the matter here.

The first filter cap negative lead is a somewhat magical point. This is the single place where the power tube cathode circuits need to come through a dedicated wire, shared by nothing else. Likewise, this is a great place to run the single wire from signal ground to the metal chassis, although not the only one which works. Ideally, almost all of the other ground wires should come here individually, giving a pure star ground. that's impractical, but you can easily enough be sure the rectifier currents start and end here.

Could I manage that by using a wire from each filter cap (negative) to this point while the signal grounds all go to the negative side of their respective filter cap? See attached pic 2.

Right now I've filter caps 1 and 2 (ground) tied together as well as 3 and 4 (3 and 4 go to the input jacks via solid bus wire).

The next thing to suspect for 100Hz buzz is solid state rectifier snap-off oscillations. Tube rectifiers are much more forgiving about this. When the voltage reverses on a solid state diode, the junctions are full of conductors, and they stay conducting until the reversed voltage sweeps the charge carriers out of the junction. When that happens, the junction snaps off very sharply, causing a square edge of current that makes any wiring inductance ring like a bell. A pulse of RF often gets radiated, and that happens at twice the power line frequency. The cure is to put a ceramic disk capacitor rated at high enough voltage to take the rectifier reverse voltage across the rectifier. Generally 0.001uF / 1kV will work. In hard cases, you may need to add a resistance of 10R to 100R in series to the cap to damp the ringing as well.

I'll keep that in mind.

Let me know where I've made it more confusing rather than less.

I had to read all this several times before I got it right. I think I've learned even more than I mentioned above.

Thanks again

[txstrat]

I can't believe it. I changed all the ground connections to the "star" or better "short line" around the first filter cap, except of the input and speaker jacks.
Still the same hum (100Hz as I checked with the guitars 3rd fret on the low E string ).
I disconnected the jacks from the chassis, used a wire from input jacks to the preamp filter cap and let the jacks flaoting above the open chassis.
95% of the hum was gone. When I moved the input jacks to touch the chassis the hum came back. Speaker jack touching the chassis made no difference.
I'm gonna interchange the input jacks with insulated ones and tidy up things now. Will report later.

[txstrat]

Update:
Have shortened all wires where possible, switched to insulated input jacks and changed the wire from the volume pot to the second grid for a shielded wire. There was still a bit 100Hz buzz left. Tried several other ways to lead the "ground line" to the first filter to no avail.
Then it somehow hit me: moved the volume pots ground to the PI filter cap and - all buzz gone. Regarding the 100Hz buzz I'm satisfied now. There is still a barely audible buzz left, but you got to put your ear close to the speaker to hear it. I can VERY easy live with that.

Now there's the 50Hz heater hum left, which increases with the volume pot and gets real loud when the vol is full open. I can't hear it when I play, but you know how it is, when you know it's there it's annoying.
I know the 6SL7 are prone to heater hum and that's why it's best to use the second triode (pins 4,5 and 6) as first stage - I've read somewhere, that the second triode is further away from the heaters inside the tube.
I "elevated" the heaters (virtual CT - two 100 ohms resistors) to the bias voltage (junction of the 270K resistors) and the hum is way less than without doing so.
What else can I do?

[Tom Phillips]

...I "elevated" the heaters (virtual CT - two 100 ohms resistors) to the bias voltage (junction of the 270K resistors) and the hum is way less than without doing so. What else can I do?

You can try substituting a "humdinger" pot for the two 100 Ohm resistors. Then in addition to the elevated heaters you will have the ability to adjust the pot for the lowest hum. The lowest hum point is not always at the center (perfectly balanced) setting of the pot.

[txstrat]

In one of my earliest builds I've used a small 100 ohms trim pot for this. Would that be OK?
Oh and besides, does the lenght of the wire going from 100 ohms resistors to the bias supply matter?

[Tom Phillips]

In one of my earliest builds I've used a small 100 ohms trim pot for this. Would that be OK?

Yes. I would recommend a power rating is 1/2W or more.

Oh and besides, does the lenght of the wire going from 100 ohms resistors to the bias supply matter?

I do not believe that length is critical.

[txstrat]

Tom,
I'll give it a try, gonna solder in a trim pot first thing tomorrow.
Thanks
Matt

[Tom Phillips]

...I "elevated" the heaters (virtual CT - two 100 ohms resistors) to the bias voltage (junction of the 270K resistors) and the hum is way less than without doing so.

I just realized that you have tapped right off the negative bias supply. I think you should try deriving your elevation voltage via a resistive divider with it's own decoupling cap from the B+ line. That may give a better result and the bias supply will remain pure & isolated from the heater supply.
Tom

[txstrat]

I think you should try deriving your elevation voltage via a resistive divider with it's own decoupling cap from the B+ line. That may give a better result and the bias supply will remain pure & isolated from the heater supply.
Tom

That was one of the things I tried first. At that time it didn't make any difference. Maybe BECAUSE of my faulty grounding scheme.
Elevating the heaters to a voltage divider from the B+ might have done it right BUT I couldn't hear the difference due to the heavy buzz. Gonna try it again.
Matt

[txstrat]

A humdinger pot didn't make any difference in hum (I used a 250 ohms trim pot which I had handy.) Tried both, 1st at the bias resistor junction and nd at the voltage divider from the B+. No difference.
Unsoldered ALL ground connections of the preamp and PI and used clip leads to change the position of the grounding. I believe I tried 25 different position combinations. Some were worse, some were the same as before. A slight improvement was the adding of a shielded wire from first anode to volume pot.
Maybe I have to live with the remaining hum/buzz. My son (7 years old) says it's not that loud. Maybe I'm too sensitive.