Signal coming through tube amp with volume down

So, it's been a while since I have posted about one of my assorted problems with tube amps...but here is another one driving me nuts.

PROBLEM
On several amps I have built recently (3) I am having a problem where the signal is somehow bleeding through even with the master volume all the way down.  The signal is without any low end, and when I turn the volume up the full signal increases until it is louder than the signal that bleeds through.  When the volume is up everything sounds great.

CONSISTENT PROBLEM
The amps this is the biggest problem for are my high gain amps (4-5 gain stages - very similar to the SLO), and the signal does completely go away when the gain is turned all the way down.  If the gain is up but the master down, then the problem is there and loud enough to be very annoying.
However, even on amps I am building that have low gain (two 12ax7 gain stages with a High Low tonestack before a phase splitter) and no master volume, the problem persists even when the gain is all the way down (granted the problem is way less noticeable)

OBVIOUS CONCLUSIONS/IMPORTANT POINTS
1.  This is a consistent issue for me and thus is something I am doing wrong
2.  The more gain, the louder and bigger problem it becomes
3.  Probably not the pot causing the problem since it has happened on several amps now
4.  Problem persists whether there is a bypass cap on the gain pot or not
5.  Grounding issues are an obvious choice, but I am using a star grounding scheme (maybe poorly).  I ground the input to the first cathode resistor.  Each stage is grounded to its cathode resistor (ground side) (including the pots) which is then grounded directly to it's filter cap.  Filter caps are grounded to the chassis together on a lug bolted to the chassis.  Other lugs are grounded to this bolt as well coming from the first filter cap/OT/and speaker out, and another one has the center taps from the PT.
6.  If I ground certain grids then the problem stops, but with others the problem persists, especially in the high gain designs.  If I physically ground the signal before or after the master, the problem is unchanged.  In the high gain amps, if I ground right before the PI, or the grid before the final gain stage, the problem goes away in that the signal finally stops.  So the bleed seems to be happening right before the last gain stage and possibly at the PI as well?


QUESTIONS
1.  Could this just be a grounding problem?
2.  Crosstalk from a bad layout?  I base my layouts on what I see in classic amp layouts (IE - Triodestore or Ceriatone layouts)
3.  Some of my RC dropping resistors are as low as 1K5, some as high as 10K (usually 10K is for the first stage)...could this be an issue?
4.  Is it possible to be bad pots on every amp? 
5.  Any other ideas?


Thank you for your time dealing with this noob...=)

when playing around with my mixer design I did wire the fader the wrong way round. I was feeding through the wiper and having the top connection going to the next stage, bottom contact to ground. that produced a similar effect, maximum cut was low at -40dB but everything being fine when the fader was full up. not saying this is the your exact problem, but worth checking...

- michael

audiomixer said:

when playing around with my mixer design I did wire the fader the wrong way round. I was feeding through the wiper and having the top connection going to the next stage, bottom contact to ground. that produced a similar effect, maximum cut was low at -40dB but everything being fine when the fader was full up. not saying this is the your exact problem, but worth checking...

- michael

Click to expand...

the attached picture is how I am wiring gain/volume pots....I assume this is correct

I'd do a test - connect the output to the next stage directly to ground (disconnect from the wiper on the pot).
Then see if the amp is quiet. If so then the output stage is coupled to the input through some other means.

It could be that the signal is communicated through ground. It could also be communicated through the power supply side. For instance if the preamp tube signal is causing a voltage drop in the B+ of the output stage.
You have either a design or build problem somewhere.

Tubes are high input impedance devices. Care must be paid to prevent crosstalk into these high impedance nodes. The tinny, HF nature of your symptom is typical of crosstalk.

I know guitar amp designers who end up with such crosstalk becoming part of the sound character of cascaded high gain stages.

JR

You have the input and output pot lugs swapped!

The input should be in the middle lug, so that when you turn the pot all the way down the entirety of the input signal gets drained into ground.

I may be wrong 

figuringstuffout said:

the attached picture is how I am wiring gain/volume pots....I assume this is correct

Click to expand...

Is that a view from the front of the chassis or the rear? If the rear, then full-on will mean turning the knob counterclockwise, not clockwise as is usual.

Meanwhile, are the wires going into and out of the pot running close to one another? You might be getting capacitative coupling between them. Use a wooden chopstick to move the wires around while you listen to the amp (keep your other hand in your pocket while you do this). Does moving the wires make a difference?

Peace,
Paul

You have the input and output pot lugs swapped!

The input should be in the middle lug, so that when you turn the pot all the way down the entirety of the input signal gets drained into ground.

I may be wrong

Click to expand...

I suppose you are! 
You should not short the input signal to ground, but attenuate the input signal going to the following stage!

schematic please

I suspect the problem is introduced at the pot. The wiper and ground terminals of the pot should be taken in a screened cable by the shortest route to the tube. Ground the screen at the cathode resistor.

Cheers

Ian

tomas1808 said:

You have the input and output pot lugs swapped!

The input should be in the middle lug, so that when you turn the pot all the way down the entirety of the input signal gets drained into ground.

I may be wrong 

Click to expand...

A) I'm failing to see how switching the in and out lugs on the pot will matter in this scenario
B) I'm nearly certain this is indeed how you wire gain and master pots
C) You are looking at the back of the pot

ruffrecords said:

I suspect the problem is introduced at the pot. The wiper and ground terminals of the pot should be taken in a screened cable by the shortest route to the tube. Ground the screen at the cathode resistor.

Cheers

Ian

Click to expand...

The gain pot does indeed have a shielded cable in and out....the master does not since the signal by that point is well above line level.

CJ said:

schematic please

Click to expand...

The most recent amp I have built that had this problem was almost identical to a Soldano SLO.  The only changes were to make a few of the caps coming off the plates a little smaller to tighten up the bass response

pstamler said:

figuringstuffout said:

the attached picture is how I am wiring gain/volume pots....I assume this is correct

Click to expand...

Is that a view from the front of the chassis or the rear? If the rear, then full-on will mean turning the knob counterclockwi, not clockwise as is usual.

Meanwhile, are the wires going into and out of the pot running close to one another? You might be getting capacitative coupling between them. Use a wooden chopstick to move the wires around while you listen to the amp (keep your other hand in your pocket while you do this). Does moving the wires make a difference?

Peace,
Paul

Click to expand...

I have the wires carrying signal as far away from plate and cathode wires as possible (and obviously heaters are tightly wound and nowhere close either).  Moving any of the signal wires produces zero change.

dmp said:

I'd do a test - connect the output to the next stage directly to ground (disconnect from the wiper on the pot).
Then see if the amp is quiet. If so then the output stage is coupled to the input through some other means.

It could be that the signal is communicated through ground. It could also be communicated through the power supply side. For instance if the preamp tube signal is causing a voltage drop in the B+ of the output stage.
You have either a design or build problem somewhere.

Click to expand...

I have not actually disconnected the wiper when testing...will do so this evening when I get the chance.

I'm of the opinion currently that the signal is either being communicated through ground or power supply.  Could it be that my resistors in the RC B+ network are too small for high gain amps?

You could test your grounds. Connect a jumper wire to chassis ground and check each ground to hear a change, is the master volume ground at the same potential as the gain ground? Determine where the bleed through occurs, put a capacitor on the jumper wire to ground and check the signal path for the signal. The signal should go to ground downstream of the fault.

I'm of the opinion currently that the signal is either being communicated through ground or power supply.  Could it be that my resistors in the RC B+ network are too small for high gain amps?

It's hard to say without a schematic to look at. Asking everyone here on the forum to track down the schematic for the Soldano SLO because your amp is similar is not the most effective way to ask and get help. There are people here who know a lot and are really helpful, so why don't you just post a schematic of the Soldano and say what component changes you made?

The higher the resistors in the ps, the more voltage sag there will be in the power rail. The ps works with stages of R-C or L-C, where the R is a resistor, and the L is a choke. The Capacitance filters out variations. Primarily, it filters the AC power to a DC voltage, but there will also be R-C separating stages, so the preamp tube doesn't cause a voltage sag on the entire power rail. If the whole power rail sags from the preamp stage, even with the volume down, the signal will be heard at the output stage, since the power rail sag will show up on the output. This is just an example and may not be the problem in your case.

And thats why you dont give advice when you know you are a noob  ;D

:'(

walter said:

You could test your grounds. Connect a jumper wire to chassis ground and check each ground to hear a change, is the master volume ground at the same potential as the gain ground? Determine where the bleed through occurs, put a capacitor on the jumper wire to ground and check the signal path for the signal. The signal should go to ground downstream of the fault.

Click to expand...

I have connected each ground directly to the chassis through a jumper, but there was no noticeable difference either way.  I'll try it again just to be sure though.

I didn't entirely understand the second suggestion regarding using the cap to ground in order to check the signal path for signal, going downstream.  Sounds like something I very much need to try I'm just not sure what you are asking me to do.

dmp said:

It's hard to say without a schematic to look at. Asking everyone here on the forum to track down the schematic for the Soldano SLO because your amp is similar is not the most effective way to ask and get help. There are people here who know a lot and are really helpful, so why don't you just post a schematic of the Soldano and say what component changes you made?

Click to expand...

Fair enough...I apologize.  I found a schematic I made for one of the amps I was building that also had this problem (I was exploring high gain designs)...see below for the preamp schematic.  Gain stages 1 and 2 share a B+ tap, and 3 and 4 (along with the cathode follower) share a separate tap.  The PI has its own tap as well and the PI and poweramp are identical to that found on the JCM800. 
I have attached the preamp schematic and a layout of the JCM800 PI and power section i copied.


Just a thought.....could the cathode follower be picking up the stray signal since the plate is directly connected to the B+ RC tap?

jcm800

The test I would suggest is ground the grid (pin 2) of the successive stages until you stop hearing the bleed signal on the output.
At that point you know which stage(s) are getting contaminated.
There are several things that could be causing this as this thread shows, so narrowing it down with some testing is in order.
I would guess that the problem would be in the 2nd gain stage since it has the highest gain after the vol pot.

You can short the signal to ground. Use a capacitor connected to ground for the probe. With the fault occurring, touch the probe to various points in the signal chain. Touch the input, it should kill output. Check downstream, grids, input to P.I., etc. If at some point you hear the fault with the signal grounded, this area is where the bleed through is. The capacitor allows you to probe any point without shorting D.C. to ground.

figuringstuffout said:

ruffrecords said:

I suspect the problem is introduced at the pot. The wiper and ground terminals of the pot should be taken in a screened cable by the shortest route to the tube. Ground the screen at the cathode resistor.

Cheers

Ian

Click to expand...

The gain pot does indeed have a shielded cable in and out....the master does not since the signal by that point is well above line level.

Click to expand...

I bet if you short to ground  the grid connected to the wiper of the master pot, the problem goes away. If it does then you need to screen the connection from the master pot to that tube as I described. Just because the signal level at that point is above line level makes little difference because so is the level that is bleeding through.

Cheers

Ian

I bet if you short to ground  the grid connected to the wiper of the master pot, the problem goes away. If it does then you need to screen the connection from the master pot to that tube as I described. Just because the signal level at that point is above line level makes little difference because so is the level that is bleeding through.

Cheers

Ian
[/quote]

is that because of rather high signal impedances rather then because of level that you get bleeding there?

- michael