Troubleshooting hiss with signal tracer probe

[FlatAffectCamper]

Made one of these DIY signal tracers as per this link.

I have a ab763 Super Reverb and I'm trying to identify the source of hiss. I have been through dozens of forum posts and replaced just about everything in the amp, so at this point I'm trying to just do it the right way and find the source, and if it can't be improved, just live with it.

With the signal tracer, the only spot I find egregious levels of hiss (while turning up the volume) is at the ground side of the channel mix resistors, and at a bit lower level, back to the .1uf coupling cap off v4. All these components are new, and I tried swapping those capacitors.

Listening at points earlier in the signal chain the hiss seems so low I can barely register it, but I'm not sure I'm doing this right. At this point I have tried many different tubes, grounding schemes, and components, so I just want to focus on tracing.

Is the high level of hiss at the PI inevitable due to the level of gain it has on it? I have a 12at7 in there and have even tried a 12au7. I have seen many comments that it's a bad idea to use a 12au7 in the PI, but is it worse than installing a master volume?

Would appreciate any tips on how to search for the source of hiss (given all components here are new and high quality).

 

[Jon Snell]

If the his increases with the volume control adjustment, the source is obviously before the volume control meaning before the phase invertor.
Usual culprits are anode load resistors and noisy valves.

Shiny and new doesn't mean noise free.

 

[FlatAffectCamper]

If the his increases with the volume control adjustment, the source is obviously before the volume control meaning before the phase invertor.
Usual culprits are anode load resistors and noisy valves.

Shiny and new doesn't mean noise free.

Right. What would be next to check, if both of those fail to provide relief?

By the way, what is a normal, and what would be an optimal S/N ratio on these amps?

 

[dan40]

Are you using silver mica or ceramic for the smaller, pf range caps? There have been reports of bad silver mica caps causing noise issues. The worst offenders seem to be the black, unbranded silver mica caps used in many kits. Take a look at the treble cap in the tonestack to see if it's one of these.

You mentioned changing many of the components already. How about the potentiometers on the control panel. Are they still the factory originals? Most pots use a carbon track and they can be a source of noise when they get old.

 

[FlatAffectCamper]

Are you using silver mica or ceramic for the smaller, pf range caps? There have been reports of bad silver mica caps causing noise issues. The worst offenders seem to be the black, unbranded silver mica caps used in many kits. Take a look at the treble cap in the tonestack to see if it's one of these.

You mentioned changing many of the components already. How about the potentiometers on the control panel. Are they still the factory originals? Most pots use a carbon track and they can be a source of noise when they get old.

Thanks much. I am indeed using those black silver mica treble caps, and I have heard this comment before about them being bad, but maybe I should look at that again. What's the alternative? A 250pf ceramic cap? I had hoped to avoid just blinding replacing more **** but,

Yes the pots are original with the exception of the volume pots. Maybe I will see if I can dig up a couple new 250k's and at least test on the Normal channel whether that could be an issue.

 

[bwacke]

Does the hiss go away when you pull V1 or V2? (one at a time)

 

[FlatAffectCamper]

Does the hiss go away when you pull V1 or V2? (one at a time)

Well it's like this: There is some hiss at idle.
When I pull v1 only = unchanged
pull v2 only = unchanged
pull v3 only = unchanged
pull v4 only = hiss at idle down, let's say (unscientifically) 3db. See first post about where I was poking the signal tracer. Both channels go through the reverb with Robinette's mod.
pull v6 only = no idle hiss at all, it's below the level I can hear.

This hiss goes up with the volume, and I thought, probably misguided, that the initial hiss was just being amplified too much in subsequent stages.

I tried lifting the treble cap just now, and much of the hiss goes away, but then that's the frequency range of hiss. I swapped in another 500pf ceramic capacitor I had, and it doesn't seem to change the level of hiss.

I tried taking the tone stack out of circuit by lifting the ground (not sure if that really eliminates the pots) and that has no real effect on hiss.

A crude type-3 trainwreck master volume helps with the hiss at moderate volumes, so I'm going to install the better type-2. But I think the level of hiss is much more than it should be - it's enough to distract from playing, and really shows up in recording. Would be so useful to scientifically compare to any other ab763 amp for S/N ratio.

Incidentally, hum decreases when I turn up the volume, to its lowest point about 3, and then starts to increase again.

 

[bwacke]

The volume control behavior sounds like you have an oscillation problem. I wonder if one of your electrolytic caps in the HV power supply is bad. A defective one can do this. If you have some 20-22 uF @ 500V you can try jumping the B, C and D power supply caps one at a time (parallel the existing cap and don't kill yourself - if that's meaningless to you DON'T TRY IT!!! If you aren't equipped to do it safely, it's potentially fatal to you!). Then you'll need a 70 uF @ 350V to try the plate supply caps on at a time (SAME SAFETY RULES APPLY).

This means unplugging the amp from the 120VAC line, making sure there's no voltage on the caps with a meter, tacking in the test cap, reconnecting the AC and trying the test.

The test is power up and listen for loss of hiss, of course.

A scope would be easier, but you have to have one and the highest of the voltages in the amp would doubtless exceed the scope input rating, so you'd have to build a voltage divider out of a couple of resistors to pull it off.

 

[FlatAffectCamper]

The volume control behavior sounds like you have an oscillation problem. I wonder if one of your electrolytic caps in the HV power supply is bad. A defective one can do this. If you have some 20-22 uF @ 500V you can try jumping the B, C and D power supply caps one at a time (parallel the existing cap and don't kill yourself - if that's meaningless to you DON'T TRY IT!!! If you aren't equipped to do it safely, it's potentially fatal to you!). Then you'll need a 70 uF @ 350V to try the plate supply caps on at a time (SAME SAFETY RULES APPLY).

This means unplugging the amp from the 120VAC line, making sure there's no voltage on the caps with a meter, tacking in the test cap, reconnecting the AC and trying the test.

The test is power up and listen for loss of hiss, of course.

A scope would be easier, but you have to have one and the highest of the voltages in the amp would doubtless exceed the scope input rating, so you'd have to build a voltage divider out of a couple of resistors to pull it off.

Thanks, that's helpful. You mean the volume control behavior in terms of the hiss, or the variation in hum? This is the first I've heard of the filter caps causing hiss, so I'm glad to have one more thing to try.

I am able to, very cautiously, work safely around the high voltages. The reminder is always welcome to maintain healthy respect for their lethality.

If the filter caps check out, is there any other potential cause of oscillation that would have this effect? The amp originally had the B, C, D preamp filter caps each grounded to a different place on the chassis. Some time back I changed that to match the Robinette diagram with a single ground point, and also put new resistors in the doghouse, but it didn't change anything.

 

[Jon Snell]

To stop any parrasitic oscillation on V6, remove the lead from pins 2 and 6 of V6 and fit a 47k resistor in series with each lead at the valve base end.

 

[dan40]

Incidentally, hum decreases when I turn up the volume, to its lowest point about 3, and then starts to increase again.

Is there a bright cap on the volume pot or tied to it?

 

[Peegoo]

@FlatAffectCamper

If it's a hiss or a SHHHHHH sound, it may be shot noise caused by the resistor type(s) you used in the circuit. Look here for a good primer on the topic:

Resistor Types - Does It Matter?

Guitar amplifiers

www.aikenamps.com

 

[FlatAffectCamper]

To stop any parrasitic oscillation on V6, remove the lead from pins 2 and 6 of V6 and fit a 47k resistor in series with each lead at the valve base end.

Thanks very much. I have been reading dozens of forum threads about these problems and have never seen that one, so your expertise is highly appreciated.

I tried this (with 67k - too high?) at 2 and 6 on v6 at the tube socket side and for a moment I thought it helped a lot, but the hiss came back, which is puzzling, so removed them. Now I am getting harsh distortion by the time I get to 2.5 - 3 on both channels, which is not the speakers or preamp tubes. It sounds like it is running on one power tube, but swapping a power tube makes no change. Im not sure what went wrong.

Is there a bright cap on the volume pot or tied to it?

Yes but lifting it didn't solve it.

 

[bwacke]

A few questions:

(1)
The fact that removing all the tubes in the preamp chain one at a time had no affect on the hiss suggests a common element in all of them, i.e., the power supply; particularly the electrolytic that supplies the "D" lobe which runs V1, V2 and V4. When you pull the phase inverter the whole signal chain is disconnected from the power tubes. Silence would be no surprise unless the power tubes themselves were oscillating.
(2)
In your first post you say the signal tracer showed hiss was loudest on the "ground" side of the channel mixing resistors. Since both channels' mixing resistors (the 220K's ?) are in the plate circuits and go to a grid circuit, please clarify what you mean by ground, since all of this is fairly high impedance.
(3)
Did the amp hiss before you added the reverb mod?
(4)
You also say lifting the "treble cap" reduced most of the hiss. What is the "treble cap"? Are you speaking of the 500 pF cap that couples the channel output(s) to the reverb driver?
(5)
Does the Reverb depth control affect the hiss?


Comment:
Not sure of Jon Snell's recommendation to put grid stoppers on pins 2 and 6. V6 should be the phase inverter, a 12AT7, which has grid pins 2 and 7. Pin 6 is a plate and doesn't need another resistor. In a long-tailed-pair phase inverter one triode has the grid grounded by a 0.1 uF bypass capacitor and receives its signal through the cathode, so a grid stopper in that grid would create a filter that would lift the grid off ground for some audio frequencies and cancel the phase inversion at those frequencies. You should only need a 300 - 470 ohm resistor in series with the grid (pin 2 on V6A in an AB763 schematic) which is fed from the preamp V4B. With tens of thousands of the LTP circuits functioning without oscillating every day, it's doubtful that the stopper is needed, but it won't hurt.

 

[FlatAffectCamper]

A few questions:

(1)
The fact that removing all the tubes in the preamp chain one at a time had no affect on the hiss suggests a common element in all of them, i.e., the power supply; particularly the electrolytic that supplies the "D" lobe which runs V1, V2 and V4. When you pull the phase inverter the whole signal chain is disconnected from the power tubes. Silence would be no surprise unless the power tubes themselves were oscillating.

Thanks very much. I took your suggestion and jumpered the filter caps one by one with some caps from another set, and didn't notice a difference. But maybe I will go back and actually solder one into "D" and see if there is a difference. The ones installed are only a couple years old.

Can I ask you this:? The preamp filter caps are grounding just like Robinette's diagram above, with the ground terminating in about the same place, at a ground rail. Is this potentially wrong?

(2)
In your first post you say the signal tracer showed hiss was loudest on the "ground" side of the channel mixing resistors. Since both channels' mixing resistors (the 220K's ?) are in the plate circuits and go to a grid circuit, please clarify what you mean by ground, since all of this is fairly high impedance.

My bad in miscommunicating this. I did the Robinette reverb mod, and took out the normal channel mixing resistor, and put the recommended 47k resistor from the vibrato resistor to the point where the NFB resistor meets ground with v5 cathode resistor.

The reverb mod made no difference whatsoever in the hiss, which has been there since I got the amp. More than once I reversed the mod to double check. The 47k resistor actually lowered reverb hiss a hair, I think.

So I was picking up the hiss at a high level on the top side of that 220k and 47k, from the perspective of the diagram, above. But also further upstream, at V4.

(3)
Did the amp hiss before you added the reverb mod?

see above.

(4)
You also say lifting the "treble cap" reduced most of the hiss. What is the "treble cap"? Are you speaking of the 500 pF cap that couples the channel output(s) to the reverb driver?

I meant the 250pf cap in the tone stack, referenced by Dan40. I recently replaced the 500pf you're talking about.

(5)
Does the Reverb depth control affect the hiss?

Yes, but not as much as the volume knob, which I hope illustrates about how much hiss I have on the volume alone, without reverb.

Comment:
Not sure of Jon Snell's recommendation to put grid stoppers on pins 2 and 6. V6 should be the phase inverter, a 12AT7, which has grid pins 2 and 7. Pin 6 is a plate and doesn't need another resistor. In a long-tailed-pair phase inverter one triode has the grid grounded by a 0.1 uF bypass capacitor and receives its signal through the cathode, so a grid stopper in that grid would create a filter that would lift the grid off ground for some audio frequencies and cancel the phase inversion at those frequencies. You should only need a 300 - 470 ohm resistor in series with the grid (pin 2 on V6A in an AB763 schematic) which is fed from the preamp V4B. With tens of thousands of the LTP circuits functioning without oscillating every day, it's doubtful that the stopper is needed, but it won't hurt.

I have a 5w 680ohm resistor at hand that I could put on V6 pin 2 upon recommendation.

I'm going to install a type-2 master volume tomorrow. The hiss is considerably lower when I put a 12au7 in the PI, so I'm thinking if I can't solve this problem, I will just dial it down with a master volume.

Thanks again for your help so far.

 

[bwacke]

Can I ask you this:? The preamp filter caps are grounding just like Robinette's diagram above, with the ground terminating in about the same place, at a ground rail. Is this potentially wrong?

The ground connection shown between the electrolytics and the preamp ground bus should be fine.

A 12AU7 in the phase inverter just drastically reduces the gain of the PI circuit because of the low mu and the lower impedance of the tube, so the master volume control will accomplish the same thing, but give you some adjustment and won't distort.

What do you hear with both V1 and V2 pulled?

"My bad in miscommunicating this. I did the Robinette reverb mod, and took out the normal channel mixing resistor, and put the recommended 47k resistor from the vibrato resistor to the point where the NFB resistor meets ground with v5 cathode resistor."

Rob's preferred reverb mod was the first one where you keep the normal channel isolation resistor and add another for the vibrato channel. Tying two tube plates together at 180V DC form two different plate resistors and extracting a common signal as shown in the second reverb mod, to me, is bad engineering practice.

I disagree with removing that 220K resistor in the grid of V6 and moving the vibrato to the negative feedback circuit of the PI. There is a potential noise source. I'd try lifting the 47K resistor you added and see how it affects the hiss. This mod I didn't see on the Robinette site. I'll have to look for it.

"I tried lifting the treble cap just now, and much of the hiss goes away, but then that's the frequency range of hiss. I swapped in another 500pf ceramic capacitor I had, and it doesn't seem to change the level of hiss."

On which channel did you try lifting the treble cap? Try it on the other one, too. Any difference? And both?

One problem with the reverb mod is that both channels now feed reverb driver and the final preamp before the PI, so both channel can contribute noise to both that final preamp, V4B, and the V3/V4A reverb circuit so it's harder to come up with the right combination of tests to isolate things.

What happens to the hiss if you pull both V1 and V2?

I'm tired of thinking, so over to you. :>)

 

[FlatAffectCamper]

Back in there with the signal tracer. Results at the points here;

1 and 2: High level of hiss, goes up with volume turned up
3. High level of hiss, goes up with reverb level
4. Consistent level of hiss, much higher when reverb output line to tank is unplugged. Tried swapping this capacitor to no effect.
5. Hiss discernable with volume increase but lower level

I guess these are all to be expected, it's just a matter of how high the level of hiss is.

I'm unable to pick up the hiss earlier in the circuit - too low level to detect with my tracer anyway.

The Robinette reverb mod is easily reversible - doing so confirms no impact on the level of hiss - except that it needs that 47k resistor. Robinette did not include that in his circuit, but through other forum posts I discovered that it is not optional. Not using it results in much higher hiss than baseline without it. Other than that, I installed (and uninstalled) the Robinette version to a T.

If you try it on an ab763 I'm pretty sure you will get the same results. As they say, what works in practice doesn't always work in theory. I think there is a clear "availability heuristic" in play when someone mentions a modification, but in my case, it's not the problem.

What happens to the hiss if you pull both V1 and V2?

When I pull 1&2 (or 1,2 and 3) the hiss at idle is the same as with them in.

However, when I pull 4 (also 1,2,3) - there is no hiss at idle.

With 4 out, and 1&2 back in, the hiss at idle is barely discernable - much lower than with 4 in - but turning up the volume dials in the same significant level of hiss. (Normal channel without the reverb mod).

Thanks to all for comments so far. I am trying to do this empirically. If anyone has ideas for further testing, I'm all ears.

 

[bwacke]

"I am trying to do this empirically."

I am trying to assist you analytically, too, but some things don't make sense, yet. I'm at fault for not specifying that when pulling tubes and listening, the tone and volume controls should be at least 50% of max. The method here is to use the amp itself as a signal tracer by eliminating sections of the signal chain and evaluating the noise level to see if there is noticeable change in what is heard.

"If you try it on an ab763 I'm pretty sure you will get the same results. As they say, what works in practice doesn't always work in theory. I think there is a clear "availability heuristic" in play when someone mentions a modification, but in my case, it's not the problem."

I put the Normal/Vibrato channel reverb mod on a Deluxe Reverb Reissue (AB763 circuit) about a month ago with no appreciable added noise of any kind despite my suspicion that it would get noisy because of the extra lengths of signal wire running close to other signal-carrying printed circuit traces. I guess 60 years of experience with RF and AF signal layout occasionally pays off. The mod did, however retain the 220K resistor you removed and did not include an additional 47K anywhere. However, you have stated that the hiss was present when you got the amp, so the mod is a non-issue.

How do you define "idle"? No signal at the inputs, no cable and instrument plugged in, and settings of all controls at ...what?

"With 4 out, and 1&2 back in, the hiss at idle is barely discernable - much lower than with 4 in - but turning up the volume dials in the same significant level of hiss. (Normal channel without the reverb mod)."

I don't understand the bold part of the above sentence. Are you saying that the idle condition is with the volume controls turned down to zero and that increasing the volume adds hiss with the V1 and V2 reinstalled?

If removing V1 and V2 doesn't affect the hiss level, I can't figure out how taking the 250 pF cap out of the tone circuits could reduce the hiss or how reinstalling those tubes could affect the level with V4 out - because the removal of V4 opens both the normal and the reverb signal paths to the Phase Inverter...except through the power supply.

That would take us back to:
1. Bad solder joints
2. Filter caps
3. Metal film 1W resistors in the plates of V1 - V4 & V6
4. New preamp tubes - or known good ones; not switching around old, unknown-condition ones, although, sadly, I've had several noisy new tubes lately.

 

[SoK66]

Made one of these DIY signal tracers as per this link.

I have a ab763 Super Reverb and I'm trying to identify the source of hiss. I have been through dozens of forum posts and replaced just about everything in the amp, so at this point I'm trying to just do it the right way and find the source, and if it can't be improved, just live with it.

With the signal tracer, the only spot I find egregious levels of hiss (while turning up the volume) is at the ground side of the channel mix resistors, and at a bit lower level, back to the .1uf coupling cap off v4. All these components are new, and I tried swapping those capacitors.

Listening at points earlier in the signal chain the hiss seems so low I can barely register it, but I'm not sure I'm doing this right. At this point I have tried many different tubes, grounding schemes, and components, so I just want to focus on tracing.

Is the high level of hiss at the PI inevitable due to the level of gain it has on it? I have a 12at7 in there and have even tried a 12au7. I have seen many comments that it's a bad idea to use a 12au7 in the PI, but is it worse than installing a master volume?

Would appreciate any tips on how to search for the source of hiss (given all components here are new and high quality).

View attachment 1056602

From your description I suspect the hiss you're hearing is coming from the plate load resistors, particularly if they are carbon composition type. A change to 1 watt metal film type resistors will usually exorcise the hiss. Another source can be the common modification(s) to the negative feedback resistor value, like Fender does in some of the reissues. The stock AB763 value will usually cure that one

 

[Thumper]

From your description I suspect the hiss you're hearing is coming from the plate load resistors, particularly if they are carbon composition type. A change to 1 watt metal film type resistors will usually exorcise the hiss. Another source can be the common modification(s) to the negative feedback resistor value, like Fender does in some of the reissues. The stock AB763 value will usually cure that one

Thank you, I was about to suggest the same thing.