Parasitic Oscillation in an otherwise fine amplifier...

[dehughes]

I just built a fine amplifier. Very fine, if I do say so myself. However, I swapped out the "test the amp" tubes with some nicer ones as I THOUGHT it was all done, and noticed some microphonic squeal when the preamp volume on the 12AX7 channel was turned up past 3/4. Odd. So, I put the old "test the amp" tubes back in, and what do you know, the squeal was there, again. It was just fine before the tube swap, so you'd THINK it would be just fine after the original tubes went back in. Weird...

The amp is basically an AC30 with an EF86 channel and a 12AX7 Top Boost side, the top boost side having the master volume (pre-phase inverter...). The first thing I'm thinking of doing is putting shielded wire from the master volume pot to the PI, as I ran out of shielded wire in the build and never got around to replacing it. The grids of V1 and V2 (both 12AX7s) have shielded wire run to them.

Any other ideas? It is basically an AC30 with a few twists... I'll post some pictures of the layout and chassis again, but I'm pretty sure the components are laid out well enough to not cause issues like this.

The odd thing is that there was no parasitic oscillation BEFORE the tube swap, so why would it be there again when those same tubes were put back in? Can tube sockets go bad and cause this? I'm stumped... The squeal is diminished when the cut and/or the midrange knobs are turned down, or if the guitar volume is turned down, OR if I just back the preamp volume down below 3/4 open. The master volume knob doesn't seem to increase or diminish the squeal...

THANKS IN ADVANCE!

[dehughes]

Pics can be found here:

https://tubeamparchive.com/viewtopic.php?t=1120

They are older pics (I've done some updating of components, etc...), but they lead dress is the same.

Your help is greatly appreciated!

[dehughes]

Well, here is the latest:

I put it all back to stock today (put the power and filtering switches back to their normal settings...etc...) and put the original tubes back in, and....lo and behold....no issues....no screech.... Odd....

So, I slowly but surely went to tinkering. Preamp volume up full, let's see what happens....

I put the power and filtering switches up to their high settings (full voltages, 33uf filter caps of the recto and on the 1st gain stage...), no issues. Good.

I replaced the EH 12AX7 in V1 with the Tung Sol reissue 12AX7, no issues. Very good.

I replaced the EH 12AX7 in V2 (cathode follower) with the Tung Sol reissue 12AX7, squeal is back. BINGO.

So, I put the EH back in there, no squeal. Good again.

I then replaced the EH EL84s with the nice JJs I just purchased, no squeal. Very nice!

I switched the power setting to "low" and dropped the filtering on the 1st gain stage and right off the recto back down to 16 uf (low setting...), and still no screech or squeal. Super cool!

So it seemed to be something to do with the cathode follower circuit or tube (V2).

Then, I noticed that the Cut knob was a bit scratchy at the "near off" setting, and that the preamp volume (gain) knob made a low "whump" when turned up past 9 or so...and the master volume also made a bit of a "bump" sound when turned up past off. Interesting. Maybe this is something in the PI area...

So, I replaced the nice Ei 12AX7 in the V3 spot (PI) with a known good JJ 12AX7, the pop and whump are gone when messing with the knobs. Interesting....

So, the summary so far: The cathode follower (V2) and PI (V3) seem to be particular about the tubes used in them, as gainy and/or bright tubes seem to get things oscillating.

So, what does this mean? Should I be concentrating my efforts in the area of the cathode follower and PI wiring? Is there something else to be considered?

THANKS SO MUCH!

[MarkB]

The "bogey" value for gain - 100 for the 12AX7 - is a target figure for that model tube. Depending on mfg. quality and testing, you may get quite a different value. It could be that your circuit puts 12AX7s at the edge - if you happen to get an extra-gainy tube, the amp starts to complain. You might want to get a set of tubes from Lord Valve - he claims to test tubes for microphonics with a hammer.

[dehughes]

The "bogey" value for gain - 100 for the 12AX7 - is a target figure for that model tube. Depending on mfg. quality and testing, you may get quite a different value. It could be that your circuit puts 12AX7s at the edge - if you happen to get an extra-gainy tube, the amp starts to complain. You might want to get a set of tubes from Lord Valve - he claims to test tubes for microphonics with a hammer.

INTERESTING.... My Tung Sols tested out at 110/110, and they were in V1 and V2....AND the Ei 12AX7 that I was using in the PI is rated at 110/110 also. Interesting.

Would an AC30 circuit push a tube "to the edge", especially if it has a 500k master volume pot installed and 500k preamp volume pot? Most other component values in the tone stack, etc., are the same as the AC30 schematic...and the voltages are a bit higher than the old schematics, but nothing crazy.

THANKS!

[MarkB]

I was thinking that the layout may be causing the oscillation when the gain is pushed a little higher. Even a low gain amp will oscillate when the right wires get too close together. Your build looks great, but there could be some position issues somewhere. I have a 2xEL84 amp with a pentode preamp - 5879 - and an AC-15 power section, and I had problems with squealing too. [/list]

[dehughes]

I was thinking that the layout may be causing the oscillation when the gain is pushed a little higher. Even a low gain amp will oscillate when the right wires get too close together. Your build looks great, but there could be some position issues somewhere. I have a 2xEL84 amp with a pentode preamp - 5879 - and an AC-15 power section, and I had problems with squealing too. [/list]

Interesting...

What would be the first issues to consider in regards to layout?

What did you do to resolve your squealing issues?


THANKS!

[MarkB]

When I changed over from EF86 to 5879 I built a new board, and that took care of it. It might show up at full volume, but I never get it up that high. The usual suspects are grid wires - they're high impedance, so they're very sensitive. You want to keep higher signal wires away from each grid wire in the preamp, so the plate wires from a later stage should never be near a grid wire. Chopsticking - pushing wires around with a chopstick - will sometimes find the offender. Any output tube wires need to be kept away from preamp wires of course.

On the other hand, if you look at enough chassis gut shots, you'll see amps that violate any rule you can think of - long runs of parallel wires are common. Those amps must sound fine, so there's more art than science in layout. Fender and Marshall basically had guys moving parts and wires around 40 hours a week, until they came up with a stable build. Less rules, more persistence.

[dehughes]

When I changed over from EF86 to 5879 I built a new board, and that took care of it. It might show up at full volume, but I never get it up that high. The usual suspects are grid wires - they're high impedance, so they're very sensitive. You want to keep higher signal wires away from each grid wire in the preamp, so the plate wires from a later stage should never be near a grid wire. Chopsticking - pushing wires around with a chopstick - will sometimes find the offender. Any output tube wires need to be kept away from preamp wires of course.

On the other hand, if you look at enough chassis gut shots, you'll see amps that violate any rule you can think of - long runs of parallel wires are common. Those amps must sound fine, so there's more art than science in layout. Fender and Marshall basically had guys moving parts and wires around 40 hours a week, until they came up with a stable build. Less rules, more persistence.

Very interesting. I have many areas where wires cross, though the grid wires (with the exception of the PI grids) are all fed with shielded wire. Again, the only long run where there probably should be shielded wire is from the master volume pot to one side of the PI. This run is next to some tone stack wires and a ground or two. I'll fix this once my shielded wire arrives from Apex Jr. (Steve is great, BTW...)

Here is a picture of the PI (left socket) and Cathode Follower (right socket). Pin 1 is at the top... Anything look suspicious?

[Ron]

In making your wiring look neat, you wound up with alot of wires in parallel. I haven't traced them back using a schematic, but I see a potential danger. For some of the high-freq oscillations, it takes only a small amount of capacitance between the "wrong" parts of the circuit. BTW, if you think that shields don't radiate then I have a bridge for you!

To answer a previous question about squealing, the first thing I check is whether I got the phase of the feedback correctly or not. After that, it's the "chopstick" test and tube changes, as well as a ground check.

One thought about the sensitivity of some tubes to squealing, different tubes may have different amounts of inter-electrode capacitance that can stabilize or destabilize the amplifier, depending on the circuit. I'm guessing that high inter-electrode capacitance will cut the high frequency response and probably help quash very high frequency squealing.

[dehughes]

In making your wiring look neat, you wound up with alot of wires in parallel. I haven't traced them back using a schematic, but I see a potential danger. For some of the high-freq oscillations, it takes only a small amount of capacitance between the "wrong" parts of the circuit. BTW, if you think that shields don't radiate then I have a bridge for you!

To answer a previous question about squealing, the first thing I check is whether I got the phase of the feedback correctly or not. After that, it's the "chopstick" test and tube changes, as well as a ground check.

One thought about the sensitivity of some tubes to squealing, different tubes may have different amounts of inter-electrode capacitance that can stabilize or destabilize the amplifier, depending on the circuit. I'm guessing that high inter-electrode capacitance will cut the high frequency response and probably help quash very high frequency squealing.

Alrighty. That makes sense, especially the last part about high frequency response. I noticed that the Tung Sol 12AX7s, while being the most chimey and airy of the 12AX7s I tried, are also the ones that cause the most squeal when the preamp gain is maxed. Does this point to anything in particular?

How can I check if I have the phase of the feedback correct? I've attatched a pic of the Master Volume and Cut pots, as well as the bypass caps in the power section, etc... Lemme know if anything looks funny to you...

Also, what was is specifically that made you say that you "see a potential danger"? Was it the wires in parallel? If so, which ones should I look to move first?

THANKS!

[Ron]

Yes, the high-frequency response and the squealing go together in the Tung Sols.

To change the phase of the feedback, you can either (1) swap the (usually blue and brown) plate leads going to the output tubes, or (2) swap the grid leads going to the output tubes (these have 1k5 resistors connecting to the output tubes in an Express).

You can easily check cross coupling in your wiring by simply separating wires and observing/listening to what happens.

[MarkB]

The NFB squeal is a screaming thing - you know it when you hear it. If your oscillation only occurs as you turn your amp up, then that's not your problem.

Take a look at this Cornform chassis. Parallel wires all over, and evidently no problems. Just goes to show that it's not paralleling wires that's the problem, it's paralleling the wrong wires.

[Ron]

Sometimes the NFB squeal goes away when the presence control is turned way down. My comment was about squealing when first turning on a newly-built amp.

[dehughes]

Mark and Ron, thanks so much. So, being as my amp mostly just squeals with certain preamp tubes, and only when the preamp gain is MAXED, then I don't have a NFB issue, correct? Plus, my amp is basically an AC30 circuit, and that doesn't use NFB, right?

Then, I most likely have an issue with some wires in the preamp being a bit too close or running together in the wrong way, right? That I can check, by poking and prodding with a chopstick and seeing when the squeal goes away. The only trick figuring out how to get constant signal to my amp while I poke, so I can hear it if/when the squealing stops. Maybe hook up a CD player?

THANK YOU!!!