Arrgh! Can't track down problem: Oscillation??

[greg]

I have an amp which I built about a year ago. It started out as a Lightning clone. From the onset, I had an oscillation problem which I could never track down. I swapped OT leads, moved the OT to various positions on the chassis (held it in hand and moved it to the quietest spot, etc), swapped the OT totally, re-routed the OT wires, moved around all of the wires in the amp while playing. I have swapped the tubes, repositioned the speaker jack. Checked the voltages.

The amp is built on terminal strips, which is making it rather difficult to just "rebuild". I am determined to pinpoint the cause rather than to gut it and build it from scratch on a board. Looks like I have built myself into a corner on this one!

Anyway - I had two main issues - a high frequency squeal and a ridiculously bad distortion coming on even at very low volumes.

I put the amp under the bench and finally, a year later, decided to give it another shot. I started by converting this to a 15 watt DC-30, which has given me the opportunity to have a second channel to play with. I get the same result on this new channel (with the original channel completely disconnected by pulling its tubes). So that is good - looks like the preamp itself is not the issue. It must be in the PI or power amp.

I have moved one of my terminal strips farther from the power supply section and have resolved the high frequency squeal, but the terrible "crapping out" distortion issue still is present.

The long and short of my question is this:
Assuming that the problem is in the PI section or power amp - What components can cause this?

Is it only those which carry the signal? Could bad caps be the culprit? a bad resistor? Has my OT been damaged from the high frequency oscillation? Or, is it likely component placement causing an interaction resulting in a low frequency oscillation that would cause the amp to distort unnaturally?


Thanks in advance

Greg

[Mouthurst]

Probably not this simple but I know if you're using old pwr xfrmrs with an added 3 conductor source and occilation occurs; you have to reverse the primaries sometimes because they never payed too much attention to that back in the old days.
.

[greg]

That is interesting - I am using an old hammond amp for this and the PT was 2 conductor. I haven't heard of this before (swapping leads on the PT). I'll give it a shot.

Thanks

[Tom Phillips]

Reversing the phase of the PT
Sounds like some kind of urban legend. Makes no sense to me but I'm willing to listen if you care to elaborate.
Tom

[Tom Phillips]

Greg,

Since you have always had the problem with the amp their could be a wiring mistake or something like a bad decoupling cap in there from the start. Also could be lead dress in an area where a high level signal gets close to a low level signal. You said that you ruled out the pre-amp section. Does that mean the problem is not affected by any volume or tone control settings?
Do you have a scope available?

Do you have photos of your layout that you can post? A schematic of your current circuit would also be helpful.

Tom

[greg]

Well, no - it definitely is affected by bringing up the volume on each channel (there is no master volume) but as soon as the amp is audible, you can hear the problem. The amp and the problem get louder with volume, but really not too much worse. The schematic I am using is a stock DC30 schematic (less 2 power tubes) with the only change being one set of caps switched from .01 to .02 to match the values seen in the Lightning schematic.

When I say I ruled out the preamp, I suppose I should say that I ruled out the original preamp components (the 12AX7 Lightning channel) by adding in the EF86 channel, which is physically very separated from any of the original components. It too provided the same crap-out tone. So, I am surmising that it is in the PI or power supply. I'm going to go clip in some caps to see if I have a bad one now.

While I do have a scope, I have yet to learn how to use it. Picked it up at a garage sale and still need to buy leads for it. I assume that even a mis-calibrated scope would probably do wonders for me once I learn what the hell I am looking at! I did take it to a local builder who spent some time with the scope on it and he was unable to find what was going on. At that time, it was simply too much of a mess to have much faith that it even should work. Much of this has been resolved and I may look into getting some pics if that will get me an answer.

Tom - can you define what you mean by a high level signal interacting with a low level signal? More directly - what in the amp would be considered a low level signal other than the unamplified signal coming from the guitar?

Thanks

[Mouthurst]

Reversing the phase of the PT
Sounds like some kind of urban legend. Makes no sense to me but I'm willing to listen if you care to elaborate.
Tom

Well I heard that from a guy ...maybe he meant the OPT...

[Tom Phillips]

Well I heard that from a guy ...maybe he meant the OPT...

I think that's probably it.
Remindes me of Tim "the tool man" Taylor mixing up the story when he tried to repeat Wilson's advice.
Tom

[Tom Phillips]

...When I say I ruled out the preamp, I suppose I should say that I ruled out the original preamp components ... So, I am surmising that it is in the PI or power supply. I'm going to go clip in some caps to see if I have a bad one now.

Greg,
It's more probable that the problem is caused by layout, faulty grounding or inadequate decoupling between power supply stages than a bad component. (Unless its a bad filter cap)

Tom - can you define what you mean by a high level signal interacting with a low level signal? More directly - what in the amp would be considered a low level signal other than the unamplified signal coming from the guitar?

One simple way to think of it is:
High level signals are on the plate leads (several volts in amplitude. Maybe several tens of volts) and Low level signals are on the grid leads. (less, sometimes much less, than a volt) I am talking about the AC portion of the signal, not the DC power supply level.
However, it's all relative, since the AC signal on the grid of a power tube is larger than the AC signal on the plate of the pre-amp tube.
High level signals can couple through the air over to the low level line if they are close to each other. Especially if the wires are run parallel to each other and the low level line is high impedance. If those signals are in phase then your amp can oscillate.


Hope you make a breakthrough soon. It’s harder with a new build than with a mature design that did work at one time. If you get some high res clear pictures to post that may help.

Regards,
Tom

[greg]

OK - I went in and played around with the wiring a bit more last night. Moving around the wires and components while playing does not appear to have any affect on this - positive or negative.

I noticed that on the 12AX7 channel, if I crank up the Bass control, the amp becomes a very bad sounding bass-kazoo.

On the Ef86 channel, turning off the Volume does not take the signal out fully - there is still some signal going through. This Ef86 channel is the new channel and is very separate from all of the other components physically. I clipped the ground lug of this volume pot to the chassis (assuming I had a bad ground here) and it still comes through. Very strange. I have grounded pin 2 of the EF86, but I also read somewhere that pin 7, which is also unused, should also be grounded. Does anyone know why these should be grounded??

Additionally, I paralled each of the coupling caps (individually) to see if I could find one that was bad (letting DC through). It just occurred to me that running a bad cap and a good cap in parallel would probably still let the DC through on the bad cap, so this test was a waste of time. Am I correct there?? I suppose I should just measure it for DC, huh.

Anyway - thanks for the help thus far. I'll use this information and map out which components are likely to be the culprit before I go carving into this thing.

So close, yet so very far...

Greg

[greg]

Question for Tom Phillips

You mentioned that "High level signals can couple through the air over to the low level line if they are close to each other. Especially if the wires are run parallel to each other and the low level line is high impedance. If those signals are in phase then your amp can oscillate."

Can you give me an example of what would be a high impedance low level line?

Thanks
Greg

[greg]

Question 2 for Tom

So is oscillation less likely to occur between two components (or two leads, or a lead and a component) if they both carry a high level signal or both carry a low level signal? I am obviously trying to be methodical and direct in my next steps at fixing this thing (instead of cutting out/rebuilding entire sections of the amp).


Thanks again

Greg

[Tom Phillips]

...Can you give me an example of what would be a high impedance low level line?

Good examples are the grid lead of the first stage pre-amp tube or the grid lead of the reverb recovery stage.

[Tom Phillips]

So is oscillation less likely to occur between two components (or two leads, or a lead and a component) if they both carry a high level signal or both carry a low level signal?

Yes. It's less likely if both are high level or both are low level. Less likely but not impossible. It still depends on how tightly coupled the leads are and if the signals are in phase (worse) or out of phase (not as much of a problem)

I suggest that you try to isolate the problem by disconnecting some of the circuitry. Like pull the pre-amp tube in one channel and see if things work better.

I'd still like to see some pictures.

Also note the information previously posted. Your power supply could contribute to the problem in several ways:
1) Bad ground
2) Inadequate decoupling do to (1) or a bad cap
3) Two stages of the amp that are in phase and hooked to the same power supply node. Note that people frequently talk about the power supply filtering and ground loops when they are tracking down hum sources. However, they don’t remember, or haven’t yet learned, that the filter caps also serve to decouple the stages by providing a path to ground for any AC signal that gets on the power supply line. If that function is inadequate then AC signals can sneak between stages on the power supply line. That’s bad. Another reason to have a scope so you can see if the B+ lines are clean.

Don’t give up. This is a learning experience. If you get too frustrated then set the amp aside and read some of the classic audio design texts.
There’s lots of good stuff here: http://www.pmillett.com/technical_books_online.htm
I’d suggest Norman Crowhurst’s literature like the basic audio course.

Cheers,
Tom

[Matt T.]

greg,
FWIW, I have an amp that has an EF86 channel and a 12AX7 channel and my EF86 channel is just like yours regarding the 'can't turn it all the way down' phenomenon. So either that's normal...or we both made a similar mistake. I can't remember the pin numbering on the EF86 but if pins 2 and 7 are that internal shield, I only grounded one of the pins.

I've always wondered if I'm supposed to not use a regular tube shield with a tube that has an internal shield like an EF86.