Dear god not the Selmer again... & PPIMV...

[EFK]

Well to anyone interested, I took all the advice and information I received here as well as what I was able to understand via Aiken, Geofex and Valve Wizard sites, and completely rebuilt the Selmer T'N'B from the ground up using the MKII schematic. DC elevated heaters, combination star and buss ground, and this thing is as silent as the grave!!!! Quietest amp I have ever heard. (Thanks Steve and MWJB especially) BUT....

...could not leave well enough alone. After all, how often can you really crank a 50 watt amp? So I decided to add a PPIMV, the common Fischer 'type 2' - 250K dual pot etc. I have installed this previously in a cathode biased 6L6 amp with nary a problem at all, and did not need to use shielded cable. So, I did the same thing here. (I do twist pairs of wires very tightly, and I lay them down against the chassis, at right angles to other wires etc.) It works as it should, and does not add any extra noise EXCEPT!!! ---> Imagine the MV pot as a dial numbered 1 to 10. I can run it from 1 to 9.5 with no problems at all, but if I turn it ALL the way up that last .5 to "10" it squeals and screeches uncontrollably. Does not matter where the normal volume knobs are - it does this even if they are all the way down. If I let it screech a little and prod around with a stick, I can hear some slight changes but it still squeals and makes all kinds of noise even if I move the wires around pretty dramatically.

This amp has NFB off the 7.5 ohm tap, through a 100 ohm NFB resistor. I have heard of problems with PPIMV and NFB, but never tried it on a NFB amp. Also, it does use an adjustable bias pot (added) but this does not seem to be affected - it works as it should. Why could this be oscillating (which has to be what this screeching is)? Any ideas?

oops almost forgot - here's the schematic I'm using although this is diode rectified instead of tube rectified. Any problems w/ using the PPIMV with this circuit?

[Chuck H]

What's standing out for me is that the amp worked fine without the PPMV. And the problem is happening as the amp approaches it's stock parameters (this is the nature of the Fischer PPMV, the amp is stock with the master cranked) so it's not a matter of the PPMV being within the loop unless there was already a problem with the loop before the PPMV, right?. So there must me something about the new installation that is causing the problem. Is the PPMV pot too close to the preamp controls on the front panel? Lead dress is always suspect in a case like this.

Just for kicks you might try swapping the PI leads at the pots in case you got them mixed up during the installation. This would turn your NFB loop into a PFB loop and manifest with the higher gain as the master is advanced. also...

You mention that the amp does this without regard for where the volume controls are set. That says to me that there is a PFB loop AFTER the preamp. And it's probably cross talk due to lead dress since the EMF of the leads tied to the PPMV increases as you advance the master. So... What leads are close to the new circuit? Is ther a better way to isolate the circuit, shorten leads, rerout leads away from the preamp, etc.

I'd definitely try swapping the PI leads at the pots.

Thinking out loud.

Chuck

[EFK]

Thanks for responding Chuck. The only place I could mount the pot was through an extra hole in the back; it is well away from the preamp (other end of amp, in fact) *but* it is near the preamp power supply lines and can caps. However, none of the leads are - they run down against the chassis. I double and triple checked the hookup - everything is going where it is supposed to go and like I said, it works perfectly until that last 1 degree of turn to full. So, I think you are absolutely right - I think there's a lead dress oscillation happening here and I either need to shorten and reroute or go with some shielded cable. Frankly, I have heard many report that shielded cable was unnecessary here but due to the longer leads involved and the fact that the chassis is quite busy, maybe it will help.

[pdf64]

Best practice is to fit grid stoppers to the el34 grids, around 4k7, one end of the resistor directly to the grid terminal on the tube socket. None there on the schematic, so it might have been marginal before the lead dress was disturbed.

[EFK]

You know I wondered about that! What kind of wattage rating do they need to be? I've got 1W in various values - doesn't seem like a location that would require much, given the low voltage.

[EFK]

Actually what's even more interesting about your suggestion of 4-5K grid stoppers (there are none on there now) is that the oscillation does not occur until the residual resistance across the pot drops below 3K (I measured when it started squealing). I have no idea what I'm talking about, but that *seems* like it should mean something in relation to the grid stoppers being a very comparable value.

[Enzo]

Maybe, but probably not. Today is 12/15, and if I measure my 12v bench supply and it is 12.15v, would they be related?

You added a volume control and now the amp shows instability. Where are the wires running to and from it dressed? and what sort of wires are they?

Looks like CHuck already pointed this out.

[pdf64]

The grid stoppers won't dissipate any power, so use whatever is convenient. As you've seen, a short section of a 1/2watt pot track is coping with it.
Good lead dress is very important - keep high impedance grid wiring with a lot of gain after it well away from wiring and components which might carry high signal levels.
Critically analyse your grid wiring, think about the gain and signal inversions which follow each one, and take special care if there's significant voltage gain and >1 signal inversion between that grid wire and anything else.
If you can't keep good seperation or avoid long parallel runs, then use screened cable for the grid wiring.
Having done that, I fit grid stoppers to every voltage gain stage.

[EFK]

Nice one Enzo!

I used tightly twisted wiring and the runs, while they are admittedly not short (there is simply no way to keep them so unless I drape them over the filament wiring), do not run anywhere near the preamp or PI wiring. Not even close. I kept them against the chassis almost entirely however the wiring from the PI stage had to run near power wiring and the wiring from the pot to the grids of the power tubes had to run near the output jacks as well as the OT common. There were no grid stoppers at all.

I do have 4K7 on-hand so I'm going to use them on the socket, and I think I'll try shielded wiring. Is there any problem with the capacitance of the wiring robbing higher frequencies? I know this is debated insofar as too much shielded wiring used in the input stages. Can it be an issue here? Do I need to be checking the capacitance of the varied shielded cables I have on-hand to use?

[txstrat]

I've read somewhere the best thing for grid wires is to keep them AWAY from the chassis (same for cathode wires, but keep plate wires close to the chassis). Might be a matter of discussion, but that's what I've read while searching the net for good layout practices due to reducing hum and/or noise (aside from other things like crossing wires -if possible- at 90 degrees etc.).

[Enzo]

the wiring from the pot to the grids of the power tubes had to run near the output jacks

That right there is asking for trouble. I would not be so worried about proximity to the PI or preamp. Where the MV to grid wires are exactly what I was concerned with.