Preamp filter stage question

[deci belle]

stokes said in "The truth about filter caps" thread:

This sounds to me like filter cap issues.I have seen different noises that you may not think were caused by filters cured by using good quality caps.Another consideration when building or changing filter schemes,is not to use too big a cap for the preamp stages,or adding an additional cap in a circuit that uses one cap for two pre stages.I experimented with a 5E3 by adding an additional cap for the first stage and even with Sprague caps it was not good.

I have a problem in my Showman amp that I haven't been able to trace yet (still learning how to use a scope). There is a metallic overtone (that stokes describes) when I play the small guitar strings at various frets (A and C notes are the worst) especially when using the neck pickup w/ tone knob low for jazz.

When I changed the normal channel to a 5f6 (stole the vibrato tube) and added an extra gain stage with the extra triode— I added an additional filter cap/R for that normal channel mod. If stokes' comment is relevant to my problem, and it is also recommended to limit the number of stages per power node, I am at a loss what to do.

The metallic overtone is clearly audible in both channels. I have tried different guitars, cabs, speakers, cables, and preamp tubes. New filter/bias supply caps.

The only thing I haven't tried is different power tubes.

Any thoughts on Stokes' comment about not adding add'l filter caps to the preamp?

[Chuck H]

It's a careful dance, and specific to each amp. Really tough answer overall because there are so many things at hand with your particular scenario. What Stokes was expriencing was likely the additional filtering preventing small amounts of coupling between stages that eliminated certain undesireable frequencies due to negative feedback. So in his case the added filtering was a detriment to the overall early EQ of the preamp that of course sets the EQ for what will be overdriven. In some amps just changing filter brands can have a similar effect due to a lower ESR (equivalent series resistance). The higher this resistance the more coupling can occur. But... It's not always negative feedback. In amps where the rail promotes more positive feedback (it's all about whats couples together at each node and how things are grounded) a lower ESR is needed for stability and using a cap with higher ESR can result in oscillation. When adding a gain stage that raises gain above what is typical in your average vintage amp it's always advisable that it has it's own filter. So no avoiding that.

If the metalic overtone is there when the amp is played standard AND when the extra gain stage is engaged I would speculate that it's not a filtering issue. If it were a filtering issue you would only hear the effect when the extra gain stage was used. It could be a layout issue. Sometimes a small amount of coupling in an amp can make it resonate at certain frequencies. You could try moving some leads around inside the amp to minimize coupling. IMHE most Fender style amps aren't layed out well enough to accept extra gain without instability problems. These would include sensitivity at certain resonant frequencies. Even when the extra gain is not used, the extra lead length and layout differences due to modding can affect the amps stability.

Another issue that sometimes causes the phenomenon you site is poor quality ceramic capacitors. They exist in old Fender amps. Some ceramic caps are somewhat microphonic and can cause resonant issues as well. If you trouble to do mods or rewire an amp go ahead and replace the lousy ceramic caps in the signal chain with film or mica caps.

Since you say it happens with any guitar or speaker I'm inclined to think it's either a microphonic tube, a microphonic ceramic cap or signal coupling due to layout.

[deci belle]

Thank you so much for that comment, Chuck!! Perhaps when I get off square one in the scope department, I'll find some clues in the suspect elements you have pointed out!

Is it ever necessary or advisable to shield the B+ wire from/to a power node or its associated gain stage(s)?

[melvin]

I wouldn't be surprised if it was a parasitic oscillation from your change in layout, even though it's happening on both channels. A pic and a schematic of the changes might help with diagnosis.

However, if you changed all the filters, I would make sure you used the same or higher voltage rating, the same values, and you changed the filter resistors. This as well as using quality caps.

Of course, it could be something totally unrelated, and could be diagnosed as such.

And by the way, when you convert a Fender amp into a higher-gain circuit, it's been my experience that you have to double the first filter value, not add to the preamp side.

[deci belle]

Wow, melvin! Thank you so much for your observations!! I did put a 10k cathode resistor without a bypass cap on the extra triode before the 5f6 circuit to keep gain down, but it's still definitely in Marshall territory. Ya, I will find out where it's coming from, soon enough.

[deci belle]

I took a huge bite out of whatever is responsible for that high-frequency ghost-note issue by creating a "shield' around all the wires coming out of the filter-cap "doghouse". I tightly wrapped a closely-spaced spiral of solid, insulated 22ga wire around the bundle of filter-cap wires from the hole where they enter the open chassis eyelet-board side and grounded it at the PT chassis ground tab.

A major change in overall tone and smoothness was achieved (less harsh and less trebly). Only a tiny hint of the ghost-note issue is still apparent— and only when I play the "A" note at the 18th fret.

Thank you for your help, everyone!

[pdf64]

My guess is that you've had parasitic oscillation, caused by gain/phase shifts/capacitive coupling; capacitive coupling is mainly due to the long parallel wire runs along the control panel (a weak spot for BF / SF Fender type designs).
The work described in post #6 may have reduced that coupling maybe by providing additional coupling to ground (ie the shield) or by otherwise altering the layout/coupling of the wire runs.
My experience is that 10k-22k grid stoppers on each common cathode 12AX7 socket grid terminal really helps; along with careful analysis of the circuit and its implementation, ie adding an overlay of all the coupling paths, considering which are significant in regards of their potential for creating instability and then doing something to reduce the significant stuff.
eg consider a 12AX7 whose sections are cascaded common cathode stages; between the grid of the 1st and the plate of the second there may be a voltage gain of ~3600. If that grid and plate circuit have poor lead dress then the combination of gain, 2 signal inversions (=360 degree phase shift) and a feedback path (by means of capacitive coupling) may result in oscillation. Not necessarily free running oscillation but may be oscillation at certain signal levels (even signal levels in other parts of the circuit), frequencies etc.
It can be really frustrating and difficult to solve but thinking about the whole circuit, including the invisible capacitors / non-ideal components and the prequisites for oscillation, can really help.
At the moment you seem to have reduced the problem to an acceptable level but it may not have adequately addressed the root cause; change a tube, move a wire and it may return. Pete.

[deci belle]

It's true, and I'm glad you said so. So I'll keep moving in the direction you've described, Pete— because I haven't really done anything about the sources of oscillation you have noted.