[Morris Slutsky]

On Jan 24, 8:24*pm, MooseFET wrote:
On Jan 24, 3:39*pm, Morris Slutsky wrote:



A two device amplifier with the feedback like shown is not a
novel idea. *It has been used quite a bit.

The wiring of the switched to ground will add some capacitance etc
to that node when the switch is open. *This will cause peaking of
the high frequencies. *It is better to run wiring at the low
impedance points of the circuit.

With the switch closed, the first section is running the tube at
its full gain. *This means that unit to unit variations and the
nonlinear effects of this part will show up in the results.

Moose,

I just checked SPICE about capacitance to ground at the footswitch
point. *There is a small peak at about 100 KHz from a 100 pF
capacitance there, and a really huge one from 1000 pF. *I'm hoping
that my actual footswitch is closer to 300 pF *- it probably is,

Did you measure it with a low frequency capacitance meter or use a
high frequency one. *I would expect that the capacitance may be higher
at higher frequencies and that there could be a large increase in the
damping factor.

assuming that most of this is due to the capacitance of the footswitch
cable itself (about 6 feet) - which SPICE expects to produce a 4 dB
boost peaking at about 165 pF. *I suspect that the actual boost will
be less than this, due to stray capacitances (wiring to ground plane?)

Any extra capacitance to ground at that point will increase the peak.
Capacitances at other points may make it get very ugly. *There is a
phase shift as you go around the loop.

lowering the overall high frequency response in ways that SPICE cannot
readily take into account. * But yes, you're right, I should expect
peaky behavior here. *It's outside the audio band hopefully, which is
good because you won't hear it, bad because it could cause inaudible
parasitics, mitigated by the fact that this peak is still less than
the response at this frequency in high-gain mode and if it doesn't
oscillate there why would it oscillate in low gain mode?

In the low gain mode, you have a feed back path that doesn't exist in
the high gain mode. *This means that as far as the tubes are
concerned,
the system gain (closed loop gain) is higher in the low gain case.



Thanks for your thoughts and advice.

Hear you, Moose, so I SPICEd it again. I've been using LT-SPICE with
Duncan Munro's tube models, it's easy to use and the price is right!
Anyhow, modeling an open switch as a 300 pF capacitor, which is
reasonable from what's known about guitar cables, I can snub out the
ultrasonic peak pretty much completely with a 30 pF cap across the 680
K grid resistor on the second tube. The cost is the loss of 1 dB or
so, at 20 KHz in high-gain mode. Probably I could live with that.
Although some guitar amp purists would complain about loss of
'sparkle', guitar amp lovers have some hatred towards snubber caps due
to their abuse by vendors in the past to compensate for really bad
wiring layouts. This ain't a huge snubber though. I could put one
in. The thing is, though, it's such a small capacitance value that I
have no idea how much stray capacitance is currently there and what
it's doing, I don't have a spectrum analyzer setup here, I make amps
in my kitchen mostly. But it's a reasonable thing to do. If I get it
open again, I might put a small snubber there.