bugfixes and adjustments...

 

I determined that the lower-than-expected fx loop gain was due to a wiring error: effectively, I was only running through one of the two gain stages on the return, so instead of a gain of 100, I had a gain of 10; given the send attenuation of 20:1, this led to an overall loop gain of 0.5, which was commensurate with the lower level I was hearing when the loop was engaged (-6dB).

After fixing the wiring error so that both stages were operational, the gain (as expected) was too high.  Theoretically, the overall loop gain would have been 5, but my ears told me it wasn't quite that high (for complicated math reasons, the gain of these negative-feedback triode stages isn't quite as high as the simple resistor ratios would suggest; akin to how cathode followers have a true gain slightly less than 1).  Nonetheless, I first tried substituting a 1M resistor for the 4.7M feedback resistor in the second stage.  This should have theoretically resulted in a perfect gain of 20 on the return, corresponding to the 20:1 attenuation on the send.  However, indeed, the overall gain was audibly a bit less than 1.  So I changed the resistor to 2M (1M + 1M).  The gain is ever so slightly above 1, but close enough.

In other changes: I put a 5:1 resistor attenuator in front of the boost stage, to tame its gain a bit.  Now it provides just about the right amount of boost, rather than the overwhelming, squealing jump in gain it was contributing when running at its intrinsic gain somewhere near 100.

And I put 10pF capacitors across all of the 12AX7 anode-cathode gaps (8 in total).  This led to an immediate improvement in stability and noise.  My plan was to remove some of the caps, progressively, to get rid of ones which are not necessary for stability (probably ending up with caps only on the high-gain stages), but actually I think I will keep them all in the circuit for "good luck", unless I find too much HF rolloff in the upcoming playing tests.

So far, I have only been using my new signal generator as audio source in making these changes: not actual guitar.  Next step will be to test with guitar.  I still haven't evaluated the v-mid EQ, beyond basic functionality (I may wish to shift its operating frequency range).  And now that I have the gain of the fx loop trimmed properly, I need to actually try plugging in an external effect unit, to make sure the signal levels are where they need to be.

Hoping to get this unit buttoned up, tested, burned in, and shipped by the end of this year.  Although we are about to leave town for two weeks, I think this schedule is plausible.  Which is very exciting!