Umble Strange Microphonic Feedback Problem

[Jdansti]

I built the Umble tone stack (shown below) and was testing it through my Noisy Cricket amp. I get an oscillation sounding midrange pitch at certain volume and master volume settings.  At first I thought it was a biasing problem, but all of the JFETs were properly biased.  I noticed several things:

1) The PCB was microphonic. Tapping on the PCB caused a sound at the the same pitch as the oscillation.  The 330pf cap near the second transistor from the left seemed to make the loudest sound when tapped. 
2). The oscillation changes from a mid-frequency pitch to a very low pitch when I blow on the board!  No joke!  I don't know what made me think to try blowing on it, but it definitely changes the pitch temporarily (if I could reproduce this effect, I'd have invented a new wind instrument  ). 
3) The oscillation goes away when I reduce the volume on the Noisy Cricket. 
4.  There is no oscillation when I unplug the speaker cab and use headphones.  Everything works perfectly. 

All of this makes me think that a component or multiple components are sensing the sound waves from the external speaker and are setting up a feedback loop. 

What could cause this?  What ever it is, it is sensitive enough to react to the pressure or temperature of my breath (or maybe the odor ).

[amptramp]

Some ceramic capacitors are naturally microphonic because their construction is similar to a crystal mike.  These capacitors generate a voltage difference across their leads and are usually worst near zero bias.  When you use headphones, there is no acoustic feedback, so the unit runs as intended.  It is rare (but not unknown) for a semiconductor item to be microphonic, so I would suspect capacitors first.

[Jdansti]

Thanks Ron. I'll see if I have any polys that will work. If I have to stay with ceramic, I'll try a different style or brand.