I need help understanding this one.

[Chico]

I have been hacking a small clone build to learn about how it works. :twisted:

Anyways, one hack was to add feedback or regeneration control.
I tapped off a point after the the delay section just before the mixer section (the emitter of the transistor at the output of the MN3007 before the 1uf cap) and fed it back in to the input section of the first op amp.  I think I used a 50k pot as a volume control, and feed the signal back from the wiper of the pot through a series fixed 2.2k resistor and a .01uf cap.

When adjusting the control, the effect goes from barely noticable to  squeal.  

Then, I added a Mosbooster (Jack Orman's Mos Boost design) before the input to the small clone to see if buffering would have an effect  as this is what is done in many flanger schematics (Thanks Mark H for pointing me to some great references).  At any rate, when the Mosbooster is kicked in, even at unity gain, the squeals go away, and I have a usable range of feedback control.  (Not quite the effect I was hoping for, but still a usable control that adds varying degrees of subtle to metalic to oscillation)

Can someone explain what is going on here?  Why does buffering the input to the small clone have such a dramatic effect on the regeneration control?

I understand how boosters can stop treble loss etc. but this one has me stumped.

[Boofhead]

I tapped off a point after the the delay section just before the mixer section (the emitter of the transistor at the output of the MN3007 before the 1uf cap)

This tap-off point seems OK.

and fed it back in to the input section of the first op amp.

You need to *mix* the feedback signal with the *input signal* of the *delayed* signal path - look how the Boss BF2 does it.

The problems you have are:
- You are feeding back to the dry signal area, at the input.
-  Yhe input to the first opamp will not allow mixing.  In fact when you place the booster stage before the opamp it is *not acting as a buffer at all*. What's happening is the booster has a high-ish output impedance and this allows some signal to feedback- effectively forming a resistive mixing point.
- To make things worse the first stage has pre-emphasis and this effectively boosts the highs in the feedback loop.  Feeding back to the delay input prevents this altogether.

[Chico]

Boofhead:

Thanks again for the explanation.   Back to experimenting.