Piezo pick up feeding back

[Dave Yelverton]

Can anyone help with this problem? A couple of months ago a customer asked me to install an Allparts undersaddle p/u and EQ in his Fender acoustic. I did so taking great care to ensure that the bottom of the saddle slot was square and flat and the same for the saddle itself. I didn't notice any problems at the time, but my customer reported later that the low E was dominating at higher volumes and that it was far too easy to drive into feedback.He's tried using the EQ to lessen the dominance to little effect.

I inspected the guitar again and this time tried checking the body cavity resonance at the soundhole. It came out so close to E that I instantly suspected a wolf to be the problem. A number of things suggested themselves including lowering the resonance by attacking the braces, or putting in a sound port on the upper bout to try and raise it.
Has anyone experienced a similar problem and if so, how did you tackle it?

[Ry Charters]

Is the E string always loud? Or just the E pitch at different positions?

[Dave Yelverton]

good questiion, Ry and one I should have checked. The two open E's are very harsh and out of proportion, whilst the middle E is less so. I had a problem with a violin once which howled on G at all pitches. Thought it was a wolf but it turned out to be a longitudinal crack at the f-hole. However, the guitar seems robust enough, so that's probably not the problem.

[Ry Charters]

I've noticed that sometimes, if the bridge pin holes are "curved" in a semi-circle, and if the saddle is low (not very much above the bridge) the middle strings won't have enough down pressure onto the saddle and can cause this type of problem. Easily fixed by ramping the string slots a little more towards the bridge, or just sanding down the bridge if appropriate. This is a steel string right?
Also, sometimes some fuzz from drilling the hole can hide out in the slot and give you grief. I use a small rattail file and put a little bevel around the hole.
I encountered a similar problem a while back on a factory guitar that looked like it had a nicely machined out slot, put the transducer in, same problem. I eventually discovered that there was a slight concave curve to the slot, probably due to forcing a gap together in the middle of the bridge during the glueing process.
If it were me, I'd route that slot out just to eliminate that possibility from the equation.

[Rob Mercure]

This is a not uncommon problem with any P/U attached to the instrument's table which the saddle P/U untimately rests on. But since you made sure that it was nicely square if you slightly uncouple the E string by sticking a little sawdust or perhaps a speck of dense polystyrene foam then you can probably solve the problem. While the under the saddle P/U helped prevent this problem which was prevalent with the Barcus Berry "Hot Dot" and other first generation piezo P/Us this is one of the primary reasons that EQs are used with piezos. The only thing that feeds back more IMHO is a clip on condenser mic which also provides the most accurate acoustic guitar signal. While I know that removing and reinstalling the saddle a few times can be a real PIA I'm pretty sure that if you can decouple a little bit you'll succeed.

Rob

[Dave Yelverton]

Thanks for the comments and apologies for not replying earlier - much time spent attending to family needs over the school holiday! I've finally tracked down the cause. The problem barely existed on my amp but something about my customer's amp caused it to resonate excessively on E with this particular guitar, so it looks like his amp needs attention or changing. Nonetheless, I suspect the underlying cause is still the guitar cavity resonance rather than the p/u. However I wil try your ideas Rob and if it still feeds back at least I will have eliminated the p/u as the culprit.

[Tim Mace]

Stand closer to the amp,you get more feedback that way!

[Rob Mercure]

Interesting, I've been an electronics tech since HS (am 55 now) and have worked professionally repairng and building guitar amps. While not totally uncommon resonance at a particular frequency in an amp is still quite rare (unless it's a feedback loop from the guitar as suggested). Most guitar amps utilize "negative (degenerative) feedback" where a portion of the output signal is fed back into preceding stages out of phase to even out frequency response. If the signal at a certain frequency is amplified more than others then more of that frequency is returned to preceding stages antiphase which then cancels out more of that signal and creates a more linear response. Ah - but since guitar amps don't reproduce music like a hifi amp but produce the music then eash amp colors the signal to a certain degree and so some amps will have greater or lesser or different frequency responses - not to mention any EQ settings. Amps fall into various circuit classes and not all use the same strategies and some amps, usually lower powered, eliminate NFB so that the response is wilder and more under the musician's control (or lack of). But some things that can throw the NFB off is using a speaker of a different impedance than designed or, almost the same thing, setting an impedance selector switch to a different impedance than the speakers used. Some amps have variable NFB controls and if this amp has one then use it to increase the NFB (names vary by manufacture but almost none actually lable the control NFB - but I do!). If the amp has more than one speaker make sure that they are in phase (causes all sorts of weird symptoms depending on the amp/speaker/enclosure). And check out any signal modifiers in the chain between the guitar and amp such as EQs or phase shifiters or digital delays, etc.,

Lastly is it the amp or the customer? Does s/he stand in a certain way that produces feedback? Or strike the stings in such a manner that the E gets whacked harder? You might want to observe the musician with the commonest set up and note anything that strikes you as causing regeneration.

Rob