Negative feedback vs. no feedback

[AusTex64]

So I built some mics using the Gefell M94 capsule, which has a substantial HF rise. I built mine without feedback on the advice of Oliver Archut. Real simple circuit, much like a U47 buffer. On the Gefell bodies and also on the Neumann KM54 (very similar capsule), negative feedback is employed to even out the frequency response, among other things. I find the mics I built to be bright, as one would expect. So I built a buffer where I added the negative feedback scheme used in KM54. There's a resistor that's used to set the amount of negative feedback. Uwe Sattler, formerly of Neumann, provided me with the setup spec for choosing the resistor. I used a trimpot instead to exactly dial it in. It worked. Then I did comparisons between the mic with and without feedback. The mic without feedback sounds much more lively and exciting, but often requires outboard EQ to tame the high end. The mic with feedback sounded dead in comparison, but the HF issues were tamed.

So the question is which is the better approach? I ended up with no feedback and using EQ (usually a Pultec or Neve inductor EQ) to dial back HF when necessary. I felt the loss of liveliness and character with feedback wasn't worth it. I'm curious if others here have had similar experiences. Thanks.

 

[emrr]

Matches some of my experiences.  Usually a balancing act. 

 

[Delta Sigma]

On a mic that has a capsule with a high 'f' rise, boosting with eq with filtered feedback feedback vs cutting with an eq on flat mic amp sure sound different. Definitely benefits to both.

I built a U67 with Max's transformer and heard the same "dead" sound you heard with KM54 style feedback. Once I replaced that transformer with Moby's, I no longer heard the deadness of the feedback taming all the highs but lost the harshness of a flat amp with a K67.

The U67 feedback is a little more complex than the KM54 and I think I just hit the right mark with Moby's transformer being quite close to the original that the feedback circuit was designed around (err, vice versa).

Have you tried changing the value of the cap in your feedback circuit? If you reduced the value, your feedback would affect the higher frequencies. You could end up with a little bump somewhere if your feedback filter cutoff is above the 'f' the capsule starts to jump up at. You could also increase the value of the cap and increase the resistor. That would lower the reaches of the filter but reduce its impact (hopefully I got all that right).

I wonder, though, if it's the loss of the "air" frequencies that you miss when feedback is engaged. Hard to dial in with a simple RC filter like that. This sounds like it would be a lot of fun to experiment with though! I love these feedback circuits, mostly because the design is mostly way above my comprehension.

 

[abbey road d enfer]

On a mic that has a capsule with a high 'f' rise, boosting with eq with filtered feedback feedback vs cutting with an eq on flat mic amp sure sound different.

It shouldn't. If there's a difference, it's because the resulting response is different.
The main problem is that the compensation in the KM54, like in many other mics, and in most of the usual hacks, is a 1st-order low-pass, which tames the over-sibilance but also kills the higher frequency response.
This has been discussed to death. The only proper compensation for diaphragm resonance is a notch filter. It is possible to use a passive LCR filter but it would be a tad bulky.
The best solution is using an active notch filter, but it must be inserted after the head amp. I don't know of any mic that does it. So actually, the best solution is to use the non-EQ'd capsule and apply EQ at a later stage. Particularly using parametric EQ, either analog or digital, has the big advantage of perfectly dialling the amount of cut and frequency and adjust the width.

I built a U67 with Max's transformer and heard the same "dead" sound you heard with KM54 style feedback. Once I replaced that transformer with Moby's, I no longer heard the deadness of the feedback taming all the highs but lost the harshness of a flat amp with a K67.

This tends to demonstrate that both the original xfmr and feedback compensation were not adequate. Your judgment shows that you value "air" more than the absence of sibilance. The feedback compensation, combined with a dubious frequency reponse of the xfmr attenautes too much the upper frequencies.

The U67 feedback is a little more complex than the KM54 and I think I just hit the right mark with Moby's transformer being quite close to the original that the feedback circuit was designed around (err, vice versa).

Actually, the U67 feedback circuit applies both LF cut and low-pass filtering of the diaphragm resonance. The diaphragm resonance compensation is no better than on the KM54.

I wonder, though, if it's the loss of the "air" frequencies that you miss when feedback is engaged.

That's my analysis.

Hard to dial in with a simple RC filter like that.

Indeed! I would say impossible. Analysis* of the response curves of many microphones just demonstrates that.

* I mean the real frequency response curves, not the Photoshopped ones that are found in most brochures.

 

[Gus]

Have you tried a plate to ground capacitor?

 

[AusTex64]

Have you tried a plate to ground capacitor?

I have not. Might need to try it sometime.