Electret capsule powering

[mhelin]

Interesting way to power an electret capsule, wonder if it's any better than the usual powering scheme:

http://www.dvdplaza.fi/forums/attachment.php?attachmentid=10713&d=1236075775


Capsule is to be connected between In and GND.

 

[Dan Kennedy]

I guess I'd be concerned about the load the impedance follower sees.

 

[PRR]

> concerned about the load the impedance follower sees.

2-wire electrets are universally "open plate". The gate is biased to ground, the source is tied to ground (rarely through a resistor), and the drain is the other wire. Naked grounded-drain JFET.

The drain doesn't really care what its load impedance is. Its "plate resistance" is very-high. The usual 2K2 load is sub-optimal but adequate. You could tie 100K to a high voltage and get significant voltage gain, but small Idss variations could put you past the JFET's voltage rating. You could load it into a short and get perfect transconductance source. Call it a JFET-TL072 "cascode".

As I see it: this plan offers a very-solid DC drain voltage, an order of magnitude more output than the usual capsule load, and fewer parts than the obvious normal capsule connection through a cap to an opamp.

A pitfall is that high-current capsules will lever the opamp output to the rail. The trend is to low and lower currents, but I doubt they reject occasional high-current parts because capsule current is rarely a critical specification.

I guess the noise performance could be a bit better (no 2K2 shunt to raise noise gain), but the noise is really about the JFET and whatever wheezy gate-bias trick they use.

It does not solve the usual problem of cheap electrets around music performance: the maximum gate voltage must be less than JFET threshold voltage, which is typically a few tenths volts. For the usual capsule sensitivity, this passes loud speech but not loud close music. I've done a lot of far-miking with inexpensive electrets, but close-miking usually comes out odd. To pass high levels through these things you must cut the source and pull it down negative of the gate. (I used to get 3-wire electrets but they are vanishing.)

 

[mhelin]

> concerned about the load the impedance follower sees.

It does not solve the usual problem of cheap electrets around music performance: the maximum gate voltage must be less than JFET threshold voltage, which is typically a few tenths volts. For the usual capsule sensitivity, this passes loud speech but not loud close music. I've done a lot of far-miking with inexpensive electrets, but close-miking usually comes out odd. To pass high levels through these things you must cut the source and pull it down negative of the gate. (I used to get 3-wire electrets but they are vanishing.)

Thanks, that actually was the answer to the guestion I didn't make but was thinking. So the "Linkwitz mod" is practically the only way to go if you want use electret as drum mikes?

Another question on the capsule powering, this time with +48V phantom: is it ok to use just a single resistor (alternatively  you could use a 1 - 2 mA constant currect source) to ground to lower the phantom about 10V lower in the (-) input compared to (+) input so that the capsule would just be floating there between say 48V and 38V voltages. This obviously works only with AC coupled inputs (transformerless usually).  I've tried that powering scheme also but it too distorts at high levels (drums). Maybe that would work with additional source resistor with Linkwitz modded (3-wire) capsule better?

 

[mshilarious]

Thanks, that actually was the answer to the guestion I didn't make but was thinking. So the "Linkwitz mod" is practically the only way to go if you want use electret as drum mikes?

Depends on the capsule.  The 6mm cardioids can take fairly serious SPLs without the mod, and even higher with.  Source resistors are mandatory.  145dBSPL might be possible at an acceptable THD.

 

[mhelin]

Depends on the capsule.  The 6mm cardioids can take fairly serious SPLs without the mod, and even higher with.  Source resistors are mandatory.  145dBSPL might be possible at an acceptable THD.

How about the Knowlels 9745 omnis, I think the DIY kit you used to sell did come with those (I have two kits here waiting for build)?

I also found a source for a Sennheiser 3-wire capsule model KE 4-211-2. It's a little bit expensive (~55 EUR) but guess worth it:
https://www1.elfa.se/data1/wwwroot/webroot/Z_DATA/03010100.pdf

It's available at Elfa (http://www.elfa.se).

Max. operating voltage is 15V (to 22k recommended impedance), and max. SPL 130 dB (1% THD) which is not too much however.

 

[mshilarious]

I don't see anything special about the Senn capsule, other than it's prewired.  For example, Panasonic WM65 can match the specs otherwise (with cut trace), although that's cardioid and not omni.  WM65 in a design like my kit will do >126dBSPL at 1%.  But then that kit is a simple circuit and it's possible to get a couple more SPLs with a lower distortion circuit (or if you simply pad the capsule output), and a few more above that with a cut trace.

Basically for high SPLs you want smaller capsules because they have lower sensitivity, and also the cardioids are lower sensitivity than omnis.  As PRR notes, less signal at the gate is a helpful feature for loud sounds. 

1% THD is not noticeable at all on fast transient sources like drums.  For example, that Knowles capsule in the kit circuit should do 1% at 120dBSPL.  It will be clipped by 130dBSPL though, and that's not good for a mic you want to put an inch from a tom.

It's not too hard to cut the trace on that capsule though, and run a jumper to a ground point.  That should get another 6dB of max SPL and a corresponding decrease in sensitivity, which is helpful if your preamp doesn't have a pad.  Otherwise I would build a pad into the circuit.

1% at 126dBSPL is just enough for toms if the overload behavior above that is well behaved (it should be).  You might have clipping on the initial peak of a very hard hit, but that should be over within 1msec.

To avoid that fate, get ye some WM65!