'Nearly free' SDC from spare parts?

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

k brown

Well-known member
Joined
Jul 7, 2021
Messages
1,250
Location
California
Since I have more Oktava MC-012 capsules than bodies, I wondered if it would possible make another pair of mics by putting 2SK660 FETs into the gutted pad bodies. The K660s are very handy for DIY projects because their built-in diode on the gate negates the need for a 1 gig resistor from gate to ground (also, I have a bunch of them already). So all that would be needed in the mic body would be a 1 gig from 48v to the capsule, and cap between the capsule and the FET. I figured to put everything else in a battery box at the end of hardwired 15' cables, with 5 9v batteries (9v to the FET, and 45v to the capsule).

Could this work? - am I missing something obvious?
 
Last edited:
I know how to load/power/interface the K660 - I've built many mics with it, but all were electrets; I was mainly curious if there might any capacitance incompatibility between the Okatva capsule and the FET; any reason to expect they wouldn't play well together?
 
Well, the Oktave capsule is a 1/2 inch, I believe, so it's similar to what the 2SK660 is optimized for.
Seen from the input port of a head amp, there's no significant difference between an externally biased capsule and an electret
 
Thanks - can't wait to try it; actual diaphram diameter is just over 5/8" - just awaiting arrival of the 1 gig resistors.
 
If you take a look at the datasheet of the 2SK660, you will see that there is already a high value resistor between gate and source:

Includes diode and high resistance at G - S

So, IMHO there is no need for a 1 G.ohm resistor.
 

Attachments

  • 2SK660_NEC.pdf
    38.3 KB · Views: 21
The 1 gig resistor is for 45v polarization for the Oktava capsule. As I mentioned, that resistor/diode built in to the FET is what makes them attractive for DIY mics.
 
Just going by the original '012 circuit - capsule grounded, and center pin gets 48v through a 1G resistor after a bit of filtering, then cap between capsule and FET (which has a 1G from gate to gnd, which the K660 doesn't need) - no?

Rogs, you're saying just put the 45v on the center pin with no resistor?
 
Last edited:
I have to confess I haven't seen the actual schematic you're using, so I hadn't realised that the capsule is grounded - as Abbey has pointed out.

So it looks as if you will need your 1G resistor to feed the bias voltage.

One of the excellent features of the Schoeps schematic is its 'minimalist' approach to component use on the front end, IMHO.
But it does rather rely on one side of the capsule not being grounded, to get away with using only a single 1G resistor!
 
Makes sense - thanks.

I also like the fact the Schoeps circuit has no cap between the capsule and FET - gotta love one less cap.
 
Last edited:
Yes, of course a resistor is needed for the polarisation voltage!
But I had the impression that it was the idea to use a 1 G.ohm resistor to the gate of the FET.
A grounded capsule has the disadvantage that it takes longer for the circuit to stabilize.
The capsule will have a capacity in the order of picofarads, But the coupling capacitor to the gate of the FET will have a much higher capacity. Until this capacitor is fully charged, there will be a positive voltage on the gate.
Maybe the protection diodes will help in this case.

In the Schoeps circuit, the gate will be positive referred to ground.
This voltage will be subtracted from the polarisation voltage, in the case no coupling capacitor is used.
 
In the Schoeps circuit, the gate will be positive referred to ground.
This voltage will be subtracted from the polarisation voltage, in the case no coupling capacitor is used.
I have always read the Schoeps circuit as the FET gate being effectively grounded by the 1 G resistor, with the P1 adjustment allowing the FET to be biased so that the source voltage is positive to ground?......
 
I have always read the Schoeps circuit as the FET gate being effectively grounded by the 1 G resistor, with the P1 adjustment allowing the FET to be biased so that the source voltage is positive to ground?......
Actually, the gate is situated about halfway between 0V and the source voltage, via the variable resistor that sets the operating point. In the original Schoeps, the source is set at 1.5V and the gate ends up being at about 0.5-1V. the 1G resistor goes to the wiper of the variable resistor.
 
But with the 2SK660 there is nothing to 'set'.
And it was mentioned that he didn't want to use a coupling capacitor.
(I wonder how correct the bias and/or distortion will be with this fixed setting.)
 
Actually, the gate is situated about halfway between 0V and the source voltage, via the variable resistor that sets the operating point. In the original Schoeps, the source is set at 1.5V and the gate ends up being at about 0.5-1V. the 1G resistor goes to the wiper of the variable resistor.
Yes of course -- my mistake!
I have used a configuration where the gate is grounded, putting the FET into a 'self bias' mode, but that was for my RF mic infinite impedance detector - so a bit of a special case!
 
In the original post you'll see I fully planned to use a coupling cap; I just said that one of the nice things about the Schoeps circuit is that it avoids one. But if that requires an ungrounded capsule, that just complicates the build.
 
Last edited:
Back
Top