Maker Pro
Maker Pro

Electret biasing using active load / source follower

K

Kai Yum

Jan 1, 1970
0
Hi all,

I have a question about biasing the stock standard electrets (eg
Panasonic WM-61s). I'm guessing most regulars on the audio groups
also frequent this group.

* Rehashed stuff, background material *
One pin is to the FET drain, that's fine. The other one, as I
understand it,
is the FET source, which is also shorted to the cathode of a diode.
The anode
is connected to the FET gate, to allow a tiny gate biasing current.
(The
direction of the diode may depend on how the electret is charged- DC
polarity?).
Googling shows that professional mics/preamps use a source
follower
configuration to reduce distortion at higher SPLs- those same mics I
presume
have a separate lead for the gate biasing device, whatever it maybe,
and one
is for the source. That's the entire point of the Linkwitz mod, I
gather.
* end background *

My question is, can I use an active load (say a quiet current
mirror
set to 500uA) to sink Id on the 2-terminal electret, and take the
output
from the source? Obviously you would want to integrate a BJT stage
right there, but thats later.
If it really is that simple, somebody must have done it
already...?

The current sink would adjust the voltage on the FET source node
to the point where Vgs is consistent with Id. I guess the trick would
be to make the current sink ultra-stable so that the DC op point
doesn't wander too much due to noise or even normal operation because
of the diode's exponential I-V relationship.
As noted in AoE, the DC op point will float around (for
different copies of the same circuit) but that can be worked around...
I think perhaps by referencing currents instead of voltages. (Haven't
got to that stage yet- just speaking out aloud)
Is my analysis right and/or I'm missing some parasitic effects?

Incidentally, is the lifetime of an electret due to the leakages
of the semiconductors or dielectric absorbtion of the film itself?

Hope to hear from someone,
Kai
 
W

Walter Harley

Jan 1, 1970
0
Kai Yum said:
My question is, can I use an active load (say a quiet current
mirror
set to 500uA) to sink Id on the 2-terminal electret, and take the
output
from the source? Obviously you would want to integrate a BJT stage
right there, but thats later.
If it really is that simple, somebody must have done it
already...?

Absolutely. Take a look on Rick Chinn's web site (google for "uneeda
audio") to see some circuit examples.
 
K

Kai Yum

Jan 1, 1970
0
Much appreciated, missed it in my first search.

Thanks
Kai
 
R

Robert Baer

Jan 1, 1970
0
Kai said:
Hi all,

I have a question about biasing the stock standard electrets (eg
Panasonic WM-61s). I'm guessing most regulars on the audio groups
also frequent this group.

* Rehashed stuff, background material *
One pin is to the FET drain, that's fine. The other one, as I
understand it,
is the FET source, which is also shorted to the cathode of a diode.
The anode
is connected to the FET gate, to allow a tiny gate biasing current.
(The
direction of the diode may depend on how the electret is charged- DC
polarity?).
Googling shows that professional mics/preamps use a source
follower
configuration to reduce distortion at higher SPLs- those same mics I
presume
have a separate lead for the gate biasing device, whatever it maybe,
and one
is for the source. That's the entire point of the Linkwitz mod, I
gather.
* end background *

My question is, can I use an active load (say a quiet current
mirror
set to 500uA) to sink Id on the 2-terminal electret, and take the
output
from the source? Obviously you would want to integrate a BJT stage
right there, but thats later.
If it really is that simple, somebody must have done it
already...?

The current sink would adjust the voltage on the FET source node
to the point where Vgs is consistent with Id. I guess the trick would
be to make the current sink ultra-stable so that the DC op point
doesn't wander too much due to noise or even normal operation because
of the diode's exponential I-V relationship.
As noted in AoE, the DC op point will float around (for
different copies of the same circuit) but that can be worked around...
I think perhaps by referencing currents instead of voltages. (Haven't
got to that stage yet- just speaking out aloud)
Is my analysis right and/or I'm missing some parasitic effects?

Incidentally, is the lifetime of an electret due to the leakages
of the semiconductors or dielectric absorbtion of the film itself?

Hope to hear from someone,
Kai

Consider using a JFET, gate resistor to ground, and a second (same
type) JFET as the source "load" wher that second JFET has its gate
connected to its source (and the supply).
A matched pair would then set the Vgs of the input FET to zero, making
a good source follower.
 
Top