selecting the right FET for mic pre

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

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

jah

New member
Joined
Apr 13, 2005
Messages
4
Location
east coast usa
Greeting; I want to separate my mic capsule from the body (preamp) such that the capsue will have a fet pre to dirve the extension cable. I do not know what level (min/max) of signal would be coming from the capsule. Is there a way to know this w/o measuring it? Also I need to locate a FET manufacter that carries "can" style fets.

thank you
jah
 
I guess the obvious question first is why. But...I guess you realize that you will need more than just a FET with the capsule, unless it is an electret type. The other high impedance components, like the capsule coupling C, high-meg bias resistors etc. should all be in the immediate vicinity of the capsule.

Vishay/Siliconix still make a few JFETs in metal cases, including the 2N4416A, which has the minor advantage of the case being isolated so it will act as a shield. I don't know how well-made they are anymore, since they long ago abandoned diffusion-based processing in favor of ion implantation and annealing. Why do you want a metal case? I suspect the good Toshiba parts in plastic would outperform, based on their typical low gate currents. Note that the shot noise in 50pA of gate leakage is about equivalent to the current noise in a 1Gohm resistor. Good 4416's used to be about 1pA at relatively low drain-gate voltage.

I'll defer to PRR on the question of max. levels---although it all depends on how much SPL you bring to bear.
 
Thanks for the replies and sorry about the late return...I have been away the last 2 weeks...plan on going on some more travel ;-(
BUT...WHY you ask is so that I can have recordings for the school that are non obtrusive.
I am aware of the required componets ;-) big gate R, couple of tuning/isolating caps and the drain and source R's....I have seen both gate giasing and "self-biasing" circuits.
I am not aware of the capsule coupling C though...a design I have seen did not have one and I thought it was not required since the condenser capsule is a C. hmmm. is this required?
It does not have to be a true metal can the Plastic styles will wrok , I just didn't want smd's.

Adrianh; thanks for the toshiba tip as I think the old one I once saw used is no longer available.

thanks both of you ;-)

I hope prr wil chime in on the max input question too...I realized that it depends on the environment subjected to but don't want it to saturate...I guess I need to determine the max spl...

cheers :guinness:
jah
 
"I am not aware of the capsule coupling C though...a design I have seen did not have one and I thought it was not required since the condenser capsule is a C."

Some have it, some don't. I just wanted to make sure you were aware of the super-high-Z parts being close to the capsule, whatever they happen to be for the given design.

The SK170 is a honey of a part, although some complain about its high capacitance as being in some way detrimental to the sound. From a signal-to-noise ratio standpoint it is very good---even though the capacitance will attenuate the voltage the overall S/N should be better than using the low C devices like the 4416 and its ilk. At high frequencies for a given doping profile/construction/gate width the best noise match is where the FET capacitance is equal to the source C, hence the 170 is pretty reasonable for ~30pF capsules.

The other nice thing about the 2SK170 is its low pinchoff voltage and low noise at relatively small drain current.
 
> I do not know what level (min/max) of signal would be coming from the capsule. Is there a way to know this w/o measuring it?

Count the number of musicians on stage. Multiply by two if there is a jazz/rock drumset.

> I want to separate my mic capsule from the body (preamp) .... so that I can have recordings for the school that are non obtrusive.

Bigger mikes mean you can charge more.

Or if a client REALLY wants invisible micing, I make them pay even more than my big-rig.

Get a couple lavalier mikes. AudioTechnica makes several with wide response, some cardioid, XLR on a beltpack several feet away.
 
Look for the real Schoeps circuit on the web, it has a SF schematic for the separate buffer for the microphone capsule from the preamp. I forget where it is on the web.

You will need a good cable, ground, shield, capsule voltage, SF voltage and signal. Small good multconductor cable might be hard to find.

What microphone model do you want to do this to?

Why do you need a can fet? Did someone tell you you needed it?

If this is for a small diameter capsule (not omni) you might have a big problem. The volume and its shape behind the capsule backside has a BIG effect on the sound.
 
[quote author="bcarso"]"
The other nice thing about the 2SK170 is its low pinchoff voltage and low noise at relatively small drain current.[/quote]
This would be requird for a clean signal at low SPL's.

Is this a replacement for the 2sk1180? As this is what I was basing my design on but now that I am ready to buil, I don't see them readily available....I saw one company that sold them as discontinued parts at $4 a pop.....foolish ;-)



The volume and its shape behind the capsule backside has a BIG effect on the sound.

My design will not affect the back pressure vent of the small diaphram interchangeable (removable) cardioid capsule.

Cable is not an issue as I have a few choices and, I'll have to search for the schematic you speak of.
It does not have to be a "can", it can be plastic. I just didn't want a surface mount or a chip.

As for the max spl....I'm told that 20mV/Pa is appropriate.

thanks and cheers :guinness:
 
"Is this a replacement for the 2sk1180? "

That's a big momma Sanken power MOSFET it appears. You probably meant the 2SK118 with the "O" drain current bracketing, 0.6 - 1.4mA. It's a smaller geometry, lower capacitance device which looks like it has about 5nV/root Hz noise voltage spectral density at 1kHz and 1mA drain current.

For the same current an SK170 does 1nV/root Hz. Your signal voltage will get smaller due to the FET's 30pF capacitance adding to the capsule capacitance, but not near enough to offset the noise advantage.
 
Yes, you are correct ...sk118 "O"..thank you.

on another Question....using a self biasing technique (gate R only..not a V divider) and setting up the Drain as the output (common source)....having a cap (1 Pf) in front of the load resistor 220 ohms (bewteen drain and load R) going to ground...would that only take to ground hi freq and allow the rest to pass thru Load R? I first thought that all AC would go to ground and none would pass thru the load but am now thinking that it is freq dependent. Is this the case?

cheers :guinness:
jah
 

Latest posts

Back
Top