KM84 style circuit with electret capsule!?

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

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

xeawr

Well-known member
Joined
Dec 14, 2016
Messages
166
Location
Germany / Austria
Ok I know this one will sound really odd. Let me explain:

I have an old microphone (from the 60s?) that once sported a dynamic capsule. But it seems to be impossible to get a fitting replacement, besides I'm not really interested in a restoration job. I just want this mic up and running with a decent quality. Good enough for some guitar amps, scratch vocals, etc. - u know - as a "color" mic.

Since space is really limited I need a circuit with as few parts as possible. I decided to take an electret capsule and an output transformer (the mic previously had a tiny transformer installed).

I put this little circuit together, it looks almost too simple, so, question, am I missing something? (I know I put the polarity of the output coupling cap wrong...)

km84-style-w-elect.png



here's the J113 data sheet for reference:

https://www.onsemi.com/pub/Collateral/MMBFJ113-D.PDF
 
xeawr said:
am I missing something?
You'll need to decouple the Zener with a 100uF cap, else it will be terribly noisy.
You need to power the electret capsule with a 10k resistor.
Finally you have to DC-decouple the capsule and the FET's gate with a 100nF cap (preferrably film or C0G) and a 100k resistor from the FET's gate to ground.
 
abbey road d enfer said:
You'll need to decouple the Zener with a 100uF cap, else it will be terribly noisy.
You need to power the electret capsule with a 10k resistor.
Finally you have to DC-decouple the capsule and the FET's gate with a 100nF cap (preferrably film or C0G) and a 100k resistor from the FET's gate to ground.


thanks for your feedback! Awesome! I've updated the schematic....:

bildschirmfoto-2020-05-18-um-03-32-30.png


Yes, I really could've thought of the zener noise. As for the DC-decoupling: The capsule came out of a mic circuit where it was directly coupled to the input of a 2SK660, so I'm not sure if I need the powering and the decoupling... any ideas?
 
xeawr said:
thanks for your feedback! Awesome! I've updated the schematic....:

bildschirmfoto-2020-05-18-um-03-32-30.png


Yes, I really could've thought of the zener noise. As for the DC-decoupling: The capsule came out of a mic circuit where it was directly coupled to the input of a 2SK660, so I'm not sure if I need the powering and the decoupling... any ideas?
Aah! When you mentioned electtret capsule, I assumed it was the standard type with built-in FET. Now it's a very different story. You don't need to power the raw electret, but you need to present it with a super high impedance, typically 1 Gigaohm (1000 megohms). thses 1Gig resistors are not too commonly available and are expensive. that's they are often replaced with a back-to-back pair of diodes that is connected between the gate and source of the FET. For it to work, the FET must be of a type that works with near(zero voltage between gate and source. Semiconductor mfgrs have produced such transistors, and decided to integrate the diodes directly on the same chip as the FET.  That's what the 2SK660 is. Check the datasheet.
You may want to take the 2SK660 from the existing mic or buy one and put it in your proposed circuit.
If you don't want to do that, you have to replace C1 with a 1nF film cap (styroflex recommended) and R6 with a 1G. You won't need to power the capsule then.
https://www2.mouser.com/ProductDetail/Ohmite/SM102032007FE?qs=%2Fha2pyFadujN9ULdix8LMrmsnWiZxbj9N6ylzAP2a1k%3D
 
abbey road d enfer said:
Aah! When you mentioned electtret capsule, I assumed it was the standard type with built-in FET. Now it's a very different story. You don't need to power the raw electret, but you need to present it with a super high impedance, typically 1 Gigaohm (1000 megohms). thses 1Gig resistors are not too commonly available and are expensive. that's they are often replaced with a back-to-back pair of diodes that is connected between the gate and source of the FET. For it to work, the FET must be of a type that works with near(zero voltage between gate and source. Semiconductor mfgrs have produced such transistors, and decided to integrate the diodes directly on the same chip as the FET.  That's what the 2SK660 is. Check the datasheet.
You may want to take the 2SK660 from the existing mic or buy one and put it in your proposed circuit.
If you don't want to do that, you have to replace C1 with a 1nF film cap (styroflex recommended) and R6 with a 1G. You won't need to power the capsule then.


Awesome feedback - it's worth gold!!!! Thanks so much! I have an 1GOhm resistor, so I think I will move into that direction, I've updated the schematic, please let me know if it looks good like this.

bildschirmfoto-2020-05-18-um-09-36-28.png
 
xeawr said:
Awesome feedback - it's worth gold!!!! Thanks so much! I have an 1GOhm resistor, so I think I will move into that direction, I've updated the schematic, please let me know if it looks good like this.

bildschirmfoto-2020-05-18-um-09-36-28.png
The 1G must be from the gate to ground.
The 330r is not really useful there. You could use two decoupling caps, one across the zener and one after the 330r.
 
RuudNL said:
From what I read, he is using an electret capsule without a built in FET.

Ahh.. my mistake -- sorry about that.

I've removed my post above to avoid confusion...
 
abbey road d enfer said:
The 1G must be from the gate to ground.
The 330r is not really useful there. You could use two decoupling caps, one across the zener and one after the 330r.

can you help me understand why the 330R doesn't work this way? 330 Ohm and 100uF gives a corner frequency of 4 Hz, shouldn't that remove all the noise created by the zener?
 
I will work. This principle is often used in commercial microphones.
The series resistor will have two effects: the supply voltage will be a bit higher than the zener voltage, but the zener noise will be effectively eliminated.
In fact, you create a zener diode with a high 'internal resistance'.
 

Attachments

  • example.PNG
    example.PNG
    11.9 KB · Views: 63
RuudNL said:
I will work. This principle is often used in commercial microphones.
The series resistor will have two effects: the supply voltage will be a bit higher than the zener voltage, but the zener noise will be effectively eliminated.
In fact, you create a zener diode with a high 'internal resistance'.

Thanks so much for your explanation! Really helps me understand the concept!!!

So if there is nothing else wrong with the circuit I would run off and build it (later today). Any objections?
 
xeawr said:
Thanks so much for your explanation! Really helps me understand the concept!!!

So if there is nothing else wrong with the circuit I would run off and build it (later today). Any objections?
Better arrangement, space permitting).
 

Attachments

  • 2caps.jpg
    2caps.jpg
    123.3 KB · Views: 52
Relying on only  the 2 x 6k8 resistors as current limits for the 9V1 zener will draw  around  11mA from the phantom power, and dissipate around 230mW in each resistor......
Would a further series resistor not be useful here?
 
rogs said:
Relying on only  the 2 x 6k8 resistors as current limits for the 9V1 zener will draw  around  11mA from the phantom power, and dissipate around 230mW in each resistor......
Would a further series resistor not be useful here?
There must be something wrong in your calculations; I see less than 60mW in each resistor.
 
I'm not sure drawing 11mA of phantom-power is possible while still ending up with the Zener voltage - do recall the pair of 6.8k resistors inside the preamp through which the power is fed :)

Most schematics i've seen use a pair of 2.2k's or so.

rogs said:
Relying on only  the 2 x 6k8 resistors as current limits for the 9V1 zener will draw  around  11mA from the phantom power, and dissipate around 230mW in each resistor......
Would a further series resistor not be useful here?

Any particular reason for recommending a physically large (15mm pitch) as well as pricey (3.50 euros!!!) resistor there? Wouldn't a cheaper one do just as well? :)

https://www.mouser.fi/ProductDetail/Murata-Power-Solutions/MHR0317SA108F70?qs=sGAEpiMZZMtlubZbdhIBIEfLfUtJicLTSiY%2Fd5C4E8I%3D

abbey road d enfer said:
Aah! When you mentioned electtret capsule, I assumed it was the standard type with built-in FET. Now it's a very different story. You don't need to power the raw electret, but you need to present it with a super high impedance, typically 1 Gigaohm (1000 megohms). thses 1Gig resistors are not too commonly available and are expensive. that's they are often replaced with a back-to-back pair of diodes that is connected between the gate and source of the FET. For it to work, the FET must be of a type that works with near(zero voltage between gate and source. Semiconductor mfgrs have produced such transistors, and decided to integrate the diodes directly on the same chip as the FET.  That's what the 2SK660 is. Check the datasheet.
You may want to take the 2SK660 from the existing mic or buy one and put it in your proposed circuit.
If you don't want to do that, you have to replace C1 with a 1nF film cap (styroflex recommended) and R6 with a 1G. You won't need to power the capsule then.
https://www2.mouser.com/ProductDetail/Ohmite/SM102032007FE?qs=%2Fha2pyFadujN9ULdix8LMrmsnWiZxbj9N6ylzAP2a1k%3D
 
Khron said:
I'm not sure drawing 11mA of phantom-power is possible while still ending up with the Zener voltage - do recall the pair of 6.8k resistors inside the preamp through which the power is fed :)
As it is, it draws a total of about 6mA, most of it going into the zener, since the FET hardly draws half a milliamp. It's not ridiculous, since that sort of zener wants to see about 5mA for best performance.

Most schematics i've seen use a pair of 2.2k's or so.
It would slightly increase the current draw with no benefit. 8mA instead of 6, but the FET would still be operating at leass than 500uA.

Any particular reason for recommending a physically large (15mm pitch) as well as pricey (3.50 euros!!!) resistor there? Wouldn't a cheaper one do just as well? :) Not a recommendation, just an example. Indeed there are probably cheaper ones. I don't care since I'm not the one who pays.  :D
 
Khron said:
I'm not sure drawing 11mA of phantom-power is possible while still ending up with the Zener voltage - do recall the pair of 6.8k resistors inside the preamp through which the power is fed :)

My fault -- misreading the schematic! ..  :-[    I was treating the 6.8k resistors as if they were the  actual phantom power feed resistors ... which are of course normally 6.8k  (or perhaps I should say 6.81k 1%)

Stupid of me.....my apologies.
 
abbey road d enfer said:
There must be something wrong in your calculations; I see less than 60mW in each resistor.

My apologies -- I was misreading the schematic, and reading the 6.8k as if they were the internal phantom power feed resistors... stupid of me  :-[
 
guys, thanks so much for your valuable feedback. Also sorry for the confusion I caused with the 6.8k resistors, in my most recent schematic I changed them to 2.2k. As some of you pointed out this is more common and also used in the KM84/U87 circuit, it shouldn't really change the circuit is working much, from what I gather.

Anyway, I still have three problems:

1. part count is still pretty high (space!)
2. the little NTE10-3 tranny lacks shielding, so I'm not sure how reasonable it is to go forward with it (the mic itself has a plastic housing)
- I would probably need to implement a permanent pad, otherwise I might overload the little transformer. What do you think?

Latest version:

bildschirmfoto-2020-05-19-um-13-02-34.png
 

Latest posts

Back
Top