FET847 Microphone Project Circuit By Jonathan Burtner PCB Layout By Poctop

[G-Sun]

Thank you for answering  Vac11,
but I don't understand.

Is it C1 to FD you mean? Why would making the connection floating from the other side make any difference?
As is, I've connected C1 to FD on the back.
Together with the connections:

FO to C1; R1 and Q1G
Bck to C11 and R2

 

[Vac11]

https://groupdiy.com/index.php?topic=63691.msg827963#msg827963

There is proper connection on the photo in linked thread ...

 

[G-Sun]

https://groupdiy.com/index.php?topic=63691.msg827963#msg827963

There is proper connection on the photo in linked thread ...

Thanks a lot!

Ok,
FD to C1; R1 and capsule front (not to fet gate? I see a connection going down there as well, don't I?)
Bck to R2 and capsule back, (not C11? or is C11 soldered to that R2-Bck-connection?)
?

(The pictures shows 2 mics, I'm not sure how identical they are)

When inspecting the picture, it seems quite identical to what I've done.
Or, what is the difference?

 

[G-Sun]

Here's the markings for back

 

[Vac11]

The second leg of the R1, R2, C11, C1 you have to solder direct to teflon PIN...from your pic it looks like you solder those to the PCB and then to teflon PIN...that is wrong.

 

[G-Sun]

The second leg of the R1, R2, C11, C1 you have to solder direct to teflon PIN...from your pic it looks like you solder those to the PCB and then to teflon PIN...that is wrong.

Thanks, but I have soldered them to the teflon pins.
Oh, you mean it's an error using the pcb? Well, the ones I've used are meant for this, and not connected to anything else, right?
So, it's more a matter of convenience?
(Thanks for the drawing btw.)

 

[Vac11]

Try this...

 

[G-Sun]

Try this...

Thanks a lot!
Yet, I still don't understand any electronic difference in the two ways.

Anyway, I'll be waiting some more weeks for some capsule-mount screws to arrive.
So, should be time enough to get this right

 

[0dbfs]

Hi G-Sun, I saw a flurry of activity here regarding this project so I wanted to comment on what many refer to as "the high impedance nodes" or similar....

These following junctions or nodes are all considered extremely high impedance. In this case this means that surface conductivity on the circuit board is likely even less resistance (eg; 1Gohm) than the values in the circuit. In other words current flow takes a shortcut and makes noise spurts / fizzles / frying-sounds etc... Any dust or oils from your hand can contaminate this hi-z area in the circuit and lead to noisy audio.... This is why all of these nodes are connected in the air and not on a circuit board (air is the best insulator). The teflon standoff's are extremely hi-z tolerant so they provide a solder point and electrical path between front and back of PCB for connectivity. This is why we also use alcohol on cleaning swabs or acid brushes to remove any finger oils / etc on the hi-z points.

- Connection or node between FET gate/input and capsule front diaphragm.
- Connection or node between R1 (1G) and capsule front diaphragm
- Connection or node between C1 and FET gate/input.
- Connection or node between capsule backplate and C11.
- Connection between R2 (1G) and backplate (this is the 48V line providing capsule voltage)

If you solder these nodes on the PCB it may well work for now but it will become more and more susceptible to noise over time and may even have noise issues right out of the gate...

In condenser microphones the capsule connections are always very hi-z and are typically either floating in the air or connected to a teflon standoff which is also very hi-z. The electrical difference between the two connection methods is that the one where the hi-z nodes are connected to the PCB instead of floating can and do suffer from short circuit over PCB surface resistance...

One way to illustrate this is to measure the resistance of your body by holding your meter probes in a hand and measuring resistance- what is it measure?? 3 or 4M or so? That's way less than 1G... Just a point of reference regarding contamination from finger oils and dust or similar...

Cheers!
-jonathan

 

[G-Sun]

Thanks a lot Jonathan!
Now things get clearer for me.
Sorry Vac11 for not understanding the difference in your method compared to what I did.
This was a part of electronic design that was outside my radar,
hence my mistakes.
Now, I just need to see how to best correct this.

So, Fet-gate, Q1G, should have the middle leg bent to air for this to be done right?

 

[Vac11]

So, Fet-gate, Q1G, should have the middle leg bent to air for this to be done right?

Correct...

 

[Ampampam]

Can anyone tell me why did they choose an RF unit for microphone amplification? What should be explanation? I just can not understand it.  ??? Thank you!

Regards

 

[sarakisof]

Hey Everybody!
Long time lurker. First time poster. I'm seriously contemplating building this FET847 as my first Poctop board mic.

I have a question about biasing the FET. In the thread starter Poctop said:
So in Andrew Shaap's build (http://www.vintagemicrophonepcbkit.com/Neumann%20U87%20Build.htm), he says:
Connect your mic preamp of choice, preferably one that works, and turn on the phantom power.  We’re going to calibrate the drain on the microphone to 11.5 volts.  Take your multimeter and put the red probe on r6. This is right next to the FET.  Take your black probe and put it on the ground hole.  Take a small flat-head screw driver and tweak the pot until you measure 11.5 volts.

I've read using an oscilloscope and injecting a 1K tone into the circuit is the best way to bias the FET, but I don't have an oscilloscope so I was going to do what Schaap did. My questions are:

a) If R6 is the red probe point of contact for the U87, what is the corresponding probe point for the FET847?
b) Would I still be measuring the drain to 11.5v on the FET847?

Thank you in advance for answering my noobie questions. I really did reading this board. There are a lot of smart and interesting people here. I have a serious interest in tinkering with mics and want to learn as much as I can.

Any reply to this? I am interested in building this as i have some BM 800 unused bodies.
I have the same questions as Wordsushi and would appreciate if have some help  here. I talked with Wordsushi and he aslo waits for a reply.

Thanks in advance. 

 

[TillM]

Look at the picture to see how you can bias the mic with a DMM.
Same method with a oscilloscope like on the u87.

Note: it doesn't matter if you change the red and black probe, if you change them your multimeter measure "- Volts".

 

[Wordsushi]

Look at the picture to see how you can bias the mic with a DMM.
Same method with a oscilloscope like on the u87.

Note: it doesn't matter if you change the red and black probe, if you change them your multimeter measure "- Volts".

Awesome. Thank you for that. Much appreciated!
Am I still trying to adjust to 11.5v like with the 87?

 

[TillM]

Look at the original schematic.
It shows 10v at the drain.
So I think with DMM Method 11,5v should be a good starting point.
Biassing with scope is always a better method, but you can use your DMM.

 

[sarakisof]

Look at the original schematic.
It shows 10v at the drain.
So I think with DMM Method 11,5v should be a good starting point.
Biassing with scope is always a better method, but you can use your DMM.

Thank you TillM!   Adjusting to 10v for better and accurate results maybe?
I have seen also a method described in U87 built, without oscilo, inserting 1k sine wave from daw, but didnt understand clearly.
If you dont mind, could you please explain it in details? Where do i have to inject signal,how the connections will be done?I have to listen to it with hadphones? etc...

 

[0dbfs]

Think of adjusting this voltage as an offset bias such that when signal clips it clips symmetrically on the positive and negative waveforms inside the FET.

It depends on the specs of that particular FET. 11.5VDC or 10VDC is only a little arbitrary but when the FET is selected such that 11.5V or 10V or whatever is on their original schematic presents symmetrical clipping then the selected FET will have the intended and desirable attributes regarding symmetry, operating points, headroom, and SNR / distortion, etc...

The FET input is the gate leg. The FET output is the drain leg. Measure that voltage on the source leg and adjust source pot for symmetrical clipping.

In addition to the audio (AC signal) on the drain leg there is also about 22VDC on the drain leg to power the FET. You can tap after the output cap to get just the AC or audio part of the signal into your measuring device.

Cut the end off an XLR and use the pin1 wire for GND and pin2 wire as input to your DAW... Then use either a software scope or just record the audio and adjust/measure.

Apply signal to the FET gate leg (and disconnect the capsule during these tests)... Adjust gain such that you can gain the signal up and as clipping starts it is symmetrical

Adjust the source resistor up/down and increase/decrease gain of the signal into the FET until you can get symmetrical clipping at the fet.

Since this is simply measuring and setting the operation of the FET it is advised to disconnect capsule and trafo to reduce interference.

****
On this FET847 there is a spot to install a precision source resistor and once you have your FET biased correctly you can measure and order that part or just use the pot.

Ciao!
-johann

 

[Khron]

Are you sure that's the case for the bolded node?

At least to my (admittedly feeble) mind, that would be the case, if the input circuitry looked more like it does in the case of SDC's (KM84, Schoeps etc), where you're forced to have the capsule connected with one end (the backplate) to the grounded case / body, and the other end, through a coupling cap, to the JFET gate.

The arrangement used here reminds me of the "ChinaMicMod" schematic of Zapnspark from the Yahoo micbuilders group, where you actually get to remove one of the 1G resistors (R2, in this case) and the input coupling cap, by applying the bias voltage to the backplate - just like here.

And besides, 1G + 100nF gives you a 100-second RC constant, which means the capsule bias voltage will only reach its maximum in about five times as long (over 6 minutes)

Hi G-Sun, I saw a flurry of activity here regarding this project so I wanted to comment on what many refer to as "the high impedance nodes" or similar....
<snip>

- Connection or node between FET gate/input and capsule front diaphragm.
- Connection or node between R1 (1G) and capsule front diaphragm
- Connection or node between C1 and FET gate/input.
- Connection or node between capsule backplate and C11.
- Connection between R2 (1G) and backplate (this is the 48V line providing capsule voltage)

On a side-note, on my BM800 / sE2200A "retrofit" boards, i "guilded the lily" and also inserted a "mask stop" polygon on the top & bottom of the board, in the hi-Z area, to not have the soldermask cover, "just in case" ;D Not to mention, striving to minimize connection distances, and not have a spaghetti of traces all over the place...

 

[Khron]

"Just for kicks", and after some PM's between myself and zaraxisof, i took the liberty of putting my own spin on the circuit (re-drew the schematic so that it was clearer and made more sense - to me, at least), and whipped together a quick board design. I'll gladly receive comments / critique on it

The thick-hashed areas at the top & sides are soldermask-stop areas. And whaddya know, all the connections fit on a single copper layer, AND with no jumpers ;D

I'm not 100% sure the positions of the two mounting holes on the sides are totally accurate, i just eyeball-ballpark'ed them. And couldn't be bothered to unpack one of my BM800's and measure...

<Edits = updated the screenshot a couple times>