FET847 Microphone Project Circuit By Jonathan Burtner PCB Layout By Poctop

[karloff70]

Ok, so the mic didn't work when it came back from the tech so I re-did/tidied up the hi impedance area as recommended so it is all airborne and on the teflon pins, including the fet gate leg. And I redid the transformer so it is 5:1. I tried with the transforner spare ground lead to xlr pin 1 but only got buzzing noise no signal. Tried without the transformer ground connected and now nothing much at all. No signal l.

This is getting frustrating.

Any help massively appreciated.

 

[TwentyTrees]

Does the transformer test OK out of circuit?

 

[karloff70]

It shows continuity and about 320 ohm on one side and about 50 ohm on the other. I don't know how to test it further.

 

[Ricardus]

Did you upload any hi-res pics of the board? I've build 3 of these already and they all just work.

 

[karloff70]

Here are some pics.

 

[Khron]

Are those the absolute biggest 100nF capacitors you could find? What construction / chemistry are they even supposed to be?

At this point i'd even suspect the capsule bias supply not being in working order.

I don't suppose you have access to an oscilloscope..?

 

[karloff70]

They are ERO KT1800's. Polyester I think but I have liked them before.

Capsule bias supply?

I don't have access to a.scope, no.

 

[Khron]

If i remember correctly, those 100nF are only filtering the voltage that ends up getting applied to the capsule. You won't ever "hear" them.

What DC voltages are you measuring between circuit ground, and the common connection of R9 & R10? What about at the "Drain" test point? And across each of those humongous capacitors?

You probably won't measure anything across that top-most one, due to an overly-ambitious design brainfart that's "baked in", but the other two would be good to know.

 

[karloff70]

If i remember correctly, those 100nF are only filtering the voltage that ends up getting applied to the capsule. You won't ever "hear" them.

What DC voltages are you measuring between circuit ground, and the common connection of R9 & R10? What about at the "Drain" test point? And across each of those humongous capacitors?

You probably won't measure anything across that top-most one, due to an overly-ambitious design brainfart that's "baked in", but the other two would be good to know.

Ok, to drain is 11.5V. To R9 leg on top of board 40.9V and R10 is 41.3V.
The top cap indeed nada and the second one down 2V and the next ERO down 3.8V.

 

[Khron]

3.8v... That's actually believable, considering a voltmeter input impedance of 1Meg, and the series 10Meg resistor that's part of the bias voltage filtering. These form a resistive divider, so you're measuring at best 10% of the actual voltage at that node (closer to 5% "downstream" of that second 10Meg). Although for the sake of testing, you could solder something like 1k resistors in parallel with R7 and R8.

due to an overly-ambitious design brainfart that's "baked in"

Then again, because of this, it takes about 5mins for the mic to reach FULL sensitivity. For the first 30s at least, you probably won't hear much of anything through it.

 

[Ricardus]

Just stick the the parts list in the BOM and get rid of those "exotic" caps. I can't even see what's happening with the wiring since those weather balloon caps are blocking my view.

But what I can see seems to look OK.

The soldering isn't awesome but it looks good enough to function.

 

[Ricardus]

Then again, because of this, it takes about 5mins for the mic to reach FULL sensitivity. For the first 30s at least, you probably won't hear much of anything through it.

Mine produce listenable audio after about 20 seconds.

 

[Khron]

Mine produce listenable audio after about 20 seconds.

Ok, and how much longer until it reaches its max sensitivity then? At least with the component values in the schematic, it should take its sweet time...

 

[Khron]

"Listenable" means pretty much nothing; at least i have no idea how to (objectively) quantify "some" amount of signal

Ran a quick sim, but went with 1000x smaller capacitance for the last stage, and virtually none for the first two, to have the sim run within human (as opposed to geological) timescales.

1st attachment - Let's be generous and say the output voltage (N3, red) reaches full value, in the sim, in half a second; with 100nF instead of 100pF for C3, that's 500(!!!) seconds, which works out to almost 8 and a half minutes. And that's even before taking into account the effect of the two other capacitors.

That's the design brainfart i was talking about, which i believe i've mentioned in this very thread, and had an argument about in another thread. Oh, and on a related note, neither C11 nor the backplate connection need to be isolated in any way. C11 is an AC short to ground, and the only node needing isolation is the front diaphragm-to-JFET-gate connection. It's been "only" a measly six years since i've pointed that out (complete with the how's and why's). Still, "lovely" that none of the people involved in this project can be bothered to give a toss about supporting it in any way...

But back to the technical side of things...

2nd attachment - Removing the influence of the last capacitor, and focusing on the first stage (10meg + 100nF). Voltage at N1 reaches full value in... call it 5 seconds? That's not even all that bad.

3rd attachment - Let's compound the effect of both 10meg + 100nF combos. Thiiiiiiis might take a while........ Oh, there we go - "only" about 15 seconds all in all.

4th attachment - Going from 1pF to 1nF for the last capacitor adds another few seconds, getting up to 20 or so.

I won't bother simulating the entire thing with real-world values because it would take literally forever, at the 40msec/sec simulation speed i'm seeing here.

Oh, and of course, measuring the voltage across any of those three capacitors WILL drain them in a relative hurry, only to be replenished at the above-mentioned speeds.

 

[Gus]

5RC for a simple RC circuit without leakage discharge paths

 

[Ricardus]

Ok, and how much longer until it reaches its max sensitivity then? At least with the component values in the schematic, it should take its sweet time...

You missed the point of my post. I have no idea how long it takes to get to max sensitivity. I was trying to let the original poster know that he should be able to hear audio fairly quickly with a "test 1...2..."

So if he's not hearing audio within 20 or 30 seconds, he's never going to hear any and the mic needs more diagnosing.

 

[Khron]

You missed the point of my post. I have no idea how long it takes to get to max sensitivity. I was trying to let the original poster know that he should be able to hear audio fairly quickly with a "test 1...2..."

So if he's not hearing audio within 20 or 30 seconds, he's never going to hear any and the mic needs more diagnosing.

My apologies for the full-on rant, in that case I guess this has just really become a pet-peeve of mine, or something

 

[Ricardus]

My apologies for the full-on rant, in that case I guess this has just really become a pet-peeve of mine, or something

No worries.

 

[karloff70]

Kinda stuck. Might try with another transformer.

If you had to pick a UTM transformer for this, which one would you go for? The 6:1 U47? The U87 type?

 

[karloff70]

Ok, so fondling the mic around looking for issues, open under power with headphones on I notice that whilst no actual output (pre at 30dB gain) when I touch the 5th lead from the beyer 5:1 transformer (ground?) , which is not connected at the moment, with a screw driver, I get a loud burst of noise in the silence. At what seems like a good level for a signal.

Starting to think it might be something really stupid. But I am too stupid with mics to see what it is. Does that ring any bells?