Getting 6-8 fewer dB out of a new DIY KM84 clone in a stereo matched pair

[joulupukki]

I’m wondering if someone might help me figure out some things to check in a pair of KM84 clone mics I’m building. They both produce sound, but the second one I built (let’s call it “Mic B”) outputs about 6-8dB lower volume than Mic A. Unpowered and in-circuit I’ve tested about as many components (resistors and caps) that I can and nothing is that far off from each other. I‘ve tried swapping the capsules and that doesn’t make any difference.

Other things I’ve checked:
- continuity from the spring pin to the JFET drain (and a resistance of 0 ohms)
- the terminal posts aren’t touching the nose cone
- all solder joints look great as far as I can tell and I reflowed them just to be on the safe side
- voltage measurements between the two mics seem very close at the FET source and drain, after the main diode, the input voltage, etc
- no components appear to be touching the mic body
- the resistance between the primary windings of the transformer are within 5% of each other. Mic A has a resistance of 703R and Mic B 668R. That wouldn’t be enough to cause a volume difference, would it?

I was so excited to have a pair of excellent mics for recording a new album. I can adjust the preamp on the quieter mic and they sound close to the same, but with the increased gain on Mic B, there’s also increased noise (which is still low). It’s just disappointing and annoying to have to not just “grab one of them” and expect them to be the same.

Anything else I should look for?

 

[Khron]

Have you swapped capsules between them?

 

[joulupukki]

Have you swapped capsules between them?

Yes. The capsules both seem to be good. It’s just Mic B’s circuit. Something with it isn’t right.

 

[kingkorg]

Yes. The capsules both seem to be good. It’s just Mic B’s circuit. Something with it isn’t right.

Measured the actual values of 4pf capacitors out ot the circuit?

 

[ricardo]

Another possibility is that one 'side' of the centre tapped transformer 'output' winding might be OC or damaged. Whether this leads to 6dB less output or much less depends on your preamp.

 

[joulupukki]

I’m suspecting the JFET and am just gonna send them back to micparts and have him diagnose, fix, and send them back to me. I don’t have the equipment to do further testing.

Measured the actual values of 4pf capacitors out ot the circuit?

This circuit uses 3 22uF caps and in-circuit I’m unable to get them to test with my DMM.

 

[joulupukki]

I'm suspecting the JFET and am gonna just send them back into micparts for diagnosis. I don't have the right equipment to troubleshoot any farther.

 

[Voyager10]

Do Mic A and Mic B have the same noise floor if the preamp gain is set the same? i.e. mic B is noisier purely because you need more gain for the same signal level?

If so it could be mic B had reduced polarisation voltage to the capsule. This will be hard to measure directly, but you could compare the voltages on the filter caps between A and B.

 

[ZHOUJUN]

Several of my Neumann KM86s also had a phenomenon that I thought was very strange at first， Their output gains were different, their codes were very close, and the current and voltage of the detection circuits were not very different， Later I found out that the output parameters of the FET tubes were different， I think it needs to be paired， I'll give you another idea.

 

[TLRT]

Change the 4pf "pad" capacitors until they match,anything is good between 2-12pf

 

[MicUlli]

Probably it helps: KM84 output transformer can be wired to 50 or 200 ohms output impedance. Check whether the two transformers are wired in the same manner...
This would exactly explain the 6 dB gain difference.

 

[RuudNL]

It is a microphone parts kit.
As far as I know the 6,5:1 transformer has four wires, so there isn't an option to select an output impedance of 50 ohms.

 

[Voyager10]

My reason to suspect the capsule voltage (rather than FET or transformer) was the observation that the SNR is noticeably worse on Mic B. Assuming it's a decent design the FET/transformer will add little noise of their own, and the noise you hear will be just the unavoidable Johnson noise (determined by capsule capacitance). In this case, the SNR is determined purely by capsule sensitivity versus the Johnson noise - changing the gain at the FET or transformer stage will affect the noise in the same ratio as the signal, and the SNR won't be affected.

So, under reasonable assumptions, 6dB worse SNR => 6dB worse capsule sensitivity, which points to polarisation voltage as a possible problem.

 

[joulupukki]

Do Mic A and Mic B have the same noise floor if the preamp gain is set the same? i.e. mic B is noisier purely because you need more gain for the same signal level?

I believe so, yes. The only reason why it gets louder (and you only hear/see it with a graphic EQ during silence) with Mic B is that the preamp is hotter.

If so it could be mic B had reduced polarisation voltage to the capsule. This will be hard to measure directly, but you could compare the voltages on the filter caps between A and B.

When I measured the voltage at the drain of the JFET these are the readings this is what I got:

Mic A:
JFET Drain: 5.27V
JFET Source: 1.271

Mic B:
JFET Drain: 5.27V
JFET Source: 1.311V

Several of my Neumann KM86s also had a phenomenon that I thought was very strange at first， Their output gains were different, their codes were very close, and the current and voltage of the detection circuits were not very different， Later I found out that the output parameters of the FET tubes were different， I think it needs to be paired， I'll give you another idea.

That’s interesting yours were different.

In my case, I ordered these as a stereo-matched pair so I expected them to be nearly identical. I suspect it’s some sort of user-error in the way I built them and hopefully I didn’t fry the input cap or JFET – though they both measure (see above) the same voltage on the JFET drain and the 470pF input cap across the terminals measures 0.47nF on Mic A and 0.47nF on Mic B. I left the optional 8pF pad cap off on both mics.

Change the 4pf "pad" capacitors until they match,anything is good between 2-12pf

I guess doing this I’d have to change on Mic A to essentially make it so it’s not as sensitive, right? I’d rather get Mic B to have the same good output volume as Mic A.

It is a microphone parts kit.
As far as I know the 6,5:1 transformer has four wires, so there isn't an option to select an output impedance of 50 ohms.

That’s right. There are only 4 wires.

The resistance on the primaries of the transformers is marginally different:

Mic A resistance on the transformer primaries: 703R (secondary: 49.1R)
Mic B resistance on the transformer primaries: 668R (secondary: 49.5R)

I don’t suppose that’d make that much difference? Should I change the transformers between the two mics just to rule that out as a possibility?

 

[joulupukki]

My reason to suspect the capsule voltage (rather than FET or transformer) was the observation that the SNR is noticeably worse on Mic B. Assuming it's a decent design the FET/transformer will add little noise of their own, and the noise you hear will be just the unavoidable Johnson noise (determined by capsule capacitance). In this case, the SNR is determined purely by capsule sensitivity versus the Johnson noise - changing the gain at the FET or transformer stage will affect the noise in the same ratio as the signal, and the SNR won't be affected.

So, under reasonable assumptions, 6dB worse SNR => 6dB worse capsule sensitivity, which points to polarisation voltage as a possible problem.

How exactly would I measure the polarization voltage? Sorry I’m still trying to learn all of this. I’ve built 5 guitar vacuum tube amps from scratch, but this is definitely quite a different animal.

 

[TLRT]

How exactly would I measure the polarization voltage? Sorry I’m still trying to learn all of this. I’ve built 5 guitar vacuum tube amps from scratch, but this is definitely quite a different animal.

Thru 1gig resistors only with electrometer,but i'm sure you have 48v. The different response and sensitivity caused by bad matching of the capsules.

 

[TLRT]

And don't forget to measure the capsules capacitance too

 

[joulupukki]

Thru 1gig resistors only with electrometer,but i'm sure you have 48v. The different response and sensitivity caused by bad matching of the capsules.

Yeah, unfortunately I don’t have an electrometer. The capsules are theoretically “matched” and when I swap the capsules between the two mics, Mic B still has lower output.

 

[TLRT]

Yeah, unfortunately I don’t have an electrometer. The capsules are theoretically “matched” and when I swap the capsules between the two mics, Mic B still has lower output.

Check the fet's(measure idss and other parameters),and check mic B doesnt have a short with the capsule mount screws

 

[joulupukki]

Check the fet's(measure idss and other parameters),and check mic B doesnt have a short with the capsule mount screws

I’ll have to check the mount screws, but I’m assuming there’s no copper in the board under their mounting locations to short with. The nose cone doesn’t touch any of the components and there’s definitely an air gap there.

In order to test the JFETs, I’d need to remove them from the circuit, right? It’s my understanding that these JFETs are very sensitive to heat and I worry that by removing them (and re-installing them), I’d potentially do more damage.