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

[TLRT]

Yes you have to remove the jfets. Did you go thru the calibration steps with signal generator and scope? Then you can match the outputs with most of the modern 2channel digital scopes.

 

[joulupukki]

Yes you have to remove the jfets. Did you go thru the calibration steps with signal generator and scope? Then you can match the outputs with most of the modern 2channel digital scopes.

I did not do any calibration as these were purchased as part of a stereo-matched pair DIY kit from mic parts. I do have a 2-channel digital signal oscillator that I’ve used when troubleshooting guitar vacuum tube amps I’ve built.

I’m not familiar with how to calibrate the JFETs, but each came with its own bias resistor, one at 4.12K and the other at 4.22K.

I’m presuming the JFETs in these kits are some sort of special kind that are hard to come by? I mean … presuming it’s the JFET, input capacitor, or the 1G resistors, wouldn’t I be able to just order some replacements from some place like Mouser with a handful of different-valued resistors to then match them and then re-install? I dunno. The alternative is shipping them back to MP, paying $50 or so to have him troubleshoot and fix them and pay for shipping both ways.

 

[Voyager10]

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.

I don't know how closely the Micparts circuit follows the Neumann original (copy here: http://recordinghacks.com/pdf/neumann/KM83-84-85-SM84_Schematic.pdf) but in the original, the polarisation voltage is fed from the R9/R10 junction, through R5 and R8 and filtered by C6 and C10, and then goes through the 1G resistor R1 to the capsule connection.

First thing to check is the voltage at the R9/R10 junction - it should be 46V or so as per the Neumann schematic. If that's OK check the voltage across C6 and C10.

It's going to be hard to measure these accurately with a DMM because the current taken by the DMM will cause a noticeable voltage drop on 10M resistors R5 and R8. The best you can do is to compare readings between Mic A and Mic B - you may find the reading starts high and then drops (that's the capacitor discharging via the DMM), so wait for a stable reading.

If Mic B is reading lower voltages than A, this is very likely to be the source of the low output.

If you have a signal generator, then before desoldering the JFETs you can do a quick test by removing the capsule and injecting a signal into the amplifier input (the R1/C9/C2/C1 junction in the Neumann schemo) with the mic connected up to the preamp. 50mV 1KHz sine wave should be plenty. Compare the output level from Mic A with Mic B (without changing the preamp gain). If you get the same level in both cases it's unlikely the JFETs are at fault.

For a more accurate test, if possible, connect the signal generator output with a 50pF capacitor in series to the input. You'll get a lower output level than the previous tests, but if it's the same on Mic A and Mic B there's nothing wrong with the FETs or the transformer. If it's different, we've learned something

(If the Micparts circuit has different component labels to the Neumann one, can you post a picture of it?)

 

[joulupukki]

I don't know how closely the Micparts circuit follows the Neumann original (copy here: http://recordinghacks.com/pdf/neumann/KM83-84-85-SM84_Schematic.pdf) but in the original, the polarisation voltage is fed from the R9/R10 junction, through R5 and R8 and filtered by C6 and C10, and then goes through the 1G resistor R1 to the capsule connection.

First thing to check is the voltage at the R9/R10 junction - it should be 46V or so as per the Neumann schematic. If that's OK check the voltage across C6 and C10.

It's going to be hard to measure these accurately with a DMM because the current taken by the DMM will cause a noticeable voltage drop on 10M resistors R5 and R8. The best you can do is to compare readings between Mic A and Mic B - you may find the reading starts high and then drops (that's the capacitor discharging via the DMM), so wait for a stable reading.

If Mic B is reading lower voltages than A, this is very likely to be the source of the low output.

If you have a signal generator, then before desoldering the JFETs you can do a quick test by removing the capsule and injecting a signal into the amplifier input (the R1/C9/C2/C1 junction in the Neumann schemo) with the mic connected up to the preamp. 50mV 1KHz sine wave should be plenty. Compare the output level from Mic A with Mic B (without changing the preamp gain). If you get the same level in both cases it's unlikely the JFETs are at fault.

For a more accurate test, if possible, connect the signal generator output with a 50pF capacitor in series to the input. You'll get a lower output level than the previous tests, but if it's the same on Mic A and Mic B there's nothing wrong with the FETs or the transformer. If it's different, we've learned something

(If the Micparts circuit has different component labels to the Neumann one, can you post a picture of it?)

Oh interesting. Thanks for posting this and giving additional input. I was having a hard time finding the original Neumann schematic for some reason. Thank you for sending the link. The MP board looks like it uses the same labels for the schematic. I'm not sure about posting pictures of it since I don't have his permission to do so (though I haven't asked, but it does have a copyright on the board). I don't have a copy of his schematic ... and I have asked for that.

In terms of the test...

I would guess that the test I did yesterday would accomplish the same thing as injecting a 1kHz sine wave directly into the circuit, especially since swapping the capsules between the two mics didn't have any change in the result.

Essentially, I rubber-banded the two mics together and placed them 1ft away from a speaker where I was playing a 1kHz sine wave. Then, I set the preamps on my Apollo Twin X the exact same and made recordings of them in Logic Pro. Here's a video of me showing the result:

Test 1: Baseline
Test 2: Capsules are swapped
Test 3: Cables are swapped (notice the lower output follows Mic B)

I'll pop them open again and compare the different voltages at C6 and C10 and check the voltage at the R9-R10 junction as well.

 

[Voyager10]

The point of using the signal generator would be to use a AC voltage source which isn't affected by presence or absence of polarisation voltage (unlike the mic capsule, whose output level is directly proportional to the applied voltage).

So with the signal generator we would be comparing just the FET and transformer parts of the circuit between mics A and B.

 

[joulupukki]

I'm not sure if this would affect anything, but after opening them both up to compare again, I was just testing continuity between different parts and on Mic A, both of the attachment screws on the nose cone have continuity to the mic body ground. On Mic B, only the bottom set screw has continuity to ground. The top one does not. Is that significant at all?

 

[joulupukki]

Ok, measurements of C6 and C10 voltages are nearly identical:

Mic A:
C6: 21.06V
C10: 13.63V

Mic B:
C6: 21.05V
C10: 13.64V

The point of using the signal generator would be to use a AC voltage source which isn't affected by presence or absence of polarisation voltage (unlike the mic capsule, whose output level is directly proportional to the applied voltage).

So with the signal generator we would be comparing just the FET and transformer parts of the circuit between mics A and B.

Ah, roger that. Let me gear up to do that. From my signal generator (just an iPhone app sending out a 1kHz sine wave at about 33mVAC) do I need to connect the ground of that signal to the XLR ground of the mic?

 

[Voyager10]

Yes, XLR pin 1 or clip to the case should do nicely.

 

[joulupukki]

With a 11mVAC 1kHz sine wave signal injected into the mic, this is what I'm getting:

 

[Voyager10]

Is C3 soldered in correctly on mic B? It wouldn't affect any DC levels but would change AC gain.

 

[Voyager10]

If it's not C3 the next thing to try is connecting the signal generator across the transformer primary, comparing levels for mics A and B. No particular need to disconnect it from the rest of the circuit, an iPhone headphone output will drive it anyway.

(Will be offline for a while now, sorry...)

 

[joulupukki]

Is C3 soldered in correctly on mic B? It wouldn't affect any DC levels but would change AC gain.

As far as I know it's soldered in correctly. It's in the same orientation as Mic A so + and - terminals are correct, if that's what you mean.

I'm gonna trace the signal with my oscilloscope of the JFET's drain and compare the two readings to isolate the transformer out of the picture just to see if I can spot a difference.

The two pins of C3 are circled here. Those solder joints look good to me.

 

[joulupukki]

Ok, based on this, I think it's safe to rule out the transformer. Measuring the signal of Mic B vs Mic A at the JFET drain, there's already a large difference in output.

 

[kingkorg]

Ok, based on this, I think it's safe to rule out the transformer. Measuring the signal of Mic B vs Mic A at the JFET drain, there's already a large difference in output.
View attachment 119604
View attachment 119605

Have you checked the source capacitor? Maybe remove it on both and see if the difference is still present.

 

[joulupukki]

Have you checked the source capacitor? Maybe remove it on both and see if the difference is still present.

I assume you're referring to C1? In circuit, they measure the same: 0.49nF

 

[Voyager10]

Ah, you have a scope!

In that case, measuring the signal at the gate & source of the FET would be useful. Gate should have close to 11mV, same as the inout, and source should be basically DC.

 

[RuudNL]

It is unlikely that the source capacitor is 0.49 nF, I would expect 4.7µF or higher.

 

[kingkorg]

I assume you're referring to C1? In circuit, they measure the same: 0.49nF

Yeah like Ruud said, that doesn't sound right, but anyways you have to take them out to measure. They can still measure good but be bad caps. Just take them out and test the levels withou them. I don't have your schematic, it's the cap across source resistor. Now i see your image. No, not C1.

 

[joulupukki]

It is unlikely that the source capacitor is 0.49 nF, I would expect 4.7µF or higher.

I'm referring to C1 in this schematic. But perhaps I'm misunderstanding. What capacitor specifically should I be referring to?

Ah, you have a scope!

In that case, measuring the signal at the gate & source of the FET would be useful. Gate should have close to 11mV, same as the inout, and source should be basically DC.

Using the scope feeding in 25mVAC I'm seeing approx 20mVAC at the gate on both mics. Same near 0-3mvAC on the source on both mics.

 

[joulupukki]

Yeah like Ruud said, that doesn't sound right, but anyways you have to take them out to measure. They can still measure good but be bad caps. Just take them out and test the levels withou them. I don't have your schematic, it's the cap across source resistor. Now i see your image. No, not C1.

Ah, so C3 in the KM84 schematic? I was having trouble getting that cap out for some reason. Maybe my solder braid isn't that great. but it wasn't working for me. Maybe carefully using a solder sucker would work.

...and it's fine to power up the circuit with C3 removed?