KM84 DIY Body & PCB kit – CLOSED

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.
I have a question that I suspect will out me as the mic novice that I am. 😅 I haven't yet biased the FETs in my pair, but I have watched through the video suggested in the guide. I do not have bench testing equipment -- I typically use my interface and DAW for tone generation / meters / etc. It seems simple enough to do that here as well, except that I'm wondering if a line out from my interface would cause problems in the measurements due to mismatched impedance... I know that mic output impedance is quite low, but I don't know what injecting a line level signal in would do. Or perhaps I'm overthinking it.

Also, I've got a TS-to-alligator clip cable. Would this work here, assuming that the above is not a concern? Or would I need to figure out something with a TRS connection?

Also, still no word from the powder coating shop. I'm too impatient. :LOL:
I don't think you need to worry about that. The Capsule portion of the circuit is a high impedance zone so the difference between output impedances is going to be negligible. Just put a capacitor in between your line-out and the microphone to prevent damage to your interface. There is a post somewhere (on this thread or maybe the build thread) which goes into a little more detail on this.

TS alligator clip is fine, you don't need a balanced signal to the mic. Just keep the runs short.

Also, anyone correct me if I am wrong (I am a relative novice particularly with microphone circuits).
 
Thanks! I’ll go through this thread and/or search the forum for info on putting the cap between the line out and capsule in. Really appreciate the info!
 
It gets mentioned a few times (e.g. post #294), but it’s basically putting a 1000pf capacitor between your signal generator and the mic to stop the 48V phantom power that you’re using to power the mic from continuing on to your signal generator / interface.
 
I finally got my mic bodies back from the powder coating place (they had them for over a month, and when I finally got someone on the phone they said, "oh yeah, those have been done for weeks" lol). They look great. Anyhow, I decided I should finally get around to the calibration.

I ran into a couple snags. The first was similar to what has been described previously with the trimmer. However...

Before I even worry about that, though, I'm trying to figure out the other issue -- the output is crazy low. With +48v engaged and the preamp on my Clarett interface turned up to 3:00 o'clock, I just barely get any sound (in the DAW it's reading around -55dBfs). This is happening on both mics to more or less the same degree, so I'm inclined to assume that I made a build error somewhere. It's a simple enough circuit, so obviously I'll go back and double check component values and orientations, but are there any other explanations for why I could be getting such a wimpy output?
 
Bias the FET and then test it. With the original trimmer turned all the way down, the output will be low and distorted.
 
Man, I think I'm really missing something in the calibration procedure. I'm sure it's simple, but I'll be honest, it kind of seems crazy to me to have this guide that spends pages and pages telling you how to follow the paint-by-numbers silkscreen and solder resistors and take coffee breaks (cute as that is) but then absolutely jumps ship at the calibration procedure. Just a thought for future versions of the project, maybe the guide could spend more (or any) time describing that process. I didn't find the video that is linked to be very helpful either, since in addition to being overly long for this purpose, he's injecting the signal at a different spot (capsule wire) than is described earlier in this thread (according to post #294, FET gate, other side of C1) -- I don't know if that makes any difference here, but it just has me feeling confused and uncertain. This is all probably just my own problem, since I see many others have finished the process successfully.

I'm only venting a little, that's all. I'm just frustrated. It feels like I'm missing something obvious.

I don't have bench top equipment like an oscilloscope, so I'm trying to do this using my DAW and interface. I put the 1000pF cap between my line out and the mic to prevent the +48v from traveling to my interface. However, when I inject signal (assuming I'm doing it right) at the FET gate, it comes through looking very noisy (in addition to harmonic distortion, there appears to be lots of other frequency content both above and below the sine wave fundamental), and adjusting the trimmer only does so much. When I inject signal where the capsule wire and the board meet as shown in the video, I get no signal whatsoever coming back into my DAW. I've double checked all of my component values and orientations, and my soldering is good and I cleaned the board well, so I don't know what I'm doing wrong (or perhaps more accurately, I don't know what I'm doing).
 
Okay, little update:

I took a little breath and then went through checking my dc voltages both mics. I'm getting some readings that are not in line with the schematic, and I can't explain them.

I get +46v as expected between R9 and R10

On the part that runs through R6 / R5 / R4, I get roughly each of the expected voltages +/- 2v until I hit the junction of R4 and C2, where instead of +10v I'm getting +1v on one mic, +4ish on the other. Now you say, "well dummy, you've got the wrong resistor value in there." Except I've checked with my meter, it's 47k, I bought the resistor in the BOM, the code printed on it is 4702F. So on one side of that resistor, I have the expected +21.5v (well, it's actually more like +20 or so), and on the other side, +1.

Now, on the part that runs through R8 and R7, I get a big drop, but I'm struggling to understand why. At the junction of R8/R7, I get approximately +23v, and between R7/R1 I get roughly +15.

I'm guessing these voltages are symptoms of whatever I've done wrong, but I'm really struggling to spot it. I'll keep looking, but if anyone has any clues based on the above, it would be appreciated!

EDIT to add that there's a small inconsistency in the docs. The JFET that I have is a 2n3819, which I believe is correct, but the note on the schematic suggests a 2n3891. That must be a typo since that appears to be a rectifier of some sort, but at this point I'm just being as meticulous as possible.

EDIT again: pressing the capsule in solves the 10v problem (gets me approx +9, so close enough). Was I supposed to be doing that the whole time? It does nothing to solve the discrepancy at the junction of R7/R1.
 
Last edited:
The FET number is a typo. Here is some instructions that I hope will help you.
1) Bias the FET with the capsule off. It should not be in the circuit during the bias setup.
2) Make all connections with power removed from everything to avoid blowing the FET
3) Connect an XLR cable to the mic (from your preamp that has phantom) - don't turn it on yet
4) Take a 1000pf capacitor and connect one side to a signal generator and the other side to the tip when the capsule connects. I used an alligator clip. Make sure the cable is either very short or shielded so you don't get excessive hum. You will also need to connect the shield/ground from your signal generator to pin 1 of the XLR.
5) Set the signal generator to put out a 1KHZ sine wave (keep the volume all the way down for now)
6) Now that everything is connected, turn on the preamp and make sure you have >40vdc on one side of R6 and about 24V on the other side.
7) Connect your meter to the drain of the FET. (easier to clip onto R4) and check the voltage. When the FET is properly biased, that voltage will be somewhere around 10V. If it's really low, start spinning the trimmer until you get it up to 10V. This is just to get in the ballpark as the actual voltage may differ somewhat when done.
8) Now, turn on the signal generator and watch the waveform on the output of the mic (in your daw or a scope). What you want is for the top and bottom of the sine wave to start clipping symmetrically. Keep raising the signal until it clips and then adjust the trimmer. Just go back and forth a few times and you'll get it. Once you find the trimmer setting for symmetrical clipping, you are done. Shut it all off and finish putting the mic together.

NOTE: Due to the change in what FET was sent with this second batch of mics (different brand), you may not get the FET biased at its best with just the 20K trimmer. In my case, one mic needed 18.5K and the other needed 21.5k to bias properly. So, don't be alarmed if the trimmer is spun all the way to get the bias correct. Once you get the BIAS set, you can replace the trimmer with a resistor or just leave the trimmer in there. Removing that trimmer is a PITA. Tight spacing and one leg is attached to a ground. Tough to het it enough to remove without damaging a trace. TIP: Though not totally the best solution, making sure you are around 10V at the drain of the FET is going to get you very close to biased. The closer to the symmetrical clipping you can get, the higher signal levels you will be able to get without distortion. 10V is the starting point. You tweak it from there by watching the waveform at the output.

In my case, (once I realized I was never going to get the right bias with only the trimmer) I removed the trimmer and put a 4.7k resistor in series with it. Then I tack soldered some legs on it to the board. Once bias was set, I removed the trimmer/resistor and measured it. I then soldered in a resistor of that value. As long as the FET doesn't get replaced, the bias should be set for life. The trimmer is no longer needed.

I hope this is of some help.
 
Last edited:
Thanks for that very thorough reply! While this feels like exactly what I've been doing, I am going to try it again following step-by-step what you wrote out above and see if perhaps I was simply missing some little detail.

I do wonder if my FETs are shot, but I'm trying not to let myself create problems where there may be none. I just can't figure out why I struggle so much to get signal to pass. Anyway, I'll report back (hopefully with good news) once I've spent some more time with it and tried out your procedure above.
 
I hope this is of some help.
This was incredibly helpful, it turns out! I still don't fully comprehend what I was doing differently, but following your description made everything click.

The trimmers definitely weren't high enough resistance. On both mics I had to max them out, and I still couldn't quite get a full 10v (somewhere in the 9.5 area) or symmetrical clipping. I may try to replace them with trimmers of a higher value. Having said that, I'll be using these to record some acoustic guitar in a few weeks with friends, and they sound good enough even as they are that I think I will wait until after to do any further messing around.

One of the two mics is a bit noisy, with a considerable 60 cycle hum, so I'm assuming there's a grounding issue somewhere. I ran out of time today, but I'll try to hunt that down when I get a chance. I have a suspicion that the powdercoating on the body is interfering with a good connection and as such is creating the issue, but we'll see.
 
Make sure to do an ohm check from pin 1 of the xlr to the body of the mic. it should be under an ohm. if the value is quite a bit higher, that could explain the hum. If you get static noise, I'd reclean near the high impedance sections at the top of the board. Any residual flux is going to allow some current to flow between points it shouldn't.
 
No static noise. I clean my boards pretty carefully, so I'm good there.

This is definitely a ground connection thing. I get more like 2 ohms from pin 1 to the base of the xlr's shell (I don't know what you call it...the metal portion that encases the xlr pins). What's odd is that this does not appear to be due to the solder joints (which I would have expected) -- instead, this is the resistance between the large grub screw and the tab that it is immediately connected to. Anyway, going to touch up a few solder joints in the ground plane where my resistance seems slightly elevated (1.5ohms or so), and scrape off a bit of the powdercoating on the interior of the body where it touches the xlr connector to try to get better continuity, and see how that goes.
 
the case is a weak point in the long term viability of these mics. The ground for the capsule is thru several weak connections, from the grub screw to the base shell, from there thru the three small screws to the brass shell (that could become problematic in the long term) and then to the screw on capsule. The body being brass will tarnish and brass tarnish is non conductive.
The KM84 is similar in some ways, but is nickel plated on these connections and use grub screws that press out on the case instead of going thru it for a tenuous connection. Even these can require cleaning over time.

Jman, you had the cases powder coated, have you checked the threads to the capsule and the inside surface of the lower holes for conductivity and to be free of coating?

Cheers
Alan
 
the case is a weak point in the long term viability of these mics. The ground for the capsule is thru several weak connections, from the grub screw to the base shell, from there thru the three small screws to the brass shell (that could become problematic in the long term) and then to the screw on capsule. The body being brass will tarnish and brass tarnish is non conductive.
The KM84 is similar in some ways, but is nickel plated on these connections and use grub screws that press out on the case instead of going thru it for a tenuous connection. Even these can require cleaning over time.

Jman, you had the cases powder coated, have you checked the threads to the capsule and the inside surface of the lower holes for conductivity and to be free of coating?

Cheers
Alan

Grounding of the body isn't weak by any means, it's made by direct mechanical connections.

– Purpose of the grub screws is only to hold the body tube in place. Your mic should have a solid ground connection from Pin 1 to the capsule head without the grub screws installed test by pushing the mic together). Make sure the inside of the body tube and bottom edge where it sits on the connector housing aren't covered with paint.

– Provided grub screws are stainless steel. Won't corrode, tarnish or strip.

– How you finish the body tube is your choice. You can polish, paint or plate. Professionals offer these services for small quanities at low cost. You can do all three for between €10 to €30/body.

– More importantly, no mic is designed to be continuously taken apart. During testing use a single regular head M2 screw to hold the body on the connector housing. Install the grub screws only once you're happy with the sound and you're ready to leave the mic alone. Or you can use a single regular M2 screw for the entire lifetime of the mic. Makes no difference.
 
Last edited:
Back
Top