MXL 603 -> KM84 Replacement PCB

[jp8]

I'm working on a charge amplifier circuit, but not identical to the KM84. Why I did not take a proven KM84 style circuit? Well, just because I want something different. I'm a hardware designer and there's little fun in it for me to just copy and build known, standard circuits. Whether it will be any better? Don't know yet. We'll see. Worst case, it might be a total failure .

I'm designing two versions. One version will be optimized for the T8 transformer with some "flavor" and the other should be a transparent, low-distortion version for the Lundahl transformer. Both should be less sensitive for load variations. Expect the T8 version to have distortion levels between KM84 and the Lundahl version. And I expect both to have lower noise and more headroom than KM84. Whether SDCs will benefit from the lower noise level is yet to be seen (don't think so), but maybe nice for LDCs? May take a while before I have finished the PCBs and tuned both versions for the right balance between noise, distortion and max output level.

Jan

 

[MicMaven]

... high resolution detailed photos of the Takstar CM-60,

I POSTED PHOTOS OF CM-60 circuit in a Media ALBUMn - some here, more there. Compressed jpg to save save space. I would link to the Media ALBUM here if I knew how. I hope someone finds them helpful. My road to Heck is well paved with all my good intentions. James

PS - Thanks to @GraemeWoller for instructing me on how to get into the dang mic body ... screw them in, not out he said ... um ... Bingo! JHR

 

[joulupukki]

Just got my two CM-60s in the mail yesterday. Looks like the nosecone is just wedged onto the board and only really held on by the wire soldered directly to a pad on the board. My PCBs haven’t yet arrived, but it’ll be interesting to see if I can still attach the XLR pins in a similar fashion for the PCB made for the MXL mics.

Incidentally, what’s gonna be the easiest way to get those XLR pins unsoldered without melting the black plastic mount? Just use a larger soldering iron tip on high heat and try to go quickly … maybe with a solder sucker and/or solder wick?

 

[klem]

For the XLR pins, I'd put some tiny alligator clips on both sides of the XLR pins. It's not fool-proof, but it can buy some additional seconds which can be the difference maker. If you can get your hands on a Hakko FR301 etc that also adds precious seconds.

 

[joulupukki]

I ordered a 10-pack of 2N3819 JFETs. Using the 2nd option listed on this page (Greatly Improved JFET Matcher) I tested them all out using a 9V battery. Assuming I did this right, it looks like there are 4 pairs with 3 of them being within 1% measurements. Generally, what's the common practice for choosing which ones to use? Use the one with the highest Idss & Vp? or the lowest? My inclination would be to use the 4-6 JFET pair.

 

[MicMaven]

Incidentally, what’s gonna be the easiest way to get those XLR pins unsoldered without melting the black plastic mount?

I would CUT and/or MELT the plastic piece that holds the XLR pins before de-soldering the pins separately from the board. Then, heat and remove each pin from where it is soldered to the board, lifting the pin off the board with a pair of needle nose pliers once the original solder melts, doing each one separately, allowing the board to cool between removing each pin, seriatim. They will come off easily and quickly once they are free of the original plastic XLR holder. By removing the plastic holder first, you minimize the time you must heat each solder pad. The only negative is you can no longer use the XLF plug, although you can replace it for low cost from a number of online sources.

This may not be as crazy at it first appears . . . James

 

[joulupukki]

I would CUT and/or MELT the plastic piece that holds the XLR pins before de-soldering the pins separately from the board. Then, heat and remove each pin from where it is soldered to the board, lifting the pin off the board with a pair of needle nose pliers once the original solder melts, doing each one separately, allowing the board to cool between removing each pin, seriatim. They will come off easily and quickly once they are free of the original plastic XLR holder. By removing the plastic holder first, you minimize the time you must heat each solder pad. The only negative is you can no longer use the XLF plug, although you can replace it for low cost from a number of online sources.

This may not be as crazy at it first appears . . . James

Yeah, in this case, I’m not that concerned about wrecking the board and would rather not have to order replacement XLR inserts. I guess ordering new inserts is a great backup plan if I can’t get the first ones off easily.

 

[jp8]

As an alternative to the alligator clips, you can also plug on a female XLR. When you remove the solder using solderwick, use a wide tip and solder wick to quickly remove the solder. You may not be able to remove all solder, especially the XLR pins that are flush with the PCB surface are likely to still have a solder bridge between pad and pin. But when the odds are with you, pin 2 will be free from the pad. If that is the case, you can heat one of the other pins and push the PCBA away from the pin. This might work for 1.0 or 1.2mm PCBs. CM-60 has 1.2mm PCB thickness. Good luck!

Replacements are not hard to find on Aliexpress. I have these:
https://a.aliexpress.com/_EHsyfp7
Better ones with ground tab can be found here:
https://a.aliexpress.com/_EGvpF5L
The best one I have ever seen is in the CM-63. It has a spring loaded contact between pin 1 and the mic body, which gives the shortest possible connection between shield, mic body and input ground on the PCB. See picture. This improves the RF suppression.

Jan

 

[jp8]

I ordered a 10-pack of 2N3819 JFETs. Using the 2nd option listed on this page (Greatly Improved JFET Matcher) I tested them all out using a 9V battery. Assuming I did this right, it looks like there are 4 pairs with 3 of them being within 1% measurements. Generally, what's the common practice for choosing which ones to use? Use the one with the highest Idss & Vp? or the lowest? My inclination would be to use the 4-6 JFET pair.
View attachment 121736

The answer you don't want to hear: there is no general guideline and you'll have to select the FET which best matches your circuit. Or design the circuit such that it best matches your FET. Perhaps this article will help? Or google for JFET biasing.

http://sound-au.com/articles/jfet-design.htm
Jan

 

[joulupukki]

The answer you don't want to hear: there is no general guideline and you'll have to select the FET which best matches your circuit. Or design the circuit such that it best matches your FET. Perhaps this article will help? Or google for JFET biasing.

http://sound-au.com/articles/jfet-design.htm
Jan

Wow. That article is intense! I think I’m gonna just grab one of my closer-matched pairs and just go for it and see how it turns out.

Incidentally, the CM-60 actually sounds pretty darn great comparing against my MP SDC-84 (on acoustic bluegrass instruments) and I’m wondering how much, if any “improvement” I’ll actually get over the CM-60 with this build. I think I’ll just start with one and see what’s different about it (once all the remaining parts arrive).

 

[joulupukki]

If I left off C2 (the 4pF cap) in the build, would that just remove a little bit of padding in the mic or does it serve a larger purpose than just that? I'm just wondering if I'm recording acoustic instruments (guitar, mandolin, banjo) maybe it'd have more sensitivity with it not there?

 

[Chrisfromthepast]

This explanation of the km84 circuit may help. The feedback capacitor c2 is important to this topology. You can change the value to fiddle with the gain, but taking it out entirely will cause your output to be much lower than the expected output level.
Hopefully a real designer can chime in, as Im just a sound guy with a soldering iron.

 

[joulupukki]

This explanation of the km84 circuit may help. The feedback capacitor c2 is important to this topology. You can change the value to fiddle with the gain, but taking it out entirely will cause your output to be much lower than the expected output level.
Hopefully a real designer can chime in, as Im just a sound guy with a soldering iron.

Oh cool, thanks for that link. I was just curious and I'm sure it'll be fine with it present.

 

[jp8]

This explanation of the km84 circuit may help. The feedback capacitor c2 is important to this topology. You can change the value to fiddle with the gain, but taking it out entirely will cause your output to be much lower than the expected output level.
Hopefully a real designer can chime in, as Im just a sound guy with a soldering iron.

I'm afraid your statement is incorrect. The gain and output signal will be much higher without the feedback capacitor. With 4pF, the gain is ~7x. Without feedback capacitor ~25x. And you'll get a lot more 2nd harmonic distortion... An LTspice simulation with 70mV RMS input revealed that the 2nd harmonic jumped from -65dB with capacitor to -43dB without capacitor. A ~3.6 times gain increase, accompanied by a ~12.5 times increase in distortion. Without the capacitor, you're basically using the JFET without as an open-loop amplifier. Only acceptable for small signals in e.g. tiny electret capsules.

Jan

 

[jp8]

I'd like to try several transformer based circuits in the CM-60 an CM-63 mics. But there's only 26mm between the low cut switch and XLR terminals in the CM-60. See picture. With a length of 25mm for the transformers, it may only be just possible to shoehorn the trannie in between the switch and the XLR. I think I'd better start with the CM-63 PCBs.

I'm still working on my FR measurement setup. I do not yet fully trust its results, so I'm not confident enough to share FR measurements them right now. Will share other measurement data later next week.

Jan

So here are finally some measurement data and observations on the CM-60 and CM-63 electronics. The attached files contain all the data, remarks and observations. The yellow-highlighted items could be things you (or Takstar) would typically want to modify (at least, IMHO). E.g. the CM-60 has low headroom due to a low 7.2 V supply voltage powering the JFET stage. And the CM-63 has the wrong JFET or wrong value source/drain resistors for optimum bias. Not very well engineered. On the other hand, the RF suppression of the CM-63 is exemplary.

Unfortunately, I can still not do any reliable FR measurements. I'll share those once they're available. And also DFD measurements and measurements on the residuals of the bias generator signal on the outputs, which can cause IMD. I hope this information can help you make informed choices about what modifications you would like to make to these microphones.

If you'd like to see other measurements being done, or you have any questions, just let me know.

Jan

 

[kingkorg]

So here are finally some measurement data and observations on the CM-60 and CM-63 electronics. The attached files contain all the data, remarks and observations. The yellow-highlighted items could be things you (or Takstar) would typically want to modify (at least, IMHO). E.g. the CM-60 has low headroom due to a low 7.2 V supply voltage powering the JFET stage. And the CM-63 has the wrong JFET or wrong value source/drain resistors for optimum bias. Not very well engineered. On the other hand, the RF suppression of the CM-63 is exemplary.

Unfortunately, I can still not do any reliable FR measurements. I'll share those once they're available. And also DFD measurements and measurements on the residuals of the bias generator signal on the outputs, which can cause IMD. I hope this information can help you make informed choices about what modifications you would like to make to these microphones.

If you'd like to see other measurements being done, or you have any questions, just let me know.

Jan

How did you get the equivalent spl/out V figures for the THD? Or in other words, how do you know 128mV output equals 110 dB SPL for CM60?

 

[jp8]

I placed my UMIK-1 mic 20 cm from the tweeter in my FR measurement setup and adjusted the level of a 1 kHz sine wave until I measured 94.0 dB SPL in REW. I then replaced the UMIK-1 by the CM-60 (or any other DUT for that matter). I use a fixture to align the front face of the microphone with the centre of the speaker and to get exactly the same distance from the tweeter as the UMIK-1. I then read the output voltage of the mic from the RMS Voltmeter in my Audio Analyzer. This is the value listed in the first line of the table. Admittedly, the 19.3 mV is quite far off from the spec (12.6 mV). In most cases, what I measure is quite close to the specified value. E.g. the CM-63 was quite close, actually. With "close", I mean within +/-2 dB, which AFAIK is a normal tolerance of (budget?) microphones. This makes me believe that my method is OK. Or at least acceptable for DIY measurements.

At 0.1% THD, the output voltage measured 128 mV, which is 20*LOG(128/19.3) = 16.43 dB above 94 dB SPL. This amounts to 110.43 dB, which I rounded to 110 dB.

If you think my measurements or calculations are flawed, please let me know. I know you are far, far more experienced in measuring microphones than I am.

Jan

 

[MicMaven]

So here are finally some measurement data and observations on the CM-60 and CM-63 electronics. The attached files contain all the data, remarks and observations.

Quick Sidebar Question:

What do the numbers tell the average consumer using either of these two microphones - First, vis-a-vis each other, and Second, vis-a-vis the competition, such as models, such as AKG P170, MXL-603/604, Behringer C-2, Behringer C-5,Golden Age Project FC4, et. al.?

I am NOT looking for an argument, or hijacking the thread - just wondering if there is a helpful takeaway comparing these to each other and similar low cost models. For example, would one prefer the Takstar CM-60 or CM-63, or some other model at this price point, and which specification turns the key? Thanks, James.

 

[jp8]

Hi James,

I consider your questions relevant, so no problems with that.

The relevance of the measurements and observations depends completely on what you consider to be important features of a microphone. As this is all about DIY, I assume that many forum members' aim is to improve or just alter (budget) microphones and build something to their taste. In my opinion, any information that could get you on the way to making an informed decision on which microphone to buy and what are the weak spots that are worth improving. Being an electronic hardware engineer, you can expect me to have a focus on the electronics of the mics. Of course, the acoustic properties and qualities are by far the most important ones to consider when buying a microphone, but I'm afraid I'm about the least experienced person on this forum to judge, let alone give advice. Though I have some listening experience with HiFi and PA speakers, my ears are a perfect 9.5 kHz LPF. So this limits my ability to judge the top octaves of microphones. I also don't play an instrument, nor do I do any recording/sound engineering. When subjectively reviewing microphones, I have to rely for 80% on my son's ears. He is a live sound engineer, plays some guitar and is able to hear the tiniest distortions, which my hearing system leaves unnoticed.

Back to the measurements and some examples of what I think can be taken from them:

• If the mic noise level is several dBs above the noise level of the electronics, the capsule self-noise is dominant. Then it makes little sense to aim for ultra-low-noise electronics.
• If you experience frequent RF disturbance issues, you may want to have a microphone (stock or self-built) with an RF filter.
• If you're living in an environment with constant high RH, you'll want well-protected High-Z circuits.
• etc. etc.

I hope this answers your questions. BTW, I plan on doing the same measurements and more on an SC140, B-5, MV692, P170, MXL991 Alctron T02A and maybe some more SDCs later. Will publish those on my website I'm still working on. Long time, slow progress project.... LDCs will come later, maybe when I'm retired .

Jan

 

[joulupukki]

Still waiting for some of Graeme's PCBs to get on with my builds (they seem to be stuck forever in LAX in customs or something). All of my other parts have arrived (3U cardioid capsules & transformers, CM-60 donor bodies, and small parts from DigiKey and Mouser).

In the meantime I've discovered the CM-60 seems to be a pretty nice mic in comparison ... at least for recording acoustic bluegrass stuff I'm doing. My goal is potentially to build the KM84 diy project into the CM-60 donor bodies and get as close as or better than the MP SDC-84 (and if that happens, maybe sell off the MP pair to recoup some $).

So, I know this isn't really precision testing, but just to get an idea of how well matched these capsules are and what the differences are, I set up the following test...

In Logic Pro I created a track with the bulit-in Test Oscillator plugin, set it to pink noise, and then had it play through my studio monitor speakers (JBL 305P Mk II). I then put the capsule about 1" away from the center of the tweeter of one of the speakers and recorded a few seconds of the pink noise. Then, in post, I adjusted the gain of each track so that they are peaking at roughly the same amount. I then added the stock EQ plugin on the Stereo Output track just to observe the analyzer readouts. Panning one capsule hard left and the other hard right I was able to compare them and see where certain peaks are. The ligher line is the Left channel and the darker line is the Right channel.

Again, I dunno how accurate this is, but it seems like:

1. The MP SDC-84 stereo pair seems nicely matched
2. The MP SDC cardioid capsules are matched really well
3. The 3U cardioid capsules are matched really well
4. Surprisingly, the CM-60 capsules (if used on the same mic circuit) seem matched
5. The capsule has more effect on the frequency response than the mic circuit. See how the same capsule in both mic circuits is nearly identical?
6. This small sample of MP and 3U capsules differ a bit in high-end response
7. The CM-60 circuit seems to have more low-end response than the MP SDC-84 and the CM-60 capsules more low end than the MP SDC and 3U capsules (that's either good or bad depending on your preference)