Banzai's KM84 DIY Body & PCB kit build thread

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.
Adjust C2 to set the sensitivity you want. You can replace it with any value you prefer.

Circuit and transformer decide what makes it from the capsule to your preamp, not the other way round.


Not the same capsules. These don't need taming, but... chopping up metal for science is always a good thing (y)

If anyone ends up with an oddball zingy capsule, get in touch and we can work out a replacement.
Read my posts above - suggesting C2 as the culprit for output discrepancies is incorrect and would be a band-aid solution. The gain of all of my preamp/transformer assemblies is the same +/- .3dB
The issue is the capsule and the varying capacitances of the capsule, i.e. poor quality control.

I'm not certain they are different capsules - take a look at the photos I took of the capsule and those linked here of the MXL 991/603. I'm not seeing any differences. Maybe better tolerances on the Maiku ones, or perhaps different diaphragm thickness. Based off the photos, the Maiku capsule is more of a MXL clone than a Neumann clone :ROFLMAO: :ROFLMAO: :ROFLMAO:

As far as zingyness goes. Well, I have 7 capsules that are very zingy in comparison to my KM86. Care to help?
 
I mentioned that my capsule fell apart, you blew it off and then recommended that I use an entirely inappropriate glue to fix it. So which is it, you will support capsule problems or no returns.
Where did I dismiss your problem? I'm sorry if it came across that way.

You didn't contact me asking for a replacement. You asked here which glue you could use to reattach the mesh grill to the top of the capsule.

Because the capsule disassembles from the rear, there is no problem in using a permanent glue in the front. Superglue can also be dissolved in acetone. If you use a light layer as suggested, you won't even need to dissolve it. It will come off with pressure alone.

If you'd rather we replace the capsule, please send an email so we can work out a solution.

(to clarify this further, returns and replacements aren't the same thing: of course we replace faulty parts)
 
Last edited:
Read my posts above - suggesting C2 as the culprit for output discrepancies is incorrect and would be a band-aid solution. The gain of all of my preamp/transformer assemblies is the same +/- .3dB
The issue is the capsule and the varying capacitances of the capsule, i.e. poor quality control.
Not suggesting it as the culprit. C2 serves a purpose. You can use it to adjust the sensitivity for each one of your builds.

I'm not certain they are different capsules - take a look at the photos I took of the capsule and those linked here of the MXL 991/603. I'm not seeing any differences. Maybe better tolerances on the Maiku ones, or perhaps different diaphragm thickness. Based off the photos, the Maiku capsule is more of a MXL clone than a Neumann clone :ROFLMAO: :ROFLMAO: :ROFLMAO:
Backplates and bodies are from China. Never been a secret.

As far as zingyness goes. Well, I have 7 capsules that are very zingy in comparison to my KM86. Care to help?
Did you buy your kits from me through the white-market?

Your kits are from the first batch in 2018.

I am sorry our kit doesn't compare well to your $3000 microphone. That's a genuine apology, since I don't think this is something we can resolve by replacements. If you have the same problem across all 7 of your capsules, it's unlikely to make any difference.

(PS: if you feel the capsules of this kit are zingy, then you've probably never heard an MXL 603 capsule)
 
Last edited:
Read between the lines: he never built the kits, let alone compare them or build an adapter.
If you disagree, feel free to explain how you build kits without PCB's:

Don't talk about things you know nothing about Banzai!
I built the microphones with my own (transformerless) circuit.
Indeed, when someone was looking for a PCB I sold my copies because I would never use them.
The first listening impression was quite good, but when I had the opportunity to compare the microphones with a KM84, a clear difference could be heard.
With the K84 capsule on my own electronics, the microphone sounded like a KM84 again.
Which makes it clear that the microphone capsule makes the difference!
 
You are totally correct Banzai,
you never called this project "Banzai KM84".

You called it "Neumann KM84", although there's no one that seem to have been able to achieve a microphone with your kit that sounds even similar to a Neumann KM84.

So what name do you think it's the most correct "Neumann KM84" or "Banzai KM84"?

My first Post will be for you , i diy for years now, this kit is in the exact way of diy mean, grouping to buy base componants at low price , and enhance it at you own sauce !

who in this world would build a real 84 for 140 bucks? if you wanna a real one, buy it!

i personally own a pair of KM84 and this kit rocks, i did a real diy on it, using my own conponents (can't stand wima etc..) adjusted C2 and C3 to the excellent maiku capsule , to my taste and now it 's part of my recording SDC , even bettah than one of my km84 , my taste!
this is all about Diy building, builder's specs and addapting to your taste

thx Banzai for your project , cheap and quick sending!
 
I did a bunch of experimenting today with tightening and loosening the retaining ring in the capsule. As expected I could easily change the capacitance of the capsule and this had direct correlation to the sensitivity of the capsule. I tested the capsule with capacitance between 30pf-38pf. While is did alter the frequency response independently of sensitivity, it wasn't much, maybe 1-2dB deviations in small areas. See graphs. I also experimented with removing the front plastic grill. I did FR measurements as well as an acoustic guitar test and I didn't notice too much of a difference. I've considered replacing the grill anyway though because I think it makes the capsule look cheap. Any ideas on where to get some very fine mesh like the stuff that covers the vent slots?
I took photos of the capsule parts if you're curious too.

I think I would like to incorporate a de-emphasis circuit like the one in the U87 to tame the high end harshness of this capsule. If that is feasible, could I also incorporate the low cut in the U87 circuit too? Would the de-emphasis be as simple as connecting the drain of the FET to the backplate by way of a 1M resistor bypassed by a 220pf. Would I need that C9 0.47uF cap from drain to the 1M resistor?
You can get all sorts of mesh and from different materials at ebay.
 
Just thought I'd mention that I noticed what I thought were drastic differences between capsules (I have 3) but I notice that taping two mics together, to try to align capsules for recording a single guitar, one is always in a better position. I found it hard to get around this as that one inch or so difference matters. Anyway, I hand held them and sang into them back and forth, and concluded I have two reasonably similar sounding capsules, although one has a little bit more low end. The other capsule is def. louder than the two.
 
Just thought I'd mention that I noticed what I thought were drastic differences between capsules (I have 3) but I notice that taping two mics together, to try to align capsules for recording a single guitar, one is always in a better position. I found it hard to get around this as that one inch or so difference matters. Anyway, I hand held them and sang into them back and forth, and concluded I have two reasonably similar sounding capsules, although one has a little bit more low end. The other capsule is def. louder than the two.
Yes, I think the subjective tests are better done from farther away where that one inch doesn't matter as much. If you have a capacitance meter, you could measure the capsule that is louder - I would guess it's in the order of 38pf+ where the other ones might be closer to 30pf. You could easily end up with a 2dB difference.
 
I am, please let us hear how they behave if you want to share, thanks. Let's leave all this grievance for once already :D
I apologize for the lack or science in this post. I hate being another anecdotal plebian on this thread.

After a long session with absolutely no time for experimentation I can say that the KM84 clones were excellent. I can tell you for certain that they are ALMOST flat.

They aren't as flat as a real KM84 and it's a disappointment to see the internal of the Maiku capsule isn't even trying to clone the KK84, BUT, the mic performed great in the context of cymbals, with a mostly flat response above 300hz. I cut everything below 300hz. Top end seems to have a bit more going on in the 6-8k range, but it's gentle and I'm not too worried about it to be honest.

I don't have daily access to KM84's and these are much flatter and sweeter sounding than the other SDC I have access to so I'll be reaching for them quite often.

Definitely interested in a future KK84 clone that I could just pop into this body. But for the price I paid for this project I'm not too woried about it.
 
Question. Is the 4pf cap (or whatever you put) to be adjusted according to individual capsule capacitance to obtain a golden 30 something pF? What is this golden capsule capacitance? And I'm in no danger of testing the capsule with my ratshack capacitance meter?
 
Question. Is the 4pf cap (or whatever you put) to be adjusted according to individual capsule capacitance to obtain a golden 30 something pF? What is this golden capsule capacitance?
No, the 4pF cap (C2) forms the gain ratio in conjunction with the capacitance of the capsule. That's why the original Neumann's add a 15pF cap in parallel with C2 to bring the gain down 10dB or so. Put simply: 36pF/4pF=9. 9 divided turns ratio of the transformer (7 for the Cinemag) drops the voltage back down for closer to gain of 1. Take a look at charge pumps for why it works: Charge pump - Wikipedia

The original KM84 schematic lists the cardioid KK84 at 36pF, strangely the KM86 KK84 is listed at 33pF. So anywhere in this range seems reasonable. The cap value you place at C2 determines the sensitivity of the capsule/JFET combo.

And I'm in no danger of testing the capsule with my ratshack capacitance meter?
No danger - the capsule is usually seeing 46V across it so whatever your meter puts out is likely much lower.
 
Last edited:
One of my mesh/screens came off one of my capsules as well. Unfortunately it's the best sounding capsule so I don't really want to send it back. Is there a trick to applying glue that isn't going to get it on the diaphram? Not much room to work with.
 
I was curious about the very high R3 value for the Central 2N3819 FETS. Mine had to be adjusted for 18.5K for one and 21.5K for the other. That made me wonder if there wouldn't be enough drive current to drive the transformer. They obviously worked, as I posted sound clips comparing it to a real KM84. Both mics worked fine but, heck, this is DIY so I decide to try another FET in this kit. From my research, many said a Linear Systems LS846 would be a good low noise replacement. I put it in and biased it. The new value of R3 is 3.9K. The voltage at the drain is around 8.5V. That was what it came out to be once I adjusted the trimmer for the highest output with the lowest distortion (symmetrical clipping). The bias resistor change is pretty huge. Speaking of biasing the FET, it kind of makes me wonder about some people who have had distortion and low output issues. That would lead me to believe they didn't bias the FET correctly. Anyway, I also replaced C4 with a 1uF film cap. I couldn't find any Neuman schematics that had C4 as 4.7uF, so I made it the same as the schematic. I left C3 alone (4.7uF) for now. I've seen Neumann schematics with C3 being anything from 1.5uF - 4.7uF. I will test these two mics out and report back. One of them, I just modified. The other has the original cap values and the Central FET.
 
Last edited:
Ok, I finished messing around and can give a few findings. In my tests, I used a KM84 and two of the kits. One of the kits I lets totally stock. Both the kits have the 3U transformers. I changed the values of C3 and C4 in one of the kits. I also changed out the FET from a Central 2N3819 to the LS846.

Capacitor changes: Honestly, the changing of the caps from 4.7uF to C3=1.5uF and C4=1uF didn't make much difference. I thought I would hear some big changes but it really didn't do much. I think the original value of 4.7uF is fine for both. Lowering them did not reduce the low end in any noticeable way. possibly the cap change accounts for the 45HZ roll-off I experienced. Not much of an issue unless recording a bass.

FET change: he tone didn't change at all with the FET change but I was able to get another dB of clean headroom from the LS846. The bias resistor value dropped to 3.3K-3.9K on my two kits. The Central FETs had to use a R3 value near 20K. Big bias difference.

Capsules: My capsules have about a 2.5-3dB difference in output from each other. The tonality is nearly identical though. The KM84 is 8dB softer than one of my kits and 6dB softer than the other kit.

Matching EQ tests: I used Ozone 9 matching EQ to see what the real difference is. Comparing one kit to the KM84, the Maiku capsule drops off around 45HZ. The KM84 remains somewhat flat. I don't consider that to be an issue. Next, I see the Maiku as flat from about 45HZ-700HZ. From 700HZ - 8KHZ it is down 1-2 dB. 8KZ appears to be flat again. Starting at 10KHZ, there is more of a decline in upper frequencies. Compared to the KM84, it's down about 3dB from 10KHZ-20KHZ. Because of this (to my ears) the KM84 sounds open and clear and the Maiku sounds darker and smoother. The Maiku doesn't sound muffled so don't think that. It just doesn't have the presence boost I always hear in a KM84. Doing a quick EQ, I could get them to sound nearly identical with a shelf EQ at 10KHZ 2-3dB and a 2dB bump at 4KHZ. They were so close at that point that it wouldn't matter which I used.

Off-axis sound: This is probably the biggest difference. The KM84 has such a wide polar pattern that I think it is nearly an omni mic. Tonality of a KM84 doesn't change very much no matter where you aim it. That's why most people love it. The Maiku acts as any other cardioid mic does. Off axis, the high end starts to go away. It can feel quite boomy at 90 degrees. I don't tend to mic like that so it's not very important for close micing.

What does this tell me? It was a fun project and worth doing. On a harsh or thin source, I think I would prefer the kit. On acoustic guitar I much preferred the KM84. Is one better than the other? Not really. They both have uses. For instance, I don't much care for a KM84 on violin. It picks up too much of the abrasiveness of bow scrape. The kit may work much better in that area and I will be trying it on the next violin session. Lately I've used a MC930 on violin and like it a lot. The KM84, though thought to be flat, has an upper mid bump to my ears which make sources poke through a mix. That can be good or bad depending on the source. Each mic has it's uses and I think the kit is worth making if just to have a different flavor.
 
Ok, I finished messing around and can give a few findings. In my tests, I used a KM84 and two of the kits. One of the kits I lets totally stock. Both the kits have the 3U transformers. I changed the values of C3 and C4 in one of the kits. I also changed out the FET from a Central 2N3819 to the LS846.

Capacitor changes: Honestly, the changing of the caps from 4.7uF to C3=1.5uF and C4=1uF didn't make much difference. I thought I would hear some big changes but it really didn't do much. I think the original value of 4.7uF is fine for both. Lowering them did not reduce the low end in any noticeable way. possibly the cap change accounts for the 45HZ roll-off I experienced. Not much of an issue unless recording a bass.

FET change: he tone didn't change at all with the FET change but I was able to get another dB of clean headroom from the LS846. The bias resistor value dropped to 3.3K-3.9K on my two kits. The Central FETs had to use a R3 value near 20K. Big bias difference.

Capsules: My capsules have about a 2.5-3dB difference in output from each other. The tonality is nearly identical though. The KM84 is 8dB softer than one of my kits and 6dB softer than the other kit.

Matching EQ tests: I used Ozone 9 matching EQ to see what the real difference is. Comparing one kit to the KM84, the Maiku capsule drops off around 45HZ. The KM84 remains somewhat flat. I don't consider that to be an issue. Next, I see the Maiku as flat from about 45HZ-700HZ. From 700HZ - 8KHZ it is down 1-2 dB. 8KZ appears to be flat again. Starting at 10KHZ, there is more of a decline in upper frequencies. Compared to the KM84, it's down about 3dB from 10KHZ-20KHZ. Because of this (to my ears) the KM84 sounds open and clear and the Maiku sounds darker and smoother. The Maiku doesn't sound muffled so don't think that. It just doesn't have the presence boost I always hear in a KM84. Doing a quick EQ, I could get them to sound nearly identical with a shelf EQ at 10KHZ 2-3dB and a 2dB bump at 4KHZ. They were so close at that point that it wouldn't matter which I used.

Off-axis sound: This is probably the biggest difference. The KM84 has such a wide polar pattern that I think it is nearly an omni mic. Tonality of a KM84 doesn't change very much no matter where you aim it. That's why most people love it. The Maiku acts as any other cardioid mic does. Off axis, the high end starts to go away. It can feel quite boomy at 90 degrees. I don't tend to mic like that so it's not very important for close micing.

What does this tell me? It was a fun project and worth doing. On a harsh or thin source, I think I would prefer the kit. On acoustic guitar I much preferred the KM84. Is one better than the other? Not really. They both have uses. For instance, I don't much care for a KM84 on violin. It picks up too much of the abrasiveness of bow scrape. The kit may work much better in that area and I will be trying it on the next violin session. Lately I've used a MC930 on violin and like it a lot. The KM84, though thought to be flat, has an upper mid bump to my ears which make sources poke through a mix. That can be good or bad depending on the source. Each mic has it's uses and I think the kit is worth making if just to have a different flavor.
would you suppose that a more authentic capsule is the only thing that prevents a more KM84 like response?
 
would you suppose that a more authentic capsule is the only thing that prevents a more KM84 like response?
Yes. My personal belief, based on tests I have done, is that the capsule is going to account for at least 75-80% of the sound of a mic. That is taking for granted that the circuit is well designed. I don't think there is anything in the circuit that is attenuating the >10KHZ frequencies. It just seems that Maiku has decided to make the capsule a little smoother and dare I say "warmer" than the original KM84 capsule. I've used 4 different KM84s over the last few years on a somewhat regular basis. They all sound similar. The two capsules I got from the kit also sound very similar to each other. They just don't sound similar to a KM84. That isn't to say they sound bad. They actually sound good and with a little EQ, they can sound very much like a KM84.

One thing I'm curious about. I wonder how an Oktava MK-012 capsule would sound with the kits body. I have not checked the threads but I have a feeling it wouldn't be a direct replacement but could be a fun experiment. I've used the Oktava mics (modded) that I have in place of KM84s all the time. When doing close micing, they sound closer than anything else I've ever tried.
 
Back
Top