KM54c SDC Tube Mic Build / Gefell mv691 Mod

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.


Well-known member
GDIY Supporter
Jun 21, 2016
Santa Cruz, CA
Hey Y'all,

I'm excited about my KM54 build and wanted to share it with you all here. The plan was to build a pair of SDC tube mics.  Years ago I was inspired by the idea of a tube SDC DIY project.  I didn't see much info about it... After a lot of research and some conversations with Stephen of Cathedral Guitars, Dennis over at AMI, the kind folks at RMS audio and AusTex64 (thanks guys), i moved forward on the project build... It took over a year from conception to production but the day has come!

I put together a tube mic based on the KM54 circuit built into an old Gefell MV691 body with the m94 capsule, TAB t8 6:1 traffo, and raytheon 5703 sub mini tube. I didn't change any of the component values as per the neumann schematic except I substituted the 3.9k Ohm Cathode Resistor (Rk) so I could experiment with the cathode bias point (thanks to RMS audio for that tip).  I set the pot to 3.9k for initial testing and listening and plan to take it from there.

I did choose to incorporate the NFB network just incase I wanted to tame the 8k peak of the m94 capsule.(IIRC m94 was intended as a distant cardioid pickup pattern so the peak was to compensate for HF loss at distance). I was thinking of adding a NFB bypass switch that would break the connection  C3 (.01uf cap) has with ground on pin 6 at the PSU.

The MV691 +  M94 combo is the perfect starting point for the KM54 build. The m94 capsule is supposed to be very similar or the same as the rolled nickel capsule of the KM54. I had to remove the dropping resistor from inside the capsule assembly to match the km54 circuit specs. Also the mv691 body has way more space than most SDC which is helpful to fit all of the components (thankfully because they barely fit as it is)...

The MV691 I found to be way better than the MV692 because it has metal rails and threads as compared to the similar mv692 which was all plastic inside. Not ideal for a DIY mod, originally I was studying my KM254s and was planning to do a point to point that was the same as the 254 but decided to just go with protoboard (uxcel 2cm x 8cm) that I ordered from amazon. This made it way easier to conceive and build the circuit, at least for my first go around as I've never done a project like this before.

The MV691's board is attached to the rails by solder via the little metal tabs coming off of the rails. This is the securing method as well as the grounding point for the mic circuit. I desoldered these points along with the capsule assembly wiring being careful to not damage the original board incase I want to revert back to FET 48v head amp. Anyway,  I utilized the solder mounting method with my build, drilling holes in my protoboard and lining them up with the tabs of the mv691 rails. This is where I started.

For the tube I ended up using the 5703 as I heard good things about its similarities to the AC701. This is a 6.3v heated sub mini tube. But I got a bunch of other tubes while planning this mic; ec1000s, 5840s, EC92s, UC92s. I decided to use the 5703 because they were inexpensive and a good place to start. Maybe i'll swap tubes in the future or build the second mic with a different tube to compare their sound. We shall see... Does anyone have experience building around a 5703 tube that may want to impart their wisdom upon my build?

I have not built the power supply for the pair of mics yet, still researching the best way to go about that. I was planning to just build two of Dany's m49 pcb PSU into one chassis (skipping the pattern section of course). Right now I am just using my KM254 PSU from Peak Power (Bill Bradley) turned up to 6.3v to test the microphone out before building the PSUs.

Unless there is a better idea, i was just going to put the 4 traffos and two chokes necessary to power two mics with the m49 psu circuit... Thats a lot of space and weight. Does anyone have a better idea about building the PSUs for this mic pair?  I'd love to be efficient with space and parts.

After a day of arranging components, soldering, wiring and staring at schematics for hours and hours i finally had a (hopefully) finished mic circuit. I fired it up with my peak power PSU and to my delight heard my voice through the new tube mic build! YAY! I definitely heard some white noise but knew better than to judge before a 72hr burn in. After the burn in the noise went away.

Anyway, the microphone sounds really nice! I have not messed with the pots for the NFB or cathode bias yet... That should be fun. Just finished burning in the tube yesterday. Made a recording of a Larrivee OMV3r acoustic guitar and it sounds beautiful to my ears. I will attach the recording I made. I have a pair of km254 that I will eventually compare against these DIY km54. I won’t be able to do a side by side until i build a PSU for the DIY pair. Stay tuned. :)

(link to recording of mic )

Considerations and questions:

The backside of the head amp functions but is not beautiful... I will tackle this design challenge on the second mic build.

Should I cut ventilation holes for the sub-mini tube?

The circuit polarizes the capsule at 40v for the kk54. The MV system can also have other capsules screwed into the circuit but these are designed to run at 60v. How will this effect their performances should I decide to use different capsules with the microphone?

Another weak point of the build is the cheap and flimsy 7 pin XLR that RFT/Gefell utilized in the mv series of microphones. Wow, so light and untrustworthy as compared to a tuchel or binder system. The threading is large and cant be interchanged with a tuchel/binder part. I am considering my options for this problem. I learned that Oliver was using a metal lathe to modify the binder/tuchel chassis mount connectors by removing metal and tapping screws to match the internal rails of the mv691 so that we can use tuchel/binder cable connectors rather than the cheap unreliable RFT ones. I have yet to tackle that operation but perhaps in the near future.

Also, as mentioned above… I want to build the most efficient (cost and space) Dual PSU for the mic pair. Ideas?

If there is interest, I can document the build of the second one of the pair! :)

Well, Thanks for checking out my post. I am so grateful to this microphone community. Its been so wonderful reading, learning and sharing on these forums. I started getting into building mics only a few years ago and now its nearly all i think about when I’m not making music It makes my life better, so thank you all. :)

Best, Eric

Circuit front:

Circuit board back:

Capsule Side:

There's no pics of your build ;)
Week ago finished work on two KM54C.
Circuit is pretty simple and pretty nice.
It ended up on 470nF capacitor at the output since i had problem to find any other film capacitor which would fit KM54 body. No worry no difference between 470nF and 1uF.
Biasing is crucial here. I had to match both mics so it ended up on potentiometers installed inside instead of 3.9K. This allowed to much better match two little bit different AC701k.
For AC701 plate voltage should be around 43V so i would for the start try to get similar range for 5703.
For the start, if you didn't make it already, ground C3 for normal use of microphone.
Higher polarisation voltage vs. lower?!?! 40V will give you more headroom, 60V lower noise :) I would go in the the middle and try 50V :D  :D  :D Or just 1M linear potentiometer instead of r8/r9 divider and try what will fit best ;)
R2 in both mics was 25k.

Weird, I had all the photos on my thread. Let me upload them separately. :)

Thanks for reply... I am getting 41v at capsule and plate.  I grounded c3 at the PSU and eventually a thinking of making a bypass switch for NFB network by cutting the ground at a switch in the PSU box. So far the mic sounds great... ill turn up the r2 to 25k and listen. Its at 20k right now. Did you settle at around 3.9k for cathode bias resistor? Will you replace the pots with resistors?

Piotr, Were you working on a build of a new km54 style mic with ac701s? or were you repairing an original vintage mic?



  • KM54CircuitFront.jpg
    322.4 KB · Views: 308
Looks nice :)

Exactly i had to rebuild one KM54a to KM54c and other, previously, nasty reworked KM54c.
I had to make a pair of 54c. Both "vintage" mics.
Few things, original transformer is around 10:1 and i think you will get much better result with higher ratio than 6:1.
If you will go for higher ratio transformer then i think you could set 50V of polarisation voltage without any problem.
C3 i would rather ground inside the mic. Then it works as filtering capacitor. For the feedback responsible is C4. Disconnecting C3 from ground doesn' affect feedback at all.
For example in KM54a not grounded C3 introduced noise when it was floating,
No i didn't replaced potentiometers, since there was no option to find exact values of resistors.
Doesn't Vishay 1839 MKP (white) is for pulse applications?
Since there's feedback in the circuit it is definately less sensitive for changes, but still is. In pair with cathode capacitor usual you are changing senstivity and low end response. For KM54c here's just more need to set most optimial working point for tube and sensitivity.  In the circuit without feedback, the higher resistance, then you can use lower capacitance and get the same response. 
Definately you will get much different result with 10:1 transformer.  This is most important in building pair of microphones, to set the same sensitivity for both. In the pair i worked on for example both potentiometers were set to different values.
If you would rearrange lower part of the circuit above transformer, you could fit 10uF film capacitor for cathode. It's enough here.
I’m wanting to test out the original mv691 transformer in the mic. It looks like two separate transformers on one core. I’ve been reading that it is a 4.5 to 1. I am not sure how to measure this for myself but the resistance of one side of coil is 45ohm and the other side is 670ohm.    There are to identical coils of each resistance value. I was reading that if you wire the primary in series we will have a 9:1 ... if I go by resistance ratio the transformer is 14:1

Can anyone help me work this part out?

there is somewhere in this knowledge pool here the answer to your question...thats what i remember
HarmonyUnited said:
I’m wanting to test out the original mv691 transformer in the mic. It looks like two separate transformers on one core. I’ve been reading that it is a 4.5 to 1. I am not sure how to measure this for myself but the resistance of one side of coil is 45ohm and the other side is 670ohm.    There are to identical coils of each resistance value. I was reading that if you wire the primary in series we will have a 9:1 ... if I go by resistance ratio the transformer is 14:1

Can anyone help me work this part out?


For example apply 1kHz/1V and measure with voltometer or oscilloscope what's on the second side of the transformer. This will give you voltage ratio.
Yes, i worked on two original KM54 with AC701k. Currently working on M49 build, not really a "clone", rather quintesence of few M49 circuits but this build is based also on NOS AC701k. There's not much problem to buy AC701, price is the problem :D
So I just swapped in the Gefell bv130 into the circuit where the T8 used to be. Im still trying to wrap my head around this... the 10:1 gives more output signal than the 6:1 with the same input signal correct?

I will see how it sounds in comparison...
OK so I answered my own questions here. I did a transformer shoot out!

Using the same mic circuit with the same mic preamp at the same gain level....Same guitar and same mic placement as well. The variable here is the transformer. ( I discovered on my own that the lower ratio transformers have hotter outputs... but seem to get saturated easier than higher ratio transformers (just a guess from listening) )

After recording I gain matched each of the regions so that I could A/B/C with better judgement.

The transformers shot out are: (with links to the recordings)

AMI T8 (6:1 ratio)

AMI T13 (9.5:1)

Gefell BV130 wired as a single side of secondary (4.5:1)

Gefell BV130 wired with secondaries in series (9:1)

For control and curiosity I also recorded a genuine unmodified KM254c, the identical circuit I followed when building the modded mic.

I noticed that the Neumann had a much higher output than the mic I built. And to no surprise sounded like it had more mojo magic than the rest. I have not touched tweaked the cathode resistor in my design... Perhaps I need to tweak the capsule voltage or cathode resistance to get closer to the real thing.

So, I found the AMI transformers to have a much more detailed top end with less bass.  The Gefell bv130 sounded more "vintage-y" and the 4.5:1 recording sounds a lot more low mid heavy with lots less top... I think of all the trannys i liked the T8 the best... but ill listen again with fresh ears tomorrow. :)

Let me know what y'all think if you get around to listening to those files. :)


Make this carefully and give you some time.
First i would set bias as best as possible for 5703. 
This not necessary is transformer saturation, note that between 9:1 and 4.5:1 is pretty big difference in output impedance of microphone. This also can give you different results in sound with different mic preamps. Yes lower ratio higher output but you need remember about impedance. For example if microphone have 200ohm output for 9:1 ratio, then for 4.5:1 it will be 800ohm.
Harmony United,

Great post!  I have the m94 capsule and was going to do this same mod.  Its currently being used in a royer cathode follower setup but I'd like to hear this circuit to compare.

From your 1st post-
" I had to remove the dropping resistor from inside the capsule assembly to match the km54 circuit specs. "

Was this complicated?  I have been warned that this can damage the capsule if done the wrong way.
Can you shed any light on this process?