KMV254c: Gefell MV691 to KM54c modification.

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

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

SparkleBear

Well-known member
GDIY Supporter
Joined
Jun 21, 2016
Messages
308
Location
Santa Cruz, CA
Hey all,

I have embarked on rebuilding my km54c DIY projects. Unfortunately, in 2020, our community in the Santa Cruz mountains of California suffered a devastating fire that took many homes including mine. So, many parts, so many tubes, so many vintage bits and bobs lost to the transformation of fire... but it's time to rise from the proverbial ashes start again.

I actually took a long hiatus from building audio equipment and am feeling really inspired having returned recently. I thought this would be a great place to restart. I had gotten pretty involved in the portable video game console modding scene which turned me on to designing custom circuit boards with kicad. Using fusion 360 to model cases and elegantly wiring everything together... I knew that one day i'd return and apply it to microphone building, my first electronics passion. :)

Here is the plan:

- build it around the 6S6B-v tube... last time I used the USA made NOS 5703 tube.
- 10:1 transformer. Possibly 3U Audio GZT-87 Output Transformer or TAB's t13...
- Ideally, find a pair of m94 in good condition with the mechanical body of the mv690/691... If I can source some. Or as a more expensive alternative, convert the m294 mic with a tube buffer.
- Build a PSU out of the original mv69x psu.
-Use Leo Schaal's custom binder tuchel connectors. OR machine a binder bottom bell to fit into the bottom of the mic in place of the stock part. (maybe thats too much work if leo's works great)

Ill be looking back at my previous work log from the rev 1 build to glean what wisdom I can as the build is gone with the wind.

( Rev 1 project: KM54c SDC Tube Mic Build / Gefell mv691 Mod )

I took some time to study the vintage schematics of the km54c and built the circuit in the kicad eda program. Im going to try to build it around the 6b6s-v tube... last time I used the USA made NOS 5703 tube.

1667286151959.png

I will be looking into how to design PSU circuits. @RuudNL had some cool posts about ac701 tube psu circuits that I might glean from. I did study full wave bridge rectifiers and post filtering... I don't yet have all the theory totally nailed down to design one from scratch all on my own.

I already sourced a few mv690/691 and a pair of m70 cardioids and m93 omnis. I even found an M71 cardioid LDC capsule that I am in negotiations for purchase right now. That same gentleman may have some m94 for me back in Germany when he returns for family matters.

1667292758833.png

As a side project and perhaps another prototype for this build i also got a few MV102 omni little cuties to experiment with.

I need to figure out footprints and layout for PCB once I get the MV69x mic bodies and can take measurements for the dimensions of the inner circuit. And consider lifting/floating the hi-z section of the circuit. FUN!! Or potentially do a p2p layout like the original circuit. I do have a stellar example pair that my family passed down to me... they are not for regular use hence wanting to build these bad boys...

1667293440557.png

Curiosities:
What composition caps and resistors to choose? Carbon comp / wirewound / metal film...

Do all of those capsules have the same polarization voltage requirements? I know that the m94 has a dropper resistor inside iirc...

Switchable / tunable NFB section for sweet spot on different capsules?

How to tie down calibration input pin 3 for accurate circuit reproduction. 200R to ground? Straight to ground? Floating?

I just copied the values of the original circuit... but am planning to put a trim on the NFB network and also on the cathode resistor position until i find the sweet spot and replace with actual resistor. Does anyone know of that tube prefers a different plate resistor or grid resistor?

I am feeling very open to feedback, opinions and ideas... Thanks for checking out the post!

I think this is going to be a really sweet microphone system! Stay tuned for more...
 
Last edited:
Progress! :)

Today I picked back up with the kicad development. Busted out the calipers and measured components... Desoldered the mv691 pcb from the mechanicals. It was fun to sketch it out...Mechanical drawings.jpg

Next I took those dimensions and built up some footprints in kicad so i could start the Layout process for the PCB. Along the way I added a few test points and pads into the schematic... Little by little built it up.

Kicad KM54 layout pic.jpg
1669280904460.png

Added Screen to ground for transformer, changed a few footprints and designations and cleaned up the silkscreen too... Here is where I ended up. I had fun drawing the silk for the tube!!

1669281122752.png 1669281007791.png

1669281749230.png

I also will be using isolated turret posts for the grid to be floating. I am wondering if anything else needs to be floating. Is it just grid side of grid to ground, the grid wire and the capsule front side? Is it necessary to float the capsule backplate?

Parts are on the way from Mouser... Couldn't get 160pf for the feedback section but could get 150pf and 180pf polystyrene. Got a couple different types of output capacitors too.

Also wondering if i should up the polarization voltage for the m70/m94/m93 capsules i will be using for this system...

Im open to any advice or wisdom from y'all. Really excited to have made progress.
 
I also will be using isolated turret posts for the grid to be floating. I am wondering if anything else needs to be floating. Is it just grid side of grid to ground, the grid wire and the capsule front side? Is it necessary to float the capsule backplate?
Isolated turrets needed
mmic.png
 
Isolated turrets needed
View attachment 100921
Hmm... i will have to rethink layout if i need to lift the bottom side of R2 as its a 25k trimmer. If i remember correctly, this trimmer will be controlling the amount of NFB that hits the anode to tune frequency. Is that right? What is the purpose of the 250k (R1) resistor? Thanks for the advice! :)
 
For R2, only soldering one leg to the PCB could give you troubles when you adjust the screw (strain on the solder connections). What I would do, is while you're soldering the single leg to the PCB, put a couple small dabs of E6000 or similar between the PCB and the pot. Could take a toothpick, and cover the area of the PCB under the pot without touching the solder pad. This gives is some mechanical stability when you're adjusting the pot; reducing the strain on the solder connections.
 
I would use both sides of the PCB, one for the big components like the tube, trafo and output cap and the other for the circuit resistors and small caps. Also you can get rid of the 22uf and a resistor going with a negative voltage bias from psu (for now I quite like it). Also take your measurements with trimpots but after use a small resistor instead, it will save space.

Do that Gefell capsules have two connections like the KK54? in opposition to most sdc capsules with just one connector (to the backplate?) and grounded diaphragm
 
Last edited:
Since nobody asked, I will do it and and answer my own question:
Q: "Why is the 6S6B-V you use black?
A: Because it is a special version, selected for low noise and it is coated (like the AC701) to avoid photo electric effects.
Ok, thanks for the explanation... But there is no light in the microphone ? What kind of photo electric effects we talked about ? Hum, crackles or other noises ? Would be benefits to paint tube in black for microphone use ? What kind of paint ?
 
Did you take into account that the body tube is round, and thus can cause problems with higher components on the side(s) of the PCB? (You probably did, but still...)
I thought about that and this microphone body tube is pretty big for an sdc. For the big components, I followed my layout from the prototype that I made many years ago (my profile photo). I don't think I could've fit them all if I centered the components. Also note that the pcb does not mount centered in the rails, it is very much like sitting on one edge of the rail laterally so the components are kind of tucked in the middle between the rails.

mv692 side view.jpg
For R2, only soldering one leg to the PCB could give you troubles when you adjust the screw (strain on the solder connections). What I would do, is while you're soldering the single leg to the PCB, put a couple small dabs of E6000 or similar between the PCB and the pot. Could take a toothpick, and cover the area of the PCB under the pot without touching the solder pad. This gives is some mechanical stability when you're adjusting the pot; reducing the strain on the solder connections.
Thats a great idea. I was thinking something like that too... e6000 is the best stuff too. Thanks for the tips DeltaSigma ;)

I would use both sides of the PCB, one for the big components like the tube, trafo and output cap and the other for the circuit resistors and small caps.
I didn't use both sides for the reasons above with where the board does not sit clear in the middle of the rails. Gefell only put components on one side too. But that might be due to the single layer pcb they were using... with connecting traces only on one side. *shrug*. Once i get this into the mic i can re-assess if I need to use the other side. But my hunch is I will not. There are already so few components on this design, but we will see...

Also you can get rid of the 22uf and a resistor going with a negative voltage bias from psu (for now I quite like it). Also take your measurements with trimpots but after use a small resistor instead, it will save space.
Yeah, im doing the trim pots right now so that I can tune the NFB and cathode bias so i can match two mics together eventually. Perhaps I will swap out the resistors with actual ones in the future. I did put a space for a real resistor next to the bottom trim pot so that when I discover the actual value that I am set on I can just use the pads next to it and take out the trim pot… But I am also recognizing that I could just use the same holes that the trim was using. *Shrug again*
Do that Gefell capsules have two connections like the KK54? in opposition to most sdc capsules with just one connector (to the backplate?) and grounded diaphragm
3 connections actually… Front diaphragm, back plate, chassis?… I have not investigated fully but that's my guess. There were three wires coming off of the board to the capsule assembly pins. *Shrug a third time*

Since nobody asked, I will do it and and answer my own question:
Q: "Why is the 6S6B-V you use black?
A: Because it is a special version, selected for low noise and it is coated (like the AC701) to avoid photo electric effects.
Thanks man… I thought the black paint was for minimizing microphonics as it's a bit rubbery and would absorb some vibrations. I never thought about photo electric effect… Very curious. Were these tubes painted by the factory? Or did a post production person outside of the "company/ Russian government" select and paint?
 
Is this pcb compleet selfmade? Nice.
Yeah, KiCAD. Its free! :) Over the last year or two, I got really into the Wii Portablizing community; it was there where I cut my chops to use those cad programs. Many reasons to build custom PCBs. Very fun, very interesting... great skill. But I'm glad I am back over on the audio side of things now.. hehehe.
 
Tonight I started working on the PSU. I pulled the design for the D49 PSU; I chose this as a starting point as I will be using an ac701 style mic and B+ / H+ match up to the design for the km254. I figured this could get down to 4v and up to 6.3v, perfect to test out the range for my 6s6b-v tube... I studied d49 psu schema for values and used only what I needed from it as my km254-V design does not have switching patterns. After studying I rebuilt it in KiCad... I incorporated the calibration scheme, both grounded and through 200R with jumpers. Eventually I would like to make a dual psu, but one is fine for now. :)

KM254v PSU Schema.jpg

Next is to figure out the layout. I am thinking to try and build this into one of those Alctron PSU as it the mechanical bits and bobs for such a thing... and its more of a prototype anyway. Too bad the transformer is 200v/20v, not 120v/20v like this project needs. Also, I have to think of how to do the dual psu in the future. If im using that chassis i need to break out the ol' calipers and trace out the edge cuts for the custom PCB.

Thanks to @poctop for the inspirations!

CHEERS! :)
 
Last edited:
Lots of progress! :) Only good news to report.

IMG_0220 (1).jpeg 69164532560__9C1FA9EA-1FB7-4CC0-95FF-2459EA9A9690 (1).jpeg

Sound samples:

This is my 1927 martin size 5 tenor guitar with M70 small diaphragm cardioid Nickel capsule.

View attachment KM254 M70 1k8 RK, 12k28NFB Tenor guitar.mp3


with the M71 large diaphragm cardioid M7 head.

View attachment KM254 M71 1k8 RK, 12k28NFB Tenor guitar.mp3


a bit of vocals with the m71: (just went through the range... lol funny)

View attachment KM254 M71 1k8 RK, 12k28NFB Vocals.mp3


Built the mic and it sounds great!! SO thrilled with the results. :)!!!
Rev 1 is a success, but needs some corrections.

IMG_0218.jpeg

First note for second revision was to make the mounting holes bigger... i had to file them larger as well as trim/ file the posts them selves to allow it to fit in there... but I got it on there.
69164447037__D84C0B0F-BEFF-41A9-8A5D-E6D549F9AA96 (1).jpeg

I also forgot to convert the decimal in the data sheet of the PTFE turret to MM so it was TIIINY. Had to drill and file those bigger. REV 2 error corrected :)

IMG_0228 (1).jpegIMG_0226 (1).jpeg

High Z section was too close to the side of the board, had to bend it down carefully. Rev 2 has those bridges more central where there is more clearance...

IMG_0219 (1).jpegIMG_0222 (1).jpeg
I had a lot of fun twisting and heatsinking the tube leads... oddly satisfying.

I ended up changing the value of the plate resistor from 250k to 100k. And the divider to 300k series and 300k shunt so its a clear 60v capsule polarization.

IMG_0223 (1).jpeg

I ended up having a bit of a clearance problem with the tuchel pins and the transformer. For this rev i ended up trimming the pin heads down to almost the base of the nub and soldering the wires at a right angle. It worked out but i made corrections. There was still some wiggle room... I moved the decoupler cap up a bit and scooted the whole footprint of the transformer up to make the space, this shortened the board a good 5 or 6mm. Hopefully that does the trick.

Ended up going with the Mic Parts VPS power supply for this build. Its small and quiet and works really well... didnt take many photos of that process... nothing special just a great supply.
69164469802__010B5D21-ACA3-4AC6-8E9E-074F21CC3F06 (1).jpeg

Here is a model of the revision 2. I just ordered another batch from JLC :) A few more notes for the next revision:

-Mounting holes made larger
- Component values adjusted on silk
- Turret holes correctly sized
- Fixed clearance issue with transformer and tuchel pins.
-moved high z bridge components centralized for clearance.
-ADDED jumper to bypass/engage Negative feedback circuit.
-ADDED double jumper to tie CAL to ground or tie it to a 200R to ground.

1670314157249.png1670314173834.png1670314316960.png1670314333165.png

Really excited to put the second revision together. I want it to be really easy to assemble. I already disassembled an mv690 in prep for the boards showing up next week.
 

Attachments

  • IMG_0221 (1).jpeg
    IMG_0221 (1).jpeg
    1.7 MB · Views: 1
  • IMG_0224 (1).jpeg
    IMG_0224 (1).jpeg
    1.8 MB · Views: 0
  • IMG_0225 (1).jpeg
    IMG_0225 (1).jpeg
    1.7 MB · Views: 1
  • 69164411020__4EF2E2A3-1418-4EDF-AA03-E3D55BE34C20 (1).jpeg
    69164411020__4EF2E2A3-1418-4EDF-AA03-E3D55BE34C20 (1).jpeg
    2.3 MB · Views: 2
  • 69164460975__7D748181-AEB4-4B30-8FD9-2A5C15F99D85 (1).jpeg
    69164460975__7D748181-AEB4-4B30-8FD9-2A5C15F99D85 (1).jpeg
    1.3 MB · Views: 1
  • IMG_0238.jpeg
    IMG_0238.jpeg
    1.5 MB · Views: 1
  • IMG_0239.jpeg
    IMG_0239.jpeg
    1.5 MB · Views: 1
  • IMG_0217.jpeg
    IMG_0217.jpeg
    2.4 MB · Views: 2
Last edited:
Sounds a little smiley (holo) but something the km84 is also a bit holo. Very good. I like it. Congratulations.
Can we sign-up for a pcb. :)
 
Sounds a little smiley (holo) but something the km84 is also a bit holo. Very good. I like it. Congratulations.
Thanks!
Also… The room I have available in this condo sounds like trash… But I can hear through it lol. Rev 2 two will more easily be able to swap in and out negative feedback as well as lift ground on calibration to check tone differences

Can we sign-up for a pcb. :)
Maybe after revision two is done and all the kinks are ironed out I can make these available
 
I'm really sorry that you lost your house in the fire, that was a big catastrophe.

Congratulation for your work, amazing stuff you're developing.
Wish you much success

By the way, what software are you using to design the PCB and do the 3D modeling?
 
I'm really sorry that you lost your house in the fire, that was a big catastrophe.

Congratulation for your work, amazing stuff you're developing.
Wish you much success

Thank you for the kindness I appreciate your words…
By the way, what software are you using to design the PCB and do the 3D modeling?
I am doing everything in KICAD. If you use the stock library it is already configured to render 3-D models of the board and components… I have not gotten around to designing my own models for things yet… Hence why you don’t see the tube or the transformer in the renders
 
I am doing everything in KICAD. If you use the stock library it is already configured to render 3-D models of the board and components… I have not gotten around to designing my own models for things yet… Hence why you don’t see the tube or the transformer in the renders

Well it looks great for sure,
I'm amazed how a Free software can be so good
 

Latest posts

Back
Top