Official C12 Clone - Build and Support Thread

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.
Balanced halves are good for things like phase splitters:  I can't imagine it being very critical for microphone use.  It may be there is some correlation between balanced sides and noise but I can't imagine what it would be.
 
Thank you for a great support thread on this build and an awesome mic! Finished mine with your headbasket, PCB and parts kit, I used a Microphone Parts RK12 Capsule and a Electroharmonix 6072A tube. PSU case and microphone body is custom built and transformer wound locally South Africa. Everything worked 100% the first time! I want another one... :).
 

Attachments

  • C12 (768x1024).jpg
    C12 (768x1024).jpg
    394.8 KB · Views: 92
I just finished the PSU and have the following readings:

B+ : 162V
Heater: 7.67V
Bias: -0.86

Wrong?

/M
 
mica said:
I just finished the PSU and have the following readings:

B+ : 162V
Heater: 7.67V
Bias: -0.86

Wrong?

/M

Looks correct.  If you are using the 6072A tube, please install the test 180K resistor and you'll be able to measure close to the real operating conditions.  You can stick the resistor right into the appropriate pins of the 7-pin XLR connector (between B+ and ground).
 
Matador said:
Category 5 said:
Awesome!  please make a PSU only kit available.  In the mean time can you peruse my post from late last night and give your thoughts.  If I need to make a component order I'd liked to try and get it in with another order tonight.

Thanks a million guys!

I am highly suspecting C4.  Can you remove it and try the loaded and unloaded readings of all of those points as before?  Unloaded voltages seem very low, which implies leakage current.  The bleeder resistor R6 is sized to bleed away (120V/470K) = 0.25mA, but your unloaded measurements indicate 40V lost across R1 (91K) which is 0.5mA, which is 0.25mA too much.  R4 is the only cap not protected by a series resistor and would suffer the effects of a B+ short the most.

Did some troubleshooting tonight and ended up not quite where I wanted to .  ;)

Replaced C4 as you recommended but got the same results.  I then replaced c1, c2, c3 and the 100K pot and I'm back where I was before the short.  B+ is adjustable to about 108 before I run out of turns.

After replacing C4 loaded volts were, diodes-268 -> R1-200.8V - > R2-134.2V -> B+ 98V

After replacing C1, C2, C3 and R4  diodes-267V - > R1-186V -> R2-108V -> B+ 108V

Not sure why the drop across R1 went up so much.

Checked the "good" PSU and indeed the transformer is different.  I get 225VAC out of the secondary wheras the one labeled 175V gives me 198.3VAC

The voltages (loaded) for the good PSU are

diodes-303.2V - > R1-215.4V -> R2-125.5V -> B+ 120V

This is with the same 180K resistor simulating load.  I'd swap PCBs and see if the tranny is to blame but I'm reluctant to go through that just yet.  I get a solid 120VAC from the line.  One transformer is giving me 12.5% more volts at the input to the PCB.  If that 12.5% translated all the way through to the output of B+ that would explain the difference.  If the transformer is to blame, then I assume I could just use lower values for R1 and R2 to get the required output.

If some of these PSUs are using different parts without making it clear, I can see where the regulated PSU boards that are in the works would be an even greater benefit to this project.
 
You have a 30V difference post regulator:  if you measured and only get 198VAC that would explain the difference.  It looks like there is a lot more spread from transformer to transformer.

I redid some maths for you based on your readings, and I think you can safely adjust R1 and R2.  Assuming your readings hold, and you want to target 120V at B+ with the test resistor of 180K, and R4 dialed to 50K (to give maximum adjustment), then changing R1 and R2 to 56K each will do it.  The filter action is essentially unchanged, having a pole below 1 mHz.

The nominal B+ voltage at R4=50K should be about 125V.  R4 at maximum would give about 80V, and R4 at minimum would give about 170V.
 
Matador said:
You have a 30V difference post regulator:  if you measured and only get 198VAC that would explain the difference.  It looks like there is a lot more spread from transformer to transformer.

I redid some maths for you based on your readings, and I think you can safely adjust R1 and R2.  Assuming your readings hold, and you want to target 120V at B+ with the test resistor of 180K, and R4 dialed to 50K (to give maximum adjustment), then changing R1 and R2 to 56K each will do it.  The filter action is essentially unchanged, having a pole below 1 mHz.

The nominal B+ voltage at R4=50K should be about 125V.  R4 at maximum would give about 80V, and R4 at minimum would give about 170V.

Thanks a lot Mat.  Would Carbon composition resistors 1/4 watt be okay here?  I think I have some metal films somewhere but I know I have some carbon laying around.  I can measure them to get closest to actual value.

This was bumming me out.  Assemble is impeccable, soldering joints are shiny and look great.  I just couldn't imagine what had gone wrong here.

One of the problems with sourcing overseas is that they will do these component swaps when convenient and not mention it.  Bring on the regulated version. ;-)
 
Hi!

Does the PCB's fit an Avantone CV-12 body?
Any need for changing the PSU (polar patterns, pinouts etc)?

I'm planning a build here and got a Avantone very cheap :)
And by the way: Beautiful PCB's!

Thanks!

Edit: I see now that they most likely do!
 
It looks like it will likely be compatible.  Depending on how cheaply you got your Avantone, you might consider selling to at profit and then using a cheaper donor.  As i understand it the Avantone has upgraded capacitors, better Q/C'd capsule, and possibly a few other tweaks onboard that make it a nice mic someone else may be able to use as-is.  It would also have extra "ventilation" in the body as it has pad and low-cut switches that our clone does not utilize and there would not be space inside to retain the stock switch pcb.

Conviction said:
Hi!

Does the PCB's fit an Avantone CV-12 body?
Any need for changing the PSU (polar patterns, pinouts etc)?

I'm planning a build here and got a Avantone very cheap :)
And by the way: Beautiful PCB's!

Thanks!

Edit: I see now that they most likely do!
 
Matador, I wanted to PM you some questions, but your box was full the last time I tried, so I will just put my questions here. I am gearing up to buy the PCB as I now have a nice tube and my Tim Campbell capsule, but had these remaining (newbie) questions before I purchase the rest of the components and begin...

When I look at pictures of the PCB build in this thread, I am confused by the Grounding scheme. I do not see a connection from the unused side of the tube to ground, and I see several ground points on the PCB that don't seem to be connected to each other. Am I missing something, or are there multiple floating ground points? I may have a misunderstanding of ground, but I thought all ground points would be connected. Maybe I am just not able to see the trace in the pictures?

Do you have an opinion on using the teflon standoffs for the capsule connections as is, versus drilling holes in the PCB to push the standoffs in, and wiring the capsule point to point?

Somewhat unimportant, but do you know exactly how much wider the Alctron mic is compared to a vintage C12 body? Is it about 4mm?

I have never soldered a PCB before, but I have built one mic point to point. Do I need to take extra precaution to clean the board with alcohol, since it is a PCB? I did not do any cleaning really on my first mic, and it sounds great.

Lastly, I saw your answer about the new regulated C12 PSU having a low impedance fixed bias feature. My question is: the older PSU PCB is more "true" to the C12 design, correct? I know it is not exactly the same, as you've mentioned specifically that it utilizes a half wave rectifier instead of full wave. What effect, if any, does that have? I just want to be sure that the older design is "truer" to the original C12 PSU design so that I know which to order.

Thanks a lot!
 
Melodeath00 said:
When I look at pictures of the PCB build in this thread, I am confused by the Grounding scheme. I do not see a connection from the unused side of the tube to ground, and I see several ground points on the PCB that don't seem to be connected to each other. Am I missing something, or are there multiple floating ground points? I may have a misunderstanding of ground, but I thought all ground points would be connected. Maybe I am just not able to see the trace in the pictures?

The default config doesn't have the unused tube side grounded.  You certainly can if you want:  I noticed no difference in the prototypes.  Grounding all of that stuff makes it much more difficult to swap tube sides later however.

Can you be more specific about the other points?  Which points do you think aren't connected?  There is a ground plane on the bottom of the board which connects all ground points:  there aren't individual traces for ground.

Melodeath00 said:
Do you have an opinion on using the teflon standoffs for the capsule connections as is, versus drilling holes in the PCB to push the standoffs in, and wiring the capsule point to point?

Today we have nice, well controlled manufacturing processes, and have well-characterized PCB materials, so we can apply science to see where we will get a benefit and where we won't.  I ran through all of the calculations and went back and forth with the PCB manufacturer to make sure that we could use PCB tracks (with a good-quality soldermask) and not suffer noise or other filtering penalties as a result of running the connections right up to the capsule connection points.

Provided the PCB is cleaned well with IPA, there isn't much to be gained by performing surgery just to re-establish the same connections (only this time up in the air).

Melodeath00 said:
Somewhat unimportant, but do you know exactly how much wider the Alctron mic is compared to a vintage C12 body? Is it about 4mm?

I'm not sure:  perhaps others can answer this?

Melodeath00 said:
I have never soldered a PCB before, but I have built one mic point to point. Do I need to take extra precaution to clean the board with alcohol, since it is a PCB? I did not do any cleaning really on my first mic, and it sounds great.

Per above, absolutely.  In return for good cleaning, you'll have an easier time assembling and results from mike to mike will be much more consistent.

Melodeath00 said:
Lastly, I saw your answer about the new regulated C12 PSU having a low impedance fixed bias feature. My question is: the older PSU PCB is more "true" to the C12 design, correct? I know it is not exactly the same, as you've mentioned specifically that it utilizes a half wave rectifier instead of full wave. What effect, if any, does that have? I just want to be sure that the older design is "truer" to the original C12 PSU design so that I know which to order.

Thanks a lot!

Yes, the old PCB is more "true" to the original design:  in fact, the bias supply is identical.  Full-wave puts the PSU hum at twice the frequency (120Hz vs. 60Hz), so the filters are 18dB more effective at quieting this hum.

In short, yes, the older design is truer to the original design:  however both the old design as well as the new "fully regulated" design are both full-wave rectified.
 
Matador said:
You have a 30V difference post regulator:  if you measured and only get 198VAC that would explain the difference.  It looks like there is a lot more spread from transformer to transformer.

I redid some maths for you based on your readings, and I think you can safely adjust R1 and R2.  Assuming your readings hold, and you want to target 120V at B+ with the test resistor of 180K, and R4 dialed to 50K (to give maximum adjustment), then changing R1 and R2 to 56K each will do it.  The filter action is essentially unchanged, having a pole below 1 mHz.

The nominal B+ voltage at R4=50K should be about 125V.  R4 at maximum would give about 80V, and R4 at minimum would give about 170V.

Thanks again for taking the time to help here Matador.  The 56K resistors did it.  With the 180K dummy load all my voltages were easily adjustable and well within the middle of the adjustment range.

If anyone else has this problem (B+ Adjustable to about 110 maximum) double check the AC voltage at the output of the transformer.  The good one output 225VAC.  The bad one was only outputting 198VAC.  FWIW, the bad one was labeled 175V whereas the good one was labeled 200VAC.  56K for R1 and R2 instead of the specified 91K (as Matador kindly calculated for me) did the trick!


I am betting I am not the only one who got a lower output transformer.
 
Matador said:
Melodeath00 said:
When I look at pictures of the PCB build in this thread, I am confused by the Grounding scheme. I do not see a connection from the unused side of the tube to ground, and I see several ground points on the PCB that don't seem to be connected to each other. Am I missing something, or are there multiple floating ground points? I may have a misunderstanding of ground, but I thought all ground points would be connected. Maybe I am just not able to see the trace in the pictures?

The default config doesn't have the unused tube side grounded.  You certainly can if you want:  I noticed no difference in the prototypes.  Grounding all of that stuff makes it much more difficult to swap tube sides later however.

Can you be more specific about the other points?  Which points do you think aren't connected?  There is a ground plane on the bottom of the board which connects all ground points:  there aren't individual traces for ground.

Are you saying you did not notice a sound difference when you grounded the unused side versus not grounding? I agree grounding it would make switching sides more difficult, but I asked only because in the schematic the unused side is grounded.

As for the parts that don't look grounded: P6 at the bottom of the board looks like it is not connected to anything else. Same with FC at the top, and C10 and C11 near the tube deck. Am I just not seeing the ground plane?


Matador said:
Melodeath00 said:
Do you have an opinion on using the teflon standoffs for the capsule connections as is, versus drilling holes in the PCB to push the standoffs in, and wiring the capsule point to point?

Today we have nice, well controlled manufacturing processes, and have well-characterized PCB materials, so we can apply science to see where we will get a benefit and where we won't.  I ran through all of the calculations and went back and forth with the PCB manufacturer to make sure that we could use PCB tracks (with a good-quality soldermask) and not suffer noise or other filtering penalties as a result of running the connections right up to the capsule connection points.

Provided the PCB is cleaned well with IPA, there isn't much to be gained by performing surgery just to re-establish the same connections (only this time up in the air).

Thanks for the info. To be perfectly honest, I'm not even sure I understand the point of the teflon. Why not just use normal keystones?

I ask about drilling and making p2p connections because in another thread you mention that your "PCB is flat down to roughly 40-50Hz," when it seems like it should be flat down to 20Hz. However, you do mention increasing the value of the output coupling cap and using the AMI T14 would help to flatten the response. Which transformer was used in your freq sweep test? Does increasing the C12 cap not just lower the high end roll off point? Does it also extend bass response?


Matador said:
Melodeath00 said:
Lastly, I saw your answer about the new regulated C12 PSU having a low impedance fixed bias feature. My question is: the older PSU PCB is more "true" to the C12 design, correct? I know it is not exactly the same, as you've mentioned specifically that it utilizes a half wave rectifier instead of full wave. What effect, if any, does that have? I just want to be sure that the older design is "truer" to the original C12 PSU design so that I know which to order.

Thanks a lot!

Yes, the old PCB is more "true" to the original design:  in fact, the bias supply is identical.  Full-wave puts the PSU hum at twice the frequency (120Hz vs. 60Hz), so the filters are 18dB more effective at quieting this hum.

In short, yes, the older design is truer to the original design:  however both the old design as well as the new "fully regulated" design are both full-wave rectified.

The filters of the original are 18dB more effective at quieting 120Hz than your half-wave is at quieting 60Hz? I guess I'm confused.
I'm looking forward to hearing your thoughts on the new regulated PSU since such a big deal is made about regulated versus non-regulated among mic gurus. It's especially interesting since the new regulated PSU has the full-wave like the original as well.

I read David Bock state that the original C12 used polystyrene caps and paper in oil. Does anyone know which were which? Was it only the coupling cap that was PIO?

I apologize for so many questions. I am a newbie and trying to learn. Thanks so much!
 
Melodeath00 said:
Are you saying you did not notice a sound difference when you grounded the unused side versus not grounding? I agree grounding it would make switching sides more difficult, but I asked only because in the schematic the unused side is grounded.

As for the parts that don't look grounded: P6 at the bottom of the board looks like it is not connected to anything else. Same with FC at the top, and C10 and C11 near the tube deck. Am I just not seeing the ground plane?

I assume you mean the original schematic?  Mine shows these nodes floating.  And yes, I didn't notice any difference after grounding them.

Flip the PCB and look at the bottom:  everything that is yellow is the ground plane. ;)  All of those nodes are connected to the plane via thermal reliefs.  If you look very closely, you see 4 little yellow stubs on the bottom that stitch the nodes to the plane itself.

Melodeath00 said:
Thanks for the info. To be perfectly honest, I'm not even sure I understand the point of the teflon. Why not just use normal keystones?

I ask about drilling and making p2p connections because in another thread you mention that your "PCB is flat down to roughly 40-50Hz," when it seems like it should be flat down to 20Hz. However, you do mention increasing the value of the output coupling cap and using the AMI T14 would help to flatten the response. Which transformer was used in your freq sweep test? Does increasing the C12 cap not just lower the high end roll off point? Does it also extend bass response?

Telfon is used because it is damn near a perfect insulator (something like 1018 ohms per centimeter).

There are many filters:  the capsule to tube, the output cap into the transformer, the tube capacitance w/Miller effect, etc.  With stock values, C12 (at 0.5uF) into a worst-case 150ohm mike pre has a 3db down point at roughly 40Hz.  You can move this lower in frequency by making C12 bigger, finding an output transformer with more Henries, or both.  I measured with a Cinemag CM-2480.  High end roll-off is not really modulated by the capacitance of C12, it's by the transformer primary inductance and winding capacitance.

Melodeath00 said:
The filters of the original are 18dB more effective at quieting 120Hz than your half-wave is at quieting 60Hz? I guess I'm confused.
I'm looking forward to hearing your thoughts on the new regulated PSU since such a big deal is made about regulated versus non-regulated among mic gurus. It's especially interesting since the new regulated PSU has the full-wave like the original as well.

I read David Bock state that the original C12 used polystyrene caps and paper in oil. Does anyone know which were which? Was it only the coupling cap that was PIO?

I apologize for so many questions. I am a newbie and trying to learn. Thanks so much!

The filtering action is different between the regulated PSU and the original.  The original uses 3 cascaded RC filters to try and make the B+ line look like "ground" to AC signals, the strongest of which will have a 120Hz signature due to the full-wave rectification.  It's a brute force approach, however a time-tested one that's been used successfully in countless designs.  It's also cheap and repeatable and pretty much indestructible.

The regulated version uses a modern IC, which has a bandwidth about six orders of magnitude faster than the ripple frequency.  It's so fast in comparison to the ripple that it can turn 120Hz ripple into a constant voltage by applying fast corrections to the internal pass element(s).  It's a different way to solve the same problem.

The key:  the IC uses feedback to make the PSU look like an ideal voltage source, with much lower output impedance.  This means that the PSU can essentially supply any current at fixed voltage, and makes it so it doesn't care which tube (or which tube type even) is used (within reason).

So which is "better"?  That's up to our ears to decide. ;)
 
So let me spend some time to write about why older designs used isolated high-impedance sections, and why we don't always have to anymore.

Let's boil this down to basics:  the capsule and the input impedance of the active element (in this case, a tube) form a high pass filter.  Given a worst-case "typical" capsule of 50pF, and a 1M input impedance (e.g. a 1M grid resistor, or R15 on the C12 schematic), the -3dB point is 3.1kHz.  This is obviously no good, unless you only close-mike coins falling on a table or perhaps whistling.

So R15 has to be higher:  if you do the math, in order to have a -3dB point at 20Hz, the grid resistance must be at least 150M on the worst capsule.  So the C12 uses 250M, probably because it was a part they could source at the time.  250M with 50pF gives about 13Hz, so life is good.

But what else is going on here?  Some of the capsule connections sit at the polarization voltage (FB and RB, for instance) or 60V (or higher, up to 120V!).  At these impedances, small leakage currents start to look like resistors that are of the same magnitude as R15, so now we have to look at our math again.

Let's pretend we have one of these connections to 60V on a dirty, untreated PCB.  A "hole" (or lack of electron) at this high potential sees all of the potentials that are around it that are less (like ground!), and it wants to go there.  If the PCB has a bunch of flux on it that is slightly conductive then we might have a leakage current there.  Let's pretend we have 1uA of leakage from the 60V potential.  Ohms Law says this leakage path looks like a 60V/1uA = 60M resistor:  no different than if we had soldered a 60M resistor there!

But we were counting on the 60V node seeing only the grid impedance, which was 250M above.  But now this new leakage impedance appears to the capsule to be in parallel with this grid impedance (which is ground referenced as well), and being that this hole doesn't want to violate Ohms Law it says "I see 60M in parallel with 250M, not just 250M!".  Knowing that two resistances in parallel always net to something less than the lowest of those resistance, there's no way the hole sees 250M any more:  in fact, it sees something (by definition, without doing any math) slightly less than 60M!

So now the leakage current has formed a phantom resistor, and demolished our carefully calculated high-pass filter.  Instead of a 13Hz high-pass filter, the capsule and the grid + leakage current appears to the system as a high-pass filter of 53Hz.  All from just 1uA of leakage current.  You can do more analysis and see that moving to 2uA leakage current makes the total resistance fall below 30M, which puts the HPF above 100Hz!  Our bass is melting away!

If we could lower the polarization voltage, then the leakage current would shrink, and the phantom resistor would increase in value.  But then sensitivity would suffer.  So we must increase the impedance between these high impedance sections and ground to minimize the magnitude of the leakage currents.

So air is a good insulator:  1014 ohms per centimeter or thereabouts.  Teflon is better (about 1000 times!).  But even 1 cm of air looks like a 1000G resistor, which means only infinitesimally small leakage currents are possible, certainly nothing to move the location of a filter pole.

But modern soldermask is a good insulator as well:  typical UV cured polymerized formulations (Taiyo UVR-150G is a common one) have insulation resistances on the order of 1012 per centimeter:  still 100 times worse than air, but still two orders of magnitude higher than the circuit impedances in question, provided they are clean.  Obviously flux sitting on top of the soldermask is in effect "in parallel" with the soldermask itself, and can cause problems per above.

These technologies didn't exist when the original C12's were produced:  air is a good free insulator in any case. ;)  But part of the challenge of these things is balancing easy-of-assembly with the electrical requirements.  In my opinion, having a stable, repeatable, and easy-for-the beginner layout and design took precedence over gaining some extra giga-ohms, given that I knew I could arrange a design that rendered the differences manageable and inaudible.

So in case it hasn't become clear from my protestations, and Chunger's 14134234982452 detailed photos of a toothbrush and a plate of 95% IPA:

CLEAN YOUR PCB's, and you'll have great sounding, repeatable results without all of the component lead gymnastics!  ;D
 
Great explanation Matador!! But when I built my first acrylic mounted point to point C-12 and compared it to my modded apex 460 sonic differences were mainly in the firmness of the sound and higher air band frequencies rather then the lows as you describe I also did the same with my 251 P to P and had the same results.This is not to say that your board would suffer the same issues as I'm sure you did your homework.  ;)

It's interesting to note that AKG used acrylic for the 251 and C-12 circuit boards and so too did Neuman in the high impedance section of the U47when they could have easily just used the standard circuit board material of the time.


Here is a comparison.

https://soundcloud.com/open-planrecording-studio/c12-point-to-point

https://soundcloud.com/open-planrecording-studio/c12-apex-pcb-mod
 
OPR said:
Great explanation Matador!! But when I built my first acrylic mounted point to point C-12 and compared it to my modded apex 460 sonic differences were mainly in the firmness of the sound and higher air band frequencies rather then the lows as you describe I also did the same with my 251 P to P and had the same results. This is not to say that your board would suffer the same issues as I'm sure you did your homework.  ;)

Here is a comparison.

https://soundcloud.com/open-planrecording-studio/c12-point-to-point

https://soundcloud.com/open-planrecording-studio/c12-apex-pcb-mod

It is important to reiterate that this posted test was performed on the OEM stock Alctron PCB's the specific properties of which we do not know, but likely they are "lowest cost possible".  They are NOT the boards we are peddling in this thread.  Matador's boards are manufactured with tightly controlled  specification, materials, and processes. 

Plus, you put a yellow PCB next to a green one of the same composition, and no question, yellow will win hands-down  ;D

Also, teflon standoffs are provided in the parts kits.  It is a simple matter to drill out the capsule connection solder lugs and press fit the provided teflon standoffs.  Silver teflon wire is also available for making your high impedance connections if you choose to drill.  It's DIY!  You get to choose your own adventure.

I suspect if there is a sonic difference, it would be non-existent or so slight that outside of a head-to-head test with 2 identical microphones (capsule, matched tubes, transformer, electronic components) the only difference being air isolated high impedance vs. board soldered, you would not be able to detect the difference.  I do encourage someone to make that head-to-head test.  If proven wrong and sonic advantages (even slight) are derived from floating part of the circuit, I will gladly concede and change the board or at least the official build documentation/procedure.

Short of conclusive, objective test results, I defer to Matador's design specifications.  There are now hundreds of these mics in service.  Among them, a few have gone head-to-head with point to point vintage originals and held their own admirably.  Also, given the high number of "new" diy'ers building this kit, I am happy with the success rate and that has much to do with the intuitive kit layout.
 

Latest posts

Back
Top