Official C12 Clone - Build and Support Thread

[Matador]

Still fighting the noise...

My PSU provides 160V instead of 200V, so I dropped the resitors from 91K to 48K... by doing so, I realize that I have also reduced the ripple rejection by about 10dB. Is the combined ripple rejection of the PSU so incredibly low that it really does not matter? Or should I exchange C1 and C2 for 200uF capacitors?

I doubt this is the cause:  even 48K into 100uF is -3dB at 0.033 Hz, as we are talking three cascaded stages!  There isn't even a hint of 120Hz PSU ripple in your plot.

The low-frequency response peak is not unusual - the output impedance is roughly 20K, and with 0.5uF (C12) into 50H (output transformer), you end up with an RLC resonant circuit  with a resonance frequency of about 30Hz.  What transformer are you using?  Are you using the stock 100K plate resistor (R17)?

There are many tweaks that can be done to change this (in order of difficulty):

1) Engage the HPF on your preamp
2) Decrease R15 down to 100M
3) Increase C12: 1uF moves the resonance down to 22Hz, and 2uF moves it down to 16Hz.
4) Decrease R17 (e.g. move to a different tube like the 12AT7)
5) Move to CDDA topology

Perhaps post a new recording of the noise after cleaning?

Grounding the grid (pin 2 if you used the "first" triode, pin 7 if you used the "second") should have rendered the mike silent (modulo the tube self-noise) - if it hums there may be other problems.

 

[dmnieto]

I doubt this is the cause:  even 48K into 100uF is -3dB at 0.033 Hz, as we are talking three cascaded stages!  There isn't even a hint of 120Hz PSU ripple in your plot.

The low-frequency response peak is not unusual - the output impedance is roughly 20K, and with 0.5uF (C12) into 50H (output transformer), you end up with an RLC resonant circuit  with a resonance frequency of about 30Hz.  What transformer are you using?  Are you using the stock 100K plate resistor (R17)?

There are many tweaks that can be done to change this (in order of difficulty):

1) Engage the HPF on your preamp
2) Decrease R15 down to 100M
3) Increase C12: 1uF moves the resonance down to 22Hz, and 2uF moves it down to 16Hz.
4) Decrease R17 (e.g. move to a different tube like the 12AT7)
5) Move to CDDA topology

Perhaps post a new recording of the noise after cleaning?

Grounding the grid (pin 2 if you used the "first" triode, pin 7 if you used the "second") should have rendered the mike silent (modulo the tube self-noise) - if it hums there may be other problems.

There is actually several noise contributions:

a) 1/f flicker
b) ~28Hz RLC resonance
c) hum at 60Hz (-96dB), 120Hz(-102dB)

I am already using a 1uF at C12, which is a bit weird because it shows resonance at almost 30 Hz... but I agree that that particular one is not a problem.
I also have approximately 5dB of NF higher that expected, but this seems to be related with the tube, which I intend to upgrade in any case.

My biggest concern is the hum at 60Hz, when I grounded the grid the mic started to hum quite a lot, but that was also when I did not put the metal sleeve over the microphone, which may be the reason it was doing it. I am checking the PCB for bad solders around ground.

Something I realized is that I was so concerned with burning the styroflex capacitors that I solder them very poorly, I measure continuity, but they are not cute solders. Could the hum come from them? C13 is not bad, but C11 and C10 are a bit on the crappy side.

Would you recommend giving them a bit more of solder?

 

[dmnieto]

And C10 was floating on the ground pad....

I don't need to publish the sounds anymore, just the pictures. Same chain as before, same settings. The peaks at around 3KHz are the halogen in the room.

Matador, yours is a very very good pcb design. Hats off.

 

[Matador]

I'm glad it's sorted out.

Tube mikes don't exactly have the best noise performance, but in general this should be pretty quiet.

 

[dmnieto]

One last question...

I have read somewhere that the C12 should be biased to -1.1V, but in this design it seems to be biased to -1V... or there is some current somewhere that drops an additional 0.1V before the grid?

 

[Matador]

One last question...

I have read somewhere that the C12 should be biased to -1.1V, but in this design it seems to be biased to -1V... or there is some current somewhere that drops an additional 0.1V before the grid?

There should be almost no current causing drops, aside from grid leakage.

You can select whatever operating point you feel is best for your particular tube:  -1V seems to be about the sweet spot for most tubes (at least the ones I have checked).  Given the grid can go a bit positive and still remain linear a -1V bias gives a reasonable input range.

The best way to check it for your tube in your mike is to do a signal swing test, where you inject a signal into the grid and check for symmetric signal swing at the output.  You may find a few hundred millivolts up or down on the bias gives better symmetry and less THD if you want to squeeze out every last drop.

If you are close miking very loud sources, causing overload of the input section, we can discuss further.  You can tweak the load-line to reduce gain and increase headroom in these cases (basically you lower R17, lower R1 and R2 to compensate, then re-tweak the bias more negative).  You can increase the input headroom by 500mV or so with some simple changes, however gain does drop too.

 

[ding]

On my end, I will begin the official build documentation for the ELA M251 configuration very soon as well as modify the C12 documentation to reflect the 12AT7 option.

Looking forward to this!!!!!! Once done I will be hitting you all up for a M251 kit.  ;D

 

[crossed.out]

Hi there,

I tackled this build last week. Things went smoothly, but the microphone has incredibly low output.. next to nothing at all, but if I max out my pre's, I can hear that it is passing audio. I've double-checked and re-flown joints, as well as double checked pin layout, etc.. A couple of weird things I encountered though:

1. I couldn't get a consistent read on the transformer leads.. this leads me to think this is where the problem lies.
2. When measured, Bias starts out at -1 then continuously falls.. I assume this has to do with strapping a meter across it?

All other voltages can be set and read correctly. Heater at 6.3v, B+ at 120v.. I'm using a Peluso CEK-12 right now.

Anyone have any thoughts? Did I get a bunk transformer? I realize the wires on these can be quite delicate, but I took extra care with them.

Cheers.

 

[Matador]

Bias can only be read right at where it emerges from the PSU.  If you tried to read it right at the tube grid all bets are off.  Otherwise it should be stable.

There was a problem in many builds (also in many U87 builds) where the transformer leads were swapped and/or mislabeled.  This will cause extremely low output.

The two pairs of leads should be separate colors:  one will read high(ish) resistance, perhaps 500-1K ohms, and the other will read low(ish) resistance, perhaps 50 ohms or less.  There should be no continuity (extremely high resistance) between any two wires of different colors (primary and secondary should be isolated).  The high value goes to the primary (XP1 and XP2), and the low value gets soldered to the output pins (pin 5 and 6).

 

[crossed.out]

Bias can only be read right at where it emerges from the PSU.  If you tried to read it right at the tube grid all bets are off.  Otherwise it should be stable.

There was a problem in many builds (also in many U87 builds) where the transformer leads were swapped and/or mislabeled.  This will cause extremely low output.

The two pairs of leads should be separate colors:  one will read high(ish) resistance, perhaps 500-1K ohms, and the other will read low(ish) resistance, perhaps 50 ohms or less.  There should be no continuity (extremely high resistance) between any two wires of different colors (primary and secondary should be isolated).  The high value goes to the primary (XP1 and XP2), and the low value gets soldered to the output pins (pin 5 and 6).

Hey Matador, thanks for the reply. I measured the bias from the terminal blocks inside the PSU.

I've tried swapping the direction of the transformer as well, thinking I must have had it backwards.. no dice. I was getting weird readings on my multimeter from the transformer as well though.. I couldn't really get it to display the readings I was expecting.. it would not give me a solid reading of either side, and they both ended up being within the same range as eachother.. both in kohm according to my meter.

 

[chunger]

Bias can only be read right at where it emerges from the PSU.  If you tried to read it right at the tube grid all bets are off.  Otherwise it should be stable.

There was a problem in many builds (also in many U87 builds) where the transformer leads were swapped and/or mislabeled.  This will cause extremely low output.

The two pairs of leads should be separate colors:  one will read high(ish) resistance, perhaps 500-1K ohms, and the other will read low(ish) resistance, perhaps 50 ohms or less.  There should be no continuity (extremely high resistance) between any two wires of different colors (primary and secondary should be isolated).  The high value goes to the primary (XP1 and XP2), and the low value gets soldered to the output pins (pin 5 and 6).

Hey Matador, thanks for the reply. I measured the bias from the terminal blocks inside the PSU.

I've tried swapping the direction of the transformer as well, thinking I must have had it backwards.. no dice. I was getting weird readings on my multimeter from the transformer as well though.. I couldn't really get it to display the readings I was expecting.. it would not give me a solid reading of either side, and they both ended up being within the same range as eachother.. both in kohm according to my meter.

Make sure your meter is suitable for this sort of work and use xlips and see if you can get a stable reading on the transformer.  When I am touching the transformer, sometimes it is hard to get the reading to settle down.

When did your kit ship?  Due to unusually high number of AMI failures recently, I have been individually testing all transformers before packing.

Failure is usually on the blue wires ~950R there will be no continuity.  The red wires should be ~22R.

 

[stelin]

About backplate bridging:
In the build instructions, you show Tims capsule with the backplates bridged together at the capsule. So, Tims capsule have separate terminals for each backplate. So does the Beesneez CK12 (I just got two yesterday).
My questions are:
Why are the backplates isolated in the first place, if they have to be bridged at the capsule? Are there other designs that use both backplates in there "isolated" state?
Why is it important to bridge the  backplates at the capsule end? Does that have anything to do with the high impedance of the circuit?

I have a basic understanding of electronics, but I'm lost when it comes to the specifics of microphones... Any input would be appreciated!

 

[trans4funks1]

Take a look at the ELAM 251 circuit and you can see the back plates of the C12 capsule being used independently.

 

[stelin]

Thank you trans4funks1! I see it. The "back end" can be switched of in the ELA M.
I'm still curious about why the bridging must be done at the capsule end and not at the circuit bord.

 

[Category 5]

Thank you trans4funks1! I see it. The "back end" can be switched of in the ELA M.
I'm still curious about why the bridging must be done at the capsule end and not at the circuit bord.

I did mine at the board in case I want to re purpose the capsules ever.  No I'll effects yet.  ;-)

 

[Matador]

It can be bridged at either end:  the electrons won't care.

I made the layout separate in anticipation of an integrated ELA M251 version that I never got around to doing.

 

[stelin]

Thank you, guys. I will do the bridging at the bord and leave the capsules alone.

 

[Dennyp]

Hey Matador,

I finally got the power supply to where it should be pre connecting the mic but I was wondering about connecting the transformer.  in the guide it says to determine which side to use for audio - (pin 5) and audio + (pin 6 ).  is this done arbitrarily? I know that the higher resistance side is going to the PCB and the lower to the XLR but how do I know? sorry if you answered this before, seems like i should know this  :-[

Denny

 

[Matador]

Hey Matador,

I finally got the power supply to where it should be pre connecting the mic but I was wondering about connecting the transformer.  in the guide it says to determine which side to use for audio - (pin 5) and audio + (pin 6 ).  is this done arbitrarily? I know that the higher resistance side is going to the PCB and the lower to the XLR but how do I know? sorry if you answered this before, seems like i should know this  :-[

Denny

The answer is above in post 1228.

 

[Dennyp]

Hey Matador,

I finally got the power supply to where it should be pre connecting the mic but I was wondering about connecting the transformer.  in the guide it says to determine which side to use for audio - (pin 5) and audio + (pin 6 ).  is this done arbitrarily? I know that the higher resistance side is going to the PCB and the lower to the XLR but how do I know? sorry if you answered this before, seems like i should know this  :-[

Denny

The answer is above in post 1228.

right, thanks matador!  the cinemag transformer has 4 different colors and although it is easy to isolate the pairs with high and low impedance my question is does it matter what color i chose for the - audio or the + audio. should i stick with what sides they are coming from or does it not matter?