C12A Sony C37a Hybrid

[tardishead]

I borrowed a pair of C37a for a while and I thought they were amazing mics. Very different from the Neumann/AKG tube stuff.
I particularly like the way they handle very large SPL
So I've got a box of 6AU6WA tubes and a HK47 capsule so I thought I would build a hybrid.
And then I thought about being able use both sides of the HK47 and use a polarising scheme like the AKG c12A
C12a I have also used a lot and I think its totally underrated. Again great for high SPL duties like trumpet or guitar cab. Personally I think cathode followers are very useful and can sound huge.
I understand that the K47 style capsule gives out about 10db more voltage than a C37 capsule. So I thought its probably a good idea to keep the headroom potential of the 6AU6 and wire it like it is on the C37a with 250v HT and biased at the mid point of the load line.
Eric mentioned that HK47 capsule can handle 80v polarisation and upwards.
So heres my schematic with these ideas.
I would appreciate any help with the design - especially ideas for the values of c1, c2 and c3 or if someone could point out what the design criteria is for these capacitors. Also I would appreciate if someone could advise me how to calculate what voltage the capsule would give at a given loudness - say 110 decibels. I have not found any info on this.

 

[Deepdark]

Interesting. I thought of something like that some time ago

 

[Matt Nolan]

If you're going to connect the diaphragms and backplates the way you have it in your diagram, you don't need C1 (look at a U47).

Alternatively, you could connect the diaphragms and backplates the way they are in the C12A and also save a capacitor (your C2). In that case, I would decouple your rear diaphragm switchable polarisation voltage a bit more. Again, look at the C12A schematic.

 

[tardishead]

Of course
Take the signal from the backplate! Doh!
But I still need c1 especially in my case because I don't want to mix the grid leak voltage and the capsule polarisation voltage

 

[Matt Nolan]

Of course
Take the signal from the backplate! Doh!
But I still need c1 especially in my case because I don't want to mix the grid leak voltage and the capsule polarisation voltage

Ah - good point. I'm sleepy today!

 

[tardishead]

So the grid resistor is 100m
What is the purpose of the other 2 100m resistors in series with the pattern and polarisation voltages?

 

[Matt Nolan]

Simplistically, we use very large resistors in these positions because we need to get a DC voltage on to the capsule, but we don't wan't to be creating a significant loading on it / path to ground for "hot side" of the signal.

Over and above that, these resistors form RC lowpass filters which help us with noise and RC highpass filters which help us with the LF frequency response of the mic. The resistors need to be large so that the cutoff frequencies are suitably far below the audio range.

I think you'll find they're 200Meg, not 100Meg. In more modern mics, you might find 1 to 3 Gig in those positions, but not in the grid bias of a cathode follower like this because the effect of that resistor for AC signals is multiplied by the bootstrapping effect of the divider in the cathode.

 

[tardishead]

Thanks Matt
What would you recommend in this case?
I took the 100m as grid resistor from the c37a.
With a k47 style capsule (let's say 80pf although I haven't measured it or asked Erik) that makes a HPF at about 20hz
Surely that's low enough?
I think the original c37 capsule is about 50pf so it's lower than that. I was under the impression that the other resistors mirrored a similar impedance. Do you think they should be higher?
Can you explain what filters the other 2 create with what other components?

 

[Matt Nolan]

C3 and the 100Meg form the LPF that keeps the front-end noise level down. The other resistors are involved too, but they are negligible compared to the 100Meg.

As you correctly say, the other 100Meg combines with the capsule capacitance (C1 is negligible if big enough) to give the HPF of the front end of the mic (the output cap and trafo may impose a higher overall corner frequency anyhow).

 

[Gus]

The nice thing about the C60 circuit is there is no grid to capsule cap and the voltage at the grid charges the capsule

I don't understand using a tube as a follower.  I would build a solid state follower.

 

[Matt Nolan]

The nice thing about the C60 circuit is there is no grid to capsule cap and the voltage at the grid charges the capsule

The Sony C37 is the same and, as far as I understand, 2 years before the C60. Though I'm sure somebody probably did this polarising trick even before then.

 

[tardishead]

I started building this circuit which has a polarisation scheme like the c12A

I'm stuck having worked at it for hours. I used an EC70 which I had around instead of a 6AU6.
Voltages are all testing good but I have no bass. Massive roll off. Like everything below 3k or something. Highs sound good. Very frustrating.
I must be missing something in trying to transplant the C37a style circuit onto the C12A.
This is the exact schematic how I have it hooked up at the moment. I have elevated the heaters just like the C37A.
Tried a couple of different output transformers. This one is a gapped UTC 10k:600 H series I think. At first I thought I needed a bigger output cap but a massive 10uf PIO cap had little if any effect.
Checked over my work again and again - I think I have probably made a basic design error.
Do I need a plate resistor like in the C12a schematic?
Can anybody help?

 

[Matt Nolan]

I think the issue may be with the use of the EC70. I'm not sure it has a high enough grid impedance to be used in a condenser mic front end.

 

[tardishead]

Really? Oliver Archut recommends EC70 as a replacement for the AC701
Ac701 was used in CU-2 version of C37a by Sony as a low noise broadcast option.

 

[Matt Nolan]

Really? Oliver Archut recommends EC70 as a replacement for the AC701

I did say "I think" and "I'm not sure" ;-)  Sounds like I'm wrong. In which case, I don't know what the problem is with your build.

Try taking the 10k across the output trafo primary out.

Reduce your circuit and then gradually build it back up. Disconnect all the rear diaphragm stuff, including the 5nF and see if that gets you anywhere. Remember clean fingers, gloved fingers or no fingers around these parts is best! Is it possible you have high parasitic resistance to ground around the grid / capsule backplate area? Are you using a PCB, tag-board?

 

[bockaudio]

I understand that the K47 style capsule gives out about 10db more voltage than a C37 capsule.

based on what?

 

[tardishead]

I understand that the K47 style capsule gives out about 10db more voltage than a C37 capsule.
based on what?

Only what I have read. Based on a greater distance between membrane and backplate. And can handle slightly higher SPL because of this. I will try find where I got that information.

 

[Matt Nolan]

Capsule sensitivity depends upon a lot of factors.

If everything else stays the same, then a higher capacitance capsule will have a higher output, a capsule with smaller membrane to backplate spacing will have a higher output, a larger effective diameter capsule will have a higher output and, within limits, a capsule with lower diaphragm tension or more elasticity will have a higher output (but maybe not at all frequencies). A capsule with a higher polarisation voltage will have a higher output.

Correct me if I'm wrong someone. I'm just thinking from first principles, theoretical and all that.

I'm not sure how you compare a C3 capsule with a K47. The C3 is usually run at up to 145V polarisation, the K47 at anything from 50 to 80V or so. So, how do you level that playing field? For a strict comparison, you should have them running into the same head amp. Differences there will have an impact on how much voltage from the capsule is actually amplified. When I say amplified, that doesn't necesarily mean gain > 1.0

 

[tardishead]

I did say "I think" and "I'm not sure" ;-)  Sounds like I'm wrong. In which case, I don't know what the problem is with your build.

Try taking the 10k across the output trafo primary out.

Reduce your circuit and then gradually build it back up. Disconnect all the rear diaphragm stuff, including the 5nF and see if that gets you anywhere. Remember clean fingers, gloved fingers or no fingers around these parts is best! Is it possible you have high parasitic resistance to ground around the grid / capsule backplate area? Are you using a PCB, tag-board?

Yes I think my build must be too "quick and dirty".
I wanted to build a quick mock up to see if it would work. I have built quite a few mics based on PCBs and/or kits but not designed my own board and layout. Definitely need practise. I thought I would be more likely to lose HF through parasitic capacitance - did not really consider losing so much low end.

So correct me if I'm wrong but when you say "parasitic resistance to ground " that could be a finger print between grid and ground or between capsule diaphragm wire and ground? As opposed to parasitic capacitance which would be components/wires too close to each other. If I understand correctly a finger print can be up to 40-50Mohms which if in parallel with any of those high impedance parts to ground could be the culprit.

I presume I will be able to check the performance of the amplifier (and board parasitics) by swapping the capsule with a 80pf capacitor and sweeping an oscillator and checking the results?

When Matt mentioned that the EC70 might not have enough input impedance I had a look at some data sheets of typical tubes for microphones.
Max grid resistor
AC701 = 180M
EC70/6778 = 0.5M
EC71/5718 = 1.2M (also recommended as sub for AC701)
EF86 when Pa > 0.2W = 3M
5840 = 1.1M
6072 no information in data sheets
6AU6 no information in any data sheets
In this thread http://www.neumann.com/forums/view.php?bn=neumann_archive&key=993770179&v=f Oliver explains that grid resistor in tube mics are 60 or more times the recommended max grid resistor. Probably explains the high value for the AC701. But EC70 does seem a bit low even with the "60x" rule. Maybe EC71 would be a better idea - also has higher max Vhk of 200v so would not need the elevated heaters.

 

[Matt Nolan]

Yeah, fingerprints on components, fingerprints on boards, etc.

That's a nice link to Oliver's post regarding grid leak resistors. Thanks for that.

If you want to test your mic head amp then replace the capsule with an 80pF cap (or thereabouts) in series with the signal generator.