U47EF Oliver Archut version

[terry setter]

Three things:
1) Reducing C1 to a very small value will attenuate low frequencies below a cutoff point that is primarily determined by the interaction between C1 and the grid resistor. Single-pole, 6dB/Octave.
2) If you try to do the equivalent by lowering R3 as was suggested, there will be noise-related consequences (depending on how low you go those consequences will be kinda small or very small, but real). Some folks are sensitive to how the grid resistor value affects the overall sound of a mic, but that's much harder to pin down than changes in circuit noise. You can find a lot of info online about the "optimum" grid resistor value, but generally, you'll see that few folks go higher than 1 gig (yes, there are notable exceptions) because Shot noise gets significant when the grid resistor gets very, very large like that. And, as mentioned earlier in this thread, going below 60 Megs might possibly take you into non-Neumann sounding results. I don't know how low you'd have to take R3 to get the reduced bass you're seeking, but the Elam 251 used a 30 Megs (and were not known for being overly quiet - for a number of contributing reasons). Summary: making C1 smaller has got far less baggage associated with it than reducing R3.
3) Your measured frequency response is so rough that I'm wondering if those are acoustic response measurements, rather than plain old "mic amplifier circuit-olny measurements" made with a 50-70pF cap to replace the actual capsule and fed into, say 1500 Ohms. If your frequency response chart is of acoustic pickup in a room, then peaks and valleys in your room's bass response are what you've been seeing on your measurements. Even if you are using gated measurement techniques, your results in all but the largest enclosed spaces will be suspect below about 200HZ and get more and more unreliable as the frequency goes down. I would expect the electronics of your mic to be much, much flatter than the chart you've posted.

 

[rock soderstrom]

Thank you Terry for your feedback.

Your measured frequency response is so rough that I'm wondering if those are acoustic response measurements, rather than plain old "mic amplifier circuit-olny measurements" made with a 50-70pF cap to replace the actual capsule and fed into, say 1500 Ohms. If your frequency response chart is of acoustic pickup in a room, then peaks and valleys in your room's bass response are what you've been seeing on your measurements. Even if you are using gated measurement techniques, your results in all but the largest enclosed spaces will be suspect below about 200HZ and get more and more unreliable as the frequency goes down. I would expect the electronics of your mic to be much, much flatter than the chart you've posted.

Replacing the capsule with a small cap would actually be a good way to see what is really happening circuit-wise.

 

[Roman Beilharz]

Three things:
1) Reducing C1 to a very small value will attenuate low frequencies below a cutoff point that is primarily determined by the interaction between C1 and the grid resistor. Single-pole, 6dB/Octave.
2) If you try to do the equivalent by lowering R3 as was suggested, there will be noise-related consequences (depending on how low you go those consequences will be kinda small or very small, but real). Some folks are sensitive to how the grid resistor value affects the overall sound of a mic, but that's much harder to pin down than changes in circuit noise. You can find a lot of info online about the "optimum" grid resistor value, but generally, you'll see that few folks go higher than 1 gig (yes, there are notable exceptions) because Shot noise gets significant when the grid resistor gets very, very large like that. And, as mentioned earlier in this thread, going below 60 Megs might possibly take you into non-Neumann sounding results. I don't know how low you'd have to take R3 to get the reduced bass you're seeking, but the Elam 251 used a 30 Megs (and were not known for being overly quiet - for a number of contributing reasons). Summary: making C1 smaller has got far less baggage associated with it than reducing R3.
3) Your measured frequency response is so rough that I'm wondering if those are acoustic response measurements, rather than plain old "mic amplifier circuit-olny measurements" made with a 50-70pF cap to replace the actual capsule and fed into, say 1500 Ohms. If your frequency response chart is of acoustic pickup in a room, then peaks and valleys in your room's bass response are what you've been seeing on your measurements. Even if you are using gated measurement techniques, your results in all but the largest enclosed spaces will be suspect below about 200HZ and get more and more unreliable as the frequency goes down. I would expect the electronics of your mic to be much, much flatter than the chart you've posted.

Thank you so much, Terry, for taking your time to elaborate on this.

Regarding
1) Ok, wait, maybe I really got it all wrong. I was trying to reduce the value of the transformer cap in the output path, which in my schematics is C5. And I guess you are not talking about the 1000pf one named C1 here, are you?



2) So when the side effect of a further reduction of R3 might implicate more noise, then I'd rather stick with the circuit as it is now - given that replacing the 1G mechanically right now would be no easy and quick operation to accomplish. As I am used to pretty quiet modern FET designs; the current state is usable: no hiss, no hum, no audible artifacts, but the spectrum shows me quite some sub bass flicker/rumble around the -60dB area. Hence my plan to insert a LC filter somewhere before the transformer. I would not be happy with more noise, as I am not recording Heavy Metal vocals - so my SNR really matters.

3) Yes, this is a true overall acoustic response with all the bumps and notches the mic (body, headbasket, capsule etc.) is introducing. It's a miracle that this mic sounds so amazing given that dragon tail of a graph. But as I have divided the actual measurement by a calibrated omni mic measurement resulting in the plotted graph in my post, the room and speaker should not be overly present here. Compared to other standard mic measurements done under identical circumstances in relation to what I can hear in my audio, the measurement truly reflects the sound of the circuit within the mic:

I am not getting less bass choosing smaller values than 1uF for C5, I get more of it. Transformer resonance or not, the maths seem to be off. But maybe the explanation sits in I tampered around with the wrong cap (see 1)?!

This being said I am extremely happy with the mic now, but sure and still I would want to understand what was going on there.

P.S.: I have just lowered H+ further to about 5.4V by the way. Thanks for that tip!

 

[Murdock]

You did not tamper with the wrong capacitor. C5 is one of the spots to change for bass response.
I don't know if Terry meant that there is also a interaction with R3 and C1. That's possible.
But I think modifing C1 does not really change the response.

Replacing the capsule with a capacitor is a great idea to evaluate the circuit response.
Like I said, lowering C4 to something between 10uF and 22uF should also result in a bass roll-off. Maybe worth a try. 47uF is quite high. Most vintage mics had something like 22uF.

 

[Spencerleehorton]

I’ve ordered a EF12 tube and a baker lite connector as well as this P2P version which I’m going to try in one of the U47 bodies.
Also got a few other bits off Ali babba, m7 capsule and bv08 to try and see what quality they are!!

 

[rock soderstrom]

I’ve ordered a EF12 tube and a baker lite connector as well as this P2P version which I’m going to try in one of the U47 bodies.
Also got a few other bits off Ali babba, m7 capsule and bv08 to try and see what quality they are!!

Very interesting, I'm curious how the EF12 performs compared to the EF800.

 

[rmburrow]

Thank you so much, Terry, for taking your time to elaborate on this.

Regarding
1) Ok, wait, maybe I really got it all wrong. I was trying to reduce the value of the transformer cap in the output path, which in my schematics is C5. And I guess you are not talking about the 1000pf one named C1 here, are you?

View attachment 112510

2) So when the side effect of a further reduction of R3 might implicate more noise, then I'd rather stick with the circuit as it is now - given that replacing the 1G mechanically right now would be no easy and quick operation to accomplish. As I am used to pretty quiet modern FET designs; the current state is usable: no hiss, no hum, no audible artifacts, but the spectrum shows me quite some sub bass flicker/rumble around the -60dB area. Hence my plan to insert a LC filter somewhere before the transformer. I would not be happy with more noise, as I am not recording Heavy Metal vocals - so my SNR really matters.

There are several time constants involved here. The first is the 2 1G resisters (effectively in parallel at audio frequencies) across the capsule so T=RC relates to T = 500x10exp6 x 40 exp10-12 where 40 is the assumed capsule capacitance in pf and 500M is the equivalent resistance of the two 1G resistors in parallel. The 2nd RC is the 1 uF cap in series with the output transformer primary (henries). The last RC combination is the 1k cathode resistor in parallel with the 47 uF cathode bypass capacitor. I personally like fixed grid bias derived from a divider on the well filtered heater supply since that eliminates the cathode time constant. The 1000 pF capacitors at C2 and C6 are bypass capacitors and appear as a short circuit with respect to audio at the high working impedance. C1 is a coupling capacitor and appears as a short with respect to audio at the high working impedance.

Dr. Schoeps used 180 megohm resistors in the M221a mics. Those mics make GOOD classical music recordings!

3) Yes, this is a true overall acoustic response with all the bumps and notches the mic (body, headbasket, capsule etc.) is introducing. It's a miracle that this mic sounds so amazing given that dragon tail of a graph. But as I have divided the actual measurement by a calibrated omni mic measurement resulting in the plotted graph in my post, the room and speaker should not be overly present here. Compared to other standard mic measurements done under identical circumstances in relation to what I can hear in my audio, the measurement truly reflects the sound of the circuit within the mic:

I am not getting less bass choosing smaller values than 1uF for C5, I get more of it. Transformer resonance or not, the maths seem to be off. But maybe the explanation sits in I tampered around with the wrong cap (see 1)?!

This being said I am extremely happy with the mic now, but sure and still I would want to understand what was going on there.

P.S.: I have just lowered H+ further to about 5.4V by the way. Thanks for that tip!