Thiersch PVC M7 backplate voltage

[AusTex64]

I discovered that my AMI U47 build's voltage divider supplies 72V to the backplate from 120V from the PSU. I understand 72V should be no problem, and will increase sensitivity and improve signal to noise. But how does this affect patterns when the max voltage I can get to the rear diaphragm is 120V in fig 8?

BTW, the  mic sounds great with 72V. Of course I haven't heard it with 60V, so no frame of reference. Trying to decide if messing with it is worth the effort, since it sounds so good now. Might have been a "happy accident".

 

[Khron]

Does that U47 even have fig-8 pattern? The original sure didn't Omni doesn't care what voltage it's biased at, as long it's the same as the front half.

 

[ln76d]

You can always make voice test for best, suitable, polarisation voltage.
Just need to use potentiometer instead resistor divider.
I always make that in every microphone i build or mod.
Results can be totally different.

 

[Matt Nolan]

You will get something between a hypercardioid and a figure-8, closer to figure-8 (basically, an unequal figure-8 with stronger pick-up from the front).

U47 was Card/Omni
U48 was Card/Fig-8

Is the AMI build 3-way pattern?

 

[ln76d]

Rather not

http://www.tab-funkenwerk.com/id85.html

 

[Khron]

That's what a quick googling brought up, before i posted my reply / question

Rather not

http://www.tab-funkenwerk.com/id85.html

 

[Matt Nolan]

Rather not

http://www.tab-funkenwerk.com/id85.html

The "alternate" schematic allows for multi-pattern.

 

[ln76d]

Sorry, my fault i opened site only with view on one schematic

Despite all the obvious differences - C4, permanently connected, will change sensitivity and overall sound of cardioid pattern.
10nF is also too big. I would go for few pF range in this case.

 

[granger.frederic]

Sorry, my fault i opened site only with view on one schematic

Despite all the obvious differences - C4, permanently connected, will change sensitivity and overall sound of cardioid pattern.
10nF is also too big. I would go for few pF range in this case.

1nF is fine

with remote polarity , if you achieve "perfect" same voltage on the back plate vs the front plate , you will have quite the same response than a pure cardio
if you can't , the difference is subtle but exists
the SNR is a bit less good also (depends on yours B+ and remote voltage quality)

 

[ln76d]

Not exactly, try for example, even for different pattern like hypercardioid, to use 1nF and 3pF via switch. Sensitivity is different. Not without a reason M49C had optional switch for cardioid pattern only to disconnect completely front diaphragm from the back.

 

[granger.frederic]

here's my interpretations : C4 is here to invert the rear signal phase and "mix" the rear signal into the front
3pf is too small and creates a phase issue in the audio range , 1nF min...
pure cardio switch is mainly here for SNR reason

 

[ln76d]

So how the phase inversion works when we compare M49 input to the U47 input topology?
We can compare omni pattern here - so one is with the cap and the other is directly connected, in both case there is 0V.
I did tests for cardioid pattern, polarisation voltages were matched, three positions - fully disconnected, 1nF, 3pF - this was last configuration. I also tried few different values starting from 10nF. Difference was truly audible and it was not only SNR difference, definately. 3pF was closest to the fully disconnected diaphragms.
Also difference in both diaphragms polarisation voltages, when unmatched by few volts, can change the polar pattern for several frequencies. It moves up subcardioid response to the midrange, sometimes looks much more even like omni. Figure of eight can be also much more omni for several frequnecies.

 

[Delta Sigma]

Most designs seem to use 10pF.

Wouldn't the size of C4 be dependent on the size of R1 and R10 to get the desired filtering and phase results?

 

[ln76d]

Most designs, especially "classic" use 1nF. On few modern you can see 10nF, but i don't remember any commercial design with higher cap than 1nF. Much way better is C12 input arrangement and the best is with negative voltage where you can throw out also input capacitor.
With the cap between diaphragms also higher R10 makes worse response (yes i know all the noise issues etc. but still noise isn't so audible than overall this topology arragement).  Anyway - especially for chinese capsules - with 10nF and 1G all the unpleasant bumps in the response are even much more nasty.

 

[granger.frederic]

So how the phase inversion works when we compare M49 input to the U47 input topology?
We can compare omni pattern here - so one is with the cap and the other is directly connected, in both case there is 0V.
I did tests for cardioid pattern, polarisation voltages were matched, three positions - fully disconnected, 1nF, 3pF - this was last configuration. I also tried few different values starting from 10nF. Difference was truly audible and it was not only SNR difference, definately. 3pF was closest to the fully disconnected diaphragms.
Also difference in both diaphragms polarisation voltages, when unmatched by few volts, can change the polar pattern for several frequencies. It moves up subcardioid response to the midrange, sometimes looks much more even like omni. Figure of eight can be also much more omni for several frequnecies.

u47 can do omni only , thus no  need to invert the phase as both diaphragm have the same voltage : 0V
but in figure 8 , it is different : the rear is in opposite phase, for ex: u48 uses 1nF between the two diaphragm.
as the capacitor is only coupling the signal , with 1nF or more , there will be very subtle sound variation inherent to the material (polystyrene is better) and the non linearities induced by the capacitor.
under 1nf the variation change is due to the created  HPF on the rear diaphragm signal

the HighZ resistors around the capsule provide the polarization with a near "static" voltage , increasing the value to 1Gohms decreases the capsule THD  but doesn't provide any "bump" with c4 , or i don't see how if c4 is min 1nF ...
if it's an improvement or not , often but your ears are the judges ...

 

[Delta Sigma]

Most designs, especially "classic" use 1nF. On few modern you can see 10nF, but i don't remember any commercial design with higher cap than 1nF. Much way better is C12 input arrangement and the best is with negative voltage where you can throw out also input capacitor.
With the cap between diaphragms also higher R10 makes worse response (yes i know all the noise issues etc. but still noise isn't so audible than overall this topology arragement).  Anyway - especially for chinese capsules - with 10nF and 1G all the unpleasant bumps in the response are even much more nasty.

I don't think my brain and my little typing fingers were communicating there. I meant to say "many", wondering if the cap size is related to the grid resistor. I had to give my head a shake when I read back my post. I haven't looked at enough schematics to say most and the build I just finished (251) and the one I'm starting (M49) use 1n!

In fact, I probably haven't even looked at enough schematics to say "many".

I recently sold a mic I loved and sketched out a schematic before selling it. It used 10n (unless I read it wrong, cramped in there!). I would call it a modern boutique mic, definitely not a chinese capsule and a very similar design to Oliver's alternate U47 schematic.