Circuit idea for switchable patterns?

[RuudNL]

I recently got my hands on a Behringer C-3.
I already own a couple of Behringer B-5 microphones, and these have a very low self noise and sound great, certainly considering the price.
Unfortunately, the C-3 is rather noisy. 
I took a look inside and noticed that they use a TL072 as head amplifier...
Pattern switching is done with a CD4066B. As far as I see the capsules look very much like the ones used for the B-5.
I am thinking about replacing the Behringer circuit with a new (low noise FET) head amplifier.
Of course I could use the classical Schoeps circuit, but I would like to keep the switchable patterns.
My first thought would be to use two FET input stages for the two (back-to-back placed) capsules, and do the pattern switching after the first amplifier stage.
I could imagine this is correct: (Is it???)

- Cardioid : Front capsule only
- Omni : Front and back capsule (electrically) in phase
- Figure 8 : Front and back capsule (electrically) in opposite phase.

I am wondering if there would be an alternative solution. The 'cold' sides of both capsules are connected together.  Suggestions are welcome!

 

[RuudNL]

I was roughly thinking about something like this. Suggestions?

Upper capsule = Front
Lower capsule = Back
(Capsules are cardioid)

1 = Omni
2 = Cardioid
3 = Figure 8

 

[ricardo]

You might like to use the FET sources as the main feeds to the mixer instead of the drains.  Bit better noise & more sensible cos lower output Z.

The drains have output Z which is the drain resistor in parallel with the intrinsic drain resistance.  You will have to tweak it and the mixing resistors to get correct mixing.  Do that on the Fig-8 setting.

I thought the capsules were electret and didn't need Upol

 

[RuudNL]

Yes, you are right that the impedance on the source is lower than on the drain.
But the gain of the FET stage is 1x (0 dB.).
I doubt if the noise on the source will be lower in this case, but I will check this.
Rs = Rd = 2200 ohms, so the impedance is already low.
Unfortunately I can't avoid to use both source and drain for the back capsule.

I thought the capsules were electret and didn't need Upol

The capsules look very much the same as used in the B-5, so I assumed they needed a polarisation voltage.
I will check this today! If they are electrets, that would make things even easier!

EDIT: I just checked it and they are indeed electrets! 

 

[zebra50]

If you have separate head amplifiers, why not take both stages out to the mixer. like in the Josephson C720?

You can then choose the pattern at mix down!

 

[RuudNL]

I thought about this in a flash...
Disadvantage is that you get a non-standard microphone.
I would have to replace the connector with a 5 pin XLR and use a 4 wire screened cable.
I will give it a second thought!

 

[ricardo]

Yes, you are right that the impedance on the source is lower than on the drain.  ...  Rs = Rd = 2200 ohms, so the impedance is already low.

If you are using a good FET, the source impedance will be much less than the drain (which will be a little less than 2k2).

The critical mixing is for the Fig-8 (unless your capsules are very mismatched)  In your new circuit, make the drain resistor adjustable.  It is effectively in series with the mix resistor.  (You don't need a drain resistor on the other FET).  Tweak it for a symmetrical Fig-8

With electrets, you don't need the input caps.

If you put component numbers on your circuit, it will be easier to discuss.

 

[RuudNL]

Schematic updated and component numbers placed.

 

[ricardo]

You may want to put R8 on p1 of the switch rather than the slider.

R4 is effectively in series with p3 of the switch so this would make the adjustment easier.