DC-DC Converter for Multi-Pattern Phantom powered mic.

[Khron]

A blurry image of the u87ai DC-DC board:

This is not a picture of the U87ai DC-DC board, but a picture of the (often hated) input circuit of version 04 or 05 of the U87ai. (The red wire is the input, connected to the capsule.) Later Neumann has gone back to a single 2N3819 FET. (From version 06)

Good point - i was kinda wondering what's with the four(!!!) transistors, and no transformer / coupled inductors.

http://cdn.recordinghacks.com/images/mic_extras/neumann/U87-circuit-front.jpg

 

[Drosselmeier]

Seems that in the Pimped Alice Mic the inductors are laid out end to end like so:


where as @Drosselmeier pointed out these chinese layouts with them side by side:
View attachment 108782 View attachment 108783

A blurry image of the u87ai DC-DC board:


I could not find any examples of smd DC to DC circuits with the THT inductors.

Regardless, i made another revision of the board with the 330uf 10v Electrolytic capacitor AND the standup THT inductors. Hopefully this one will do the trick.

Simple circuit, one inductor, very low current:

russian language )))

 

[Khron]

Simple circuit, one inductor, very low current:
View attachment 108794 View attachment 108795

What sort of frequency range does this oscillate in?

 

[RuudNL]

Interesting circuit. Worth to give it a try!

 

[Drosselmeier]

IIRC author said about ~1 MHz

 

[Drosselmeier]

Interesting circuit. Worth to give it a try!

Recommended PCBs

no copper around inductor! ))

 

[Khron]

Recommended PCBs
View attachment 108796View attachment 108797View attachment 108798
no copper around inductor! ))

That's an odd-looking pcb there. Is there supposed to be some intentional capacitive coupling into the base of Q2?

 

[Drosselmeier]

Yes, it's "virtual" cap. From 0.2 to 1 pF. If the transistors are bad and there is no generation, cap must be soldered.

 

[ccaudle]

Might also depend on the style of inductor.

That is a good point. I was thinking of this style:

 

[SparkleBear]

Good point - i was kinda wondering what's with the four(!!!) transistors, and no transformer / coupled inductors.

http://cdn.recordinghacks.com/images/mic_extras/neumann/U87-circuit-front.jpg

Whoopsie! I guess Ive always misidentified that thing… so sorry for presenting mistakes.

I definitely used 0201 (said 0603 metric, thats where I tripped up)... Newest revision has reasonable hand solderable SMD, bigger layout to make up for correct sizing of components for hand assembly.



Its kind of taller than i had originally wanted but it is what it is for now...


This all feels much more reasonable now.

 

[SparkleBear]

Just some ideas and questions about this project:

Doing a 0-120v polarization scheme for the back diaphragm is easy on a tube mic as the high voltage line is already available in the microphone for the B+. I am used to just tapping that off for the capsule and tube needs.

Trying to work this out for my first time on a phantom powered microphone and curious about the best scheme. I had assumed the +/- 60v with backplate at 0v would be easiest. This would mean there would be a constant +60v on the front capsule, also would need to add a dc blocking cap between diaphragm and FET input. There would also be a dc blocking cap between front and back capsule.

What would be the better polarization scheme for having both Omni and Fig 8 along with the Cardioid on a phantom powered condenser mic? Better to do a Backplate @ 0v and front back at +/- 60v? OR 0v front diaphragm with Backplate at 60v and do a voltage doubler for the figure 8 pattern? Having bias on backplate and the diaphragm at 0v would mean no cap to gate... But having the 0v backplate design would mean not needing to boost voltage to 120v for bidirectional pickup.

Just wanted to get the reflections of more experienced designers and builders here. Thanks!

 

[SparkleBear]

Does this work? Is there a simpler solution for disconnecting back diaphragm in card and connecting bias voltage in Omni and Fig8? I came up with a DP3T switch scheme... Will I have problems with Hi-Z going through switch? Float those pins maybe?



EDIT: Looking at it again there is no DC Ground reference on the backplate..... so probably take out C1 and tie BP to ground?

 

[Khron]

Why is C10 there, though?

 

[RuudNL]

Why is C10 there, though?

Because otherwise you would get 1/2 VBias+ on the gate of the FET.
(R10 / (R1+R10)) * VBias+

What concerns me more is that the backplate is 'floating' for DC...
If you put the polarisation voltage on the membranes, you could take the signal from the backplate.
In this case you could connect the backplate through a 1 G.ohm resistor to ground and connect it directly (without a capacitor) to the gate of the FET.
Maybe have a look how Neumann does it in the U87ai.
There is also no need for a double switch, in the center position the capsule is connected as 'pure cardioid'. (Rear membrane not connected.)

Attached is a snippet from a schematic, to show what I mean.
Resistor R9 is optional and is there to discharge the rear membrane in cardioid.
In this situation a capacitor between the backplate and the gate of the FET is needed, but in the situation where the negative gate voltage is created by the voltage drop over the source resistor (as in your drawing) you could omit the capacitor.

 

[Khron]

Oh right, i missed R1 there, i'm used to seeing it "pointing downwards".

I don't hate the Rode NT2-A setup of having a high-megohm discharge resistor between each bias rail to ground, across the pattern switch, actually...

That being said, R2 wouldn't even need to be there...

 

[SparkleBear]

Here is the DC-DC and polarization scheme in one schema:

Does this make more sense now?

What concerns me more is that the backplate is 'floating' for DC...
If you put the polarisation voltage on the membranes, you could take the signal from the backplate.
In this case you could connect the backplate through a 1 G.ohm resistor to ground and connect it directly (without a capacitor) to the gate of the FET.
Maybe have a look how Neumann does it in the U87ai.
There is also no need for a double switch, in the center position the capsule is connected as 'pure cardioid'. (Rear membrane not connected.)

I think this took care of the floating backplate. Can I omit c9 or is that still necessary? Does this scheme act as a disconnected rear diaphragm for pure cardioid?

Oh right, i missed R1 there, i'm used to seeing it "pointing downwards".

I don't hate the Rode NT2-A setup of having a high-megohm discharge resistor between each bias rail to ground, across the pattern switch, actually...

That being said, R2 wouldn't even need to be there...

Interesting, I’m not opposed to getting rid of a $6 resistor if I can! I don’t quite understand the rode scheme. How would I get rid of r2?

 

[Khron]

What's C1 doing there, apart from acting like a -40dB or something pad / attenuation?

Why are C11/C2 connected to R1 and the front diaphragm?

R1 doesn't even need to be there, by the way, and nor does R2 & C9, with the capsuled wired that way.

https://audioimprov.com/AudioImprov/Mics/Entries/2015/12/20_Modding_a_BM-800_Mic.html

 

[SparkleBear]

What's C1 doing there, apart from acting like a -40dB or something pad / attenuation?

Good point! I’ll take that one out.

Why are C11/C2 connected to R1 and the front diaphragm?

That was me trying to follow the pad scheme from the km84 schema.

R1 doesn't even need to be there, by the way, and nor does R2 & C9, with the capsuled wired that way.

Yah I was wondering about the 1000pf c9 cap. I can take that out then. Also, I don’t quite understand why I don’t need the high z resistors r1 and r2.

https://audioimprov.com/AudioImprov/Mics/Entries/2015/12/20_Modding_a_BM-800_Mic.html

 

[Khron]

That was me trying to follow the pad scheme from the km84 schema.

Perhaps, but in that case, the signal is taken off the front diaphragm, not the backplate.

Yah I was wondering about the 1000pf c9 cap. I can take that out then. Also, I don’t quite understand why I don’t need the high z resistors r1 and r2.

See the BM800 link.

Does the capsule wiring not look like the arrangement you're going for? You only care about having say 60v between the diaphragm and the backplate. Since you're taking the signal off the backplate, and that's grounded via R10, the capsule itself is a capacitor (and thus blocks DC). So why would then the front or rear diaphragm need to be "floating" (so to speak), biased through the 1G resistors (as opposed to directly)?

https://cdn.recordinghacks.com/images//mic_extras/akg/C414EB-schematic.pnghttps://gyraf.dk/schematics/AKG_C414_old.GIF

 

[SparkleBear]

Ok so this seems to be what the feedback has led me to...
Removed the high value resistors on the bias lines. Removed the 1000pf cap that connected front and back diaphragm. Changed the c2 to connect to the signal line. How does this look?





Are there high z connection considerations for the switch that connects to the diaphragms? Floating pins and such?

Thanks so much !