Some questions about Schoeps circuit variations + can it be used with a ribbon?

[midwayfair]

Hi, everyone. New here but not to DIY. (I build and design guitar stuff.)

I had some questions about the Scoeps circuit, Dorsey's variation of it, and a couple other ways to use it. I've read as many threads as I could find, but an awful lot of them seem to take some knowledge as read, which makes it hard to suss out a few things.

Here's the original circuit (large image):

http://www.sdiy.org/oid/mics/Schoeps.gif

Here's Dorsey's modified version, from Making Mics Better:

First, I have to say, I find this circuit incredibly elegant in the way it uses the output buffers to create the supply voltage for the input amplifier. I feel kind of silly looking for ways to change it, but some things are bugging me.

I'll run through my questions one at a time.

1) The charging voltage for the cap: Dorsey's circuit creates a 35V circuit, from which an incredibly tiny amount of current and voltage is applied to the center of the capsule. But in the original Schoeps circuit, one side of the capsule is held above ground by a 1M resistor (0v, but not GROUNDED grounded), which lets them switch in a cap for a pad (or boost?), and the center of the capsule goes directly to the gate of the FET. The center of the capsule in the original is biased with a tiny amount of current through the 1G gate resistor.

2) It seems like Dorsey's uses not only a few extra resistors to create the 35V supply voltage but that extra 1G resistor is quite expensive (I'm seeing them for $13 each at mouser). Was this likely so he didn't have to put any current/voltage on the FET's gate at all, or is there some sort of benefit to using a higher (35V) reference voltage in the Dorsey circuit?

3) The Schoeps circuit used a trimpot for the reference voltage on the gate/capsule -- could I do anything meaningful here when setting this using just my ears, or is there a nominal setting?

4) Can I use the Schoeps circuit with a ribbon mic with a transformer (either 1:35 or 1:110) as follows: Omit the 35V reference voltage; use a 1M for the gate resistor; ground one transformer lead and send the other to the input cap (to make sure there's never any DC on the ribbon). Is there any danger to the ribbon if I do this? Would there be there any benefit at all to doing that over, say, PRR's circuit here (which has enough gain for what I want and is super tiny):

http://groupdiy.com/index.php?topic=5743.80

[Tangent: Somewhere in that thread, though I couldn't find exactly where, it was suggested that the input caps could be left out of that circuit because the voltage should be identical on both sides of the . Is that safer than having them in, like if charge builds up on one side of them (in which case -- couldn't I just put a pull-down resistor from in1 and in2 to ground?

One last thing: Before someone just says, "just go experiment" -- which normally I'm happy to do -- I can't cut my own ribbons, and they're really frigging expensive and not the sort of thing I can replace at the drop of a hat if I blow one up while messing around. If there's a danger to the ribbon with something I've described above, I'd really appreciate it if someone just said so.]

Thanks for any help you can give.

Also, this forum's verification images are literally impossible to read.

 

[Monte McGuire]

The 'ribbon or condenser' part will not be addressed in detail here, but I think I can clear up a little re. the 1GΩ resistors and such. The Schoeps circuit extracts a high voltage from P48 in order to provide a high voltage to polarize a condenser capsule. This voltage is typically 40V or more, and as you noted, is essentially an open circuit load - the condenser capsule itself. You mentioned 35V, so I guess that is what the capsule is polarized with.

The reason for the 1GΩ load on the capsule is to reduce the noise of the circuit. If you look from the JFET gate backwards (towards the capsule), you have a source impedance consisting of the parallel connection of the capsule capacitance, the 1GΩ polarization resistor, and the JFET noise current. To get the mike to have a reasonable LF response, the polarizing resistor (the 1GΩ resistor) needs to be large, so that the product of the capsule capacitance and the resistance results in a long time constant. Capsules are anywhere from 15pF to 80pF or so, and this is really a small capacitance, so the polarizing resistor needs to be really large to allow any sort of LF to be transmitted from the capsule, at least 200MΩ or more.

It turns out that if you double the value of the polarizing resistor, the noise caused by increasing the size of the polarizing resistor gets larger by a factor of sqrt(2), but the resulting filter formed by the capsule capacitance in parallel with that resistor filters out (lowpasses) that noise by a factor of 2. So, double the polarizing resistor, and you get sqrt(2), or 3dB, less overall noise.

Some of the other choices that were made in Dorsey's circuit I think have to do with whether the 35V polarizing voltage is on the outside (the diaphragm) or the inside (the backplate) of the capsule. Modern mikes place the diaphragms at ground and polarize the backplates so that the capsule doesn't turn into an electrostatic air cleaner, soaking up all of the smoke and diesel soot in the air, onto the capsule. By polarizing the backplate, the capsule works just the same, but the parts exposed to the dirty air are grounded, so they don't accrete dirt and contamination over time.

I could address using this circuit with a ribbon, but as a first guess, a ribbon preamp needs a _lot_ of current to bias some low noise voltage bipolar transistors to make a  suitably low noise preamp. This is pretty much at the other far extreme of voltage/current scaling… tens of milliamps at a few volts versus femtoamps of leakage current at high voltages.

I don't know who started this whole craze of having ribbon mikes with amplifiers built in, but IMHO it's sorta silly. What the heck is wrong with a ribbon mike that has an output transformer, plugged into some cable connected to a real preamp? The fun of that is that the nature of the preamp loading gets reflected back through the transformer into the ribbon itself - it affects the dynamics and damping of the ribbon!

 

[Monte McGuire]

One final comment about the whole 35V / 1GΩ bias to the capsule, and I apologize if this is something obvious to you. A condenser capsule only develops an output voltage if it is polarized. Not to throw equations at you, but the basic idea is that if you change the capacitance of the capsule, and the charge on the capsule is constant, then the voltage across the capsule will reflect the change in capacitance. This is a result of the basic capacitor equation of q = C * V. If q, the charge on the capsule, is constant, then if you vary 'C', the capsule capacitance, which is determined (among other things) by the spacing of the diaphragm to the backplate, then you will change 'V', the voltage across the capsule, and you can amplify that with a very high impedance amplifier, such as a tube or JFET. Why does the capsule's capacitance change? Because sound pressure has moved the diaphragm relative to the backplate.

This is why the polarizing voltage has to be applied through such a huge resistor - the capsule really wants to be charged with a fixed amount of electrons, and if the diaphragm and backplate move relative to each other, the capsule changes capacitance and thus develops a voltage that can be amplified by the JFET, without being loaded down.

 

[midwayfair]

Monte, thanks so much for taking the time to type all that out. I think I have  a better handle on how it works now.

Some of the other choices that were made in Dorsey's circuit I think have to do with whether the 35V polarizing voltage is on the outside (the diaphragm) or the inside (the backplate) of the capsule.

This is a really good observation and definitely not something that would have occurred to me to watch out for while experimenting.

What the heck is wrong with a ribbon mike that has an output transformer, plugged into some cable connected to a real preamp?

I planned on having it switchable ... my main reasons are that I needed a little more gain. At a decent distance, my interface (Scarlett 18i8) needs to be pretty much wide open to pick up some instrument sources. (Originally I was also looking for a way to make the ribbon safer because if I use three active mics and one ribbon, the interface doesn't allow me to turn off the phantom on only one channel, so I needed something to run on phantom power between the mic and the interface ... I've already been informed multiple times that modern ribbon mics will be safe if I need to do this, as long as I'm careful about turning the phantom off before plugging in and unplugging.) I'll abandon the idea of using the Schoeps with it; I was more curious if it was feasible with small changes.

 

[Chris_V]

I don't really see the point in using this (but true, clever) circuit for a ribbon.
This circuit is good for converting a high impedance, relatively high output single ended signal into a low impedance differential one. It has no voltage gain, so it is useless for a ribbon mic without a high ratio transformer and with this kind of transformer you can directly plug into most mic preamps (with an high enough input impedance).

 

[midwayfair]

It has no voltage gain

Yeah, I missed before that the output transistors are emitter and not collector followers. [EDIT: Which, I see, still wouldn't have given any gain because the PP resistors are the same as the 6k8 to the zener reference voltage].

 

[abbey road d enfer]

I don't know who started this whole craze of having ribbon mikes with amplifiers built in, but IMHO it's sorta silly. What the heck is wrong with a ribbon mike that has an output transformer, plugged into some cable connected to a real preamp? The fun of that is that the nature of the preamp loading gets reflected back through the transformer into the ribbon itself - it affects the dynamics and damping of the ribbon!

I guess it all started with cheap ribbon mics using cheap transformers sold to cheap customers.
The output level of these early types was so low (as any ribbon mic) that customers thought they were defective, so manufacturers started increasing the turns ratio (but not the core or winding quality), which would not increase significantly the output, but also would make the sound even duller than it was (for the unwashed customers, ribbon mics lacked the zing of the plethora of poorly damped condenser mics that had been made available) because of the excessive loading caused by typical mic pres and cables.
Then somebody (?) figured out this could be solved by shortening the length of connection between mic and pre and by making the latter high-impedance.
The next step was to build the pre into the mic's body.

I have personally measured several chinese ribbon mics. Some present an impedance of 2500 ohms which varies by no more than 1dB at LF resonance. That means that the motional impedance is completely drowned by parasitic resistance in the transformer windings and ribbon connections, that the magnetic flux is probably not well concentrated and/or the ribbon's mechanical damping is excessive.
All theses factors contribute to poor output level.
Indeed, when such a high impedance source (for a mic) meets an input impedance of 2k (typical of many mic pres), there is an attenuation of 7dB to start with, and some HF attenuation due to the leakage inductance of the transformer and the capacitance of the cable and mic pre.
Some have chosen to build the pre into a barrel adapter; I have experimented with one, and I must admit that, although the one I have does not fully deliver the goods, it makes some of these otherwise unusable mics into workable ones. Even then, I have never used any of these because I could not find one single quality that would make them shine.

 

[PRR]

> It has no voltage gain

Strictly, it has gain of 2. The FET works as a cathodyne. If FET has high Gm, unity gain to Source and equal gain at Drain. Followers, then a differential input at the other end of the cable, we get nearly gain of 2.

 

[Chris_V]

Yes !!
Even if I never know were to put the gain 2 in single end-differential-single end conversions  ;D

 

[emrr]

Modern mikes place the diaphragms at ground and polarize the backplates so that the capsule doesn't turn into an electrostatic air cleaner, soaking up all of the smoke and diesel soot in the air, onto the capsule. By polarizing the backplate, the capsule works just the same, but the parts exposed to the dirty air are grounded, so they don't accrete dirt and contamination over time.

Just want to pin that bit for consideration; have never seen that mentioned before.

 

[abbey road d enfer]

Just want to pin that bit for consideration; have never seen that mentioned before.

Resurrecting an 8-year old thread!
This is actually quite common knowledge, however it's true of some LDC mics. The type of mounting of most SDC's does not allow floating the backplate, which is quite often screwed-in the body.

 

[midwayfair]

Resurrecting an 8-year old thread!

uh-oh, Doug's been replaced by a bot! Get him!

 

[emrr]

Resurrecting an 8-year old thread!
This is actually quite common knowledge, however it's true of some LDC mics. The type of mounting of most SDC's does not allow floating the backplate, which is quite often screwed-in the body.

8 year old thread appears to be the only time it's ever been said here. If it's common knowledge, why have I never stumbled across it in the thousands of pages of mic info here? Lots of threads in which people are reversing the polarity for various reasons without acknowledging this consideration.

Resurrecting threads is a thing people should do more often!

 

[abbey road d enfer]

8 year old thread appears to be the only time it's ever been said here. If it's common knowledge, why have I never stumbled across it in the thousands of pages of mic info here? Lots of threads in which people are reversing the polarity for various reasons without acknowledging this consideration.

Not all the knowledge in the world is in the DIY group. That part of knowledge may well be in the micbuilders group. That group and this group complement each other, because the members have different (but complementary) interests. For some reason, this specific point has never been raised in our group, but is studied somewhere else...

Resurrecting threads is a thing people should do more often!

Agreed!

 

[emrr]

That part of knowledge may well be in the micbuilders group.

I thought that group had died when Yahoo killed 'groups'. I just found it again (I think), but haven't been approved.

 

[abbey road d enfer]

I thought that group had died when Yahoo killed 'groups'. I just found it again (I think), but haven't been approved.

It has since been transferred to iogroups, as well as the LTspice group.