Understanding condenser microphone bias adjustment and its audible effects

[soliloqueen]

Some condenser capsules are loud enough to not need a preamp at all if you have an OPA that can go straight to line impedance and a relatively low noise ADC. the bv13 in the u87 takes the level down by 20dB. It really is about the impedance, which is why it's often called impedance conversion and not amplification when talking about condenser headamp circuits

 

[kingkorg]

I've heard that LDC mics can put out that much; but the capsule itself?

That is, in a typical mic circuit using an op amp as an impedance converter, the input of the op amp is seeing only a few millivolts under any conditions.

Yes, the capsule itself can put out a volt or two easily (again, next to a loud source with sharp transients, like close miked drums).

I've actually measured 7.4vpp capsule output using simple opamp unity gain circuit and measuring transient at the
output with an oscilloscope.

https://groupdiy.com/threads/ldc-k67-capsule-output.79855/

 

[k brown]

I stand fully and humbly corrected; must have mis-understood something I read long ago.

 

[ccaudle]

I've actually measured 7.4vpp capsule output

That is good to keep in mind because it will have implications for power supply bypassing in the mic circuit (assuming you want to prevent power supply droop on those large transients). That will drive a few mA peak into a common 1k5 Ohm mic preamp input impedance.

 

[rogs]

I've actually measured 7.4vpp capsule output using simple opamp unity gain circuit and measuring transient at the
output with an oscilloscope.

https://groupdiy.com/threads/ldc-k67-capsule-output.79855/

I've found that my simplest version OPIC - which uses phantom power to provide the capsule polarisation voltage - is probably a good choice for high level input signals?

There is only a single OPA1641 opamp used, and that draws around 1.8mA from the phantom power supply.
The capsule polarisation voltage essentially draws no extra current of course.

That will allow around 30v of DC to power the op-amp. With a 'rail to rail' output available from this device, that should allow for a ±15v signal swing before clipping..... That should be enough to cope with the output from most capsules!?

There is of course any extra current drain required by the low impedance of the mic preamp to factor in, and that might reduce the headroom slightly, by reducing the DC supplied to the ampifier. How significant that will be just on transients, I'm not sure?

The polarisation voltage should be 42v (in theory!) but allowing for the fact that the phantom power supply will have a finite impedance, it's more likely to be around 40v.
The sensitivity of the mic is likely to be around 3dB less than the same capsule being biased wth 60v, but when used as a drum mic, for example, this is not likely to be very important....

That should place all the clipping control requirements outside the microphone -- making it someone else's problem!

 

[Khron]

With a small enough power filtering cap, you could even include "in-mic compression"

 

[Tubetec]

The B&K cathode follower preamp with capsule can output 20 volts before clipping
Capsule polarisation is 200v , a 1 inch capsule can accept around 125db Spl, half and quarter inch capsules can take even higher SPL .

The cathode follower has no gain , so its only the capsule itself produces the voltage .

 

[micolas]

I've actually measured 7.4vpp capsule output using simple opamp unity gain circuit and measuring transient at the
output with an oscilloscope.

https://groupdiy.com/threads/ldc-k67-capsule-output.79855/

This is true. I was very surprised how much signal a capsule can produce!
In the U87ai circuit, with Bias for relatively max symmetrical response, with 2N3819 CEN from Mouser, Idss 7.9mA, with the capsule from SC1100, biased with 63V, I screamed as loud as I could in the midrange, close to the headbasket (aprox 1.5. ..1.7cm distance from the membrane). At the JFET output without the connected transformer I saw on the DSO-TC3 oscilloscope, almost 6Vpp.
The signal was strongly distorted, (I would say almost guitar fuzz ).
That got me thinking, I started to reconsider the size of the output transformer.
I really started to evaluate the transformation ratio by injecting 8..9V ac into the primary and measuring the voltage on the secondary (in parallel on the secondary I connect 3.9Kohm).
(Previously, we injected 3..5V ac into the primary).
The microphone adjustment is done at 1KHz, but I started to check the Bias jFET for other frequencies as well.
I write down my relatively correct bias, then I do a lot of tweakins with full-circuit listening tests, with the voice from 4..6 inches away, completely subjective.
I choose the settings that seem more musical, richer, that convey more a lot of emotion. I write down my settings with detailed artistic observations. In the end, the decision is mostly subjective.
I don't like the sound too clean, septic. (I'm a solo guitarist, a clean tone bores me, harmonics delight me ).
So, in the end for me it's an artistic decision, a matter of personal taste.

 

[rogs]

........ I choose the settings that seem more musical, richer, that convey more a lot of emotion. I write down my settings with detailed artistic observations. In the end, the decision is mostly subjective.
I don't like the sound too clean, septic. (I'm a solo guitarist, a clean tone bores me, harmonics delight me ).
So, in the end for me it's an artistic decision, a matter of personal taste.

When it comes to adjusting for the 'musicality' of distortion, it's very much a matter of personal taste of course..... and discussions on which distortion is preferable to create the 'best' sound can get very agitated ! ...
Distortion is of course anything that introduces non-linearity into the signal path.... That might be deliberate (or accidental!) changes to the frequency response, or perhaps different characters of overload distortion, introduced by different bias adjutment?... Not forgetting the 'nice' distortion that many transformers are said to add!
I have to confess, at the moment I'm tending towards using op-amp impedance converters, which are very linear and can have high maximum signal level outputs. They do tend to produce a very 'clean' signal.
That then gives the opportunity to introduce any desired 'non linear' effects into the signal path outside the microphone -- a channel strip, for example, can provide compression, EQ, soft clip diode limiting, de-essing etc etc....
Flexible options which can be carried out at line level, making the the problems of adding unwanted noise by introducing effects early in the signal path much less likely ...
But we all have our own favourite ways of doing these things!

 

[micolas]

When it comes to adjusting for the 'musicality' of distortion, it's very much a matter of personal taste of course..... and discussions on which distortion is preferable to create the 'best' sound can get very agititated ! ...
Distortion is of course anything that introduces non linearity into the signal path.... That might be deliberate (or accidental!) changes to the frequency response, or perhaps different characters of overload distortion introduced by different bias adjutment?... Not forgetting the 'nice' distortion that many transformers are said to add!
I have to confess, at the moment I'm tending towards using op-amp impedance converters, which are very linear and can have high maximum signal level outputs. They do tend to produce a very 'clean' signal.
That then gives the opportunity to introduce any desired 'non linear' effects into the signal path outside the microphone -- channel strip compression, EQ, soft clip diode limiting, de-essing etc etc....
Flexible options which can be carried out at line level, making the the problems of adding unwanted noise by introducing effects early in the signal path much less likely ...
But we all have our own favourite ways of doing these things!

Rogs, it seems that we both come from the far reaches of the analog era, of course each on his own path

 

[terry setter]

I've found that if I do my designs with 2.5V in mind as the highest "likely and reasonable" level that will exit the capsule, I don't get surprises. I generally polarize the capsules somewhere near 63V, but have gone as high as 130. My experience has been that the threat of diaphragm collapse become a determining factor before any problematic shift in the sound of the capsule. Yes, capsules can output more than 2.5V,, but so far I've been able to mic very loud sound sources with successful results, and TO MY PERSONAL TASTE, get a bit of excitement from the circuit at the same time.

I've wondered before whether there is something about very loud sounds that we associate with as least small amounts of distortion, and we, therefore, take one of our loudness cues from that distortion. I'm sure all of us have had the experience of being in a venue with an extremely clean sound system, not realizing at the time how loud it was actually playing. That is, not until we left the venue and could hardly hear because our ears had gone into "protection mode" - nearby sounds sounding like they're 15 feet away, or the "ring" that tells us we've been assailed by the sound system at the event. I've heard speculation that one of the reasons young folks are going deaf due to earbud-induced damage is that they have no idea how loud they're actually listening. The earbuds can play so clean, so loud.

And, related to microphones, I've heard top engineers say how they love the way tube mics start to break up a bit when a player "really leans into them". To me, they're talking about distortion. So, in my thinking, they like the kind of distortion that they connect with tube mics in loud sound environments. But heck, we all know that you can build a tube mic that will withstand extremely high SPL and not break up as much as many transistor mics would at even lower SPL. So my take-away is that engineers remember tracks that they love, and they come up with theories as to why they sound so good. Often, they overlook variables such as when the sax player played those loud notes and "leaned in to the mic", the sound of the sax itself was different than at lower volumes. They decide that it was microphone and/or preamp mojo, rather than player mojo. Of course, every recorded track is the sum of its parts: performance style, gear selected, placement, acoustics, dumb luck, personal taste, etc.

It's telling that many people love the sound of recordings from Mowtown Records, even though rather early in the development of the Hitsville Studios, they got rid of their reasonably large microphone collection and replaced them all with Neumann KM-86s. Yup, almost every famous mic'd track that came out on Mowtown records was recorded through only one type of microphone. (I wasn't there, but I've heard this for many years from people who were connected to Mowtown. Tape Op did a feature article on Hitsville's chief recording engineer who talks about it, and I see that it is mentioned in the advertising materials for UAD's recently released "Hitsville Reverb Chambers", so I'm thinking that it's true.) Just ONE type of microphone for Stevie, Diana, Marvin, Smokey, and all the others and all the instruments.

But I sure do love building microphones....

 

[Gus]

why do people use sine waves? Try a triangle wave.

 

[k brown]

Tryangle wave . . .

 

[rock soderstrom]

I've heard top engineers say how they love the way tube mics start to break up a bit when a player "really leans into them". To me, they're talking about distortion.

I think when you talk about sound in this context, you're mostly talking about distortion. I realized very early that the measurement-technical optimum is not what I want from a microphone. If you choose this route, you end up with an opamp, whose advantages are not only reduced to ultra-low THD and noise, but you also don't need a bulky additional power supply and expensive tubes that also wear out.

So why all this nonsense?

It's the sound. I favor microphones that refine their sound sources, put them in a better light. I'm not aiming for a realistic ideal, that's not my thing. I know that there are people who see these things completely differently, and I fully understand that. My way is different and completely subjective and taste-oriented. Pure technical measurement (which is very important, without a doubt) is only secondary for me when it comes to microphones. To answer the original question of this thread. If a microphone achieves sufficient technical specifications, the ear is always the deciding factor for me as far as biasing is concerned.

 

[jp8]

I've found that if I do my designs with 2.5V in mind as the highest "likely and reasonable" level that will exit the capsule, I don't get surprises. I generally polarize the capsules somewhere near 63V, but have gone as high as 130. My experience has been that the threat of diaphragm collapse become a determining factor before any problematic shift in the sound of the capsule. Yes, capsules can output more than 2.5V,, but so far I've been able to mic very loud sound sources with successful results, and TO MY PERSONAL TASTE, get a bit of excitement from the circuit at the same time.

I follow the same guidelines when designing mic circuits. Many mixers can only handle up to 10dBU (2.5V) input signal on a mic input. So why would one want to design the mic electronics for higher output voltages? To me, this seems rather pointless. Then I'd rather draw some more current from the phantom supply and use that current for other purposes, e.g. to lower THD or lower the noise. Assuming a transformerless circuit, the current can be raised until a 12-15V supply voltage remains for the impedance converter stage, which will do for low THD at 10dBU max output.

If it can be expected that the capsule output will be so high that the output voltage will get near to, or even exceed, 10 dBU, then one should simply include a pad circuit to stay below the 10 dBU output level and prevent clipping of the microphone.

Jan