EEVblog #609 - Condenser Microphone Design Tutorial

[EEVblog]

Part 4 of microphone technology with Doug Ford, former head designer at Rode Microphones.
In this video Doug discusses how to design a condenser microphone, in particular the Rode Classic II microphone.
Many things discussed include:
External polarisation voltage generation and the relationship with sound pressure.
The trade-offs involved in bias voltage vs physical construction and sensitivity.
Diaphragm deformation and its effects.
CMOS oscillator design for high voltage generation.
Zener diode noise in relation to the knee type, and how to reduce it.
Heating and dehumidification advantages of valve based microphones.
Valve based amplifier design, and blending solid state for improved performance.
http://www.rodemic.com/microphones/classic2ltd

All videos in this series are here: https://www.youtube.com/playlist?list=PLvOlSehNtuHv98KUcud260yJBRQngBKiw

[dentaku]

The Classic II comes with a  leather embossed coffee table book and a titanium warranty card 

These videos need to be seen by people outside the regular EEVBLOG community. Microphones are such everyday hardware and it's great to know how they work.

[calexanian]

We use 5  Rode Broadcaster mics and two procaster mics on the corolla show. We also have a set of their Pin Mics. So far so good and I think they donated them to us. Good video.

[just_fib_it]

How surprising, in the multi-page audiophoolery wank-filled description of the Classic II microphone Rode forgot to mention that Doug whacked in a transistor to flatten the frequency response 

[N2IXK]

What kind of tube would be used here? An exotic electrometer triode or something more mundane? A google search on the Rode Classic II shows a 6072 (Mil-Spec 12AY7), but that's a dual triode. Do they parallel the sections, or just leave one half unused?

At least the 6072/12AY7 is still readily available. The revered Neumann U47 condenser mic used a specially selected Telefunken VF14 triode, and those are pure unobtanium these days.

[calexanian]

We had discussions with telefunken to produce a new type VF14. They only wanted to pay $10 for it. I never even bothered callin back!

[SeanB]

$10 would be possible if they ordered 100k of them though. Payment up front of course.

[calexanian]

And how many old telefunken microphones are out there that use that tube? They estimated about a thousand. 10 grand don't even pay for the dies for the metal parts!

[SeanB]

At least each microphone would have a lifetime supply ( or at least 100) spare tubes.

Would probably be simpler to make a drop in replacement using some high voltage Jfet devices or design a simple transistor amplifier using low noise bipolar transistors that could retrofit the amplier.

[calexanian]

Somebody makes an adaptor system already.. Its trivial really...

[N2IXK]

Would probably be simpler to make a drop in replacement using some high voltage Jfet devices or design a simple transistor amplifier using low noise bipolar transistors that could retrofit the amplier.

Such adapters are available, both solid state and tube (using a 13CW4 nuvistor, IIRC). But neither of them gained any acceptance.  Like general audiophoolery, much of the recording/music industry is driven by hype, elitism, and magical thinking. If something isn't rare, expensive, and long out of production (therefore not available to the masses), it can't possibly sound any good. Even if TFK managed to restart production of VF14 valves using the original equipment and materials (like Western Electric did with 300B triodes a few years ago), they automatically would be considered inferior to the originals.

[calexanian]

Agreed

[Zbig]

I like that series. The only minor thing I'd have liked to be elaborated a bit more on is, why electret microphones, while using permanently charged diaphragm, still do need a biasing voltage. I think I recall Dave mentioning the built-in FET amplifier in passing, but I'm not sure I'd have got this hadn't I know that already. I remember being rather confused after learning that condensers electrets are "pre-polarized" first and later, that I have to DC-bias them anyway. Given the popularity of the tiny cheap electrets among the makers/hackers, I'd suggest the polarized / non-polarized vs. powered / unpowered aspect deserves a bit more clarification.

EDIT:
Will we hear about MEMS microphones as well? The technology seems fascinating to me.

[Richard Crowley]

How surprising, in the multi-page audiophoolery wank-filled description of the Classic II microphone Rode forgot to mention that Doug whacked in a transistor to flatten the frequency response 

No argument about audiophoolery and wank-filled. Rode has to cater to those people.   

However, the transistor was added to the circuit to convert the high-impedance out of the firebottle into a low impedance output that would be suitably insensitive to variations in output loading.  It has also been the practice of microphones to provide LF-rolloff (=high-pass) switchable variations to better match various conditions one finds out in the field.

It seems odd, though. Most microphone designers accomplish that without an extra transistor by using a cathode-follower circuit instead of the plate-output circuit Mr. Ford showed.

[Richard Crowley]

why electret microphones, while using permanently charged diaphragm, still do need a biasing voltage. I think I recall Dave mentioning the built-in FET amplifier in passing, but I'm not sure I'd have got this hadn't I know that already. I remember being rather confused after learning that condensers electrets are "pre-polarized" first and later, that I have to DC-bias them anyway. Given the popularity of the tiny cheap electrets among the makers/hackers, I'd suggest the polarized / non-polarized vs. powered / unpowered aspect deserves a bit more clarification.

Electret mics do NOT require a "biasing voltage". As you said, the electret charge takes care of biasing the condenser capsule.

But ALL condenser microphones (whether electret pre-charged, or traditional externally-charged) are ULTRA-HIGH output impedance. In the gig-ohm range. That is such a "delicate" signal that there isn't much you can do with it in its original state.  If you tried to feed that through even a meter of microphone cable, you would have no signal at the other end.  So that ultra-high impedance must be converted/transformed into something MUCH LOWER to survive transmission through the Real World. 

That is why either a firebottle or a transistor is necessary to provide the bare-minimum impedance conversion.  If you look at the spec sheet for those cheap 89-cent electret mic capsules, you will see that they have a FET inside, and transistors don't work for free. All active components require operating power.  Now a cheap FET condenser mic capsule needs only maybe 1.5V at a few mA, but you must have SOME kind of power there to let it operate.

The output terminals of a cheap electret capsule are actually the source and drain terminals of the FET. The gate of the FET is connected directly to the mic capsule.


http://en.wikipedia.org/wiki/Electret_microphone

[pgross]

Thanks for the great mic series.

I remember the story about the heat resistors – acting like dehumidifiers.

Humidity is one of the condenser mics worst enemy, where leakage in the capsule / diaphragm, pcb and wiring can lead to excessive noise and loss of sensitivity. In the 60s, some manufactures overcame this problem by using RF modulation.

Instead of the traditional high impedance capsule load, the RF microphones are still utilizing a condenser capsule, but now the capacity variations are used to modulate a high frequency signal (10 MHz typ.).

Several modulations types (AM, FM, PM), where used. The RF mic technique is still in use today in Sennheiser’s Academy Awarded MKH series.

Best Regards

Peter

[calexanian]

That's right. I forgot about the RF microphones.