I had two problems: Too much time and to few things to do. Thus I strolled the Internet and -- long story short -- started reading a master thesis on the evolution of condenser microphones. And at some point, there's been the idea that I wanted to walk Georg Neuman's road from the early 1930s. So I found myself a valve and some other small electronic parts and started soldering around.

Initial thoughts

It sort of started with skimming through Volker Meiz' master thesis ``Die historische Entwicklung von Kondensatormikrofonen'' (on the historical development of condenser microphones) which includes some circuitis used in old time microphones. At some point I came to the conclusion that I want to see whether it's possible to restage one of those old time microphones using today's everyday equipment.

For those who are not that into microphones, the CMV3 was one of the world's first condenser mics which were used in ``real life'' environments, which is, outside of laboratories. In fact, this mircophone was prominently used by Adolf Hitler as seen in the picture below. So this mic was kinda state of the art back in 1935s Germany. Also this is how it got it's American nickname -- ``Hitler Bottle''.

Excerpt from

I wanted to stay somewhat close to the original, which is why I ended up using an acaual triode tube in this project. As I went through my havings, I found an ECC83. It may not be perfectly suited for the original circuit but... well, I kinda went for trial and error here, a triod is a triod after all so there should be at least some result.

Thinking and collecting stuff

Now, the idea of a condenser mic is quite simple: two flat electrodes facing each other, one of them is thin and the other is rather thick.

Those two electrodes are nothing else than a non-standard parallel plate condensator. The trick is that one of the electrodes is thin enough so that it will move when hit by sound waves, thus causing the distance to change. Which then again in return changes the electrical capatity of that construction (it will increase a tiny bit as the distance is decreased and it will decease a tiny bit as the distance is increased).
All that's left to do then is to grab the electrical signal caused by that. Straight forward, isn't it?

The idea is that the change of capacity will cause a really tiny current though a really high value resistor (in the actual schematic it's gonna be 60 MOhm) at which it then can be as a tiny AC voltage. First and foremost goal of any following adapter circuit is threfore impedance adaption to a low impedance input of any following processing devices (e.g. ``usual'' microphone input which is probably expecting 600 Ohm rather than 60 MOhm impedance).

And now for the tricky part -- actually building it. Electrods have to be mounted as close to each other as possible, one electrod has to be massive while the other has to be flexible (yet not touching the frst one). Then there's the initial signal which is the loading or unloading current for this condenser caused by the changes to the capacity which are very tiny and need to be amplified so that they can actually be fed to some next level of processing (which in this case will be my Sony TCD-5 tape deck for I will not risk to roast any of my other sound equipment items). As a matter of fact, the electrical worries are actually not the worst problem (which must ne solvable, demonstrated by each Hitler recording created with an CMV3); my way bigger issue was finding something that would make a useable condenser. That is, the question is how to build the actual mic capsule with like no specialized equipment for tiny mechanics. Also, it's a question of materials -- what can this stuff be made of? I quickly realized that a blank copper coated printed circuit board would make a fine base to put a elestic seccond electrode onto -- distanced. Which then made me search for something suitable. Kitchen tinfoil seemed quite a bit to stirdy; parcel tape is not conductive enough and so on. I don't have the facilities to create something suitable, i.e., a thin piece of PVC with an even thinner gold layer. Not to talk about the necessary tools to craft some kind of a fitting which will keep a diaphragm in a distance of some micrometers from another electrode. Being stuck at that thought, I got hungry and helped myself by eating a german classic in terms of sweet snacks -- hanuta (which stands for ``haselnuss tafel'' -- hazelnut chocolate bar, even tough it's more of a creamy snack).

This snack comes in a tinfoil packaging featearung one very thin top layer which is in a perfectly squary shape...

I immediately wanted to try that as diaphragm after I saw the tinfoil vibrate as I yelled at it (yes, this is a story of me yelling at, among other things, pieces of tinfoil). Which still leaves me with one problem: How will I get the diaphragm into a defined distance? Which was when I saw the remainings of a recently shredded CD-ROM. A perfectly thin yet defined piece of isolator. A bit of math told me that a 5.5 by 5.5cm piece of tinfoil in distance of 1 mm to a copper surface will make a capacitor of roughly 22 pF. Which is a surprisingly good starting point for experiments.

Finally I needed some higher voltage for the anode but that's a mere no-brainer. To rule out any influence originating from AC supply, I wanted to go for solid batteries (yeah, call me a whimp for buying myself out of building a power-supply). And in terms of batteries, 9V-blocks are nice as they're stackable w/o any further equipment. A box of 10 pcs is available for roughly 15 Euros at Conrad, 2 pcs are roughly 1.50 at your local ALDI.
The heating supply is specified for either 6 or 12 Volt for the ECC83, which sucks as both are not 9 Volt, but is kind of manageable.


Everything started with the re-drawn circuit of the CMV3 as shown right below.

As found on http://www.gyraf.dk/schematics/Thermo_bottle_Tube_mic.GIF

And me soldering things together. Inspired by what was state-of-the-art back in the days, I started building it from scratch on my desk, resulting in a a somewhat hardly handable result which needed some ten square cm. To be somewhat safe that my head, headphones and the rest of my electronics survive the presence of 90 Volts (I just don't want that voltage accidentially hitting my amp... just "no" on that one), I connected it to my old Sony TCD-5M cassette recorder which I thus attached to the output of the transformer. Then the sweaty-finger-moment: Connecting the anode supply with its roughly 90V... short bump, silence. Connecting the heating -- Hum. Horrible hum. 50Hz-hum. Nothing to hear but 50Hz-hum.

Well, fuck.

On a second thought, don't be too fast on that. I just built an extremely high-impedanced amplifier, so of course the slightest hum will be loud as heck. And as I used some el-cheapo switched power supply for the valve heating which is literally the centerpoint of the whole amplifier, there's a great chance that it's the origin of the hum. Said, done, disconnected the heating power (the great thing about valves, they continue to stay in usable state for some seconds after disconnecting the heating supply) -- ha! Notably reduced hum. And, even better, underneath a fuckton of noises, I can actually hear my voice.


Now that I got onto that track, I knew that I have to put the whole thing into some kind of a Faraday cage, that is, isolate it from all kinds of elecrical fields we're having in our homes these days. In the end, I've been putting the whole circuit into a metal tea box purchased in an el-cheapo sale for a lone Euro. As I still liked to experiment with the capsules itself, I added a BNC connector to that metal cage and had the capsule as external module. Which is kind of exactly what ye olde Neuman did, only he made the capsules attachable to the top of his ``bottle'' while I kept the connector at the side for practical reasons.

Some Pictures and Some Soundbites

As a matter of fact, there are no pictures of the first setup. Nonetheless, the seccond setup looks like this:

What you see is the essential circuit being soldered around the valve's socket on a metal chassis. This cascade of 9V blocks delivers the roughtly 110V to operate the whole thing:

Now let's move on from the impedance adaptor to the actual microphone capsule. My initial attempt was using a thin piece of tinfoil from a snack and a copper surface of a circuitboard:

Which together with the 1st version amp actually was able to deliver the first recognizable pieces of recording.

The next step was made after adding a metal shielding to the amplifier, dramatically reducing induced noise. Which then made me use a metal candy can as my next attempt to build a microphone capsule; including a grid at the front which I raided from a kitchen stainer:

This capsule actually works with really low noises; however it is not only made out of a piece of can, it also sounds like that.


Of course this mic cannot hold up to any of today's $3.99 electret capsules made in China. But that's not the point. It actually does better than I expected, especially in terms of SNR. One main point of interest stays of course capsule design (that is, other materials which are cheap to get but more suitable than plain tin foil -- and how about some holes in the massive eletrode to give the capsule some kind of a pickup pattern). Also I'd like to try this with some valves suitable for lower anode voltage and see how they perform.

Sources and Resources

Volker Meiz, ``Die historische Entwicklung von Kondensatormikrofonen'', Master thesis, Berlin, 2005. https://www2.ak.tu-berlin.de/~akgroup/ak_pub/abschlussarbeiten/2005/MeitzVolker_MagA.pdf

ECC 83 Datasheet, http://drtube.com/datasheets/ecc83-philips1954.pdf

Gyraf Audio, ``Gyraf Audio's Obscure Schematics'', Aarhus, 2006. http://www.gyraf.dk/schematics/schematics.html