The original version of this text was written fairly long time ago in finnish. This DIY microphone was built before I even decided to write anything about it. Based on the dates on the original image files it was around 2005. I published the text on my personal (and very web 1.0) webpage then. A few years later I started to translate that to english, but never finished the translation. Let's see if I can finnish this in 2020... (I didn't. It's 2022 now.) Forgive me the repetition of the title. I started with what I have written earlier and only made some corrections and edits where I could remember something. It's a quite hard to follow what I did from the text, but I really don't remember it any better. I'm publishing this anyway because there may be about three or four people in the world that will find it interesting. (Six at most.)

Images will be mostly quite awfull too. They were shot on an early digital compact camera (it was old even back then) and to make it worse it was well after my first photography phase (with film) had faded and before i really got back to photography as a hobby.

Do-It-Yourself Condenser Microphone

This is written from what I remember after the project, so there might be some minor inaccuracies. I try to move chronologically, but I might remember something wrong. These are not instructions on how to build your own microphone. This is just a report of how I built one. If you are interested about building something similar, it I'll try to provide links to my sources. Most of the original links are dead by now though. I would not recommend this as a first electronic project for anyone. Additionally the project also includes some other crafts beside the electronics part.

Also, I would not recommend this project for anyone who really needs a good mic. My purpose for the project was merely prove to myself that I could make it work. Although I was a bit surprised how good this mic became, it cannot really compete with its commercial counterparts.

The Idea

I'm not sure when this idea came to me. It must have been around midnight when I was lying on my bed, waiting for falling asleep. It was something like this: "What if I built a condenser microphone myself using aluminium foil as the diaphragm..." The idea was to make a real condenser mic so that I would even build the capsule from scratch. I had already built some mics from ready made electret condenser capsules.

Preliminary Studies

Condenser microphones work approximately like this: The mic has a diaphragm and behind it a back plate. Both of them conduct electricity. These are very close to each other, so they form a capacitor (also called condenser). If the diaphragm is moved closer to or away from the back plate, the capacitance of the capacitor changes. So, when the air pressure around the capsule changes the diaphragm moves and the capacitance changes (and sound is periodic airpressure changes). If the capacitor is connected to a circuit that converts the changes in capacitance to changes in voltage, we have a working microphone.

When I was searching for a suitable circuit for the internal preamp of the mic (from internet ofcourse) I found out that I'm not the first one that has thought about this kind of project. One radio amateur had made one and had even writen an article about it for some magazine. There is some information about it on his web page. The link is dead now, but it can be found on Wayback Machine for now.

I did not want to use the circuit on his page because I wanted my mic to be able to work on phantom power. His circuit was meant for ham-radio stuff and I wanted something that would be more recording studio style, but I got good tips from his site on how to build the microphone capsule.

I was thinking that one circuit that is meant for electret capsules could work if I used a separate FET with my own mic capsule and some other parts. I still tried to find more suitable circuits but manged to only find some tube circuits and some others that use transformers for balancing the output. Transformers have some good properties, but good audio transformers can be hard to find and usually they tend to be expensive.

After some more searching I found what I was looking for. It is in an article (Link is dead, but Wayback Machine saves again, for now) that is about modifying cheap commercial condenser mics. Circuit looked very familliar and even some part values where just what I had been thinking about. :)

Let's Build!

I started building the capsule with the tips on the page I mentioned above, adapting them to my purposes. Parts I used are kind of interesting. The metal ring needed and the back part of the capsule are from an old 5½ inch hard drive (physically really big one, about the size of two normal CD-drives). They were one of the spacers and the endplate of the platter stack. The diameter of the diaphragm is about 4cm. This can be really called large diaphragm mic. :) I used a piece of a copper clad circuit board as the back plate of the capacitor as mentioned in the article. I used some foam plastic in the capsule as described in the article to hold the backplate firmly against a piece of transparency film that was acting as a spacer when gluing everything together. Before gluing I had to attach a wire for reading the signal from and connecting bias to the backplate.

Some things were a bit problematic. For example I had to make sure that the shell of the whole construction would conduct electricity so it could be grounded to reduce interference. Epoxy glue is not a conductor. The hard drive parts were aluminium, and regular solder will not stick to aluminium. In the very end I used quite hard pressure to hold things together when gluing. This caused the glue layer between parts to become very thin. After the glue had set there was a good conduction between the parts. I also used some extra tricks to make sure it conducts. (I wish I'd remember what that was. I probably made some scratches on the surfaces that were glued together.)

After this it was time to glue the diaphragm to the capsule. I took a smooth (or more like smoothish) piece of aluminium foil and taped it to a smooth base so it was a bit stretched. The aluminium foil I used is just the regular kitchen grade stuff from a budget brand. The reasoning for using a cheap foil was it might be thinner. In reality there is probably not much difference.

I used a piece of thick plastic (plexi glass or something) as the base. It is important that the base is as smooth and straight as possible. After that I spread some epoxy glue to the edges of the capsule and laid it against the foil. At first I used a screw clamp to hold them together, but the pressure was a bit too hard and the diaphragm became slightly wrinkled.

Nevertheless, measurement with a DMM revealed that the diaphragm was not shorted to the back plate, so I decided to try it. I made a simple battery powered one FET impedance-converter/pre-amplifier circuit on a breadboard and connected it to the mic input on a laptop computer. The bias of the capsule came from the same 9v battery that was powering the rest of the circuit. I set the computer to record and said something in front of the mic. I was surprised as it could capture sound and it actually did not even sound like an electric kazoo. (I really thought it might sound like a kazoo if it worked at all.) It did capture a quite a bit of hum with the sound though, but that was not a surprise as the whole mess was not housed in anything. In other words nothing was shielding it from interference. Unfortunately I did not take a picture of that setup. However... I managed to find the original test recording from my mailbox.

Something like half-dozen attempts to glue the diaphragm on the capsule properly and a lot of trashed aluminium foil later I finally believed that screw clamp should not be used. It could have worked if I got the tension just right, but in reality it was too hard to adjust for that light tension. Next I tried to hold it together with just some masking tape. Although not perfect, the first attempt revealed that the tape worked better than the screw clamp. After a few more attempts I was relatively satisfied with the result.

Testing and Searching For Components

Then some more testing. This time I even tested with the final circuit. At this point it was still on breadboard. There was a lot of hum even though I tried to shield the circuit with a piece of aluminium foil. It helped a bit, but the wire between the capsule and the rest of the circuit was unshielded. As a high impedance node it is very sensitive to external interference. If I just put my hand on something grounded (like my mic preamps casing) and put my other hand near that sensitive wire the hum level decreased a lot. So even I could act as quite effective shield myself.

All parts were not my final choises, because some of the ones I wanted were quite hard to get from the local electronics store I usually used back then. The hardest to find where 1 GΩ (Giga ohm) resistors and 2SK170 FETs.

It works with other FETs but I suspected that the 2SK170 that was the part suggested for the circuit, would have considerably lower noise. My suspicion was later confirmed. The difference was not huge, but it was noticeable.

The circuit worked with highest easily available resistor values (10MΩ), but they loaded the capsule down so that there was considerable loss of low frequencies.

In the end I did not get 1GΩ resistors at all. I could have ordered them from the that local electronics store though. From another local store that I use less frequently I found 100MΩ resistors. They were old carbon composition resistors that have quite poor manufacturing tolerances. The accuracy of the resistors values should not matter too much in this circuit though. Carbon composition resistors are also the noisiest of all resistors. That's not nice, but I thought that they will do for testing.

When testing with 100MΩ resistors, the signal got much more low frequencies. I decided the carbon composition resistors will stay. Actually at this point I first time realised that the value of those resistors affect the low frequency cutoff. In theory it moved the circuits lower -3dB frequency to ten times smaller frequency. I never tried to find out what the real cutoff frequency is though.

Later I did some research and some sources seem to state that Carbon composition resistors will make more noise only when there is a significant current through them. In this application there is almost no current at all going through them. I'm not sure if more modern resistors would have less noise in this. Maybe I can test it if I ever make another capsule part and manage to get some more modern resistors for it. For this build I decided that the carbon composition resistors will have to do.

The foil needed to be trimmed. I also left a small strip to the edge of the foil and connected it to the grounding screw at the back of the capsule just to make sure it is grounded. I also had to glue the foil strip to the body of the capsule so it can't rattle and won't brake so easily. According to my measurements, the diaphragm was electrically conected to the body of the capsule and it worked without the extra connection. This was probably again because of enough pressure when gluing.

Modularity

I had been previously thinking that if this contraption was made modular in some way, I could make different capsules for it if I wanted to. Gradually, the final concept took shape. By leaving the FET, one capacitor and the two high-value resistors on the capsule side, "body part" becomes such that an electret capsule can also be connected to it if desired. In addition, these few parts allow some variations for connecting possible replacement capsules.

Enclosure for the Capsule

I made the frame for the enclosure from a tuna fish can. The most common tuna cans would have been a little big for this purpose in my opinion but I found one brand that sold a multi-pack of three cans where the cans were slightly smaller. I also ate the tuna. I don’t remember what kind of food I made from it. I should document these projects better indeed.

After considerng it for quite a while, I chose a five pin DIN connector with a bayonet lock as the connector between the capsule and the body. The selection criteria were availability now and in the future and the fact that they seemed nice for this purpose. At least I would suspect that those five-pin DIN connectors will be available for at least some time to come, as they are used for quite a few things, of which midi is by no means the least. Well, the availability of those locking ones may not be so good, but they still work with the ordinary ones. There would have been several different models available at YE, but the bayonet lock seemed the most suitable at the time. It remains to be seen whether my choice will still be right in the future.

I punched several holes in the can. I thought that the surroundings of the capsule should be as free as possible from all obstacles, so that sounds coming from all directions can get to the capsule without interference. Basically, the directional pattern of that capsule is a omnidirectional, meaning it should receive sounds from all directions about as well. I haven’t tested how well that theory iholds in reality, but I’d guess it doesn’t, especially at higher frequencies. The size of the capsule itself places its own limitations on this. Also, no matter how I tried, the structure didn’t end up being that open after all.

I attached the connector to the capsule part with epoxy glue. I originally thought of attaching the actual capsule to its case with either strings or elastic bands, but I came to the conclusion that the case was too small for that and it would have been almost impossible to get it built properly like that. With a bigger can it might have been possible. I attached the capsule to its case with a few pieces of plastic foam and some glue. The attenuation of the handling noise may not be as good, but it will have to do. After all, this is not meant to be really hi-fi.

A colleague gave me some suitable metal mesh to protect the capsule. By that time I already had bought a different type of mesh for it, but it was less dense and I thought that the denser one works better as a protection against physical damage. I glued the mesh on with epoxy glue. The glueing didn't turn out very neat, but it's good enough for my purposes. I would have attatched it by soldering, but the solder didn't stick to this mesh. Additionally, the mesh needed to be grounded to the rest of the housing to shield the the capsule against interference. It really matters here as it had become apparent when prototyping with the half-finished casing. I took care of that grounding with some rivets, pieces of sheet metal and pieces of wire here and there. It's not very neat, but works.

When the capsule was in its case, I just had to test it again. The circuit was still on the breadboard and the importance of shielding became even more obvious when I played around with some aluminum foil.

Due to my laziness, I had put off the design of the circuit boards for quite a while. Designing circuit boards for some previous projects had made me shy away from this task. It should be noted here that I'm not very good at it and it often seems to require too much concentration. A bad workman always blames his tools, of course... I used Mikro$oft's Visio for the job, which is definitely not intended for circuit board design. I had just been so lazy that I hadn't been able to acquire and learn any better tools for this. (This was at the time of writing the original finnish text. I have now used KiCad for one other project.)

However, the circuit for the capsule part was so small, that it didn't take long to design the circuit board when I got around to it. After all, there were only four components.

When I had the circuit board etched and drilled, I soldered the other components in place, but instead of the FET I used socket. Later, I did solder the FET I chose in place without the socket, but I wanted to test different FETs first. I almost immediately tested some of the types I had, but I didn't have the 2SK170 type yet. It took me a while to get it, again partly due to my laziness, but also due to the somewhat inconvenient location and opening hours of the store that sold them. So I moved on to prepare other parts of the project.

Body part

Unfortunately, I didn't take pictures of the construction of the mic body. However, there is nothing that special about it. You will get at least some idea of it from the pictures of the finished microphone.

As the casing I used an aluminium tube that was originally packaging for some vitamin C effervescent tablets. I had already used it for one other mic-project and found handy. I just had to take my vitamins every day to empty the tube. Those tubes are quite perfect size for microphone projects. It fits in typical microphone holders nicely. In addition, it is easy to insert a standard XLR male connector intended for a cable into the cap of the tube.

It's a kind of a annoying that most of the various effervescent tablets are now in plastic tubes. Those do not offer the circuit protection from interference like metallic ones. In this build, the most interference-sensitive parts of the circuit are inside the capsule part, so the plastic one could have worked as well. However, I decided to "invest" in metal casing for minimum interference. They might be surprised if they knew on what basis I choose my vitamins at the pharmacy.

Again, it took a lot of time to design the circuit board... That is, a lot of time was spent not designing it or doing anything else for the project. When I got around designing it, I got most of it done in one evening. It's pretty simple after all. After I got the circuit board designed and etched into the actual circuit board, I of course wanted to test it again. Still working. Amazing.

You can get a hole of suitable size in the cap of the tube by cutting along the inside edges of the "silica container" part inside it. (The "silica container" has some crystals inside that absorb moisture so that the actual contents of the tube stay dry... Or so I guess.) A hole of that size is just right for Neutrik and Schulz  XLR-connectors. In my early mic project, I used Schulz because it is cheaper and the parts I consider to be the weakest in those are not even used in these mics. For this I used Neutrik, because the cheaper ones were out at the shop when I went to buy parts.

The DIN connector for the capsule part was attatched to the other end of the tube. I used two screws for mounting it because the DIN connector I used had two holes for that. I didn't make proper threads to the aluminium though. I just made holes for the screws and screwed some junk screws that I had lying around in. It turned out to be sturdy enough, but I thought that if it turns out to be weak, I could fix it with some epoxy glue. After all, I only used whole a lot of epoxy into this project. In fact, I even used two different kinds of it.

I put the circuit board in the tube and tested again. When it worked again (which still amazes me), I put a few drops of epoxy glue on the back to hold the circuit board in place. And another test... Yep. It still worked.

Oh... Before putting the circuit board in the tube, I cleaned it with a flux remover, as it was recommended on one of the mic modding pages. However, I would think that in this part it is not as important as in that the part of the circuit in the capsule enclosure. In addition, I varnished the copper side of the circuit board with a metal protection varnish I happened to have. I don't know how suitable it is for the circuit board, but at least it didn't seem to stop it working. At least the copper strips won't oxidize.

Finalizing the Capsule Part

Again, it took some time before I could get more done. I desperately wanted to test that 2SK170 FET before making the final choice. When I finally got a few of those from one local suplier, it seemed that the effort was worth it, at least in terms of noise. According to my tests performed with relatively crude methods, the noise level in relation to the signal decreased. However, it didn't drop a lot, but several decibels nonetheless. I don't remember the exact reading and I lost the piece of paper I used to write down the measurement results.

After soldering the FET to the circuit board and checking that all the wires coming to the circuit board are properly connected, I cleaned this circuit board with the flux remover. I had to be quite careful when the capsule was attached to it. However, I didn't paint the circuit board of the capsule part, but ended up coating its copper side with epoxy glue (it was praised as an excellent insulator on the side of its package). However, there was no real reason for it. I was just playing. The varnish would probably have worked just as well. After that, I made small cuts to the foam I used to attatch the the capsule to the case and set the circuitboard in those. I also ensured that it stayed in place with some glue. Then I tested it again. It still worked.

At this point there was not much else to do but connect the wires and assemble the connector. The DIN connector has a separate shell part and then a plastic lump where the connection pins are attached. The entire connector is held together with a screw. I had previously glued the connector shell to the capsule housing with epoxy glue. Finally, I attached the grille on the back side with a rivet and epoxy glue. Here again the rivet was used to ensure the grounding of the grille.

Finished. I attached the capsule part to the body part and tested. It was still working. This was a real miracle project in itself because almost every time I tested it, it worked. That doesn't happen to me often.

The most difficult part was gluing the "diaphragm" to the capsule. Developing a usable mechanical structure also produced its own problems, as always.

Pictures of the Finished Microphone

I could probably take new better pics of the finished microphone as I still have it, but I think these original ones are just fine for the purpose.

2022 addenum

At some point I stumbled upon another similar mic project by Wiesner Thomas. I had lost the link, but I found it again through Hack-a-Day. The links to the original page are now dead, but Wayback Machine seems to help here too. Unfortunately the images have not been archived from it.

In early 2020 when the pandemic hit, I started working from home almost exlusively. Before summer vacation I decided to try this self built mic for my audio/video -meetings. I used it for a week for all meetings and no one complained anything about my audio quality. After summer I went back to the mic I used previously. At some point I decided to try my DIY microphone for remote meetings again. I have used it for almost all remote meetings for over a year now. Only once I had some troubles with the mic and it was resolved by replugging the cable to it. I quess the connectors had just oxidized a bit. After that it has worked fine again.

Soundsamples

My original speech sample: Testing in finnish 

My new speech sample (about 17 years later): 2022 and still testing in finnish 

At some point I thought that somehow the frequency response had improved during the years, but comparing these two samples it seems pretty much the same. It's hard to judge from these though. I had some other soundsamples too, but they were mainly electric guitar through miked up weird self built amps etc. and thats pretty useless for a mic soundsample.

Links

Information on these articles helped a lot.

Homebrew Condenser Microphone - Microphone project of an American amateur radio operator (The original page has vanished. This points to Wayback Machine version.)

Making Mics Better: a Practical DIY Project - Article about modifying commercial condenser microphones (The original page has vanished. This points to Wayback Machine version.)

DIY Microphone Projects - microphone projects using electret mic capsules

A Dangerous Science Fair Project - interesting article about fixing a condencer mic capsule (The original page has vanished. This points to Wayback Machine version.)