Small diaphragm capsule design help needed

[Jonte Knif]

Hi folks,

Here is my initial plan and drawing for general purpose omni capsule. Perhaps ultimately leading to something M50-like. (haha, sure)

http://picasaweb.google.fi/jonte.knif/Mic#5369380081824378450

I need some help. I searched the microphone meta and found some things but obviously people are more interested in bigger capsules.

I recall somebody saying that about 1,3mm hole size and 30% hole area works. This is in agreement with photos of M50 for example. But nothing about depth. Obviously I can have different versions made, but I need a starting point. There is abundant info about holes in big capsules, but nothing definitive about small. Something in M50 thread, but not too much.

There is a school of micro mechanics a few blocks from my house, so I can get the accuracy I want and it will be very easy to communicate. I don't know the price yet , might be too high, but I'm going there on Tuesday to discuss and have to have more knowledge.

About my design so far:

I was thinking that messing with diaphragm spacing ring is stupid, and decided that it is easier to have the spacing built in another way. So, after the parts are screwed together the surface is lapped and shimming material (blue) is then put between the plastic base and outer brass "tube". I can not see any reason why this wouldn't work. Also, it is easy to try different distances without diaphragm replacement.

The wires are easily attached to the screws on back side which is nice.

The plastic? Recommendations? They have PTF which is electrically superior, but is it hard and stable enough?

Is there anything super critical concerning the capillary hole? (not in the picture) I guess the smaller the better?

Thanks in advance,

Jonte

 

[leswatts]

Hi Jonte

I thought i'd reply since you had no answers yesterday.

Yes, very little info on omnis...everyone wants to make dual diaphragm capsules.

I would just try the holes you mention, but start very shallow. It's easy to deepen them later. For flat response you'll need to basically compensate for diffraction rise.

I have a B&k 4145 free field omni capsule in the lab where you can see the backplate holes. They go all the way through a thin backplate. It's a highly tensioned nickel diaphragm though.

The old williamson article is the only reference I have seen for an omni but it is so poorly documented as to be useless. It's in the meta. Oh, my old boss wrote an omni article too, in 1966. It's in the meta too.
His mic turned out badly, but he did better later.

Perhaps start with 6 micron (gold) PET film and 25 micron spacing. 50-100 volts polarization?
An easy way to alter tension would be good too...

As far as the material...I stock and use Acetal for uses like transducers. It's very hard and stable, with low moisture absorbtion. It also machines very very nicely to close tolerances. A trade name is DELRIN.
I just made a hundred or so ultrasonic microphones from it.

PEEK is also good.

PTFE is too soft I think.
NYLON would be very bad. Too much moisture absorption.

As far as your construction...yes It would be convenient to have the spacer on the bottom as you show. Easy to change spacing. The only disadvantage is that the mechanical datum is far from the diaphragm/ backplate, so amy temperature /moisture instability would cause changes.

Les
L M Watts Technology

 

[Jonte Knif]

Hi Les,

Thanks!

Funny, I have worked with Delrin a lot, because I used to build harpsichords and it is used as quill material. In very small and thin slips. In that form it doesn't feel like very hard

What would shallow be in this case? I have no idea. 1,5mm or  2mm? I hope I have money for several proto capsules to experiment, but some kind of starting point would be nice to have.

The depth will affect the overall sensitivity. What happens to the response when hole depth is adjusted and we have a stiff membrane?

I guess the source for gold sputtered mylar has been mentioned a million times, but if you can easily point me to a good source, I would be happy. Otherwise I'll just start with capacitor material.
I have actually dreamed that ultimately I would try alu foil. I know it is gonna be difficult, but just for the sake of it.

What would be suitable thickness for aluminum diaphragm? I found a pretty good shop and they stock almost anything.

I'll keep the depth of the capsule as small as possible. Can surely be reduced, if the holes are not very deep.

-Jonte

 

[leswatts]

I'd start really shallow...like 1mm. Might be too shallow, but again easy to deepen.

Are you trying to get the sound of a Neuman M-50? I'm sure you could, if that's what you want.
From what I read that has (had) a  very thin 2.5 micron aluminum foil diaphragm and only 10 microm spacing. Probably highly tensioned...much unlike the loose resistance controlled pressure gradient dual diaphragms.

I have been using some titanium foils but they are horrible to work with. No acoustical advantage over aluminum either. Titanium=marketing mostly.

Also the big deal on the M50 is that acrylic sphere the capsule is mounted in. Smooth diffraction shelf of 6 db or s starting at 2 kHz.

I mentioned the 4145 because it's sometimes compared to the M-50. It has a highly tensioned nickel diaphragm.

I do like the delrin. It's the standard material I use for cnc machining prototypes if they need to be plastic. It doesn't melt so easily during machining like so many others. About the only problem with it is its low surface energy. Adhesives stick poorly. I surface plasma treat it with a hand held Tesla coil to improve it.

You know what? I'll bet Klaus Heynes knows every detail of the m50 backplate. He might tell you...

http://recforums.prosoundweb.com/index.php/f/33/0

Oh, about the vent hole...yeah, small. It can be a major source of pressure (johnson) noise at LF.

Neat project. I'm sure you could get an exact m-50 type sound if that's what you want...

Les
L M Watts Technology

 

[Jonte Knif]

Thanks again!

I just came from the school of micro mechanics. We had a nice chat about the parts and it seems that they will easily machine the parts. I hope they can give an acceptable price. I wouldn't like to stop the project now.

Here is what I have drawn for them:

http://picasaweb.google.fi/jonte.knif/Mic#5370951598380477538

I decided that a "tool" would be handy. It is on that drawing too. It will facilitate the drilling of all holes, exactly where they should be and I can do the drilling at home too. It will probably be made out of silver steel and heat treated to last for ever.

Would my idea of the capillary tube work? The holes in the middle would go through the whole capsule, and in the mid there is a space for some porous piece of fabric or the like.

Are you trying to get the sound of a Neuman M-50? I'm sure you could, if that's what you want.

Well, perhaps not. How shall I put it...in my knowledge there are no M50 in Finland so I will not be able to compare. I will try to get the same response, yes. I will absolutely not try to imitate the amp. I want a mike that works for Decca tree, and I want it to be tube.

I have ordered the alu foils and some mylar to practice. I also ordered the acrylic balls.

One question about skinning. Because the air gap is so thin, I guess it would be better to attach the foil to the ring first, so that no glue is needed between capsule and foil and therefore the gap won't be altered by a glue layer. But what glue to choose? Epoxy? How can I get the thinnest possible epoxy with very low viscosity? The alu foil will be so fragile that extra friction caused by glue is probably detrimental when skinning. The glue should stick fairly well to keep the foil from breaking when the holes for screws are "punched."  (with a needle perhaps?)

When I first time printed the design in real size I was shocked...damned this thing is small. But then...look at what this guy does:

http://englishrussia.com/?p=3401#more-3401

That IS small.

-Jonte

 

[burdij]

Yes, it is very small, not like one of those gigantic capsules that Dale makes 

That's a dime next to it:

 

[Jonte Knif]

We don't have dimes in Finland.

Well, seriously, have you any comments on my design? I would appreciate. I admit that my project is not a hard core DIY project anymore, but I have no space for a lathe or time to learn to use it. Perhaps some day.

Have you had time to try different back plates yet?

Should this discussion be moved into M50 thread or is it more appropriate to keep this as a SDC-design help thread? I would still like to see this as "general" topic and not restricted to M50.

If my project materializes I'll take decent photos of the process.

-Jonte

 

[riggler]

Hi Jonte,

If you check the M50 thread you can see that I've -with a lot of help- came up with a layout for the M50 amp using a 5703WB tube. Your capsule could very well be the solution for moving the project forward. I was fortunate and was able to buy a pair of KK83 capsules for myself, but we really do need a solution for the project in general. One important thing would be to get measurements of the capacitance of your capsule as you work. The KK83's I have are 68pF. If you want your capsule to "drop in" that would be important for those using the M50 circuit.

You're right that this topic is more general in nature. I am not up to speed as much as a lot of the guys here technically, however I think I am close to getting the pieces together for the package. I've got transformers coming in from Cinemag and am working on a body design that will be able to use the U47 shockmounts. I'm going to make molds to form the headbasket grilles.

If you are not doing the M50 amp, what are you planning on, if I may ask?

Thanks!

 

[Jonte Knif]

Hi Riggler,

There are too many unknowns to know whether a capsule is close to M50 or not from the capacitance only. But with only 10u air gap the C will be quite high in "my" design. We'll see. I don't care if it is 60 or 70pF. The higher the lower noise we'll get, but on the other hand the close spacing doesn't allow for extremely high bias voltages. I trust the original design to be close to ideal to get the highest output possible with good sound quality..
I don't think my capsule diameter is very close to M50 either. Might be 1mm too big. Any way, it will be interesting to play with the polarization voltage. I suppose that I'll try to get the maximum sensitivity and not care if the diaphragm gets sucked to the back plate at 140 dB SPL

Yes, I've followed the tube choice and biasing discussion. Interesting.
I will stick to my standard design with fixed bias EF804S and 10:1 Lundahl in a D=50mm, longish body. I operate the tube with a bit higher current than standard tube mics to get good linearity and low noise. Idiot proof combination and has sounded good with K87 capsules. Very transparent compared to U87. But can not be crammed into shorth body

I'm not very familiar with sub mini tubes. I guess there is a lot of good stuff. I tried one very linear triode in a mic circuit, but the grid current was way too high, and I just gave up. On the good side the tube was extremely low noise and especially had _no_ microphony, really, nothing. I guess the small size helps a lot. So if I only had too much time I'd probably try more sub minis. I hope you'll find good specimen!

Any way, lets continue with the capsule issues. Who is gonna take apart an original 

-Jonte

 

[burdij]

A US dime is 18mm in diameter, about the size of a .02 Euro coin.

Your design looks good. From my experience with attempts so far on a similar capsule, you might want to go with less area covered by holes and a somewhat deeper depth. The drawing in the other thread is to scale.

You probably won't be able to glue your diaphragm. The glue will need to stick to the side of the diaphragm that has the evaporated coating. The coating is not attached to the surface. It is being held there by relatively weak bonding to the atoms in the plastic. I built a tensioning device with a floating ring that holds the spacer in place while the diaphragm is being tensioned and the ring placed, diaphragm pierced and screws installed.

It's hard to see, but the spacer in this photo has a hole punched in it that is slightly larger than the brass area of the electrode. Another pair of rings that hold the diaphragm fit over the ring carrying the spacer shown here. The capsule itself is shown in a slightly withdrawn position. It sits on the top of a micrometer drive which is used to push the capsule upward and set the tension while simultaneously picking up the spacer which moves along with the surface of the capsule.

The biggest problem I have encountered is keeping everything extremely clean. Even a small dust particle is larger than the distance between the electrode and the diaphragm (approx. 13 microns or .015mm).

I wouldn't worry too much about the capillary at first, either, unless you are planning on recording symphony orchestras playing on airplanes. Having had some experience with vacuums now, I know it is really hard to prevent air from leaking into just about any kind of an enclosed space.

Also, I have been working on a custom transformer. It is about 30% larger in volume than the Cinemag CM2461NiCo and uses 49% nickle steel as opposed to the 80% nickle. This is closer, I believe, to the original unit. This is a picture of the prototype:

 

[Jonte Knif]

Thanks!

Yes, I see your point. Working with mylar sounds very logical that way. I'll start with something like that.
But how about aluminum. I'm worried that when piercing, raptures will start from the holes immediately. Well, we'll see. Perhaps pointless to speculate now. If there is a problem, then gluing the aluminum to the ring should pose no problems.

The biggest problem I have encountered is keeping everything extremely clean. Even a small dust particle is larger than the distance between the electrode and the diaphragm (approx. 13 microns or .015mm).

I guess I'll make the skinning in our sauna then. It can easily be made pretty dustless. Just shower all surfaces, close ventilation and wait a while. Should work. And work naked with hear cut off, I guess. Sacrifices, but for a good purpose

Mighty trannies! I would love to start making my own trannies for special purposes. I have a nice hand operated AVO winding machine, so it shouldn't be too difficult. Well, many original Neumann trannies _are_ quite complex in design, but there is a full documentation somewhere.

-Jonte

 

[leswatts]

Hey Jonte!

Looks like you are making good progress. I took a look at your new drawings.

I guess the usual drill with pencil type SDC omnis is to tune a good bit below the diffraction peak (10 db at 27kHz) and then damp to partially or completely eliminate that peak for overall flat response.

That won't be the case with an M50 type 40mm ball though, since it's difraction is much lower frequency and more spread out. I would imagine getting the microphone upper resonance a bit higher would be in order in that case. A metal diaphragm blind hole design will sure do that!

One capsule that looks very compatible with the 40mm ball is the b&k 1/2" free field designs. They are a bit different though... A thin backplate with through holes is fitted (about 30% 1mmx1mm) and it's smaller than the diaphragm to eliminate the dead capacitance from near the edge. 2.5 micron nickel with 10 micron spacing.

But blind hole might work just as well. Starting off really shallow will allow you to gradually deepen them easily at home since it's self jigging at that point. Just a little deburr would be required...

You can glue the aluminum with epoxy, but I agree that it would be no good with a gold evaporated PET film.

You should get ok output...10 micron spacing will increase it, but the higher tension will decrease it. So a bit of a wash.

Shouldn't have too much problem with suck in with the high tension. B&k uses 200v...I think the M50 uses about 60 when the attenuator isn't in.

Lots of stuff about this in  old AES papers...expensive to download but free in the library!

Nice transformers, Burdij. As you can see from another thread I am looking at mic transformers closely.
The dual bobbin is interesting. Low winding leakage capacitance, high leakage inductance. Many of the old ribbon ones seem to have been made that way.

Les
L M Watts Technology

 

[Jonte Knif]

Hi Les!

Yeah, sometimes things proceed quite quickly. It was three weeks ago when I said to a friend that I'll never try to make a capsule. Not! Never trust me.

AES E-library has a nice paper about sphere capsule history with some technical content too. I wonder why I didn't take a look there before. The original diaphragm diameter is 12mm. Hmm...I think I'll still stick to my 14mm, just to get a little bit more output. The papers about acoustical modeling of back chambers are too technical for me.

That won't be the case with an M50 type 40mm ball though, since it's diffraction is much lower frequency and more spread out. I would imagine getting the microphone upper resonance a bit higher would be in order in that case. A metal diaphragm blind hole design will sure do that!

There has been some nice suggestions for measuring diaphragm resonances. I wonder what to do with this project. To measure the diaphragm separately would not be easy, especially at those frequencies. I do have a spark gap and I've also noticed that small bubble plastic gives very good (and loud, over 120dB!) impulses and I use it routinely. But what is the trick to measure the diaphragm movement without optical devices or super quality measurement mike which does not have resonances below 50k?

Perhaps sound and flour? Too bad that tweeters in my loudspeakers don't go beyond 25k, but that is enough?  But will the movement be big enough (before tweeter blows) to move the flour?

Of course I can just build diaphragms with different tensions and try them in the whole system. But I wouldn't want to go that way.

One capsule that looks very compatible with the 40mm ball is the b&k 1/2" free field designs. They are a bit different though... A thin backplate with through holes is fitted (about 30% 1mmx1mm) and it's smaller than the diaphragm to eliminate the dead capacitance from near the edge. 2.5 micron nickel with 10 micron spacing.

What is behind that backplate? How much distance is left between ring and center electrode, I mean for the dead zone?

The dual bobbin is interesting. Low winding leakage capacitance, high leakage inductance. Many of the old ribbon ones seem to have been made that way.

Aren't there both primaries and secondaries in both bobbins? In old Neumanns they are layered "horizontally", in Lundahl stuff vertically.

Thanks again,

Jonte

 

[Tim Campbell]

Tha AIP(american institute of physics) hanbook of condenser microphones is the only book I know of written about building microphones and it only concerns omni's. As a matter of fact it's basically a textbook on how to build a B&K measurment microphone.

 

[riggler]

You could get the book via inter-library loan from Northampton Community College, in Bethlehem, PA if you don't want to buy it.

 

[leswatts]

Yeah, the AIP book is good. It's online on google books, but as usual a whole bunch of it is missing. Library sounds good!

The B&K mics have a back chamber. It's treated as a lumped compliance at LF, but in some is big enough to have wave effects at HF. You can see some pictures at the b&k site. I have the 1", 1/2",and 1/8" units, but the smaller ones are in Chicago right now...and i'm not. Can't get the 1/8 ones anymore....too bad because I ruptured the diaphragm on one. Pressure response flat to 160 kHz!

Measuring response is always a problem. I think a small spark gap is best, far away from everything. I use one made from a gas kitchen range igniter.Time window then FFT of course to reject later reflections. Only good for HF, but you'll be solidly compliance controlled at LF, so it will be flat except very low where the vent kicks in.

To get the spark (FFT) very flat I would take care to not have anything that could reflect/refract near the electrodes like brackets or holders. just two long skinny wires with a very short gap. Whatever does hold them might be covered with foam, as far away as possible.

Few tweeters have anywhere near flat response. I sometimes use a ribbon, or electrostatic. But that's when I don't care much about flat response. Mine are for ultrasonic...up to about 100kHz. That's the kind of microphones I have been making. I analyze the things by measuring impedance...driving them as an acoustic transmitter.

It's pretty hard to measure diaphragm motion directly. Amplitude is often only angstroms!

Oh, heres a link for you...it's about using factorial design of experiments to determine small omni condenser mic parameters via the Zuckerwar model. I think it's kind of silly though. I think most DOE schemes are silly, even though I studied them in school a good bit.

http://censam.mit.edu/publications/tan2.pdf

As far as the transformer windings...Don't know! Studying them though, because I'm thinking of a ribbon mic line. (Too many out there already!)

Les
L M Watts Technology

 

[burdij]

Les, thanks for the Tan article reference.

There are some interesting observations in this article that should be considered for this type of design. The only proviso, I think, is that we may not be talking about the same types of microphones vis a vis free field vs. pressure. I don't think the M50 capsule would be considered a free field microphone, but I could be wrong.

Observation one is that the air gap affects sensitivity. I think we knew that. What is important though is that it is the highest ranking variable affecting bandwidth.

Observation two is that the annular gap around the edge of the backplate is the second highest ranking variable affecting bandwidth and apparently the relationship of one and two need to be optimized for a particular design in order to maximize sensitivity and bandwidth before all other factors.

Observation three is that the damping holes need to be kept closer to the center of the backplate. The further toward the edge they are, the more detrimental the effect on the bandwidth and sensitivity.

Observation four is that the radius of the backplate damping holes has a much less important effect on the overall characteristics and particularly the bandwidth than I think we had assumed in the past.

 

[Jonte Knif]

Uh oh,  too much math for me.
The mic under simulation was pretty small so not everything in the article might apply directly to this concept here. BUT I have been, just intuitively, thinking that because the diaphragm obviously dos not move uniformly the optimal hole pattern might not be uniform either. Just how far from optimal for example KK83 is ? Perhaps not too far?

I don't think the M50 capsule would be considered a free field microphone, but I could be wrong.

Obviously we don't want any built in boost in the capsule, 'cause the ball does it 6 dB:s anyway. Gets too bright otherwise. So basically because "free field" often refers to flat on axis response that might be what we want from the capsule alone anyway. So the capsule would be specified as free field but M50 not.

Measuring response is always a problem. I think a small spark gap is best, far away from everything. I use one made from a gas kitchen range igniter.Time window then FFT of course to reject later reflections. Only good for HF, but you'll be solidly compliance controlled at LF, so it will be flat except very low where the vent kicks in.

I made mine from modified spark plug. After filling any voids with damping it seems to work, but the impulse is not exactly the same all the time. Measuring those small Panasonic capsules I get fairly nice response up to 40 kHz. The spark seems to produce rising 6dB/oct resp. But the SPL is very low.

 

[Marik]

Hi folks,

Here is my initial plan and drawing for general purpose omni capsule. Perhaps ultimately leading to something M50-like. (haha, sure)

http://picasaweb.google.fi/jonte.knif/Mic#5369380081824378450

I need some help. I searched the microphone meta and found some things but obviously people are more interested in bigger capsules.

I recall somebody saying that about 1,3mm hole size and 30% hole area works. This is in agreement with photos of M50 for example. But nothing about depth. Obviously I can have different versions made, but I need a starting point. There is abundant info about holes in big capsules, but nothing definitive about small. Something in M50 thread, but not too much.

There is a school of micro mechanics a few blocks from my house, so I can get the accuracy I want and it will be very easy to communicate. I don't know the price yet , might be too high, but I'm going there on Tuesday to discuss and have to have more knowledge.

About my design so far:

I was thinking that messing with diaphragm spacing ring is stupid, and decided that it is easier to have the spacing built in another way. So, after the parts are screwed together the surface is lapped and shimming material (blue) is then put between the plastic base and outer brass "tube". I can not see any reason why this wouldn't work. Also, it is easy to try different distances without diaphragm replacement.

The wires are easily attached to the screws on back side which is nice.

The plastic? Recommendations? They have PTF which is electrically superior, but is it hard and stable enough?

Is there anything super critical concerning the capillary hole? (not in the picture) I guess the smaller the better?

Thanks in advance,

Jonte

Hello Jonte,

I spent considerable amount of designing and making SD both cardioid and omni capsules. The very first observation (unless you have a real M50 capsule to take apart and make precise measurements and then clone them), with your arrangement it will take an awfull amount of time for you to put together and then take the capsule apart. Be prepared to take it apart some 1000s times, so what you really will need is LOTS of patience. First, I'd suggest to make a very mobile prototype, so you can check your backplates and spacers very fast. The obvious arrangement would be MXL603, where the whole thing can be taken apart and then put together in a minute or so, just with one collar nut. I would make that one first.
Very soon you will figure out how different dimentions affect the response--take good notes on that.
You will need to figure out the backplate right damping and will need to do LOTS of machining work. I'd start with more shallow blind holes, as it is easy to enlarge them.
Then you will need to figure out the right spacing. Once you are done, then you can make a final prototype, already with clamping bolts, etc., but with all internal dimentions you found.
Usually, the venting "tube" is made by the splitting the spacer, so you can worry about this part in the end, after all the major work is done.

Good luck, and keep us posted.

Best, M

 

[Jonte Knif]

Hi Marik,

Thanks for the tips!

so what you really will need is LOTS of patience

I do but only if I'm motivated

The way my capsule is now designed requires the removal of 6 screws from the back to take the backplate+delrin botton part out. Perhaps it is not obvious at the first sight but there are not 12 screws on the bottom side edge. The 30 seconds spent shouldn't be too much of a problem. Then the backplate holes can be trimmed without separating the delrin from it.

Trimming the spacing will be very easy and safe because the diaphragm doesn't have to be removed, (I hope)  but for the same reason I can not use the split spacing ring, because the spacing will be carried out with a couple of foil pieces between the delrin and capsule "body" and the delrin and body have an airtight contact. Perhaps you didn't see the first drawing which demonstrated the method. It is not there anymore.

I hope I will have money for 3 capsules and several diaphragm rings to start with. I'm also very concerned about good notes, too many times one notices that after a couple of days work some earlier version is needed again to be compared.

If and when I start with 10u spacing I don't have to mess too much before I have to lap the surfaces again to keep things in alignment. This might be the biggest problem affecting the accuracy of a series of measurements in my plan.

I really would like to have reliable method for measuring the diaphragm resonance before screwing it to the capsule. The impedance method sounds quite good, but then I have to build an acoustically transparent backplate for that job only. Worth the trouble I guess.

I'll keep you posted, for sure. From this point on it will take some time to get results.

The first capsule will be assembled in a couple of weeks or so. I hope it is not a total disaster.

-Jonte