Getting gold onto diaphragm

[dale116dot7]

I tried putting my diaphragms on a backplate I machined up today, and it worked! Not perfectly, mind you, but it worked. I glued the diaphragms on with cyanoacrylate, and the middle insulator I made with acrylic. It sounds pretty good, but the pattern is somewhere between hypercardioid and figure-8 with only one diaphragm connected - obviously I didn't get the machining quite right, or maybe the design is bad. I suspect my air volumes on the two sides of the diaphragm are not the same. I may try the other side to see what its polar response is.

 

[asm]

great work man. VERY inspiring. hopefully in the future i can mess around with something like this.
did you just take a stab at learning this process or did you have a similar process book, ect.
also, what are you getting a gold source? just scrap?


keep up the good work, dont hesitate posting more pictures. i enjoy them and im sure everyone else does too.

:thumb:

 

[dale116dot7]

http://www3.telus.net/~dulan/MyCapsuleSmall.jpg

 

[dale116dot7]

I get the gold wire (and evaporation boats, vacuum grease, etc) from www.2spi.com.

I've tried another capsule, and got a very bright HF response. It looks as though larger blind holes (larger diameter), and more of them, increases the high-frequency response - and I over-did it. I'm thinking more holes = less air film damping. My low frequency response still isn't very good (it drops off below 200 Hz), and the polar response has only about 8 dB front-to-back except at low frequencies, where that drops off. Below about 200 Hz it appears to have a figure-8 pattern, gradually changing to cardioid above 1 kHz. The mic with a lot of holes seems to have an 8dB peak at 10kHz, the one pictured on the post above has about a 4 or 5dB drop above 4kHz. I am using 55 micron spacing between the diaphragm and the backplate on both mics.

 

[Gus]

dale116dot7

From what I can find it seems the transition points
"Below about 200 Hz it appears to have a figure-8 pattern, gradually changing to cardioid above 1 kHz." are were the "art" is in design a capsule.

Some of it has to due with the diameter of the capsule . Not only do you need to get the 8 and Omni part right for the cardiod patten at lower freg to the transition point you also need to need to control the thru hole number size for the change from holes working to not working as good.

I still am tring to understand how one can get a cardiod at higher freg because the backplate air hole system mess changes.

This is from reading I have to get around to making some backplates and do some real learning like you and Tim.

I am beginng to think almost no math is used when one builds a capsule. The math gets you to a first protot type them you build a bunch writeing down the measurements, do some thinking then build another etc etc until you get it "right"

Thanks for the update and information.

 

[Marik]

[quote author="dale116dot7"]

I'm thinking more holes = less air film damping.
[/quote]

In theory, the more air damping--the flatter response--the less output of the capsule. On the other hand, if we get to another extreme--much less holes, then we will get the same effect--too much HF response and no LF.
It seems that the hardest is to get exact balance between all the parameteres, otherwise any even small deviation results in far off response.
Try to punch a little hole in the mylar, to let the capsule breath a little bit.

 

[dasbin]

So when are you going to start selling these babies to all the prodigy folks? :grin: :grin:

 

[dale116dot7]

The pictured capsule is closer to flat than the latest one. The new one has about the same number of holes as a capsule from a U87, but the through holes are implemented a bit differently - they are nestled in between the blind holes - since I'm working on a single-backplate design. All of the on-grid holes are blind. The blind holes are 3.9 mm deep, and are a #54 bit. Through holes are a #73 bit. The capsule is 6mm deep. The hole pattern is on a 2mm grid, every hole drilled, but 28 through holes are nestled in between some of the blind holes.

The capsule in the picture has blind holes 3.9mm in depth, but a #58 bit. Through holes are a #68 bit and the capsule is 6mm deep.

I might try turning down my existing backplate a bit at a time to shallow up the through holes, but I might try the existing backplate with a closer spacing first.

I'd like to get a feel for the single backplate design first, I might try a dual backplate next. For that I'd start with a KK87-like design, but with Marik's suggestion of shallower holes.

The Altec M11 (Coke Bottle) patent shows that they vary the air damping by varying the spacing of the backplate to the diaphragm, and it is adjusted for flattest response.

Unfortunately, Santa didn't bring me a digital camera so photos take their time - I need to use a film camera and get them put on a CD at a photo lab.

I doubt the rolloff is the electronics - the 797 capsule that was in there before didn't experience this rolloff. I might check to see if anything changed, though.

So far my biggest concern is trying to get the LF to not roll off - that portion I'm not so familiar with. To me it seems as though I would need more chamber volume or even smaller or fewer vent holes, but the number of them seems pretty low already.

-Dale

 

[dale116dot7]

#73 = 0.6mm
#68 = 0.8mm
#58 = 1.1mm
#56 = 1.2mm
#54 = 1.4mm

I wanted to make the current one thinner to experiment with the top end resonant part, bit while I'm working on it I might try converting it into a dual backplate design - make it thinner - about 4mm, by reducing the front side. That will make the blind holes shorter. Then I can add another plate to the back, and even experiment with or without a chamber, plus experimenting with capsule thickness fairly easily.

I got 28 holes by staring at the K47 capsule, and it has something like 30.

Unless I'm doing things wrong, these holes should resonate at about 30 kHz, or maybe I need to use a different formula when the length gets that short?

I'm running another set of diaphragms right now - the diffusion pump is just warming up.

-Dale

 

[Marik]

Dale,

How about sealing some holes, and shallowing some others with something like plasticine? My guess, this material is pretty easy to work with, and also, you will have a lot of flexibility in fine tuning.

 

[Marik]

[quote author="Tim Campbell"]You don't mention if your capsule is vented. If it is vented block this vent. This could affect the bass response.[/quote]

Tim,

For me still there are a lot of things to undestand in the capsule. Two of them:

1) From my experience cardioid two diaphragm vented capsules exhibit raise of LF, vs. LF cut in true omnies.

2) With true omni capsules I experienced raise of LF and HF cut with raise of polarazing voltage, vs. opposite behavior in two diaphragm capsules.

 

[dale116dot7]

My capsule is a single diaphragm, the through holes just go out the back of the capsule. I tried machining down my existing capsule to do two things: 1. make the holes shallower, and 2. re-set the diaphragm spacing to a reasonable number. Those two changes controlled the high end a lot. The HF response is similar to my TLM103. The hole depth is now 3mm instead of 4mm. The other thing is the closer diaphragm spacing and shallower holes lowered the front-to-back rejection. I then tried plugging some of the through holes and got even worse front-to-back performance. The low frequency response did not change even with changing the number of holes. There is a proximity effect, but the response starts rolling off at about 200 Hz.

 

[dale116dot7]

I tracked down the low-frequency problem. It was the preamp - I had mistakenly put a 100 M resistor in place of a 1G resistor. I tested the preamp with a signal generator and something like a 1000pF capacitor to couple the signal - and the preamp tested flat. I tried it with a 47pF cap to couple in the signal and it wasn't flat. Thanks for pushing me to re-check that! So now I get thunderous proximity effect on my first capsule (the one in the photo).

I've been doing frequency response plots of each change of mine, plus usually two other mics - I've been using the TLM103 and the AT4050 as a comparison. My last set of changes include blocking half of the through holes, and that made the cardioid more like an omni - I used to have 10 dB difference between front and back, and now I'm down to 6dB. I'm thinking a few more holes would make it work better, now that I've figured out my lack of low frequencies. I'm not sure if I will continue working with a single diaphragm mic, or switch back to a dual diaphragm. I think I like the sound of a dual diaphragm.

 

[dale116dot7]

I've tried two diaphragms on the same backplate, different tensions (one is about double the other) and there is almost no difference in frequency response (or sound). Perhaps the Neumann design is not very sensitive to this parameter? I was thinking this might be a way to control the top end of the mic's response but apparently not so much. It appears to be affected more with the hole depth, diameter, and the number of them. Is this what other people have found?

 

[PRR]

> It appears to be affected more with the hole depth, diameter, and the number of them.

AND diaphragm spacing. Spacing should generally be as small as practical, limited by pull-in voltage.

In a basic omni condenser, with flat backplate, small spacing means low back-volume, high stiffness, low sensitivity, some damping of the top-resonance. Large spacing increases sensitivity but gives negligible damping. Somehow you need close spacing with large trapped volume. Blind holes increase the volume while allowing close spacing.

In a bidirectional dual diaphragm condenser, you need large damping over the whole audio band for air sloshing through the backplate. Small through-holes can do that.

In a dual-diaphragm used in omni mode, the through-holes are equivalent to blind holes of half the depth. (Air at the center of the through-hole is a null.) Used in bi-di mode, the blind holes and the through holes damp each diaphragm, the through holes flatten the overall bi-di response AND set the front-back discrimination.

The Debenham article (6MB PDF) is a good starting place for holes.

 

[dale116dot7]

I looked there but I'm doing a centre-terminated capsule which is a bit different. My starting points were that article and staring at the KK47 capsule, which is also a single backplate. I have been trying a single backplate single diaphragm design starting with something of a cross between a KK47 and a K67.

Spacing has been interesting - I've started using a lapping plate and a fancy digital micrometer to get that dimension down to 30 microns. I find it easier to machine in that step on the capsule than make a nice spacer. It limits adjustability doing it this way - you can only make it closer by removing the skin and lapping some more.

 

[dale116dot7]

I ran two new capsules today. I tried one with lots of blind holes but very shallow ones. Result? Something like a carbon mic in a telephone. So perhaps the shallow holes don't give the high-frequency pick-up that is needed, but lots of shallow holes reduce the damping for the midrange? I might try removing the skin and trying one with a higher tuning to see if that makes a difference. I might try removing the diaphragm and making it closer to the backplate.

The other capsule was my M7 copy. I don't know the correct hole depth, though. I drilled this one to 3mm. The high-end rolls off rather abruptly, so I'm thinking the holes need to be slightly deeper. The pattern is pretty good, a decent cardioid above about 300 Hz or so, and slowly changing to a figure-8 at low frequencies. This one is pretty close.

 

[dale116dot7]

Here's some data if anyone is interested...

I wanted to know what weight corresponded to what diaphragm tension. I now have the data to figure it out. It takes a lot of weight to shift the frequency - I was rather surprised.

Test setup - resonant 'ping' measurement
Diaphragm mounted on a 40mm ID ring.
Test ring is 31mm (pretty close to most capsule sizes).
Weights were round washers with a 40mm ID. The mounting ring's mass was counted in these measurements.

45 g - 725 Hz
72 g - 862 Hz
98 g - 900 Hz
125 g - 1000 Hz
245 g - 1200 Hz
355 g - 1350 Hz

To record the 'ping' I used an Altec Coke Bottle (omni capsule 21D, high-tuning so its capsule resonances aren't anywhere near the band I'm interested in), and a 'scope. The ping was generated by dropping a small washer on a diaphragm. I tried two diaphragms, one with gold, one without. The results were within maybe 50 Hz. I also tried going up and down, and everything was pretty repeatable.

 

[Gus]

dale116dot7

Thanks for the information.

 

[gyraf]

Good info!

Diagphrame material and thickness?

Jakob E.