R. Williamson capsule

PPR's comments have been VERY helpful. I'm going to try it with the 2 mil gap, and if that doesn't work well, perhaps I can shim behind the backplate with a 1 mil shim--provided I can get the through holes aligned again when reassembling the capsule! On the other hand, perhaps it would work better without the through holes...

I think I would use Williamson's design, actually. A 2 mil gap is pretty wide - that kind of gap is commonly used in capsules such as the AKG CK12, and I would drill the through holes as he has marked.

I think the Schulein capsule is flawed because the diaphragm should be damped enough so the mid-band tuning of the mic does not appear as a peak in the frequency response. This is where I think the closer diaphragm spacing and not machining the step is going to help get a response you'd be happier with.

You might want to get enough material together to make a few capsules and experiment with them. I've built about 8 capsules, of which one sounds good. The others were varied but generally were either mediocre or just plain bad.

> perhaps it would work better without the through holes...

Operation without the through-holes is quite clear. It is omni. The resonance is controlled by the diaphragm mass/area and the total stiffness. The total stiffness appears to be 90% trapped air. The mass/stiffness ratio seems to point to the top of the audio band, as it should: any lower and you lose treble, any higher reduces sensitivity without audio-band advantage. If un-damped, the resonance would bump-up the top of the audio band. Damping is gotten by putting about 85% of the back-volume down in holes, thus making most air displacement wheeze through the thin gap between diaphragm and backplate before reaching a hole.

So the back-gap, and the ratio of gap-volume to hole-volume, are key parameters.

The number of holes is not, of itself, a first-order parameter. We want damping distributed all over the diaphragm, but up to 5KHz-10KHz the dimensions are "small". Some designers use hole-drag for damping, requiring a large number of small holes to also get a reasonable back-volume with a reasonable back-gap. That might even work better. But as others have noted, huge numbers of teeny holes means hours of tedious labor, drilling and breaking bits.

Now those through-holes.....

I see the basic idea. Same as a dynamic cardioid: delay the back-wave in an R-C network and let it cancel the front-wave. When the R-C delay matches the delay in the air-path from front to back, you null the rear of the pattern. Slight shifts of delay give hypercardioid or fat-cardioid, still with significant directionality. This is how the $3 electret cardioids work.

And since the stiffness in omni-mode is mostly "trapped air", the tuning naturally shifts when you open some holes into the air-space. So, like a dual-diaphram, the tuning can be about right either way. (One difference is that you can extract omni, bi-di, or cardioid from a dual-diaphram directly; the Williamson requires mechanical plug/un-plug action.)

But I don't have the intuitive "feel" to know if the Williamson can be a good cardioid, or just an omni with an attitude.

Dale and PRR, thanks for your comments on the design, which are again VERY helpful. I was only half serious in suggesting leaving the through holes out: the serious half was the thought that perhaps I'd be better to try to build a decent omni capsule, given the difficulty I foresee disassembling and reassembling the backplate in the body and getting the through holes (which have already been drilled) to line up again perfectly and with a proper shim behind the backplate: The more I consider this, the more I am convinced by your comments that the original 1 mil spacing should be maintained. Of course I will be building several capsules, not just the one, but I'm trying not to get too far ahead of myself. DIY is an addiction, you can't build just one of anything; my wife will attest to that--the house is full of guitars and other stuff I've built. The only saving grace: it would be impossible to die (wouldn't it?) before a project is completed, and if you keep starting things..

To see how things work, the last few days (since the thread started--thanks Gus for bringing it up!) I spent playing with this type of arrangement about 10 hours every day.

At first, I used a Chinese backplate to see what happens bringing things to extremes, so I used the side with blind holes, and then its flat side. I drilled 4 bigger diameter holes for a phase shift network, and then plugged certain pattern of holes to see how they influence response and pattern.

[quote author="PRR"]> I see the basic idea. Same as a dynamic cardioid: delay the back-wave in an R-C network and let it cancel the front-wave. When the R-C delay matches the delay in the air-path from front to back, you null the rear of the pattern. Slight shifts of delay give hypercardioid or fat-cardioid, still with significant directionality?..

?.But I don't have the intuitive "feel" to know if the Williamson can be a good cardioid, or just an omni with an attitude. [/quote]

Although in my latest version (see below), I took a long deviation from Williamson arrangement, I feel the last should work fine, especially if well tuned. First, with different "plugged holes" arrangement, the pattern would change significantly. In its "optimal" position, the pattern is close to fig8 in low band with close micing, turning into nice cardioid with increased distance from the mic, which actually is one of the properties of the single diaphragm capsule. In my capsule the omni was OK. Williamson might be also fine.

[quote author="PRR"]
So the back-gap, and the ratio of gap-volume to hole-volume, are key parameters.
[/quote]

Precisely!

[quote author="PRR"]> Operation without the through-holes is quite clear. It is omni. The resonance is controlled by the diaphragm mass/area and the total stiffness. The total stiffness appears to be 90% trapped air. The mass/stiffness ratio seems to point to the top of the audio band, as it should: any lower and you lose treble, any higher reduces sensitivity without audio-band advantage. If un-damped, the resonance would bump-up the top of the audio band. Damping is gotten by putting about 85% of the back-volume down in holes, thus making most air displacement wheeze through the thin gap between diaphragm and backplate before reaching a hole. [/quote]

That?s what I have read, but have some confusion, as I got completely opposite results in practice. When flipping the backplate to its flat side, intuitively I feel that diaphragm stiffness should increase significantly, due to very thin layer of air trapped between diaphragm and backplate, and top end should be flat, with a bump due to undamped tuning resonance. However, in omni mode, the top end was completely gone, whereas putting the backplate the other way around (with blind holes towards diaphragm, I got nice top end, but the low end was gone.
Obviously I am missing something here?how the thin air vs. viscosity of the air in the holes contribute to the stiffness and damping?

[quote author="PRR"] The number of holes is not, of itself, a first-order parameter. We want damping distributed all over the diaphragm, but up to 5KHz-10KHz the dimensions are "small". Some designers use hole-drag for damping, requiring a large number of small holes to also get a reasonable back-volume with a reasonable back-gap. That might even work better. But as others have noted, huge numbers of teeny holes means hours of tedious labor, drilling and breaking bits. [/quote]

I took this route, and made a new backplate. I started with smaller number of holes, and listening to the result, little by little drilled the new ones, until I felt the capsule sounds about right to my ears. Needless to say, I have never had such a frustration, and done such a tedious and hellish work?to audition each additional drilled hole requires dismounting the mic, taking off 12 screws to dismount diaphragm, drilling the hole, and then putting everything back together... just to realize it still did not sound how I wanted to....

I spent playing with this type of arrangement about 10 hours every day

Click to expand...

:shock: :shock: :shock:
We're not worthy!!!

Well, I went back and listen to the capsule again. As a reference for cardioid I used K87. Both capsules had similar tonal balance between bottom and top ends. They sound very different though. My capsule does not have kind of boxiness and compression in the sound, compare to K87. I attribute it to more "breathing" due to its vents in the back vs. enclosed volume of K87.

With omni mode evaluation I was little too fast--I was testing it late in the night, way too tired. Anyway, now as a reference I used Studio Projects C4 with omni capsule. Mine had more top end, but the bottom end suffered. I guess there are three possibilities:

1) It needs more tweaking, or
2) The omni mode does not "like" my particular backplate construction, or
3) There is no free lunch, and omni and cardioid cannot work equally well with the same backplate in LD capsule.

In dual diaphragm design omni works differently, where each half is a cardioid, and then both signals electrically combined with right phase.

But at the moment I am quite happy with its cardioid performance, which was my main concern. Next I will try it with SD capsule.

I don't have a photo yet, nor audio files, but this sounds pretty good. There are no instructions or drawings for it, I just tried a bunch of capsules this weekend and this one sounded nice. This is a pretty simple capsule and hole pattern. You can make the backplate assembly just like in the Williamson article.

Centre terminated capsule, 32mm brass backplate, 6mm thick.
Diaphragm diameter is 25.4mm.
Drill 16 holes on a radius of 10.5 mm, #58 bit, to a depth of 4mm.
Drill 4 holes on a radius of 6.25 mm, #68 bit, all the way through.
Diaphragm is spaced from the backplate by 40 microns. Use a 0.0015" shim for this, or machine it into the backplate.
Diaphragm tuning is just enough to get the ripples out, plus a little bit, or you can go for about a 700 to 900 Hz tuning.

This is very similar to a Sony C37A backplate, though I have not measured a real C37A yet. I might get a chance a bit later. The Sony is a bit different in that it uses a centre support, so from a diaphragm perspective, it acts like a centre terminated capsule. Since I am set up to easily make M7-like capsules, I adapted it to a centre terminated, glued diaphragm setup.

Thanks Dale.

I just tried a Debenham capsule, and I highly recommend it! My diaphragm spacing is a bit high, so the output is a bit low, but it sounds great. The drawings are here somewhere, check the mic meta. I just built it using the drawings in the article, except for some adhesive and plastic substitutions - I used nylon 6-6 instead of Perspex, 3M DP820 to glue the (brass) backplate to the ring, and brass for the backplate.

-Dale

[quote author="dale116dot7"]I just tried a Debenham capsule[/quote]

Damn, Dale! You're just too much! :thumb:

Any chance of some audio clips?

Peace,
Al.

Yes, as soon as I rectify the diaphragm spacing problem. I got a dial indicator set up at work and I have new diaphragms so I can take the capsule apart, lap it in to the right spacing, and try again. The diaphragm spacing mostly wrecks the sensitivity, and I get a lot of hiss because the output is a bit low.

-Dale

Hi Dale .
What frequency did you tune the diaphragm to ?
Andy

I tune most of my diaphragms a bit low, around 850 Hz. I tune the diaphragm by hanging a large washer, 110 to 120 grams, off of the ring the diaphragm is attached to.