Getting gold onto diaphragm

[Bo Hansen]

I find this Neumann capsule making links for some time ago, I don`t remember if it was in this forum, but any way here they are:

http://www.omnipressor.com/Other/NeumannBookPix/NeumannAu1.jpg

http://www.omnipressor.com/Other/NeumannBookPix/NeumannAu2.jpg

http://www.omnipressor.com/Other/NeumannBookPix/NeumannAu3.jpg

http://www.omnipressor.com/Other/NeumannBookPix/NeumannCapsule1.jpg

--Bo

 

[rafafredd]

Man, this is beaultiful.

 

[zebra50]

[quote author="Bo Hansén"]I find this Neumann capsule making links for some time ago, I don`t remember if it was in this forum, but any way here they are:[/quote]

Sshhh! Keep it quiet! I'll get in trouble with the copyright police again

:green:

z50 / omnipressor.com

 

[microx]

I am sure that one day I will read "Prodigy group solves problem of nuclear fusion using modified domestic appliance".
BTW welding shops sell DIY size Argon cannisters.
Steve

 

[Bo Hansen]

Zebra,

Sorry
(but you are :green: so I don`t know if I shall delete the post ??)

--Bo

 

[zebra50]

Hi Bo,
Just joking!

No problem. Leave it for now. I'll take the files down in a few days or weeks when everyone has had chance to look.

The book is still in print so people should go buy it if they like this kind of thing.

Stewart

 

[dale116dot7]

Update: I got my 6 cfm backing pump and vacuum gauges that measure to about 0.2 microns. The pump tests out at below 20 microns and that's enough to back the diffusion pump. Tomorrow I will be machining the jar baseplate to accept the diffusion pump flange. Hopefully in a week I will be able to test for leaks and in another week try another run of diaphragms. My setup should be able to do three diaphragms at once. I'll post photos when it's constructed.

-Dale

 

[dale116dot7]

Update: Tested everything out. The diffusion pump works, but I have a leak that prevents the pressure from falling below about 4 microns - I need about 0.2 microns, so I've pulled everything apart and I'm re-doing all of the pipe connections. There are a few pipe thread (NPT) joints and they are likely the cause of the leaks. A pressure rise test confirms that it's a pretty large leak, so getting 0.2 microns shouldn't be a problem once that pesky leak is gone.

 

[gyraf]

Hehe.. Keep us updated - this is very interesting..

Jakob E.

 

[tk@halmi]

[quote author="zebra50"]
The book is still in print so people should go buy it if they like this kind of thing. [/quote]

Where would I get this book?

Thanks,
Tamas

 

[alk509]

Man, looking at those pictures I can't understand how they keep the diaphragm tension consistent from capsule to capsule!

So is that setup used for gold sputtering or evaporation?

It all seems so easy, doesn't it? :wink:

Peace,
Al.

 

[dale116dot7]

That setup is a thermal evaporator. The gold wire in the tungsten boat indicates this is the process they are using - a sputtering system would use a solid gold disk and another electrode in the chamber.

From my calculations, the final diaphragm tension will be pretty consistent (within a few percent to maybe 10 percent at the outside) based on the weight that they place on the diaphragm before they clamp it down to the backplate, although the initial clamping arrangement (where they cut out the diaphragms out of the mylar strip) has to be pretty well engineered to have an even tension to start with - I suspect someone did a lot of work to come up with that clamping jig.

The Neumann microphone design (K67/87) is less fussy on diaphragm tension in terms of response than, say, the AKG CK12 capsule. Neumann also has the luxury of building a couple hundred capsule halves, testing them, and matching the closest pairs for pattern and frequency response, so maybe very accurate tuning isn't critical for manufacturing.

-Dale

 

[tk@halmi]

Back in school we made diodes in the semiconductor lab. The metal layer was aluminium. The evaporator was nothing more than a huge glass bell with a hard rubber seal on the bottom. There were two pumps:
1. Rough pump.
A pretty simple scroll pump was used to get down into the 50 mTorr range.
2. Diffusion pump.
This was a liquid nitrogen operated pump. Every five minutes you had to pour a pint of the cold stuff into a funnel on the side. This got you below 1 mTorr easily, but it took sometimes 20 minutes to pump down.

This was very low tech, but we could get very thin and uniform Alu layers and a many decent pn junctions out of it.

 

[dale116dot7]

That's pretty much the same as my setup, except I'm using an electrically heated diffusion pump without liquid nitrogen. Likely the nitrogen is used for a cold trap. The devil is in the details - sealing any connections, for example, the electrical connections for the heating element, plumbing for the vent valve and vacuum gauges, etc. is what takes a lot of work.

 

[dale116dot7]

I tried the setup again... after months of work getting rid of leaks, getting bigger pumps, expensive pump oil.....

Nickel coat is working!!!! I only set up one heater so I could debug my process first, and I do have a couple of things to do. First, the heater element must be absolutely flat and symmetrically supported. Any asymmetry causes the boat to bend and the sample (nickel or gold) to fall off. Second, a very slow current ramp up seems to be what's needed to help avoid this problem - if I do both, it should work better. Basically, once I got about a third of the sample melted and deposited, it fell off of the boat and onto a baffle.

I was a bit nervous pulling the jar off, all I knew was that the sample got pretty hot but nothing melted or leaked. I was rewarded by a translucent silver-grey spot in the middle of a piece of mylar.

Photos tomorrow night! I took photos but I need to develop and scan them.

 

[dale116dot7]

The nickel is supposed to be good for gold adhesion - if the surface isn't exacty perfect to accept the gold, in other words, a bit of safety. It's actually no trouble to put on the nickel - it just takes a second heating element and a second electrical feed-through.

As for the thickness meter, that would be nice, but I probably don't want to spend too much more on my setup. I've been using opacity to judge the thickness so far, and I'll see how that goes.

 

[gyraf]

Fantastic work, Dale...!

:thumb:

Jakob E.

 

[dale116dot7]

http://www3.telus.net/~dulan/deposition_system.htm

The initial glue-down and putting the diaphragm on the capsule was an earlier test capsule - it's not the deposited part. Note the capsule has wrinkles around the deposit - I need to heat-treat the diaphragm (pre-shrink it by heating to about 150 C) before putting it in the chamber to avoid that.

 

[Gus]

I like the clean room? might it be better with lead cast in a ring that you place on the ring.

What is the big gear bottom plate from?

Very cool building with stuff you find and have


150C? Tg is 80C

 

[dale116dot7]

Clean room? Just wait for the dust to settle, I guess.... I expect the odd failure but what the heck... I might clean off my bench before doing any diaphragms for real.

I was going for a shrinkage of about 1%, which is what I estimated from the ripples in the diaphragm, the Dupont data I have indicates that 150C is about 1% shrinkage - it's just before the knee in the curve.

The gear was a prototype reluctor wheel for an engine position sensor. It was to be installed on a British 469 litre engine (9 cylinders, 500 rpm, 4500 hp), but it was too big for the housing - we used the next smaller gear. One of the teeth is milled off for getting a reference.

I didn't post all of the photos, but there's a diffusion pump right underneath the plate, and a mechanical pump right by it. So far I have not had any problems pumping through the diffusion pump then heating it up. This pump was designed for that service (no roughing/high vacuum valving), as it has an additional cooling coil to get the pump turned off in about 10 minutes. The cycle time is about an hour. The top plate will have another two holes put in, that's why it's off to the side.