Electroform Suppliers

FWIW

http://garelectroforming.com/

Thanks, I'll give them a try.

Thomas-Register-listed nickel electroforming companies. A number of companies listed do optical-grade work.

I would like to ask, what is the principle advantage of electroforming these parts over simple stampings?

Quite likely the parts are complex and incorporate shapes such as seamless bellows, for example. Not readily amenable to stamping.

There are several features, the quantities involved do not justify the tooling cost, the internal features need to have sharp internal corners, one area is undercut, flatness is critical & we are trying to hold 5µm tolerances. It is hard to describe how delicate these parts are, the finished component has a mass of 0.75g.

If you research "galvano" manufacturers you might find someone. Our supplier was in CANADA making corner cube arrays for retroreflectors in lense tooling. These corner cube sets were sub micron accurate for surface finish and angle. Most sets were in hesize range that you nominated. (Think roadside reflectors or reflectors in vehicle rear lamps or trailers.)

Thanks to previous posters I now have quite a list of potential manufacturers, what I was hoping for were recommendations from people who have used electroform companies. Later today we will be talking to the Dutch sister company who were making these parts to try to understand where their problems lie & hopefully get them to continue. Even if this works out, we will still be looking to 2nd source the parts elsewhere.

http://slideplayer.com/slide/9803120/

Are you able to divulge any information about this photon detector? The energy of the photons, for example, would be a clue.

Is it like a moth's eye or fly's eye? Is it arranged in a spherical symmetry or in a ring symmetry? Is the array larger than a bread box?

In reply to you and Andrew, I can't really say much about the detector. We have three years invested in the design of a novel photomultiplier and have just about perfected it. The device is 60mm square with an active area of 53mm and a multi-anode output (currently 32x32, but we have made 64x64 and are aiming for 128). As this increases, the problem is how to make 16k+ electrical connections (for 128x128) in a 53mm² area. These devices can be stacked together on three sides to make multi arrays, each having a gain of around 1x10⁷. They are part of the TORCH project at CERN, which is described here. The multi-anode output detects photons but also allows positional and temporal information to be extracted.

From what I read in your link, the spatial and temporal resolution requirements are indeed quite high; I am thoroughly impressed by the old and the new capabilities.

Apparently, less gain is required by the TORCH on LHCDb, is that right?

The light gathering part, which our detectors with fit to, is a massive quartz light pipe. They expect to funnel most of the available light to our detectors so the gain is not so critical, but I'm sure they won't complain if they get a bit extra.

I'm guessing you're using some form of microchannel plate for the gain stage. 16k anodes in 53 mm² is <57 microns centre-to-centre. Doesn't give you much room for interconnects; clearly you can't bring the connections out the back if these detectors can be stacked, so you're pretty much stuck with edge readout. :(

Can you fabricate the anodes on silicon and multiplex them, maybe serial readout by row or column or even by partial rows or columns to minimize latency? How much actual area per anode? What I'm getting at is: could you fit the multiplexer circuitry between the anodes? Alternatively, could you back-etch the array to expose the anode dopant wells? I don't want to ask you too much detail, but interconnecting that many anodes in that little space is an intriguing problem to say the least.

These are stacked side by side to give a large area array; the connections come out of the back to pads through vias in the ceramic anode. To connect, we make a PCB with a matching anode pattern and interpose a piece of ACF material to connect and bond the two. The PCB then outputs to high density connectors. Our problem there has been to develop a bonding method that works without destroying the detector through heat or pressure.

Use graphene on the ACF?

No, connection has to be in one axis only with minimal cross-talk between pads.

The link shows a minimum spacing of 100 microns, conductive particle size of 10 microns, and that a pre-tacking temperature of 90°C under a pressure of about 10 kg/cm² should be used. That's a lot of pressure for a delicate array.

That spacing will work for lower-density arrays but not for your 128² array, at least not in a 53 mm² area where your anode centre-to-centre spacing is <57 microns.

Just a thought: for your 128² array, have you considered a 'eutectic-solder' approach using, say, gallium or gallium-alloy wetted contacts/bumps between the array and the PCB? Little pressure and much lower temperature.

Correct, we've only used this on the 32x32 so far. We've looked at bump bonding with solder & epoxy; the problem is controlling the spread of the contact material. That is essentially down to flatness, but our anode inevitably has a very small bow because it is under vacuum.

The new mulichannel optical detectors have quite high gain, even in pixel mode.

I believe that, in fact, data is off-loaded pixel in series mode row by row (if I am not mistaken). It becomes a time stream.

Additional information can be had from the signal (under circumstances where these detectors are used in camera mode) by incorporating another high dynamic range scanning mirror detector to allow measurement of the light intensity from various areas of the image plane with up to 160 dB dynamic range (far beyond outstanding). This is expected to vastly improve autonomous vehicle discernment of "targets" in the area under low contrast light conditions (think white box van against pale blue sky).

A finished detector, we now have 6 of the initial 10 we need to deliver & should have the remainder soon.