Ann: Market Update - Progress Report, page-148

What you are describing is essentially the prober. This allows every die on the wafer to be addressed rapidly and sequentially for testing purposes, and presumably is what has allowed the MEMS dies to be driven by a separate or emulated ASIC. I suggested that there is reason to have one or maybe two of these probers for redundancy or throughput, but as they probably cost hundreds of $K to build, why would you build four? Yes it would be possible to construct a device to connect to 4 MEMS in some kind of a holder, and to use 4 ASICs alongside them, so you could drive 4 MEMS in the cleanroom. But temporary and makeshift devices have a habit of giving poor connections and hard to fix faults. So why would you go to that trouble when what you want is a bunch of reliable, high quality, standalone devices that can be given to prospective customers to trial the finished and packaged chips in their own labs? They have developed these demo devices prior to COVID assuming they would be provided along with chips fitted, QC'd, and ready to play. It is not as if they will not have packaged chips in time for global customers to visit or be visited and demonstrated to convincingly. Why would they rush this step and risk handing out problems?

The still image probably shows everything you would see from a video - i.e. nothing. The MEMS move at ultrasonic speeds and are very small anyway - ~3 pixels per mm. I suggested to Danny Lewin some months ago, prior to the AGM under lockdown, that they should take a video of a MEMS being played by moving the camera around it so at least the viewer could hear the directionality of the sound actually coming from the MEMS rather than have somebody saying that it came from a bank of speakers on the ceiling for instance.