Yes, I posted up an all analog solution in that thread, but the digital solution based on an atmel or similar has several benefits such as programmable zones instead of simple vertical curtain zone, extensibility where you can have it calculate values if there is a brightness pickoff, and possibly if there is a beambrush2 feedback line it could monitor that as well. Finally, the digital sampling solution would have considerably less components, and would be extensible via software updates. There are ways to ensure the digital solution has high reliability, and in some ways they might be better than the analog solution. If there is an error anywhere in the analog solution, it may be difficult to detect. In the digital sampling solution, there are only a handful of passive, high reliability components before the sampler (MCU), and the MCU itself can have internal and external watchdog features. In this case, since they are only a few dollars each, you may want to have two parallel MCUs with their outputs ANDed or ORed depending on implementation in some way so that the case of a mistaken calculation in one MCU due to a RAM failure will be caught as well.
It's definitely doable digitally, and it's enough of a goodwill safety problem that there may be enough motivation on the board to actually complete one in open source.
Reply With Quote
