The two schematics below show the initial version of the connector board (on the top/left) and the new version (on the bottom/right).  They both do the same basic thing -- opto-isolation and high voltage switching for the strobe board and the camera trigger signals.

So, why the change?

Well, now that we’re starting to think about what needs to be done before releasing the project into the wild, there are a lot of considerations.  They include things like proper attribution of open-source, IP management, ease of building and testing, and … safety!

One issue that came up in the safety review was the fact that the original connection board design would allow the LED strobe board to be full ‘on’ continually if the power from the Raspberry Pi failed, or if the Pi otherwise became disconnected from the board.  The problem was that if a 100 watt, densely-packed LED array is on for any more than a few seconds, it starts to get really hot.  Like, hot enough to potentially even melt its #d printed plastic mounting hardware!

This was a good (if painful) lesson in identifying and considering failure modes earlier in the design process, where things are easier to accommodate and/or protect against.  The circuit in the connector board is pretty simple, requiring an non-inverting opto-isolator among other things.  One easy way to do so (other than switching from a pull-up to a pull-down transistor on certain opto-isolation ICs), was just to add an inverter to the input signal.  That’s the red-colored path in the top/left schematic.  

But, we hadn’t thought of the consequences of inverting the input, as opposed to adding the inverter after the isolator as in the purple path in the bottom/right schematic.  The right-most circuit has no other material downsides, but it also has the benefit that the strobe output to the 12 V MOSFET switch will go to low in the event of most failure modes, like a loss of power from the Sys1_Conn from the first Raspberry Pi or anything else that would cause power to the 7404 hex inverter chip to fail.  The opposite is true of the original circuit, where a failure causes the input to the isolator to go low, which causes the output to go high and switch on the strobe.  So, the new circuit inverts the output, not the input.  This should vastly decrease the chances of the LED array ever being on for more than a few microseconds at a time.

Easy fix, but required another run of PCB boards. :/  More great lessons from this project…