dkonigsberg
commented
3 months ago Another likely source of EMI issues is power line noise from nearby brushed AC motors, which some old enlargers use to control the up/down movement of the head. A good analog for these are old sewing machine motors, which have similar specifications. Such motors produce far larger transients on start/stop than the above-noted transformer spike, and can cause these same sorts of problems.
Completely solving the issue for all possible sources of noise will likely require a combination of solutions, from improvements to the main board's input filtering and grounding strategies, improvements to decoupling around specific components, software robustness to faults, and a lot of testing.
The power board contains a pair of relay-switched outlets for controlling enlargers and safelights. Sometimes, switching these AC loads can causes transients which manage to make their way across to the DC side and can then cause intermittent issues.
The easiest way to reproduce this issue is to use a 10W halogen desk lamp as a "test" enlarger load. It is powered by a wall wart which contains a transformer that converts 120VAC to 12VAC. Turning the power to the wall wart on can cause a minor spike, while turning the power off causes a noticeably larger spike. When this event happens, there's a good chance that connected USB peripherals (such as the meter probe) will stop responding. (It is unclear if this is due to those devices reacting to the spike, or some other components on the USB interface.)
Here's an example of what this looks like, when measured with an oscilloscope connected to the 12VDC output of the power board and set to AC coupling:
This problem can be temporarily eliminated by simply placing an EMC line filter between the enlarger relay outlet and the wall wart. That is good enough for development, but not a viable long term solution.
The current thoughts about how to fix this for real involve placing some sort of filter at the input to the AC/DC converter brick. This filter would likely go in place of the internal fuse holder and MOV, which are redundant components that already exist inside the AC/DC brick itself. (There is still an external fuse, which should handle most other concerns.)
The filter will likely consist of a common mode choke, possibly combined with X and Y capacitors in the usual arrangement. There will be a preference towards using such components with all the relevant safety ratings and sufficiently encapsulated to avoid exposed metal contacts on the top of the PCB.