9 years, 1 month ago.

You say it is isolated except for USB, but have you tried giving it a separate USB power supply? You can try what happens if you use the other USB port (the KL25 one instead of the SDA one) if you didn't already use that one: Then the debug interface is not powered (but you also cannot send serial data, I don't know if that lib uses it, I didn't check all code considering how extensive it is. Btw checking that code was quite funny, pretty much every major library I made is in it :D. Nice to see it getting used).

Plan B is of course simply using a separate supply for the KL25. How are you connecting it now? Are they all powered via a USB hub for example? Is that a powered USB hub? (If no, consider using a powered one).

9 years, 1 month ago.

Cary,

I think you're on the right track that it has something to do with the feedback devices.

You didn't mention what kind of devices you have hooked up to the SainSmart and LedWiz outputs, but is it safe to assume that you have some inductive loads like contactors, solenoids, or motors? Those sorts of devices can cause big fluctuations in the power supply voltages, especially when they turn *off*.

One really important step to help with this is to use flyback diodes on everything with a magnetic coil. If you're not familiar with that term, it means that you put a diode across each coil's power terminals - preferably as close as possible to the physical device - with the diode polarity *reversed*. So you connect the diode's striped end to the "+" voltage terminal of the coil. If you don't have these diodes on all of your coil-type devices, installing them would be the first thing I'd suggest. I use 1N4007 diodes, but any similar diode should work.

If you already have those, I'd suggest trying to see if you can identify a device or devices that cause the glitch. First try disconnecting all of the feedback devices, and play a while and see if the problem goes away. If so, it would support the working theory that it's one of the devices. Then start re-attaching them one at a time and test until the problem comes back. As part of this, I'd remove the SainSmart board entirely, because those relays are themselves inductive loads that could conceivably be causing the glitch. When you start re-attaching things, plug the SainSmart board back in with all of its output devices detached first, to help determine if it's the SainSmart board itself or something attached to it.

One thing you didn't say about your power supply setup is whether you have one power supply or two. If you're powering everything from one power supply, you actually can't isolate anything electrically, because everything shares a common ground connection at the PSU. I know you said you're not wiring everything directly to the common ground, but they'll end up connecting back to the same point anyway. The USB cable that's powering the KL25Z connects back to the same point as well, so the KL25Z has the same ground as the PC and the SainSmart board and the LedWiz, at the very least. When you reworked your wiring to remove the star point in the grounding, you might have seen an improvement simply by virtue of making the wires back to the common ground point in the PSU longer - longer wires can help with glitchy power surge behavior by adding incremental resistance, capacitance, and inductance to the mix. But that will only go so far; it'll help even things out for smaller power fluctuations but won't filter out larger surges.

I hope this helps. This sort of problem can be really difficult to get rid of, but if you can identify the specific device(s) causing the problem, it should be possible to find a solution.