I believe his power is continuous, but there's noise on the line.

He lives in downtown Annapolis, where they still have most of the original electrical distribution network. He's living in a townhouse that was built before the Civil War. I don't know that he has a modern 3-prong grounded circuit in the house. For all I know, he's running knob and tube wiring still.

Is there any way to duplicate that in a test environment? I mean, without spending $$$$ on a signal generator?

 

Got it.

Do you have an RFI (e.g. Corcom) filter on your gizmo?

In a work environment where a UPS wasn't necessary I used one in the series of these power conditioners https://www.ebay.com/p/ONEAC-CP1105...ditioner-4-Outlet/1034213352?iid=122799903354 and an ISOBAR outlet strip,
the ISOBAR fro the warranty on protected equipment.

ONEAC is now PowerVar. I saw a live demo and I was hooked. I had a lot of equipment connected to this and a Mac Centris 650 ran 17 years with only a floppy failure and cleaning out dust. The hard drive even lasted that long.

I even have one of their gizmos http://www.gryphon-inc.com/Spec Sheets/Power Monitoring/917010A - ONEView.pdf that can look at the power line with an oscilloscope, but never tried it.

So, three pieces of equipment eventually got this protection.

At home I had two issues and I happened to throw the right-stuff at it.
1. Noise AM radio and wall wart. I used a real commercial linear regulator for power.
2. Killing of X-10 modules. Put an expensive; $80.00 USD power line fllter at the furnace. Added a transorb to the 24 VAC supply.

 

Do you have an RFI (e.g. Corcom) filter on your gizmo?

I don't, no. I'm having my friend try one more test of input power before I buy any more stuff, but that looks like a good idea to try.

Does the ISOBAR power strip smooth the input power at all? Or is it more for covering significant power spikes and surges and what-not?

Nominally, I'd like to incorporate any necessary power conditioning on the board, if at all possible. Adding in $80 external components seems like a painful solution to me.

 

The CORCOM solution may be cheaper.

This https://www.x10.com/xpf-20a-wired-in-noise-filter.html is what I used on the furnace.

The ISOBAR does have RFI filters and Spike suppression.

The ONEAC thing tends to make the ground cleaner and the transformer won't couple the high frequencies. It's not a surge suppressor.

 

Update: It looks like there may be an issue with the 24VDC power supply. Specifically, and this is interesting: Meanwell GSM40 power supplies work. Meanwell GSM60 power supplies do not. The GSM60 provides 2A, while the GSM40 is limited to 1.6A. Aside from that, they should be identical. I have no idea why a power supply with a higher limit would fail, and the lower limit power supply would work great.

I'll try to get some o-scope scans of both power supplies, in situ, and will post in a bit. At this point, it's just for curiosity and my own edification. I'd love to understand why this fails.

 

Following up on this: I plugged in the o-scope and set up a row of power supplies that had failed back in December.

Ready for this?

Everything worked.

I swapped back and forth, double-checked the readings, but it was just as good with the "bad" power supplies as it was with the good. I'm baffled.

Going forward, board designs will all have a few more capacitors on them, as recommended here, and I'll be sure to use the power supply that I believe works best in more situations. But ... I don't know why some power supplies sometimes work better than others. Not really a question, just a statement.

Thanks again for all the info.