Here's an odd one that came into my shop recently. The product was a recent Lionel Legacy Station Sounds Diner. The issue was that it couldn't receive TMCC/Legacy commands if the lights were on! It was tried on multiple layouts and failed on all of them, but turn off the lights and no sweat, worked perfectly. OK, that's an odd one. Apparently, it had been back to Lionel multiple times and they didn't find anything wrong.
First test, I put it on my bench and fired it up, worked fine. Turned the lights on and it still worked fine. OK, let's move over to the layout. Same results! I'm starting to think the customer has a vivid imagination. After some discussion, I decided to crack it open.
The lighting is totally separate from any of the other electronics, it simply gets track power through a lighting control switch.
I wondered why the lighting could affect the TMCC/Legacy signal, and after examining the lighting board briefly, I have my answer. As I suspected, Lionel went for massive overkill on the design of the lighting for something that lights a measly four LED's! This is a switching power supply, unfortunately sometimes known for being electrical noise generators.
Note the arrow above pointing to an inductor, below is what I see with my 'scope at that point. That's a 20V amplitude signal coming from that coil, that's certainly some electrical noise being generated! The signal was 15khz, and I'm sure the harmonics go far beyond that, maybe right up to 455khz? 
When I looked on the other side of the lighting strip I found this. The arrow again points to where the coil is located on the other side. In case you don't recognize that brass strip, that's the antenna for the TMCC/Legacy signal! It's sitting right on top of that source of electrical noise. It's hard for me to believe there's any other cause of the issue than simply injecting big bursts of noise into the TMCC/Legacy antenna signal.
The ill-advised positioning of the antenna is a total mystery to me! I suspect the lighting strip might not have been an issue if the antenna was far away from the switching P/S. However, since it didn't fail for me here, I didn't feel like doing a repair that I couldn't verify and finding out that it didn't address the problem. Since I was pretty sure of the reason for the issue, I also was pretty sure that simply eliminating the switching P/S would fix the issue.
I don't understand why it only affects specific layouts, that's a research project for another day.
Since it seemed clear that it would be a PITA to hack the lighting board to fix it, I simply removed the lighting board and installed my LED lighting regulator and a strip of LED's. Better lighting anyway than four measly LED's. To minimize the impact, the same connector was used on my LED regulator board and I plugged in the existing wiring harness that supplied the old lighting with track power. Unless you crack it open, you'd probably never know it was modified, documented operation remains the same.
It's a little difficult for me to believe that other folks haven't had similar issues with these cars given this one failed on multiple layouts. In case you're wondering, yes it now works fine on the customer's layout.
