7i76e Communication errors

I converted a mini lathe using a Mesa 7i76e. I had it working really well and reliably and then it started throwing errors like this:

hm2/hm2_7i76e.0 smart serial port 0
hm2/hm2_7i76e.0 Smart Serial Comms Error: There have been more than 1 errors in 10 thread executions at least 200 times.
hm2/hm2_7i76e.0 smart serial port 0 local error = (3) Timeout
hm2/hm2_7i76e.0 smart serial port 0 local error = (13) Communication error

This usually happens within a couple of minutes of starting.
I've tried connecting the PC and the lathe on a dedicated hub so there's no other ethernet traffic but I still get the error.

I have a nearly identical configuration on both MachineKit and LinuxCNC and I see the error on both platforms.

Any idea how I can further troubleshoot or isolate the fault?

This is not a Ethernet problem but rather a sserial error communicating with the isolated
field I/O section. Most likely it indicates a 7I76E 5V or field power issue

I haven't had much time but I've continued to pick at this for months and still haven't figured it out.

I've checked out both the field power and 5V supplies and they seem fine. What kind of problem should I be looking for?

The error, if anything, has gotten more frequent. The only consistent thing I've noticed is that it never happens while jogging or running a program. Just when sitting idle.

I've attached a wiring diagram. Any clues or suggestions where to look next would be appreciated.

If its not a power supply issue, I would suspect some kind of on card fault.
If its an older card you might try reflashing the firmware (with mesaflash)
if that does not work I would request an RMA and return the card for repair/replacement.

hmm. Purchased new in October.

Is that old enough to benefit from a reflash?

Probably not, but if you try, use something really new like:

7i76e_7i76x1_bstechx2d.bit

as long as you are not using the expansion connectors

There were some fixes in the firmware related to sserial noise immunity
but your actual errors seem like a totally lost connection so I suspect a card issue
(is it maybe temperature related?)

Not temperature. Everything staying very cool and the error sometimes occurs within seconds of turning on.

Would it be worthwhile to test a dedicated 5V supply on P3?

Just noticed this comment.
I thought you were meant to use a direct connection between your PC and 7i76e (eg. ditch the hub) but I'll leave it to PCW to advise.

By temperature I dont mean high temperature, I mean a thermal intermittent, that may happen as the board warms up when first powered. The error is not Ethernet related.

PCW wrote:

By temperature I dont mean high temperature, I mean a thermal intermittent, that may happen as the board warms up when first powered.

I don't think so. I've left it on for hours and can't see any difference in behavior. My gut tells me it's hardware but not necessarily the board. Right after the conversion it was running very reliably so it 'feels' like a wiring/configuration/connection problem.

My inclination is to try to change things and see if I can get the symptoms to change enough to give a clue. That's why I asked about dedicated power on p3.