Magical disappearing 7i84 issue

OK, that makes sense. I did set up my 7i80 which happened to have a config on that looked similar enough to try, but it also didn't want to talk to the 7i84 - that could just be that it wasn't relevant for the 7i52 though.

I'll try out the pin health tests tomorrow when I'm back in. I think I have the means to make up another couple of cables that I can swap out, so I can rule that out.

I also thought of moving the 7i52 to P4 - it's unused in this setup - please, could you make me a firmware for that arrangement? The firmware in question, in case this helps you, is the one you made for me following this thread: forum.linuxcnc.org/27-driver-boards/5075...ware?start=10#286332 which has been working well up until the last couple of weeks. This would allow me to keep using the 5i23 which would be optimal.

Thank you for your support.

Kind regards,
Andy

Same as FMD config but 7I52 on P4:

 

 

Hello again.
No luck so far with the setup I need.
I have tested some things on another system with some positives:
Connecting my 7i84 through a 7i94 and using the command:

sudo mesaflash --addr=192.168.1.121 --device=ETHER --sserial

gives me a response from both of my 7i84 cards - they are detected here, showing SwRevision = 14, baud 2.5Mb, NVUnitNumbers etc. So my 7i84s appear to be OK.

I can start up a simple lcnc config using the 7i94, with one pin in use from the 7i84, with no errors.

I can use:

sudo mesaflash --addr=192.168.1.121 --device=ETHER --readhmid

on the 7i80HD, loaded with the firmware you gave me, and get the pinout I'm expecting with SSerial TXData0 on I/OPort 45, SSerial RXData0 on I/OPort 46, and SSerial TXEn0 on I/OPort 47.

However sudo mesaflash --addr=192.168.1.121 --device=ETHER --sserial gives me no response here - not even an enumeration of the SSerial ports shown on the readhmid list. Earlier yesterday the mesaflash...--sserial command used on the 5i23 was giving me a few lines as a response, but seems to have stopped doing that.


What I can't seem to do anymore, is use a 50-pin flat cable to connect a 7i52 to an fpga card, and connect to the smartserial card that way.

I've checked my cat5-to-terminal wiring a handful of times but still am not sure it's correct. It looks like this:

 
 

Also, can you confirm that that terminal block should be plugged into the bottom 8 pin space of P3 on the 7i52, which I think is COM 0?

The likely candidate really seems to be the cat5-to-terminal wiring, but I'm reasonably sure I haven't changed that since it worked!

Have also tried swapping out 50-pin cables, using cable and ext. 5V power for the 7i52, swapping out 7i52 for a spare, swapping out 7i84 for a spare, and as many combinations of these as I can. Properly stumped. I would be grateful if you could confirm the cat5 cable wiring and the correct 7i52 terminal position. 

Did you try the I/O health check?

The cable looks correct,  maybe make another cable as either end may be flakey

(you can verify with connector location with mesaflash --device ethernet --addr 192.168.1.121 --readhmid --dbname2 7i52)

 

Oh that's weird, I had replied re the health check but don't seem to have posted successfully. All three gave me the answer 00FFFFFF when I re-ran the test just now. (Oddly, that is not the answer they gave me when I tried previously - that time iirc I got 00001000, 00001004, and 00001008 respectively).

Connector location comes up as P3-5,6 for TXData0, P3-3,4 for RXData0, and P3-5,6:/Ena for TXEn0.

The cable in the photo is one I made earlier today using a fresh CAT5 cable that tested perfectly with an rj45 cable tester before I cut it in half to make the interface cable.

Any further suggestions would be extremely welcome; I have to admit I am at a loss here.

Is there any other card that will support the combination of 7i48/7i84 I am using? Or a way to test the sserial connection via the 7i52, through halcmd maybe?

A simple loopback test would be to wire:
 
RX0+ to TX0+ 7I52 P3 pin 3- -> pin 5
RX0-  to TX0-   7I52 P3 pin 4 --> pin 6

and then run

mesaflash [card specific stuff] wpo 0x1004=0               #  Set the output latch to 0s
mesaflash [card specific stuff] wpo 0x1104=0xA00000 # Set  TX and TXEN pin to output mode
mesaflash [card specific stuff] wpo 0x1204=0               # disable all secondary functions
mesaflash [card specific stuff] rpo 0x1004                     # read P3 GPIO bits
mesaflash [card specific stuff] wpo 0x1004=0x200000 # set TX data high
mesaflash [card specific stuff] rpo 0x1004                     # # read P3 GPIO bits

If working, you would expect bit 22 (0x400000) = RXData to follow  bit 21 (0x200000) = TXData

You might also want to verify that you don't have T568A cables (which have a different color code)
Your connector wiring is correct for T568B

 

Thank you again. In the meantime, I went back to the earlier firmware on the 5i23 (having removed and re-seated that), and got a response out of a 7i84. As I was a bit scattershot in trying out combinations earlier in my troubleshooting, I now don't know whether it's the same 7i84, 7i52, or 50-pin card. Going back through combinations now to see if a specific item appears to cause the fault.

Aaaand I seem to be back up and running; the mill is making chips as I type. The culprit was a 7i52 - swapped out 7i84, 50-pin cable, cat5-sserial cable, and 7i52, and it was only the latter that brought back the fault.

Thank you, as ever, PCW for your efforts to help me. +1 once again for Mesa support.

Fingers crossed it continues to work...