6i25-7i77 Comunication errors

Well if you get a sserial break error, you have a pretty severe noise issue (> 1.4V induced in DB25 cable for longer than 4 usecs)

Is it possible Estop causes a relay/contactor to open which doesn't have arc suppression?

The e-stop does open a contactor or two, and a few small relays. I could not tell you if they have arc suppression. They are in a separate electrical panel. All I know is I didn't see these problems before 2.7 (not that they weren't there) but the new tool planner is worth putting up with it.

Hi Peter,

I have been doing the following: I took linuxcnc back to 2.6.12, but then I couldn't get linuxcnc to even start up. I downloaded the latest 5i25.bit file, and did mesaflash because I thought that the 77f.bit could maybe have caused the errors, but still continued getting the following errors: dmesg output:

[ 264.379071] hm2/hm2_5i25.0: Smart serial card hm2_5i25.0.7i77.0.3 error = (4) Extra character
[ 264.380266] hm2/hm2_5i25.0: Smart serial card hm2_5i25.0.7i77.0.3 error = (13) Communication error
[ 264.380270] hm2/hm2_5i25.0: Smart serial card hm2_5i25.0.7i77.0.3 error = (4) Extra character
[ 264.381465] hm2/hm2_5i25.0: Smart serial card hm2_5i25.0.7i77.0.3 error = (13) Communication error
[ 264.381470] hm2/hm2_5i25.0: Smart serial card hm2_5i25.0.7i77.0.3 error = (4) Extra character
[ 264.382663] hm2/hm2_5i25.0: Smart Serial Comms Error: There have been more than 1 errors in 10 thread executions at least 200 times. See other error messages for details.
[ 264.382669] hm2/hm2_5i25.0: ***Smart Serial Port 0 will be stopped***
[ 292.446038] hm2_5i25.0: dropping AnyIO board at 0000:03:00.0
[ 292.446047] hm2/hm2_5i25.0: unregistered
[ 292.446081] hm2_pci 0000:03:00.0: PCI INT A disabled
[ 292.446128] hm2_pci: driver unloaded
[ 292.449696] hm2: unloading
[ 294.668352] RTAI[math]: unloaded.
[ 294.689428] SCHED releases registered named ALIEN RTGLBH
[ 294.748853] RTAI[malloc]: unloaded.
[ 294.848554] RTAI[sched]: unloaded (forced hard/soft/hard transitions: traps 0, syscalls 0).
[ 294.851508] I-pipe: Domain RTAI unregistered.
[ 294.851629] RTAI[hal]: unmounted.

After trying everything in my knowledge ( including checking for grounding loops, checking the voltages, and having someone with me monitoring the scope - voltages while I tried , I decided to go back to 2.7.3

I can also confirm that is seem to happen every time that I trigger the E-stop. I checked my E-stop wiring, but all looks good.

this is the dmesg output :

5i25.0.7i77.0.3 error = (13) Communication error
[19495.692073] hm2/hm2_5i25.0: Smart serial card hm2_5i25.0.7i77.0.3 error = (3) Timeout
[19495.696078] hm2/hm2_5i25.0: Smart serial card hm2_5i25.0.7i77.0.3 error = (13) Communication error
[19495.696086] hm2/hm2_5i25.0: Smart serial card hm2_5i25.0.7i77.0.3 error = (3) Timeout
[19495.700063] hm2/hm2_5i25.0: Smart serial card hm2_5i25.0.7i77.0.3 error = (13) Communication error
[19495.700070] hm2/hm2_5i25.0: Smart serial card hm2_5i25.0.7i77.0.3 error = (3) Timeout
[19495.704066] hm2/hm2_5i25.0: Smart serial card hm2_5i25.0.7i77.0.3 error = (13) Communication error
[19495.704073] hm2/hm2_5i25.0: Smart serial card hm2_5i25.0.7i77.0.3 error = (3) Timeout
[19495.708093] hm2/hm2_5i25.0: Smart Serial Comms Error: There have been more than 1 errors in 10 thread executions at least 200 times. See other error messages for details.
[19495.708103] hm2/hm2_5i25.0: ***Smart Serial Port 0 will be stopped***
[19815.874945]
[19815.874949] In recent history there were
[19815.874952] 7993776, 5967552, 2015376, 7996464, and 8028960
[19815.874954] elapsed clocks between calls to the motion controller.
[19815.874962] This time, there were 7999200 which is so anomalously
[19815.874964] large that it probably signifies a problem with your
[19815.874967] realtime configuration. For the rest of this run of
[19815.874969] LinuxCNC, this message will be suppressed.
[19815.874971]
[19953.969494] hm2_5i25.0: dropping AnyIO board at 0000:03:00.0
[19953.969503] hm2/hm2_5i25.0: unregistered
[19953.969536] hm2_pci 0000:03:00.0: PCI INT A disabled
[19953.969576] hm2_pci: driver unloaded
[19953.973318] hm2: unloading
[19956.228861] RTAI[math]: unloaded.
[19956.250305] SCHED releases registered named ALIEN RTGLBH
[19956.310399] RTAI[malloc]: unloaded.
[19956.410012] RTAI[sched]: unloaded (forced hard/soft/hard transitions: traps 0, syscalls 0).
[19956.413468] I-pipe: Domain RTAI unregistered.
[19956.413607] RTAI[hal]: unmounted.
[19958.881968] I-pipe: Domain RTAI registered.
[19958.881978] RTAI[hal]: <3.8.1> mounted over IPIPE-NOTHREADS 2.6-03.

I am now reformatting the pc, with weezy, as I have been running all my pc's with Ubuntu 10.04 .
I don't know what else to do at this stage, as I am running out of ideas and running out of time too.

I really hope a solution will come up soon, as I have now lost too much time,

Please try and get 2.6 running with the 77f.bit file

The 77f bitfile will work with 2.6 (there are no linuxcnc visible changes)
Please do not change bit files and ALWAYS us the 77f.bit file and 2.6 (until 2.7 is fixed)
if you have electrical noise issues, this is VERY important

Also these errors must be on a second 7I77 so make sure that any header-->
DB25 cable adapter is a short as possible (and ALWAYS us the 77f bitfile if you use the second port)

If you get errors when your trigger an EStop, this is possibly because you have
relay contacts that switch an inductive load that are missing arc suppression components

BTW a likely cause of 2.6 not starting with a 2.7 HAL file is that the
watchdog petting function has been removed/commented out in the hal file since 2.7 does not need it

aha! you're right. I did comment the watchdog out.

about the arc suppression, I have miniature relays that I am switching on, with diodes across the coils, and I have never had a problem like this with the last 32 machines that I have retrofitted, so I doubt that it could be a arc suppression problem. I think these relays draw 0.01A or less. Not sure exactly, but I know they draw very little current, and with the diodes across the coils I would think that everything should be okay.
I did a rest program, where I switched three relays on and off at a very fast rate, and the problem did not occur. Definitely something to do within linuxcnc and the E-stop/Enable.

Its not the little relays I worry about, but any relay or contactor coils
switched by relay contacts, these are where the arc suppression devices are needed

DId you try running 2.6 with the 77f bitfile? (this combination will be more noise resistant than any past setup )

Hi,

I managed to get my machine on Weezy, and brought it back to 2.6, for the time being, it gives errors, but I can continue to run linuxcnc. Thanks for all you help so far for everyone involved. At least I can continue with the machine now...