hm2/hm2_7i76e.0: error finishing read! iter=25403

More
20 Mar 2024 15:52 #296393 by bhowiebkr
Hey I'm running into the dreaded error finishing read! iter problems with my CNC

I'm tired of dealing with this and want to get it solved once and for all. I'll post everything about my machine updating this initial first post as I gather up all the details of my PC hardware etc.

Hardware:
Mesa 7i76e
JMC 400w ac servos

My INI file can be found here:
github.com/bhowiebkr/PrintNC_conf/blob/main/PrintNC/PrintNC.ini

Latency test:
github.com/bhowiebkr/PrintNC_conf/blob/m...p/latency%20test.png

ip a out:
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN group default qlen 1000
    link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
    inet 127.0.0.1/8 scope host lo
       valid_lft forever preferred_lft forever
    inet6 ::1/128 scope host
       valid_lft forever preferred_lft forever
2: enp3s0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc fq_codel state UP group default qlen 1000
    link/ether c8:1f:66:4a:c1:2b brd ff:ff:ff:ff:ff:ff
    inet 10.10.10.11/8 brd 10.255.255.255 scope global noprefixroute enp3s0
       valid_lft forever preferred_lft forever
3: wlx6c5ab0d0684d: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc mq state UP group default qlen 1000
    link/ether 6c:5a:b0:d0:68:4d brd ff:ff:ff:ff:ff:ff
    inet 192.168.1.227/24 brd 192.168.1.255 scope global dynamic noprefixroute wlx6c5ab0d0684d
       valid_lft 55963sec preferred_lft 55963sec
    inet6 fd74:8db0:9331:f2f0:6b99:8289:4e95:de6f/64 scope global dynamic noprefixroute
       valid_lft 1438sec preferred_lft 1438sec
    inet6 fe80::4fa5:3f66:9622:fcb9/64 scope link noprefixroute
       valid_lft forever preferred_lft forever


howard@debian12:~$ halcmd show param *.tmax
Parameters:
Owner   Type  Dir         Value  Name
    63  s32   RW          17700  debounce.0.tmax
    60  s32   RW          26140  estop-latch.0.tmax
    60  s32   RW           1188  estop-latch.1.tmax
    60  s32   RW          17528  estop-latch.2.tmax
    60  s32   RW          17292  estop-latch.3.tmax
    60  s32   RW          17708  estop-latch.4.tmax
    60  s32   RW          18236  estop-latch.5.tmax
     5  s32   RW          80456  fast.tmax
    52  s32   RW              0  hm2_7i76e.0.read-request.tmax
    52  s32   RW        2301968  hm2_7i76e.0.read.tmax
    52  s32   RW          67752  hm2_7i76e.0.write.tmax
    45  s32   RW          17920  motion-command-handler.tmax
    45  s32   RW          55328  motion-controller.tmax
    55  s32   RW          16204  pid.s.do-pid-calcs.tmax
    55  s32   RW          16740  pid.w.do-pid-calcs.tmax
    55  s32   RW          26388  pid.x.do-pid-calcs.tmax
    55  s32   RW          16440  pid.y.do-pid-calcs.tmax
    55  s32   RW          16608  pid.y2.do-pid-calcs.tmax
    55  s32   RW          16708  pid.z.do-pid-calcs.tmax
    46  s32   RW        2337288  servo-thread.tmax
     6  s32   RW          81300  slow.tmax
     9  s32   RW          80456  timedelta.0.tmax
     9  s32   RW          81300  timedelta.1.tmax




 

Please Log in or Create an account to join the conversation.

More
20 Mar 2024 15:58 #296394 by bhowiebkr
latency histogram:
Attachments:

Please Log in or Create an account to join the conversation.

More
20 Mar 2024 16:02 #296395 by tommylight
latency-histogram --nobase --sbinsize 1000 --show
?
Also, ping times
ping 10.10.10.10
?

Please Log in or Create an account to join the conversation.

More
20 Mar 2024 16:11 #296396 by PCW
1. Have you disabled all power saving option in you BIOS (Turbo modes, speed-step,Cstates >C1 EIST etc)
2. What is your host hardware?
3. For difficult cases, see this thread:
forum.linuxcnc.org/38-general-linuxcnc-q...ead-and-irq-affinity

Please Log in or Create an account to join the conversation.

More
21 Mar 2024 14:00 #296469 by vre
What can do if BIOS does not give you selection options to disable (Turbo modes, speed-step,Cstates >C1 EIST etc) ?
Can these disabled from linux kernel ?

Please Log in or Create an account to join the conversation.

More
21 Mar 2024 14:47 #296475 by PCW
Many can be set with the kernel command line

here is my command line (setup with grub)

ro quiet splash intel_idle.max_cstate=1 processor.max_cstate=1 idle=poll intel_idle.max_cstate=1 i915.enable_dc=0 isolcpus=3 i915.enable_rc6=0

What PC hardware do you have?


 

Please Log in or Create an account to join the conversation.

More
21 Mar 2024 22:20 #296509 by bhowiebkr
Here's the bios for this dell optiplex 3020.


I have about 5-6 of these dell boxes of different machines/specs and running the latency tests, this one came out on top. Most of the newer ones that are better spec'd preform worse on the latency tests.

i'm gonna read through IRQ affinity post later today. And see if I can get a readout of all the hardware including what network controller is in this box.

Please Log in or Create an account to join the conversation.

More
21 Mar 2024 23:52 #296520 by bhowiebkr
Here's the latency and ping times 
Attachments:

Please Log in or Create an account to join the conversation.

More
22 Mar 2024 00:55 #296522 by PCW
Is there a wireless card in the 3020? These are often the cause of network latency.
(USB WIFI dongles seem OK)

I have a similar vintage HP 8300 USDT (I5 3470S 2.9 GHz)
running cheerfully at 4 KHz.  (with the IRQ pinning trick)


 

Please Log in or Create an account to join the conversation.

More
22 Mar 2024 01:11 - 22 Mar 2024 01:33 #296525 by bhowiebkr
I've got a USB wireless card in there to access the internet from the linuxcnc machine.

I'm using Debian 12 and it seems as if irqbalance is missing on that one. So I'm unable to setup the service the way it's shown.

Would it be bad if I just set in the bios to use 1 core? would this help? When the machine is running, it's not like I'm using anything in the background.
Last edit: 22 Mar 2024 01:33 by bhowiebkr.

Please Log in or Create an account to join the conversation.

Moderators: PCWjmelson
Time to create page: 0.119 seconds
Powered by Kunena Forum