Ethercat HAL driver
30 Apr 2014 13:36 #46493
by wazza77
Replied by wazza77 on topic Ethercat HAL driver
Hi guys
well i now have a response when i send ethercat slaves. i tried what was suggested and i hooked it up to twin cat and it worked fine
went back to linux and poked around there for a while and had no luck then it dawned on me what i believe was the problem when i installed the driver for
the network card i think i had done it in such a way that it would not allow the ethercat driver to take over
so i tried to revers that and with my skills i just went and re installed in the end and did it all again but this time i --enabled- all of the modules
before i built the package. and if you have ever seen a 6'4'' bloke weighing 110kg do a happy dance well you get the picture
thanks guys
well i now have a response when i send ethercat slaves. i tried what was suggested and i hooked it up to twin cat and it worked fine
went back to linux and poked around there for a while and had no luck then it dawned on me what i believe was the problem when i installed the driver for
the network card i think i had done it in such a way that it would not allow the ethercat driver to take over
so i tried to revers that and with my skills i just went and re installed in the end and did it all again but this time i --enabled- all of the modules
before i built the package. and if you have ever seen a 6'4'' bloke weighing 110kg do a happy dance well you get the picture
thanks guys
Please Log in or Create an account to join the conversation.
30 Apr 2014 14:40 #46496
by wazza77
Replied by wazza77 on topic Ethercat HAL driver
Hi Andreas i have run it through its paces exactly as you did in your video just on the command line and it worked fine
i will start Linux cnc tomorrow and let you know how it all goes
i will start Linux cnc tomorrow and let you know how it all goes
Please Log in or Create an account to join the conversation.
07 May 2014 03:26 #46660
by roschi
Replied by roschi on topic Ethercat HAL driver
Please Log in or Create an account to join the conversation.
07 May 2014 20:51 #46688
by wazza77
Replied by wazza77 on topic Ethercat HAL driver
Hi Andreas
Thats Very encoroging to see this video congrdulations. I see what you mean when you say that it has not been optimised yet
will you use the biult in auto tune in the drive for the A axis?. I have not gone to much further with mine as the mill arrived last week
and i have been trying to familiarise myself with the files eg .ini .hal and the like i have been going throught 402 information on some
of the settings and i think i have my head around it so far.
Warren
Thats Very encoroging to see this video congrdulations. I see what you mean when you say that it has not been optimised yet
will you use the biult in auto tune in the drive for the A axis?. I have not gone to much further with mine as the mill arrived last week
and i have been trying to familiarise myself with the files eg .ini .hal and the like i have been going throught 402 information on some
of the settings and i think i have my head around it so far.
Warren
Please Log in or Create an account to join the conversation.
07 May 2014 22:45 #46691
by roschi
I have used the auto tuning of the drive but the problem was that you can increase the 'machine rigidity parameter' (ESTUN P101) for the auto tuning algorithm only up to 1 (default is 5 max. is 12). If I set it higher the axis becomes unstable.
In the Video the auto tuning was used and as mention the rigidity parameter was/ can only set to 1. That results in a bad dynamic stiffness (there were less damped resonances if you have applied a step torque to the shaft. This was done by the interrupted cutting as shown in the video. (static stiffness is only limit by the mechanical compliance because the control loop has integral behavior)
You have to take in mind that the ratio load intertia/motor inertia is in case of the A-axis (nearly 1:1 gear ratio, ESTUN P106) is a challenge for the position control.
But I solved this.
The main problem was the timing belt.. Belts can only pull. If you have a change in the direction of the transmitted torque you have to avoid to go through a region were the belt tension on the driving side is zero. I have increased the preload of the belt and now I was able to adjusted the control loops gains and integral time manually to higher values. No filters wre used. Now the dynamic stiffness is much better and sufficient. (stiffness = external torque/pos. error)
Andreas
Replied by roschi on topic Ethercat HAL driver
Hi Andreas
I see what you mean when you say that it has not been optimised yet
will you use the biult in auto tune in the drive for the A axis?.
I have used the auto tuning of the drive but the problem was that you can increase the 'machine rigidity parameter' (ESTUN P101) for the auto tuning algorithm only up to 1 (default is 5 max. is 12). If I set it higher the axis becomes unstable.
In the Video the auto tuning was used and as mention the rigidity parameter was/ can only set to 1. That results in a bad dynamic stiffness (there were less damped resonances if you have applied a step torque to the shaft. This was done by the interrupted cutting as shown in the video. (static stiffness is only limit by the mechanical compliance because the control loop has integral behavior)
You have to take in mind that the ratio load intertia/motor inertia is in case of the A-axis (nearly 1:1 gear ratio, ESTUN P106) is a challenge for the position control.
But I solved this.
The main problem was the timing belt.. Belts can only pull. If you have a change in the direction of the transmitted torque you have to avoid to go through a region were the belt tension on the driving side is zero. I have increased the preload of the belt and now I was able to adjusted the control loops gains and integral time manually to higher values. No filters wre used. Now the dynamic stiffness is much better and sufficient. (stiffness = external torque/pos. error)
Andreas
Please Log in or Create an account to join the conversation.
08 May 2014 05:24 - 08 May 2014 05:27 #46697
by willrat
Replied by willrat on topic Ethercat HAL driver
Thanks for posting the files and the posts on other drives. I've used them as a basis for setting up the A5 drive.
I've put them up here -> app.box.com/s/g7jczj0jc15cv3ztl22m
I've included the lengthy A5 ethercat spec I got from our distributor; this is not on Panasonic's website yet so may prove useful to someone
Many of the entries are the same (drives xml file states its a 402 profile type) and I've made the adjustments as per the Rexroth drive document and the OMRON example; It appears the drive does not need to have a value placed in the 6060 register to set the operation mode.
I'm using 2.6.32-122-rtai; aschiffler/linuxcnc (make install not rip)
I start the system in a empty hal session as follows
everything looks good at this point.
the slave goes to
which looks good at this point but the drive gives an error 80.4 - PDO watchdog error protection (on the drives LCD display)
dmesg gives
does anyone know what would be the next step: is the watchdog not being updated correctly or is there some other problem here?
I've put them up here -> app.box.com/s/g7jczj0jc15cv3ztl22m
I've included the lengthy A5 ethercat spec I got from our distributor; this is not on Panasonic's website yet so may prove useful to someone
Many of the entries are the same (drives xml file states its a 402 profile type) and I've made the adjustments as per the Rexroth drive document and the OMRON example; It appears the drive does not need to have a value placed in the 6060 register to set the operation mode.
I'm using 2.6.32-122-rtai; aschiffler/linuxcnc (make install not rip)
I start the system in a empty hal session as follows
loadrt threads name1=master period1=100000
loadusr -W lcec_conf A5-slave-only.xml
loadrt lcec
addf lcec.read-all master
addf lcec.write-all mastereverything looks good at this point.
startthe slave goes to
state-op = TRUE
link-up = TRUE
slaves-responding = 1which looks good at this point but the drive gives an error 80.4 - PDO watchdog error protection (on the drives LCD display)
dmesg gives
[ 190.541998] RTAI[malloc]: global heap size = 2097152 bytes, <BSD>.
[ 190.542218] RTAI[sched]: IMMEDIATE, MP, USER/KERNEL SPACE: <with RTAI OWN KTASKs>, kstacks pool size = 524288 bytes.
[ 190.542222] RTAI[sched]: hard timer type/freq = APIC/20832837(Hz); default timing: periodic; linear timed lists.
[ 190.542224] RTAI[sched]: Linux timer freq = 250 (Hz), TimeBase freq = 3000244000 hz.
[ 190.542226] RTAI[sched]: timer setup = 999 ns, resched latency = 2943 ns.
[ 190.542272] RTAI[usi]: enabled.
[ 190.568496] RTAI[math]: loaded.
[ 190.601946] EtherCAT: Requesting master 0...
[ 190.601950] EtherCAT: Successfully requested master 0.
[ 190.602234] EtherCAT 0: Domain0: Logical address 0x00000000, 41 byte, expected working counter 3.
[ 190.602236] EtherCAT 0: Datagram domain0-0-main: Logical offset 0x00000000, 41 byte, type LRW.
[ 190.602251] EtherCAT 0: Master thread exited.
[ 190.602253] EtherCAT 0: Starting EtherCAT-OP thread.
[ 190.664064] EtherCAT WARNING 0: 40 datagrams TIMED OUT!
[ 198.968024] EtherCAT 0: Slave states on main device: OP.
[ 199.076296] EtherCAT ERROR 0-0: AL status message 0x001B: "Sync manager watchdog".
[ 199.238257] EtherCAT 0-0: Acknowledged state SAFEOdoes anyone know what would be the next step: is the watchdog not being updated correctly or is there some other problem here?
Last edit: 08 May 2014 05:27 by willrat. Reason: corrections
Please Log in or Create an account to join the conversation.
08 May 2014 05:32 - 08 May 2014 06:25 #46698
by willrat
Replied by willrat on topic Ethercat HAL driver
the solution to the watch dog was to decrease the thread period... should of been obvious!
I still get configuration errors on dmesg
loadrt threads name1=master period1=10000I still get configuration errors on dmesg
[ 2133.788021] EtherCAT ERROR 0-0: Failed to receive state datagram: Datagram timed out.
[ 2133.820015] EtherCAT ERROR 0-0: Failed to receive CoE dictionary request datagram: Datagram timed out.
[ 2133.832560] EtherCAT 0: 0 slave(s) responding on main device.
[ 2133.848005] EtherCAT 0: 1 slave(s) responding on main device.
[ 2133.848008] EtherCAT 0: Slave states on main device: PREOP.
[ 2133.848052] EtherCAT 0: Scanning bus.
[ 2134.026542] EtherCAT 0: Bus scanning completed in 176 ms.
[ 2134.026545] EtherCAT 0: Using slave 0 as DC reference clock.
[ 2134.064017] EtherCAT ERROR 0-0: Failed to receive state datagram: Datagram timed out.
[ 2134.096021] EtherCAT ERROR 0-0: Failed to receive CoE dictionary request datagram: Datagram timed out.
[ 2134.108018] EtherCAT 0: 0 slave(s) responding on main device.
[ 2134.120021] EtherCAT 0: 1 slave(s) responding on main device.
[ 2134.120024] EtherCAT 0: Slave states on main device: PREOP.
[ 2134.120074] EtherCAT 0: Scanning bus.
Last edit: 08 May 2014 06:25 by willrat.
Please Log in or Create an account to join the conversation.
08 May 2014 16:46 #46717
by roschi
The unit of time is nanoseconds. I think this time have to match the cycle time set in the XML file and should be 1,000,000 ns. I tested it with a minimal vlaue of 500,000 ns.
Replied by roschi on topic Ethercat HAL driver
loadrt threads name1=master period1=10000Please Log in or Create an account to join the conversation.
08 May 2014 21:13 - 08 May 2014 21:40 #46722
by willrat
Replied by willrat on topic Ethercat HAL driver
Setting the period back helped; as did adding the second dc entry on the xml file and also correcting the pid at the top (missed the F off the end)
So now I've got the system running stable. master and salve are both in state-op
I run the shell scripts a5-2, a5-3, a5-4 to enable the drive (just for testing)
here: app.box.com/s/g7jczj0jc15cv3ztl22m
I can control the drive in velocity mode.
I've not worked out how to clear the fault when the drive goes into error yet; so its a power cycle for that.
So now I've got the system running stable. master and salve are both in state-op
I run the shell scripts a5-2, a5-3, a5-4 to enable the drive (just for testing)
here: app.box.com/s/g7jczj0jc15cv3ztl22m
I can control the drive in velocity mode.
I've not worked out how to clear the fault when the drive goes into error yet; so its a power cycle for that.
Last edit: 08 May 2014 21:40 by willrat.
Please Log in or Create an account to join the conversation.
22 May 2014 16:58 #47190
by narogon
Replied by narogon on topic Ethercat HAL driver
Has someone tried to configure lcec with beckhoff modules?
I'm testing this simple configuration:
But i obtain this error EtherCAT ERROR 0 0:1: PDO entry 0x6000:01 is not mapped.
Is there any known error in lcec_el1xxx.c or any other file of hal ethercat driver?
I'm testing this simple configuration:
<master idx="0" appTimePeriod="1000000" refClockSyncCycles="10">
<slave idx="0" type="EK1100" />
<slave idx="1" type="EL1014" />
</master>But i obtain this error EtherCAT ERROR 0 0:1: PDO entry 0x6000:01 is not mapped.
Is there any known error in lcec_el1xxx.c or any other file of hal ethercat driver?
Please Log in or Create an account to join the conversation.
Time to create page: 0.180 seconds