[etherlab-users] Configuring DC for a axis control

[Ezio DelBono]

Dear Andreas,

I know that. In fact in the past i tried to write some wrapping 
functions to avoid ioctl system call but now I'm wondering if the jitter 
is still acceptable without doing any hack to the master.
How much should be the maximum and minimum latency of the ethercat frame 
so that the DC mechanism can work properly?
Must be the 10% of the period time? In my case I have the 20%.
Can I compensate it optimizing the "shift time"? My idea is to bring the 
frame arrival in between two Sync0.
Does it make sense?

Thank you very much,
Ezio

On 28/01/2011 09:51, Andreas Stewering-Bone wrote:
> Hello Ezio,
>
> Are you really sure, you use the master under lxrt with realtime 
> capabilities.
>
> There is no lxrt or rtdm interface for the master until now.
>
> Greatings
>
> Andreas
>
> Am 28.01.2011 09:19, schrieb Ezio DelBono:
>> Hello all,
>> I am using the ethercat master version 1.5 to control some brushless 
>> motor drives with Dc support.
>> What my application does, is to read every 1ms the actual position 
>> (0x6064) and send the target velocity (0x60ff) to the drive and read 
>> the status word and write the control word as well.
>>
>> My application is executed by RTAI-Lxrt in user space with a period 
>> of 1ms and a jitter of 200us (due to the character device). I am 
>> quite sure that the dc is running well because I monitor the SYNC0 
>> generated by the first slave that is the reference clock.
>>
>> Anyway I haven't understood how I have to configure the 'shift time' 
>> in my application because the problem is:
>> Sometime (every 2-3 seconds) the actual position is not refreshed 
>> meaning that I read the same actual position in the next cycle. This 
>> occurs when my application period has the worst jitter. How can I set 
>> the shift time to compensate it? Or this is not the right solution?
>>
>> Kind regard,
>> Ezio DelBono
>>
>> _______________________________________________
>> etherlab-users mailing list
>> etherlab-users at etherlab.org
>> http://lists.etherlab.org/mailman/listinfo/etherlab-users
>
>