[etherlab-users] delays in the etherlab I/O blocks?

[Julian Stoev / Юлиан Стоев]

Dear Etherlab-ers,

We have a working setup of a prototype linear motor, operating in
torque (current reference) mode.
The power drive reference is given in analog +/- 10 V using beckhoff
EL4132 and the motor position is received using Beckhoff EL5101.

The corresponding Simulink diagram is shown in the attached small png
file (hope this is OK for the mail list)
We make the test measurements to get the frequency response for our
set-up. We get the FRD diagram in the second attached png file. The
magnitude behaves exactly as expected, but the phase indicates the
presence of a time delay, which is too big. The attached file shows
both the raw data and the data with estimated and compensated delay.
We have determined that part of the delay is changing when we change
the sampling time of the etherlab blocks (we do not change) the other
blocks sampling. In the diagram attached the blocks in green operate
at 1kHz, while the etherlab I/O block operate at 4kHz. The time delay
indicated by the phase is approximately 1.4 msec, which is above our
sampling rate for the control part. When we change the sampling time
of the etherlab blocks to 10kHz, the delay goes down to ~0.8msec.

The above makes us suspect that the etherlab blocks have internal
delays associated with them and they depend on the sampling rate.
While the source code is available and we could find this in the code,
this will take time and we would like to ask if other users and
developers can confirm this.

Thanks a lot!

--
Julian Stoev, PhD.
Control Researcher
-------------- next part --------------
A non-text attachment was scrubbed...
Name: gain_controller_simulink.png
Type: image/png
Size: 40104 bytes
Desc: not available
URL: <http://lists.etherlab.org/pipermail/etherlab-users/attachments/20091005/9c63f3ad/attachment-0006.png>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: report_02_10_2009_07.png
Type: image/png
Size: 14307 bytes
Desc: not available
URL: <http://lists.etherlab.org/pipermail/etherlab-users/attachments/20091005/9c63f3ad/attachment-0007.png>