Experimental Physics and
Industrial Control System

Kevin Peterson <[email protected]> · Wed, 28 Aug 2013 17:43:21 -0500

On 8/28/13 1:12 PM, Konrad, Martin wrote:
>> there can be a very narrow window during which the DINP calc must
>> be done.
> Sorry, I don't get it, yet. Can you please describe a bit more
> detailed how it is supposed to behave and in contrast to that what
> you think happens if it fails?
> 

Here's the scenario I was trying to describe:

1. Soft motor is commanded to move
2. The motor record sets the soft motor's DMOV to zero
3. Real motor is commanded to move
4. Real motor's position is updated (after one moving-poll-period)
5. Soft motor reads PV linked by DINP field

Two problems can occur at this point if the DINP PV doesn't accurately
reflect the state of the real motor:

1. DINP PV is one and real motor still moving

The soft motor sets its DMOV to one.  The continued motion of the real
motor is interpreted as an hard or external move and the DMOV toggles
back to one until real motor completes its move.  This additional DMOV
toggle is a problem for clients like spec.

2. DINP PV is zero and the real motor is done moving

If the real motor is using asyn motor support with idle polling, this
isn't a problem.  There would simply be an additional, idle-poll delay
for the DINP PV to accurately reflect the real motor state.

If the real motor is using old motor support (no idle polling), I would
expect the motor to get stuck in the moving state.

I've looked over the database you are using (g6test2.db) and the DMOV
gating should make problem #2 from above impossible.

Kevin

Experimental Physics and Industrial Control System

Experimental Physics and
Industrial Control System