Experimental Physics and
Industrial Control System

Luedeke Andreas <[email protected]> · Thu, 15 Nov 2007 09:44:08 +0100

Susanna Jacobson wrote:

[...] we are considering writing a custom Power Supply record. [...]

I would also welcome any advice or cautions you have, based on your
experience in this area. [...]

Hi Susanna,

I had a similar idea 8 years ago for some in-house developed power
supply at PSI.

Luckily I got some advice not to do it that way (Thanks Bob!).
If you use one record per power supply, your application will suffer
from a lack of expandability, flexibility and usability.
Just some simple example:
Expandability:

My specification at the beginning consisted of about a dozen values for
one power supply.

Within one year this did grow by a factor of 10.

I would rather have a template with 120 records connected to one power
supply, then having a record with 120 fields.
Flexibility:

If you want to read a new value-type (e.g. resistance) from your PS, and
let's say you have a serial connection to your PS.

Reading the new value should be possible by just adding a few lines in
your protocol file and adding an "ai" record.

But if you have to add a field to a record it becomes much more trouble.
Usability:

you will read back a current and a voltage from the power supply.

Is HOPR/LOPR the display range for the current or the voltage?

Of course you can have an IHOP/ILOP and VHOP/VLOP, but since there is
just one function get_graphic_double, no standard display will be able
to use those fields context sensitive.

The same applies to ADEL, MDEL, HIGH, HIHI, LOW, LOLO, ....

Other people may disagree, but I would strongly advice you not to
develop a power supply record.

Good luck with your power supplies anyway ;-)
Cheers
Andreas
--
Dr. Andreas Luedeke, SLS Operations Manager, Tel:+41-56-3104002
[email protected], http://people.web.psi.ch/luedeke

Experimental Physics and Industrial Control System

Experimental Physics and
Industrial Control System