Experimental Physics and
Industrial Control System

Tim Mooney <[email protected]> · Mon, 20 Dec 2004 12:57:04 -0600

Lawrence T. Hoff wrote:

 In my application, I have a dozen "stripping foils", only
one of which can be in use at a time. Each foil degrades with use

(related to aggregate number of particles stripped). I plan to

associated a "goodness" (or "badness") factor with each foil
(something like a battery capacity indicator in a cell phone).

Let's assume this means a dozen analog records of some sort,

one for each foil.
	Ideally, I will provide a facility where a user can say:
"insert the best foil", which will find the foil with the highest
"goodness factor", and substitute that foil for the current one.
 To do this, I need to find the index of the foil with the

highest "goodness factor", and pass that index to the foil drive
mechanism.
	I will probably use some sort of subroutine record to
achieve this. However, as a mental exercise, I tried to think
of how to do this with "standard" EPICS records. I could
not come up with what I thought was a "clean" solution!
Maybe someone can offer a suggestion about what I might be
missing?
	I.e. I realize that I can find the highest "goodness
factor" via a SEL record. As far as I know, that can provide
the highest value, but not which input that value came form.
 I also realize that I can use CALC records, and can

even use both the "goodness factors", and the SEL record output

as inputs to the CALC record. However, I cannot think of a
CALC expression to compute the index, within the constraints

of the CALC field length limit. I.e. if the raw values are

in fields A-K, and the highest value (from the SEL record)

is in L, the obvious expression is:
"K==L?10:J==L?9:I==L?8:H==L?7:G==L?6:F==L?5:E==L?4:D==L?3:C==L?2:B==L?1:
0"
which is too long (AFAIK).
 I could break up the calculation into several CALC

records, or use a SEQ record to re-use a CALC record with

different formulas, but it seems to me that once the use of
records exceeds some level of complexity, a subroutine record
is called for.
Suggestions?

I don't have any suggestion for doing this with only the records in base,
but you could do everything you've described with one transform record,
which I'll call 'T'.  Here are the field definitions:
T.A ... T.L are the goodness factors
T.CMTA ... T.CMTL are per-foil descriptions/comments, if needed
T.CLCM is 'max(a,b,c,d,e,f,g,h,i,j,k,l)'
T.CLCN is 'a==m?1:b==m?2:c==m?3:d==m?4:e==m?5:0'
T.CLCO is 'f==m?6:g==m?7:h==m?8:i==m?9:j==m?10:0'
T.CLCP is '(k==m?11:l==m?12:n+o)-1'
After the record processes:
T.M is the max goodness factor
T.P is the index of the best foil
T.OUTP is the PV to which you want to write the foil index.
--
Tim Mooney ([email protected]) (630)252-5417
Beamline Controls & Data Acquisition Group
Advanced Photon Source, Argonne National Lab

Experimental Physics and Industrial Control System

Experimental Physics and
Industrial Control System