lets just think about this...
Let us think that the PLC scan time is a process variable.
The measurement (the scan time register) is actually a
report of the 'previous' scan time.
Lets also say that you made your first rung in the program
capture this register, and average it for the last few
scans.
Now you have a non-weighted variable, and its just a little
delayed, but gives a better idea of the recent average
scan time.
We know (because AB tells us) that some output instructions
take a L O N G time to execute if they manipulate arrays,
use indirect addressing, or are just darn complicated.
We might even take a collection of these very long instructions
to manipulate meaningless data, and place them in an equally
meaningless ladder file.
So, then, we _could_ (if we wanted to get real fancy) write a PID instruction that looks at the
non-weighed recent scan time variable, and integrates up
(very quickly) an output that basically says 'how much scan
time to add'. This variable could be then used to call the
previously constructed program a 'set number of times' to
create the required addition to scan time.
(Of course, if this were a PLC-5, you could simply put the amount
of scan time to add into S:77, and you'd be done, but that's
not the question.)
The real question in my mind is: what does doing this get you?
As previous posters have opined, the STI is the hot set up for
repeatable computation timing, notwithstanding the ability of the
processor to deliver I/O data at a faster or more repeatable
rate.