Hi,
I would like to make simultaneous measurements on several instruments at a scan-rate of 5 sec.
But daily, at a chosen time, several sub-scans will run sequentially to calibrate one of these instruments. This sequence (which go on 30min) must not interfere with the principal scan.
I thought about the slow sequence instruction but do the principal scan of the program will run properly in parallel of the slow sequence? Indeed, the other measurements should run normally during the calibration procedure of the instrument.
Regards,
Hi,
The "parallel" and "simultaneous" words are very important and are related to the delays eligible by your work. Keep in mind the priorities of the executions in the program of the logger , including automatic Background calibration. Much information is available in the manual.
For example
A)Time skew between consecutive voltage measurements is a function of settling and integration times, A/D conversion, and the number entered into the Reps parameter of the VoltDiff() or VoltSE() instruction.
A close approximation is:
Time Skew = Settling Time + Integration Time + A‐D Conversion Time(1) +
Reps/NoReps(2)
1) A/D (analog-to-digital) conversion time = 15 μs
2) Reps/No Reps -- If Reps > 1 (i.e., multiple measurements by a single instruction), no additional time
is required. If Reps = 1 in consecutive voltage instructions, add 15 μs per instruction.
B)Pipeline Mode Task Priorities
1. Measurements in main program
2. Background calibration
3. Measurements in slow sequences
4. Processing tasks
C)Sequential mode executes instructions in the sequence in which they are written in
the program.
Smile
That's the theory, yes. Practically, it is not so evident.
In my program, I have only digital control and measurements (via RS232 ports or SDM).
I just wanted to know if forcing the program (with scan-rate of 5sec)in pipeline mode and add a slow sequence with scan-rate of 1 Min, there will be no interference between them.
To give you more details: in the "slowest" scan-rate (1 Min), I ask an analyser in RS232 (serialout) to entering into its calibration mode. The analyser takes 2-3 seconds to respond. My question is the following: during this time 2-3 seconds, are the other measurements in the main scan (1sec) running properly?
Hi,
In every minute, when the main scan coincides with the slow scan, will be performed, all request every 5 sec and after, in the time that remains to next 5 sec, what request every minute, but, the slow scan process will be interrupted if take more than of what remains to next 5 sec.
When this happens one error " skipped slow scan" will be added.
If the main scan takes 1 sec and slow scan 2 or 3 sec. you should have no problems since the total does not exceed 5 sec.
If you do not know there is an interesting instruction that returns for each line of program the execution time.
The InstructionTimes instruction returns the processing time of each instruction in the program.
Smile
Otemohu,
Yes. Based on the information you provided, the logger will be able to send out the serial command, jump back to the main scan to work on any processing there, and then back to the slowsequence to finish it up. It will slice the processing time up between the two sequences.
Ok thank you Sam. But why have we some skipped scans during this processing time?
May I send you the program to check?
Hi
for a test change the main scan rate from 5 sec to 10sec.
Let's see if they continue to accumulate errors.
This will give us suggestions on solving
Smile
Sutley a. Campbellsci d com
Address you can send program to. Or post to this thread