Analog Linearizer Routine For A-b 5/30

NEED A ROUTINE TO TAKE AN ANALOG VALUE FROM I/O AND BE ABLE TO GENERATE A LINEARIZER TABLE FOR THE PURPOSE OF CUSTOMIZING A PID SETPOINT VALUE ( THE SYSTEM IS AN A-B 5/30 PROCESSOR WITH 1771 I/O)

I'm not sure why you need a table for this. The SCP function will generate an engineering value directly from the analog input raw data. It can also be used to generate a linear correlation to any other data.

Is your question perhaps related to a square root function for a flow instrument? In that case, A-B has a square root function in the math capabilities.

More details, please.

This would be used for a setpoint generator for setting Backwash flow rates for a Multi-media filter. The water temperature would set the given flow rate for this backwash, and would change with seasonal river water temperatures.-----Earl Olson

I think I understand where Tom is going, is it possible to do a CPT "compute" statement with your current temp feedback then just transfer this value indirectly back to a SCP scaling value? This would allow your value to compensate immediately based on water current temp with very little logic and no need to generate an entire data table?
Your linearization would be performed by your compute function [mathematical equation] based on your current water temp?

Hey Tom, electrified3:

Earl's talking about a PLC-5/30; no SCP there.

I'm not exactly what Earl wants but if I understand correctly he wants a extract a set-point from a look-up table based on some process variable.

1) Extracting a value from a look-up table is easy. Set up a data file and use an indirectly addressed MOV instruction source to the desired destination.

2) Manipulation of the index will depend on your PV. The main question is: How many breaks do you have (or need) in your linearization?

a) If you've got only a few use individual rungs with LIM (limit)instructions as conditions and MOV instructions which move the index values (constants) into the index register. Set up the LIM instructions to frame the PV values for breaks in your linearization.

b) If you've got lots of breaks then you can set up a loop that will check the PV against ANOTHER table. This is more complicated but it's much less previlent. If you need help with this post again and we'll get you going.

3) SCP. As I mentioned at the top of this post, there is none in th 5/30. Regardless you can easily program it using a CPT (compute) instruction.

Please don't hesitate to post but please try to give us as much information as possible.

Good Luck,

(8{)} ( .)

(Yosi)

Earl,
do you want to make a linear approximation to a series of sampled analog values ?
i.e. y=f(x)=A + Bx. Then find A and B that fits best the sampled values of (x,y).
I have done that with least squares method.

That is, I have done it in an S7. If you need I could provide the formulas in a generic format so that you can implement it in your PLC5/30. But I will only do it if you really need it.

Are you wanting to compute the average temperature for the river water?
If so then you could set up a data table and load it with values like Yosi said using indirect addressing and a MOV instruction.
Then use the AVE instruction on that table to generate the average over the period of the values.
The AVE instuction is only available on the enhanced series of processors.

I think what olson100 wants to do is something similar to a thermocouple linearization table but with custom values. He has an input that produces a complex curve as a return and he needs to scale it to user units.

I think (8{)} ( .)'s suggestion is the way to go. Create a data file of the appropriate data type (integer of float). Decide on an increment value and fill in the table as required. Then indirect address into it. I can also provide a couple of compute examples that will interpolate between data points if that is important to you.

Keith