RSLogix5000 Analogues

Hi guys,

Another quick question for those in the know:

When using analogues the raw data file is produced as an INT. Throughout my program I have used DINT's for transferring and processing values. Is it possible to allow the program to compare an INT directly with a DINT? Or do I need to copy the INT's to DINT's before comparison?

thanks

Jonny

When you say compare, do you mean using instructions such as GRT, EQU, LES etcetera?

If so that is possible.

Cheers

Mark

What analog modules are you using..?

If ControlLogix (1756-) you can get the modules to produce the channel data to the controller already scaled to engineering units, so nothing is needed in the controller. You have to set the data format to "Float data" when installing the module (this is the default setting), then configure each channel signal min/max and engineering unit min/max to do the scaling.

If CompactLogix (1769-) The data is indeed produced as INT values, and you will have to convert this to engineering units.

Sadly, most (if not all, I haven't checked) CompactLogix processors don't have the SCP (Scale with Parameters) instruction, and SCL (Scale) is only available in Function Block and Structured Text languages. If you are able to use SCL, then the input parameter needs to be a REAL, for which you can simply MOV the INT to a REAL for use in the SCL instruction. If you don't want to, or can't, use ST or FB, then consider using an Add-On Instruction (AOI) to do the scaling. There are many examples of AOI's to do this scaling.

Alternatively you can "roll-your-own" conversion using the following formula, which is the formula used by the SCL, SCP instructions.

Math instructions accept operands in mixed data-types, converting as necessary.

Thanks for the replies guys.

No, scaling is not an issue for me, I just set the engineering units or work with the percentage format. It does affect the resolution a little but for what I need is perfectly acceptable.

Yes 1769-xxx modules.

I'm just comparing using commands like equ etc but I also have to carry out some simple arithmetic. Would that require the converting to DINT?

If I have to convert to DINT, should I use the COP or the FAL instruction to do this?

thanks

It is not necessary to convert to DINT. The Math and Comparison instructions can mix and match data types.

Text from the Help file on Compute (the category) Instructions:

You can mix data types, but loss of accuracy and rounding error might occur and the instruction takes more time to execute. Check the S:V bit to see whether the result was truncated.

The bold data types indicate optimal data types. An instruction executes faster and requires less memory if all the operands of the instruction use the same optimal data type, typically DINT or REAL.

OG

JonAW said:

Thanks for the replies guys.

No, scaling is not an issue for me, I just set the engineering units or work with the percentage format. It does affect the resolution a little but for what I need is perfectly acceptable.

Yes 1769-xxx modules.

I'm just comparing using commands like equ etc but I also have to carry out some simple arithmetic. Would that require the converting to DINT?

If I have to convert to DINT, should I use the COP or the FAL instruction to do this?

thanks

Click to expand...

You will not be able to use either COP or FAL.... The ChxData members of the module-defined input tag cannot be addressed as an array, which is the domain of COP and FAL instructions.

You will definitely have to address each channel data value uniquely, so you can use any math or compare instruction. And as you have been told, you can use the INT data directly to perform any math or comparison. All math and comparison instructions accept operands in any data-type, automatically converting as and when required.

You would gain better precision if you scaled RAW data coming in to a REAL percentage, or even engineering units, but if you are happy with only 100 "steps" in the incoming analog value, then feel free to continue to use the percent scaling on the card.

Hi Guys,

Thanks for all your help and insight.

You would gain better precision if you scaled RAW data coming in to a REAL percentage, or even engineering units, but if you are happy with only 100 "steps" in the incoming analog value, then feel free to continue to use the percent scaling on the card.

Click to expand...

Daba, the percentage scale is a little misleading as I did think that 100 steps would be very limited. This is my first time using a compact logix and so to be sure I tested an add-on I had built for a VSD motor control. When testing the analogue output and input I realised that the actual integer values used are 0-10000 instead of 0-100 which gives a reasonable resolution.

regards

Jonny