analog signal into register is being maxed out before reaching scaled value?

[mattyg]

I'm pretty new to PLC's, so forgive me if I use the wrong terminology and whatnot. We have an issue at work where we have a flow meter that is scaled for 0 to 60 million gallons per day (MGD).
We simulated a 20 milliamp signal to verify our signal reaching our SCADA. When the 20 mA was applied, we got a flow reading of 47.184 MGD when 60 MGD was expected. In the logic when the 20 mA was applied, our analog value was correct at 32,000, so it seems like the problem is in the logic with the way the info i being stored in the register, maybe?

Can someone help me understand what is going on here and how I can get this to read correctly?

 

[Steve Bailey]

Do you have 4 - 20 mA signal or a zero to 20 mA signal? What formula are you using to calculate the MGD value? Are you doing integer arithmetic or floating point?

 

[Ken Moore]

So...your raw counts are 32000? What is the range of the input raw value for you analog input? 6400-32000?
What does you scaling logic look like?

 

[robertmee]

So...your raw counts are 32000? What is the range of the input raw value for you analog input? 6400-32000?
What does you scaling logic look like?

Sounds like an old TI 505 PLC and the 20% offset wasn't applied for 4-20ma signal. Although either way with a value of 32000 that's full scale and would still read 60.

Would help if the OP would identify the PLC and analog module.

 

[Ken Moore]

Yeah, I'm an old TI guy, when he mentioned 32000, I reverted back to the good old days. TI 505's have built in scaling features, so not sure what their problem could be. I do not recommend it, but I have seen folks do all the scaling in an HMI or SCADA system.

 

[robertmee]

The TI scale function has a 20% offset and Bipolar selection, Yes/No, so guessing if TI, the scaling block doesn't match the signal type.

CTI bought out the rights to the TI architecture and still going. I have customers to this day using them. Slowly migrating to Logix.

 

[mattyg]

We are working on a Rx3i. We are set up to use a 4-20mA analog signal.

 

[Steve Bailey]

I can't tell from what you showed of your logic, but when you divide 47,184,000 (Gallons per day) by 1440 (Minutes per day) you get 32766.66667
That suggests to me that there is integer arithmetic hiding somewhere in your calculation.

 

[clintm]

Can you post a picture of your scaling logic with 20mA applied to the AI? The scaling looks fine. ~30% input range == ~30% scaled range. So it would appear that 20mA input would give ~41666 GPM (~60 MGD). I'd look to how your SCADA display is interpreting Inf_High_GPM. As Steve suggested, is there a clamp, or a conversion to signed int, somewhere along the way?

 

[Ken Moore]

What analog input module are you using? Are you sure it is a 0-20ma field instrument, and not a 4-20 ma?

 

[Steve Bailey]

According to the screen grab, he's using a module that can be jumpered for either 4 - 20 mA or 0 - 20 mA. In both cases, a 20 mA signal gives a raw data value of 32000.
The screen capture shows a the function block "AI_REAL" providing the correct result for a raw data value of 9464.0 if the module is jumpered for 4 - 20 mA. The OP should monitor the output of that function as he increased the signal at the analog input module. If he gets good results all the way to a raw data value of 32000, the problem is somewhere else. If the output of the function freezes at 32766.67 for any raw data value greater than 25165 (16.58 mA), the problem is in the arithmetic inside the function block.
I have not maintained my software license for the Rx3i since I retired, so the latest version I have is 9.5 which dates back to when GE still owned the product line. I know Emerson has introduced some new functionality. I don't know if that AI_REAL function is an Emerson product or if its home-grown.