L23E Compactlogix Noisy Analog Input

[desertsurf520]

I had some recent issues with the analog inputs on an L23E. At first, I did not wire the analog com back to 24 neg and that was a lot of it. However, I have since corrected the wiring issue and the analog counts are much more stable.

Still, even with everything wired correctly, the counts seem to jump a small percent but they still jump and this is at multiple sites. I monitor the mA and VDC (250 ohm resistor in the card) and both are steady and stable while the counts like to jump.

I believe this is a power or grounding issue as the panels are plastic boxes and the electrical was not installed very high quality. A couple of the sites have the correct electrical system and some are more mickey mouse style. I don't have a ground test set but I am trying to reason with the customer.

The noise is on the PLC side not from the instrument itself.

Has anyone else had issues with this processor or similar having counts jump up and down or somewhat unsteady.

Thanks

 

[drbitboy]

how many bits of noise are you talking about?

 

[OkiePC]

Is the common of the power supply tied to ground? I tend to avoid doing that and have better results when the ground is poor or noisy.

 

[desertsurf520]

I think it was around 1000 or 1500.

The wire is potted in the level transmitter but it did not come twisted shielded. It is twisted only. However, when measuring mA and VDC it is very stable even to the hundreths or thousands

 

[desertsurf520]

It is not right now but the engineer wants to try that, too.

I don't think the whole AC power system is grounded very well plus its in a poly panel and the backplate is aluminum and is not properly grounded.

 

[OkiePC]

What is the complete part number of the L23E?

What is the part number of your analog input module?

I have a recollection of those packaged controllers using the DIN rail mount as a means of grounding. I would definitely ground the backplate and ensure that the DIN rail and ground terminal on the controller are properly grounded.

 

[desertsurf520]

I don't have the complete part number at the moment.
Its not a separate analog input. Like you said, its a packaged controller.

 

[drbitboy]

I think it was around 1000 or 1500.

The wire is potted in the level transmitter but it did not come twisted shielded. It is twisted only. However, when measuring mA and VDC it is very stable even to the hundreths or thousands

is that an answer to the query "how many bits of noise?" so noise magnitude is 1-1.5k raw counts? And that is out of 16384 or 32k or 64k?

 

[desertsurf520]

32646 = 20 mA

 

[Mickey]

You also might try putting a jumper across all unused channels.

 

[desertsurf520]

really??

Just jumper between all the open AIs? and then to dc -?

 

[Mickey]

Just across the channels input terminals.

As ask above what is the catalog number of your analog module?

 

[rdrast]

Do you have any unused analog inputs?
If so, tie them directly to common.

 

[desertsurf520]

L23E QBFC1B

I tried connecting all I+ to common but that did not help.

It's jumping from 28416 to 28500. It seems to just bounce back and forth.

Re-ran cable and grounded the shield at one end.

 

[ASF]

According to the documentation I can find, the input data for the embedded analog inputs is:

Bits 0-6: not used
Bits 7-14: 8-bit analog input value
Bit 15: sign bit (always positive)

28416 in binary: 110111100000000
28500 in binary: 110111101010100
...so what you're seeing is the smallest possible change that the analog input can represent. Though I'm not sure why some of the "not used" bits are also turning on/off, unless "28500" was just a rough figure and it's actually 28480?

Is there logic in place to extract those 8-bits into a standalone integer? And is the variation you're seeing before or after that happens?