Weird behavior on analog input

Hi guys,

I just came out of the plant where I had two temperature probes (4-20mA) installed today in two exactly same hydraulic power units.

Both these sensors go to two channels in a Flex IO module.
We wired the first one, compared the reading to the sensor and program and everything was fine.

We then wired the second sensor, connected to the IO and was showing 4.1 mA... this was strange, so we plugged in the configuration unit to the probe and it was stating 6.1 mA output.
I confirmed with a meter and it showed 6.1mA output as well.

Thinking it was the channel, I swapped the two around and the behaviour moved with the probe, so it isn't the card.

I then connected a loop calibrator and checked the sensor reading without the PLC and it was the same 6.1mA.
I connected the sensor to the PLC and put the loop calibrator in measure mode in series in the loop and it showed 6.1mA while the PLC still reported 4.1mA.
I injected current in the card and it was a reasonable reading through the range (not 4.1mA).

I checked the PLC card configuration and it was showing that input configured with 4-20mA too...

I'm really at a loss as to what can be going wrong here. Any takers?

Just a guess - check the output voltage from the sensor going to the card.

Maybe it is enough for a 10-24V calibrator, but not for a 18-24V PLC? (Again, the ranges are a W.A.G.) But if the sensor is only putting out a 12V signal that may not be enough for a less tolerant input.

Aabeck said:

Just a guess - check the output voltage from the sensor going to the card.

Maybe it is enough for a 10-24V calibrator, but not for a 18-24V PLC? (Again, the ranges are a W.A.G.) But if the sensor is only putting out a 12V signal that may not be enough for a less tolerant input.

Click to expand...

The sensor is a two wire transmitter. The voltage going to both sensors comes from the card Voltage bar. I just put a brand new cable temporarily and sure enough the same happens. So the cable has been excused.

It looks as it's the sensor, if not for both the meter and the PC interface to the sensor having one reading and the PLC input seeing another.

I'll have a go with another sensor.

I was thinking of a 2-wire transmitter.

If it is connected to a 24VDC power supply but it is only passing a lower voltage that could be a problem if the input requires a voltage closer to 24V.

Some inputs for PLC's and displays can be rated as low as 10V, where a transmitter reducing the voltage to 12V would easily read the proper value but an input that requires something higher, say 16V, would read the 12V as minimum.

You could try a signal conditioner set to 4-20mA in/4-20mA out and see if that would read the sensor and output to the PLC a better signal.

Likely that the temperature probes aren't isolated so you are getting a ground loop, only solution will be to add 4-20mA loop isolators. Had a similar problem with wave guided radar, one worked fine but with two they just gave the same reading at the highest value.

BryanG said:

Likely that the temperature probes aren't isolated so you are getting a ground loop, only solution will be to add 4-20mA loop isolators. Had a similar problem with wave guided radar, one worked fine but with two they just gave the same reading at the highest value.

Click to expand...

This is my reasoning too... except I can't quite figure out how to correct it.

cardosocea said:

two temperature probes (4-20mA) installed today in two exactly same hydraulic power units.

Click to expand...

I am curious:
1. what is the distance between the two hydraulic power units?
2. are they both bonded to a ground grid in the facility?
3. is the 24V system floating, positive grounded, or negative grounded?
4. is the flex analog input module an isolated module?

I am considering a situation in which we have been instructed to float the 24V system but also use non-isolated analog input modules which I believe are mutually exclusive if issues such as the one you are experiencing are to be avoided.

Hello,



1. enough to have one person pass through them, plus some space for hoses.

2. I believe so.

3. I believe it is floating
4. I do not think it is. It's the IE8B I think.

Looking at Rockwell's documentation, I'm now not too sure on the connections as it seems the "V" connector isn't common across all the terminals.



I'll need to check all of this again.

interesting.

The manual says it is a 238 ohm impedance input so 4 mA yields a 0.952V voltage drop and 20 mA yields a 4.76V voltage drop across the PLC + (current in) and - (current out) terminals.

On a non isolated module the - (current out terminals) are essentially all common. The manual includes the warning

Use caution to prevent ground loops when using a common ground, since the
channels are not isolated.

No instructions on how to prevent ground loops though other than using caution. Are the 4-20mA cables shielded with the shield isolated on the hydraulic units and grounded in the PLC panel?

What is the voltage drop from + to - across each PLC input?
What is the voltage difference from + to + on the two PLC inputs?
What is the voltage difference from - to - on the two PLC inputs?
What is the voltage difference from - to ground on each of the two PLC inputs?

Is there any measurable voltage difference between the metal chassis or ground terminal on the two hydraulic units (since they are beside each other)

edit: yes it looks like there should be no connections on odd numbered terminals 1-15 as these are the voltage inputs. Connections for 2 wire sensors should be from 34-51 for +24V wetting supply and then even terminals 0-14 for current input.

Been a while since I looked for self powered loop isolators, these came up, price is OK and they have a dual channel version.
https://www.omniinstruments.co.uk/txisoloop-loop-powered-4-20ma-isolator.html

BryanG said:

Been a while since I looked for self powered loop isolators, these came up, price is OK and they have a dual channel version.
https://www.omniinstruments.co.uk/txisoloop-loop-powered-4-20ma-isolator.html

Click to expand...

I believe P+F do them as well (the IS barriers without blue connectors). I'll have a lok at them.

No idea if this applies, but I've seen some strange things happen when you have unused analog inputs in a non-isolated card. I have my guys wire a -24 common to the unused analog inputs to pull them down to 0. This usually keeps the weird numbers from showing up on the other channels.

Spent part of the morning on this...

The one sensor not giving us a proper reading was in an hydraulic power unit that was running yesterday.
Today I decided to run the other one showing the correct reading and minutes after the reading went from the 6.1 mA down to 4.8 or so and dropping.

Before doing this I checked the transmitter where the reading wasn't good and noticed that I had 24V between the + and - terminals, but between those terminals and the ground had 0V.
The good sensor however, showed 24 to ground as expected.

Sneaky suspicion is that the probe is damaged and causing the weird reading. I'll pull it out in the afternoon.

Tried a galvanic isolator on the signals as well without luck.

You solved your problem in the very first post, after swapping the two probes, since the problem moved with the probe, you knew all along it was the probe or the cable. Are you able to unplug the probe from the wiring and swap only the probes? That will let you know if you have damaged cable or damaged probe.