ECN Forum Forums General Topics of the Trade General Discussion Area 4-20mA headache

I have a 4-20mA pressure transducer that sends a signal 300-500' to a recording interface hooked up to a PC computer through a com port.
It is run in 4wire #18 shielded com cable.
When I isolate it from the the recorder my Fluke reads 4mA. When I hook up to the recorder the amps jump to 7 to 9 mA. The shield is bonded at the transducer end only. I can't figure out where the current is coming from.
The recorder works fine on all 5 channels with 3 other transducers from 3 other machines. When I hook up this one to any channel, the current increases.

PineMarten,

I see that nobody has answered yet, I'll take a stab at it.

By your description, it sounds like you are opening the circuit completely so that the only thing in the loop during your test is the pressure transducer, the wire, and your fluke. Is this correct? If so, you can't do it that way.

You must test the signal with the recorder left in the circuit (circuit intact), and your fluke connected in series. This should be a series loop, current is always the same throughout a series circuit. The circuit impedance would be less with the recorder in the circuit, therefore current would be greater.

You've pretty much already ruled out any problem at the recorder by getting the same results on all 5 channels.

If you test the circuit with your fluke in series like described previously, both your fluke and the recorder should read the same value. If they don't, either your fluke or the recording interface is out of calibration.

You didn't say if the transducer in question was under any amount of pressure or not when you were testing it. I assume that this transducer is suppose to allow 4ma @ zero PSI, and 20ma @ its maximum PSI rating.

With the circuit connected just as it would be during operation, and zero pressure on the transducer, install your fluke in series. Your fluke and the recording interface should read 4ma. If it does not, the transducer is either out of calibration, or defective. Subjecting the transducer to an amount of pressure at half of its range, should result in a reading of approximately 8ma.

If you work with 4 to 20ma circuits often, it might pay for you to own a milliamp driver/tester. These instruments can drive a milliamp signal of your choice anywhere between 4 and 20 ma so that you can substitute it for the source in your circuit. This will prove if the value on your readout is accurate or not by comparing the readout to the signal driven by the tester. It could also help isolate problems in the 4 to 20ma source. Of course these instruments will also measure the signal just as your fluke does.

Matt

[This message has been edited by Matt M (edited 08-24-2003).]

I should have mentioned the transducer was at 0 psi and the 24vdc power supply is at that end. Originally the transducer was in series with a PLC at the pressure tester and the interface in the QC (quality control dept.)office.
When both were hooked up, the PLC at the tester would read 10x the tranducer's max rating (65000psi) of 400bar (6000psi). I teed off the pressure line and added a second transducer to feed the QC interface.
I am keeping the fluke and interface in series. When just the fluke is in series I get 4mA steady, when I add the interface in series as well the reading jumps up to 7 to 9 mA and isn't steady. 7 for a minute or so, then 9 for the same and back. I can't figure out where the extra amperage is coming from.
The next steps are to:
-hook up a scope
-use a separate 24vdc supply
-put in a bridge rectifier
-your 4-20mA source idea (the fluke has one, and we have more flukes)
-run a new cable (temporarily over coat hangers)
I am on vacation next week. This problem started 4-6 months ago, and now it's my turn to tackle it. I have already spent quite a few hours on it, but can't seem to narrow it down. The electrical dept. keeps wanting to blame the interface, but I don't think that is the problem. There is a 300 page manual with it but no schematics.I'm afraid to go there.

I have tried 2 other transducers, and tested them by getting a 0 psi reading from the PLC readout panel (Magelis and Panelview) on the other testers.
When the fluke and interface are both in series they both read 7-9 mA. When just the fluke is in series it is 4mA steady.
I would think that adding the interface in series would raise the impedance and lower the current?

Edited because I remembering more stuff. My notes are at work.

[This message has been edited by Pinemarten (edited 08-24-2003).]

[This message has been edited by Pinemarten (edited 08-24-2003).]

[This message has been edited by Pinemarten (edited 08-24-2003).]

Going to have to think a little more on this, and yes you're right, adding the interface should increase the impedance and lower the current. Not sure where my head was earlier.

Here's a couple of things you could try:

1) all of our transducers are grounded at the panel only. I don't know what the difference would be between the transducer end or the panel end but the panel does have a dedicated ground for this purpose(green/yellow striped)

2) 300' to 500' is a long run. Don't assume the wire or connections were properly installed. Lift all connections and ensure proper torque when you retighten. I have seen entire installations have poor connections. This is a long run and will need every little free electron it can find.

3) do you have a handheld calibrator. You could try comparing machines side by side. Also comparing before and after the wire.

I'm sure after months that every little nook and crany has been checked. But when I get stuck on a hard one I try to go back to my basics:

1)Don't completely trust the operator or any technician before you. Take in all info but go with your best judgement regardless of how simple it may seem.
2)Most hard problems I solve are found by a combination of good troubleshooting and ROOTING around. Take the time to open all connections, don't just tighten, remove connections.
3)This is sometimes hard to do because of lost production but try to compare two machines piece by piece. Start at the transducer end and start comparing every little thing you find.

I hope this helps you out but I'm sure you've already tried most of this stuff. What immediately raises my eyebrow is the length of run. 24VDC can easily be affected by that length if there is a bad wire or bad connection.

Keep us informed, I'm curious.

I was thinking of bonding at the other end, that was one of my next steps, but I don't think it would make much difference in this case. It is in an office and the only bond point I have is wall recepticles.

I metered each wire at 17 ohms per by linking them all at one end and getting 34 across each pair. There are actually 2 used out of 4, but it doesn't help if I double up to parallel runs, or go single. I wan't to put the megger to them at 500v, the lowest setting on ours. (What the heck is the 200 ohm setting for, anyway?) My boss doesn't like that idea.
It's a Belden comm cable and I think it should handle it.

The run is tie-wrapped 'outside' existing pipe on a 30 foot ceiling. As far as I can tell from the floor (I should buy some new binoculars), it is one continuous run. We have a manlift, but if I go that route I will pull new, heavier, cable; plus other spares. It is a diagonal route across a good piece of the plant and wire is cheap where I work.

I am not sure what you mean by a handheld calibrator. If we don't have one, we can buy it.

The interface can be hooked up to read in mV or mA, but its 'jumpers' are set to recieve 3 volts max. (that's one chapter I glanced at in the manual). Some of the other inputs are set to mV. Would the input type (V or A) make a difference? I assume I would have to wire a mV configuration to parallel, the same as a voltmeter.

[This message has been edited by Pinemarten (edited 08-25-2003).]

PineMarten,

Is there a signal conditioning, or load range resistor (ex. 250 ohm) installed as part of this loop? If so, the actual input needs to be set for 1-5 volts.
When you say the recorder reads 7-9 mA, are you actually deriving these values from a different set of displayed engineering units, or do you have the recorder actually set up for mA display?
It sound like, as Matt suggests, that a signal generator, or loop calibrator would quickly rule out the possibility of a faulty transducer, or input channel.
They could be purchased from Omega and received fairly quickly. They are invaluable for process loop instrumentation servicing.
Also, I'm not convinced that adding the indicator, while increasing loop resistance, would necessarily lower the loop current. I believe this type of transmitter is a constant current transmitter, and automatically compensates for additional loop devices to a certain extent.

Matt,
On a 4-20 mA loop, at 1/2 process span, the loop should read 12 mA. which is the 8 you mention plus the 4 from which zero begins.

[This message has been edited by Redsy (edited 08-26-2003).]

I'm back from vacation and the worst has happened. They want a new cable run.
18/3 belden, (one spare), 625', 25' ceiling. I am sure it won't fix the problem, but it wasn't my decision.
I hope to start at one end with 1000'pull-out box of cable in the manlift, pull it through the mess in the ceiling, leaving loops every 20-30'. Then follow it back and ty-rap it neatly to existing pipes etc.
Is there an easier way, or am I doomed to a week of frustration?
Pull a string first perhaps? Start in the middle?
Help!!!!!!!!!