4-20 m amp transmitter fluctuation

W

Thread Starter

wisam .j

Hello

I have 4-20 mamp ABB 2600T two wire pressure transmitter. its signal
goes to DCS, but there is signal drop or lose in it for a short time
(fluctuate).

what can I do to avoid this problem?
please
 
First you have to identify the source of the error to know what to do to avoid it.

You can substitute a different pressure transmitter and see if the same errors occur.

If errors persist, then it is not a 'transmitter' issue, it could be either a process or an electrical issue.

Given the minimal information provided, it might be noise periodically generated by some device (like a variable speed drive) that is picked up in a section of instrument wiring that runs in close proximity to the power wiring.

If such is the case, upgrade the cabling (shielded, twisted pair for the xmtr, properly terminated, appropriate shielded cable between the drive and the motor) and/or re-route the wiring.
 
4

4-20 m amp transmitter fluctuation

Mr.David thanks a lot

I did the following:
1-replace the transmitter by a new calibrated one.
2-change the cable (36 V DC power) (4-20 m amp signal) from another location.
3-replace the (36 V DC) power supply by a new one.
4-connect earth from one side of the loop.

but the problem stay!!!!
Wisam.J
 
It's sure that it is not Transmitter problem. please check process and try to solve problem all the best.

amit
 
1) When the signal 'drops' or you lose it for a period of time,

a) does the signal go to 4mA or 0 mA?

b) how long is a short period of time? 0.1 second? 1 second? 10 seconds? 10 minutes?

c) How often is this happening?

2) "Fluctuate" means to oscillate up and down

a) Does fluctuate mean noise? Does this mean the signal is sometimes noisy and sometimes not noisy?

b) Or is the signal clean, but fluctuation is a series of drop-out, drop-in, drop-out, drop-in, drop-out . . .

3) How many junction boxes between the transmitter and the DCS AI?

4) Is there an intrinsic safety barrier in the loop?
 
Mr.David answers:

1- a) the signal drops to zero
b)drops for short period of time 0.1 sec or less
c) its not orderly about 6 times during 12 hours

2- a)it is a signal drop to zero
b)the signal is clean for awhile then it drops to zero for a short time.

3- three junctions
4- I think no

thanks for your interest.
Wisam.J
 
Given the info thus far,

A 0.1 second drop out is one, single update value.

Signal drop out infers an intermittent open circuit in the loop. Ordinarily I'd go after a loose screw on a terminal block, particularly with 3 junction boxes.

But you say you substituted cable from another location and got the same results. Was that cable substitution from the transmitter all the way back to the same AI? Different set of junction boxes? And a loose connection might likely stay loose for more than 0.1 second. Depends. It's still worth checking all screw terminal connections.

A faulty instrument is not likely by instrument substitution, a faulty connection is not likely by cable substitution, a faulty power supply is not likely by p/s substitution.

I'm suspicious of the earth connection. With a fault frequency of 6 times in 12 hours, or thereabouts, I'd disconnect the earth from one side of the loop for 24 hours and see if the fault situation stays the same. If the fault disappears, an intermittent earth potential is providing enough voltage to drive the signal downscale momentarily.

The other consideration is that the AI on the DCS could be flakey. But I'd test the earth connection first.
 
Hi!

If you can't find source of problem, what about workaround?
Is it permissible to average or filter signal?

Best regards,
Vedran
 
A

Alan Balcombe

You may be picking up some EMI onto the transmitter cable.
Is the transmitter output floating, or ground referenced?
Is the transmitter supplied with power from the DCS input channel – i.e. a 2-wire connection?
I assume the signal cable is a twisted pair with a shield/screen?
.
If you can ensure the signal is floating, then make the cable shield/screen to ground at both ends (yes, both ends!), this could be a cure.
 
Top