Current Loop Monitoring

[horsefeather]

I have several circuits comprised of a transducer providing a 4-20 ma signal to an electronic gauge. I'd like to capture that signal in a data acquisition system (DAS).

I have the design vendor looking for gauges with a data output, but in the meantime I thought I would inquire here.

My first thought was to have the transducer feed the DAS and the DAS feed the gauge, but I can't do that for safety/regulatory reasons.

I'm wary of sticking things in current loops, my second thought was sticking a .1 ohm resistance in the loop and using a high impedance channel DAS to measure the drop and extrapolating the value, but it seem to me that there's probably an off the shelf solution.

PLCs sample data, how is this handled? I don't like the idea of a second transducer to only feed the DAS, because they will eventually disagree with each other.

 

[dpc]

Use of a 1 ohm precision resistor in the current loop used to be very common. We routinely did it for every 4-20 mA loop just for troubleshooting purposes. Generally, this works well - just keep the voltage sensing wiring short and shielded.

 

[ScottyUK]

You could pass the loop through a galvanic isolator such as those marketed by MTL and P&F. They require minimal volt-drop to operate and give a fully isolated output signal which you can do as you please with knowing it won't affect the process loop.


----------------------------------

If we learn from our mistakes I'm getting a great education!

 

[danw2]

It is possible to run the current signal to two devices in series, the advantage being that each device 'sees' exactly the same signal.

Most transducers with 4-20mA output will drive up to 500 ohms of loop resistance, that resistance being the sum of the resistances of the input resistances of the devices and some small amount of resistance for the copper wiring.

 

[horsefeather]

So Danw2, you are suggesting I could run from the transducer (load cell) to my 4-20 ma Analog In Channel on the DAS and then on to the gauge. I haven't selected a vendor yet, but if that's how it's typically done I'm golden. I'll check with the vendors next week.

 

[ScottyUK]

Just check if any piece of equipment grounds the loop. Maximum of one ground per loop, and be careful where it is. Otherwise what Dan says is perfectly feasible. I had the impression you were trying to separate the DAQ circuitry from the control system.


----------------------------------

If we learn from our mistakes I'm getting a great education!

 

[horsefeather]

I want the gauge and the DAQ to always say the same thing.
I want the DAQ to be able to die (power off) and the gauge still works.

The former will occur if they use the same signal and tolerances. I suspect the latter will only occur on a vendor by vendor basis. I will explore with them next week.

Thanks for the info.

 

[itsmoked]

You are likely to never have them agree.. The devices will have different front-ends so there will always be the chance of disagreement depending on the precision you are seeking.

Keith Cress
Flamin Systems, Inc.-

http://www.flaminsystems.com

 

[horsefeather]

I suspect that if I start with the same sensor, I have a better than average chance of making the readings line up. The current practice is to add another sensor; it drives the operators nuts and reduces their confidence in the electronic equipment.

I like the idea of a gauge with a data output.

 

[itsmoked]

A second sensor is much safer as it can provide validation for the control sensor and signal a control sensor failure. Otherwise your operators can sit there happily recording and controlling by some drifting bad sensor. I would much rather have my operators worrying about sensor comparisons then sitting there being clueless.

What happens is over time the operators notice and learn that two sensors have say a 1 degree F offset with each other. That's okay! When one day it is a 3F offset they know something is failing.

Keith Cress
Flamin Systems, Inc.-

http://www.flaminsystems.com

 

[horsefeather]

Remember when your six year old first learned how to tell time and would go from clock to clock and enquire why one clock said 7:41, the one in the kitchen says 7:39 and the one in the library says 8:03?

Not that operators are six year olds, but they work in absolutes in the real world. When the company man is standing over their shoulder questioning their every move, having one set of indicators make everyone happy. If they have a gauge that says they have 11,622 feet in the hole and another one that says 11,620 feet it provides a significant conundrum. Sure I have standby indicators but they requires the operator to dismount and their contradiction is not blatant.

When the company man is happy, the operators are happy. When the operators are happy, I'm less concerned. Sometimes it's easy to fall in love with the design of the hammer and forget that to an operator it's only there to drive nails.

Thanks for the reply.

 

[danw2]

Keith's warning about non-agreeing devices sharing an identical signal is a real-life issue, particularly if the number values in question are real-life, ranging 12,000 counts, where the last digit needs to correspond. I'm under the impression most industrial 'meters' are spec'd to ±X% accuracy, ±1 count in the last digit.

However, there is the digital comm option. For example Precision Digital's (PD) Trident model is a panel meter whose front end takes 4-20mA and displays it as a digital value. A Modbus slave option is available with an RS-485 port.

A different model meter has Modbus only with an RS-485 port (no analog input), and acts as a limited Modbus master, which polls the PD slave and gets the value via Modbus. The displayed values have to agree, within the timing of the Modbus cycle, because the value coming over Modbus is a digital value in hex, not an analog value.

But I'm not sure what the precision is (3 1/2 or 4 /12 digit), nor what the accuracy spec is, and I'm nodding off after a day at the plant, too tired for web search.

If PD makes a Modbus meter, there's probably other brands as well.

Dan

 

[controlsdude]

What about this off the shelf din rail mounted 4-20ma signal spliter, its a phoenix contact product?

http://www.phoenixcon.com/products/interface/db/datasheets/mini_2i.pdf

 

[horsefeather]

Thank you all for making me smarter!

It appears there are three options to do this:

1. Use a low resistance resistor in series with circuit and measure voltage drop.

2. Wire SCE in series with circuit and measure current

3. Use a signal splitter.

Each option has its pros and cons, I'll run it by my DAS vendor for their recommendation.

Regarding the potential problem of different values displayed on two indicators, I'm going to use an analog instrument and a digital display. There's enough ambiguity inherent in the analog to assuage any concern.

 

[itsmoked]

Regarding the potential problem of different values displayed on two indicators, I'm going to use an analog instrument and a digital display. There's enough ambiguity inherent in the analog to assuage any concern.

Nice solution.

Keith Cress
Flamin Systems, Inc.-

http://www.flaminsystems.com

 

[EricRICPDAS]

A 4th solution. You can use the data acquisition sensor set to read volts. Then put that across the inidcating meter. This way both the meter and the DAS are using the same shunt resistor to detect current. You can get input devices that are isolated and differential for this, such as

http://www.icpdas-usa.com/products.php?PID=477

.

Eric R.

 

[horsefeather]

How incredibly obvious! Why did I not think of that?!

I'll check with the vendor tomorrow regarding the nature of the meter.

Thank you!

 

[roydm]

Horsefeather,
I'm not sure if I got it right but I think you want to sample and log the mA signalb. I did something similar using one of the small Modicon PLCs with limited No of analog inputs. I connected 2 diodes in series with each loop. The normal voltage drop is 1.2 V. Using a 0.2to1V analog input module and 2 8 input relay output modules sequenced to sample each pair of diodes in turn I was able to read 16 loops with 1 analog input in about 5 seconds.
The primary loop was never broken as the diodes maintained the circuit while not sampling.

Regards
Roy

 

[roydm]

Horsfeathers,
I tried to attach a sketch to explain my last post. Send me your address I will send it back

Regards

Roy

rmatson@noram-eng.com