4-20ma Passive Device Wiring between 2 boxes, with Seperate Power Supplies

theColonel26

Lifetime Supporting Member
Join Date
Feb 2014
Location
West Michigan
Posts
781
Ok, so I have 2 boxes with separate power supplies they are going to be about 300 to 400ft apart. the Main Box as the PLC in it with the Analog Inputs, The remote box has the 4-20mA devices/drivers in it. The 4-20mA device/drivers are passive so I am supplying them with the Power Supply in the Remote Box. My question is, does it matter how I handle the 0V Bond / Return?

Option 1: a Single-wire to bond 0V together for both boxes.

Option 2: a Return wire for 0V for each signal, and another wire to Bond 0V.

See attached schematics (Note: this was made as an example only. So don't nitpick it, this is not the real schematic)


Side Note: I will get all the current flowing back to the Remote Box correct? So in other words, if I have 4 signals running at 10mA I will get 40mA returned to the Remote Box. Whether it is on 1 wire or spread across multiple wires. Correct?


https://drive.google.com/open?id=1PM-H6t44War7JV-f2NeM8sHT7c0aFyAE


https://drive.google.com/open?id=1Ika79mBWnqspa_5EXakmXoGXF0iChcPc
 
Last edited:
Well first of all most 4-20 ma signals can work well with a total resistance of about 500 ohms, the analogue input will be typically 250 ohms so a cable with a resistance of 250 ohms should be ok, but that is not taking into consideration of other factors like noise etc. 4-20 ma is reasonably noise immune as it is current driven. so your cables need to be screened, sized to reduce the resistance. Normally the PSU for the converters would be supplied from the PLC cabinet.
 
I would not allow either in my plant. Just my opinion, no offense intended.

I would have the power supply in the PLC cabinet supply the loop power and run twisted pairs per device. If I have a problem down the road, I want my supply with my IO cards.
 
I would not allow either in my plant. Just my opinion, no offense intended.

I would have the power supply in the PLC cabinet supply the loop power and run twisted pairs per device. If I have a problem down the road, I want my supply with my IO cards.
+1. 2nd power supply isn't necessary or helpful.
 
I agree with Ken on this I would not use either one
Copper Ethernet is only reliable for about 300ft beyond that it would be unreliable.
4-20ma was never designed for that wire length I would consider running it that far to be unreliable as well.

I would read the 4-20ma in to the plc at the source.
I would look at using fiber ethernet at that distance or maybe a radio ethernet connection. Then transmit the converted digital value over ethernet to a remote analog output
 
The 4-20mA signal can easily be run that far, but I would still use at least 18awg wire size, because the theoretical distance and quality of termination are never exactly by the book.

I would also recommend signal isolators at the PLC end to protect the analog inputs from spikes/surges especially if the remotes are in separate buildings.

I was hesitant to respond due to the confusion I got from your side note. I am not sure what to think of this:

Side Note: I will get all the current flowing back to the Remote Box correct? So in other words, if I have 4 signals running at 10mA I will get 40mA returned to the Remote Box. Whether it is on 1 wire or spread across multiple wires. Correct?
 
I don't see any reference to Ethernet I assume he means using 485 cable and this would be a no no as the resistance would possibly be at it's limit.
I agree about remote I/O or coms. I have seen both Pt100's connected by a single multicore cable over 100m and it posed no problem (maybe scaling needed to be adjusted) and 4-20 ma at a similar distance using what we call multicore silflex screened cable again no problems I was aware of.
I once connected a PT100 and 4-20 ma head at 134 meters for remote monitoring of a CIP tank to site wide Scada, I calibrated it using a dry well and did a comparison of recorded wave forms on local Scada & site wide one, no differences they closely matched.
 
For the reference to ethernet most PLC's today communicated using ethernet so I presumed they were using ethernet if not then things would change
Normally RS485 would be an excellent choice for this the length for copper RS485 is 5,000ft it can easily make the 400ft. but most PLC's can't communicate directly over RS485 you will need a converter on both ends then you have to convert the protocol
you would need some device to pole and read both devices and convert the data that would really add to the hardware and software cost.
to go directly with the 4-20ma you will need a splitter to feed both devices and you still have the questionable reliability if running it over that distance. I just think its a poor choice in this application
 
I was hesitant to respond due to the confusion I got from your side note. I am not sure what to think of this:
The answer is "correct"


I was asking about current flow on the 0V wire back.


if there is 40mA flowing through the 4 wires to the main box, then there would be 40mA flowing back on the return.
 
As for Ethernet (Copper and Fiber), RS-485 etc. not an option. This system was spec'ed out long ago to use 4-20mA back to the main box the only question was how to do the wiring.


So I guess I can change it so that the Main Box supplies power to the 4-20 drivers, the other stuff in the Remote Box still needs the local power supply though. (I didn't mention them because they were not relevant to the original question.
 
A


So I guess I can change it so that the Main Box supplies power to the 4-20 drivers, .


Curious, what are these 4-20ma drivers. Product name , model #?


Your drawings show them as 4 wire devices, most (not all) 4 wire devices are not passive but active they produce the 4-20ma not loop powered.
 
Last edited:
What you've proposed just won't work.

1. (Echoing others who know what they're doing) Get rid of the panel 2 power supply (for anything to do with the transmitters).

2. Use shielded twisted pair cable.
- Run one conductor for each transmitter from panel 1's 24Vdc (+) power supply to transmitter (+)
- transmitter (-) back to panel 1 AI (+).
- Jumper all the AI (-) together and return them to the 24Vdc power supply 0V (-) return.

3. I'm not sure where 'Ethernet' came from in the discussion, but if you were planning on using 24G CAT 5/6 Ethernet cable with its 24G conductors, its resistance is 26 ohms per 1,000 feet, so its resistance is not a disqualifying factor for 4-20mA at 800 feet round trip.

I've seen a number of places where they've used CAT 5 cable for 4-20mA and it works, but the issue is how non-robust the terminating connections are. 24G wire just doesn't work well under screw terminals. Any vibration and it breaks. There are RJ-45-to-screw terminal adapters that I'd use if I had to.

But you should know that the industry convention for 4-20mA is 18 or 20 gauge shielded twisted pair, not because the loop resistance is less, but because the cable/conductors are more robust.
 
I would be concerned about the length of wire with any analog signal
even more in this case if you don't have a signal splitter and try to use one current loop for both you could end up with unreliable reading on both and you would never see it.
look at Redlion controls
https://www.redlion.net/products/in...nsors-and-process-control/signal-conditioners
covert the 4-20ma to a 24V frequency then transmit it to frequency to 4-20ma at the other rend the resole would be noise immune and the wire length would not be a concern you would also eliminate the problems with loading in the original sensor
look at the AFCM and the IFMA modules from Redlion
 
Option 2: a Return wire for 0V for each signal, and another wire to Bond 0V.
Yes and no. Why bond the 9V together ?
Your transmitters and PLC AI both are setup for differential output and inputs respectively.

Apart from that there is nothing special about it, neither the general setup nor the distances mentioned. 2 wires per signal, preferably individually screened. Commonly screened is also possible. Screen(s) to be terminated in one end. That is about it.
As danw suggest, you can tie the 0V to the power supply 0V, but only on one side.

Why do consider having a "common" 0V for the return of the current ?
To save wires ?
It is certainly possible to do so, but you will then have a larger current loop without the effect of twisted pairs. So because of that it is normally only done for short distances, i.e. inside a control cabinet or locally on a machine.

Other thing to consider is equipontial bonding between the two systems (not the same as tying the 0Vs together), and surge arrestors if this is between buildings.
 

Similar Topics

I need to go from a powered 4-20mA output (i.e. transmitter supplies voltage) to an Allen Bradley HART input card 1756-IF16H. I cannot set input...
Replies
2
Views
2,362
Problem: Our PLC can only output 4-20mA, but the actuators it needs to control, modulate based on a 0-135Ohm signal. Buying 4 or 8 individual...
Replies
7
Views
218
What's your go to 4-20ma isolator. I've got some 4 wire 4-20ma from several flow meters and scales that I would like to isolate before it gets to...
Replies
10
Views
1,636
Hi, I need to read three 4-20mA signals from a DCS(ABB) in a remote 6 channel analog input module with RS485 modbus port. When I connected...
Replies
2
Views
466
Hello this is my first post. Looking forward to being more involved in this community to learn and hopefully help others. Any help or guidance...
Replies
7
Views
751
Back
Top Bottom