Types of analog inputs

Assuming your PLC analog input module can accept either 4-20mA diffential or 4-20mA single ended inputs, what are the advantages and disadvantages of each...noise, isolation, what? Obviously when cost is the factor you can get 16 single-ended analog inputs vs. 8 diffential inputs (on a 16pt. module, but besides that concern.

Thanks for the responses.

4-20 ma translates into 1-5 volts. This may limit resolution compared to 0-20 ma (0-5 volts). With 4-20ma, you can detect a loop failure. A signal under 1 volt (say .5 volt or so to eliminate noise) will allow you to write code to generate an 'Open Loop or Failed Transducer' error.

Single ended inputs must all reference the same common. That can be a problem when you start mixing loop sourcing devices in with externally sourced devices.

You have a greater potential for signal noise with single ended. Since the common is shared, once noise source can influence multiple channels.

No one has mentioned any hardware in particular so far.

I had a discussion with Ron B. a little while back about single ended vs. differential for Allen-Bradley 1771-IFE analog input cards. Check out this thread.

http://www.plctalk.net/qanda/showthread.php?t=21674

Without speaking to specific hardware, I don't think you can guaruntee that there is any benefit to a differential mode analog input.

For instance, using a 1771-IFE in differential mode you can still have problems with badly behaved transmitters that reference their signal to a voltage significantly different from the common of the input card (some older drive follower signals will do this). You need a real signal isolator to solve that.

The shared common is something of a red herring. If you are wiring several 2-wire transmitters to an analog input card, they will all share the +24 and com from the power supply. It doesn't matter if you wire them single ended or differential. If they are all wired with twisted pair from the field to the PLC panel they should be fine either way.

Another advantage a 4-20ma signal has over a 0-20ma signal is a "live zero". That is, when the signal is close to 0% of scale, it's not lost in the mud at 0ma. Or to say it another way, you have a much better signal to noise ratio with a live zero.

just a quick note ... I have NOT forgotten about that discussion that mellis just mentioned ... I've got some ideas for some lab experiments and I'm just trying to find enough time to do them ... basically my idea is to hook up some samples of single-ended vs. differential inputs ... and samples of two-wire vs. four-wire inputs ... and so forth and so on ... and incorporate some "noisy" signals into the mix ... hopefully I'll be able to graph the results of these experiments and post them before too long ... I've just got a LOT of things (good things) going on right now ... if anyone has any ideas that they'd like to see tested, let me know and I'll try to make that happen ...

I just want to point out that 4-20mA and 0-20mA don't translate to 1-5v and 0-5v unless you are using a 250 ohm resister. While this is very common, it is not a safe assumption to make. Some inputs use a 500 ohm input impedance so you're now talking 2-10v or 0-10v and some use 125 ohm which is 0.5-2.5v or 0-2.5v. There are other values out there too...

In an industrial application I wouldn't use 0-20mA if I had any choice because of the s/n issues and the lack of signal failure detection. For similar reasons I also stay away from analog voltage signals whenever possible. Simple reason on that is if I put 5.23mA into a wire on one end I'm going to get 5.23mA out at the other end, doesn't matter how long the run is or how much resistance (to a point) there is, or if the resistance changes. With voltage if I put 1.23v in on one end I'll get something different at the other end every time; problem is how much lower can vary.

Ron Beaufort said:

if anyone has any ideas that they'd like to see tested, let me know and I'll try to make that happen ...

Click to expand...

Applying an increasing common mode to one signal (of several) and the seeing the effect as the common mode exceeds the capability of the input card would be instructive.

I've found that well isolated analog inputs with high common mode rejection are always differential, never single ended.

I agree with mellis that the use of single ended inputs with 2 wire, loop powered transmitters, all of which share a common power supply, seems to work well; but outside of that specific case, I always run into problems with single ended analog cards.

Dan