Analog input wiring question

Hello all,

I'm a bit of a newbie to PLCs but not basic electronics. Here's my situation:

I have a several mature TI/Siemens 545 plcs with 6208 analog input modules configured for current input (via external jumpers and internal 250 ohm res.)

The system is used to measure water flow with 24 vdc loop powered (2 wire) differential pressure transmitters. These transmitters are going to be replaced with MagMeters who's controllers/interface supply power for the 4-20 output. The instructions state that the magmeter interface is not to be hooked up to devices which also supply power. The analog input modules are not dedicated to just these transmitters so the magmeters and other transmitters must co-exist on the same modules.

A co-worker suggested removing the card's external jumper for the specific input to be used and then hooking up the magmeter's output to the module's input terminal and the return to the card's common terminal.

A "red flag" went up as I thought that was a no-no and that an isolator (such as an AGM "AUX4000-13") would need to be used to isolate the power sources.

Buying a different or additional modules to dedicate to the magmeters are not an option.

Can anyone tell me who's right, or an easier (cheaper) way to do this?

Thanks in advance...

VON

Your old 2-wire transmitters are most likely connected as follows:

The + terminal on the transmitter is connected to the +24vdc of the power supply.
The - terminal on the transmitter is connected to the PLC input channel.
The com terminal on the PLC input card is connected to 24vdc com of the power supply.

If your new 4-wire transmitters have isolated outputs (most new high quality transmitters have isolated outputs), then you should be able to connect them as follows:

The + terminal on the transmitter is connected to the PLC input channel.
The - terminal on the transmitter is connected to the com terminal on the PLC input card.

The com terminal on the PLC input card is still connected to 24vdc com of the power supply. It doesn't matter to the new 4-wire transmitters but the other 2-wire transmitters still need it.

Manufacturer and model number of the magmeters would help nail down the details. If your new magmeters do not have isolated outputs, then you will need signal isolators to have them peacefully coexist with the other transmitters.

The DP transmitters were loop powered. Was the power supply external to the PLC? If so, then the PLC input card is not supplying any loop power, and you can wire the magmeter directly to the PLC input card, + to +, - to -.

In effect, that would be replacing
- the DP with its external power supply
- with a magmeter with its own internal power supply.

There are some brands of analog input I/O blocks that can be jumpered so that the I/O input has 24Vdc tied to one its input termninal from a power supply connected to the I/O. I don't know if the TI/Siemens does that, but if your existing DP transmitter used some outside, external power supply, then the manner in which the AI card is functioning will work with the magmeter's active output.

Dan

You can download the manual for the 505-6208 from the link below, there are examples of 2-wire and 4-wire instrument connections.

http://www.sea.siemens.com/autogen/docs/plc/505/manuals/505-8120-3.pdf

Hey Guys,

Thanks (so far)
You both have good points but seem to differ on the need for an isolator based on what you've been told. I know this is hard to do without pictures or drawings but let me try to add some more details:

The AI card(s) are powered by an external P.S.

The existing (2wire) xmitters are wired as mellis described.

I don't believe the magmeter has isolated outputs, thus the warning I mentioned from the inst. manual.

When connecting the com. of the PS to the com. of the AI card, all 8 input commons are now also connected (internally).



My concern was the statement in the magmeter manual stating it supplies power for the 4-20 out so don't connect to devices which also supply power (which technically the AI card does not, the external P.S. does).

These AI cards will still have some (existing) xmitters wired as mellis wrote AS WELL as these new magmeters simultaneously.

I guess it really gets down to the fact that the PS com. and the neg. out of the magmeter will be tied together. i.e. two returns from two different sources (P.S. & magmeter output).

Does this seem to introduce the possibility the commons are at different potentials, and thus possible ground looping (or other negative effects? I'm not sure). Since now all the existing xmitters share the same P.S. common that's not currently possible.

Any new thoughts???

TIA, Von.

Von said:

I don't believe the magmeter has isolated outputs, thus the warning I mentioned from the inst. manual.

When connecting the com. of the PS to the com. of the AI card, all 8 input commons are now also connected (internally).



My concern was the statement in the magmeter manual stating it supplies power for the 4-20 out so don't connect to devices which also supply power (which technically the AI card does not, the external P.S. does).

Click to expand...

OK, just because the manual says not to wire it with an external power supply, doesn't necessarily mean it's not an isloated output. Once again, manufacturer and model number of the magmeter would allow us to nail this down.

Otherwise you seem to have a good grasp of the situation. The key point is that an isolated output will allow the - terminal of the output to "float" to the same voltage as the power supply com.

Thanks, Ken for the manual link. In answer to Von's query, "Does this seem to introduce the possibility the commons are at different potentials, and thus possible ground looping?" the answer is yes, unless the mag has an isolated output, as mellis indicates.

Von, note that page 30 of the pdf manual Ken linked to shows that the terminals for connections are reversed for 2 wire loop powered (figure 2-6) and 4 wire powered devices (fig 2-7) which is what mags typically are. The A2/A1 jumper stays put. Mag + wires to A2/A1; mag - wires to A3.

Page 45 gives an isolation spec, field to controller of 1500 VRMS (which is probably the optical isolation), but there's no CMRR or channel to channel isolation spec.

Page 46 states that the inputs are single ended, as Von has indicated, all with a common ground. The cited input protection level of 30Vdc is probably the common mode limit, meaning that if the ground differences plus the difference of any or all signal levels exceeds ±30Vdc then you have a problem requiring an isolator.
How to know whether you need an isolator or not? My experience shows that you have a better than even chance that the mag output is isolated, not requiring an isolator. If the mag output is not isolated, you can measure the common mode between the grounds with a voltmeter to see whether an isolator is needed or not.

If it were mine, I'd fire up the magmeter, put a 250 ohm resistor across the mag outputs, but leave them hanging in the air, not connected to the I/O card yet and force the mag to full scale output, 20mA, so the drop across the resistor is 5Vdc.

I'd put a voltmeter between the magmeter - terminal and the common on the I/O board and see what DC voltage was present. If I got 10Vdc or more between mag - and common, I'd put an isolator in. If the DC reading is low, like less than a couple volts, I'd connect the mag - to the common and see if any of the other readings jump or alter. If not, I'd connect the mag correctly and see how the I/O card responds, but I'd expect it to work OK.
Dan

Thanks for all yor help danw, Ken, and mellis! (in alpha order)

After reading your replies and returning to the office, I was able to confirm the magmeter's controller has an opto-isolated 4-20 output.

Not apparent in the text of the documentation but it is indicated the schematic.

I guess that closes the case on this one.


I will certainly retain and utilize some of the
testing "tricks" mentioned.

Thanks again and I'll be back for sure!

VON

VON

Thanks for the feedback on the iolated magmeter output. It's great when a thread actually closes out.

Dan