Analog output card to analog input card

Greetings,
I have been charged with a project to link data from a Beckhoff PLC using Allen Bradley Point IO to a PLC 5. The PLC5 does not have ethernet so the systems cannot "talk".
I assume I can come directly out of the Allen Bradley point I/O analog output card to the PLC 5 1771 IFE/C analog input card. The scary part is that the PLC 5 analog input card ties the 0 volt between two inputs and there is only 1 input available. However, the assumption would be with both systems using a grounded 0 volt this shouldn't pose a problem. However I am also unsure if the wiring should be + to - or + to +. Has anyone had any experience with this?
Thanks

http://literature.rockwellautomation.com/idc/groups/literature/documents/in/1771-in045_-en-p.pdf

Page 8. Tie the two zero volt sides of their power supplies together. Even at different points there can be small amounts of potential difference in a noisy grounding system that can cause issues. You want to bring them to the same potential.

Active (powered) analog output to an AI is (+) to (+), (-) to (-)

Thank you.
Sparkle, the data accuracy isn't needed to be exact. Lucky because I anticipate plenty of noise. The customers electricians ran the comms cable thru a 110vac conduit. I questioned the customer about that and they said its common practice at this plant. Let's me off the hook of using the decimal point!
I figured I would tie shield at both ends. I don't think this will create a ground loop because of the 2 different power supplies

Tommycatt63 said:

Lucky because I anticipate plenty of noise. The customers electricians ran the comms cable thru a 110vac conduit.

Click to expand...

Analog noise is one thing, its noisy data.

Noise on a digital communications line actually creates a bigger problem because digital protocols universally use the rule, "If it isn't good data, it gets discarded." One either gets a comm error or good data. Various checks are used to determine whether data is good or not, but digital protocols never pass on suspect data.

danw said:

Active (powered) analog output to an AI is (+) to (+), (-) to (-)

Click to expand...

I have to correct myself. Wiring depends on what the analog input is and how it works.

Most AO's wire to AI's + to +.

But there are always exceptions. Some analog inputs provide DC power to passive 4-20mA loop devices or passive analog outputs. The better designs allow the user to disable or enable the DC power option. But I've run into A-B remote I/O that provided DC power and that could not be disabled, requiring the use of a loop isolator with an active input for the active AO and passive output for the remote I/O input.

What's the data you're sharing? What resolution is required?

You could eliminate noise, uncertainty, and the need to filter analog values altogether by using n discrete I/O points to transmit words with 2^n possible values.

With 16 I/O, you can easily 'transmit' 16-bit floating point values, as well.

The setup would be easier to wire, troubleshoot, and expand.

Thank you Dan and Jeremy. The input card on the PLC 5 is a 1771 IFE/C. This is an isolated, non-powered analog input card
Power for the analog Loop will be provided by the Allen Bradley Point I/O analog output card from the Beckhoff. I think I'll be fine as long as both systems are using a grounded 0 volts. I do not believe there will be enough noise to completely corrupt the data. Accuracy requested is 0.1 pound per hour so the resolution is pretty broad.
I did consider going BCD with descretes but this is a very old PLC5 with few spares and real estate is st a premium in the Beckhoff panel.
The only worries I have/had was compatibility between the 2 using separate power supplies when the 1771 uses common 0V on the backplane, the uncertainty of +/+ or +/-, and if that would pose a problem.
Thanks for your responses. I'm feeling a little more confident but would sure like to hear from someone who has done this without smoking the 1771!

Is transferring data via serial modbus not a possibility?

Nope. The only option without adding hardware etc is using the existing open analog channel, which is the "why". The method is somewhat of a hack but so is most of these 3 systems. We did present an option to go Ethernet between the 3 but they don't want to invest. Among the reasons besides just wanting only that single process variable on the PLC5 HMI is they are having us quote upgrading the 5 to a controllogix for the extruder in the future, which would also run the dispensers currently using an old SLC 1-5 brick.
At this point with my research and the responses here I think I'm good. Still a touch concerned about pos to pos or vice versa but that should be easy to see on line.

They’ll find the money for proper comms after iterating through six inferior ways of solving the problem.