Which RS485 end to bond

[ModbusMan]

One of our offices' commercial power companies has offered to let us connect a data line to pull status info off the 13.2kV ATO. First time I'm dealing with this level of voltage and radiated EM, so I'm curious if anyone has opinions on whether I'll be better served connecting my shield/drain at the medium-voltage ATO side, or at the low-voltage instrumentation cabinet. The ATO will be the only device on this Modbus loop.

 

[petersonra]

I think it makes no difference. Grounding the shield does not accomplish all that much with an rs485 signal.

 

[ModbusMan]

That's been my experience so far (and probably to be expected since 485's a differential signal anyways), but until now, everything I've worked in has been 480 or lower.

 

[rlundsrud]

As long as you terminate/ground it at only one end it will adequately shield any RF. As others have said, differential signaling rejects common mode noise up to the rail limits of the transceiver. With that said, I would still want the shielding in place given the magnitude of the radiated RF.

 

[TwoBlocked]

The rule of thumb I have always seen is to ground the end at PLC (instrumentation) end, NOT the field end. I see a few reasons for this. First, you are likely to have a better (lower impedance) ground at your control cabinet. Who knows what has been done, or will be done between the field device and an actual earth connection? Second, it is usually easier to find (or make) a ground connection in a control cabinet, than in some field devices.

 

[rlundsrud]

The rule of thumb I have always seen is to ground the end at PLC (instrumentation) end, NOT the field end. I see a few reasons for this. First, you are likely to have a better (lower impedance) ground at your control cabinet. Who knows what has been done, or will be done between the field device and an actual earth connection? Second, it is usually easier to find (or make) a ground connection in a control cabinet, than in some field devices.

I should have made that clearer, that is the industry standard. I just meant you can't terminate the shield at a second point.

 

[GoldDigger]

As long as you terminate/ground it at only one end it will adequately shield any RF. As others have said, differential signaling rejects common mode noise up to the rail limits of the transceiver. With that said, I would still want the shielding in place given the magnitude of the radiated RF.

A good reason to utilize the shield when the RF (or even AF) noise level may be very high.
Many people forget that the differential signal can be degraded when the common mode noise is high enough to cause clipping of only one line at a time at the receiver input.

 

[ModbusMan]

The rule of thumb I have always seen is to ground the end at PLC (instrumentation) end, NOT the field end. I see a few reasons for this. First, you are likely to have a better (lower impedance) ground at your control cabinet. Who knows what has been done, or will be done between the field device and an actual earth connection? Second, it is usually easier to find (or make) a ground connection in a control cabinet, than in some field devices.

This has always been my way as well, just wasn't sure if the (much) higher RF environment would make using the other end more appropriate.

 

[herding_cats]

This is a really good post. We try to drain in the middle of the modbus run (shielded pair) rather than on one field end or the other. I believe that as long as you are not draining on both ends, or multiple connections, that should be ok. I don't know if the PLC end is better than any other field end. Just my FYI thoughts.

 

[ModbusMan]

We try to drain in the middle of the modbus run (shielded pair) rather than on one field end or the other. I believe that as long as you are not draining on both ends, or multiple connections, that should be ok.

It's fine from a technical standpoint (and some might argue more reliable since the length from the far end to the grounding point is halved), but the weakness of mid-point is that there's nothing to stop someone from going "Ehh, was this the grounding point? I can't remember!" and accidentally leaving the whole thing to float, or worse, multi-grounding and turning the shield into a radio loop.

 

[herding_cats]

Yeah, I could see that. For me it's either the midpoint to drain, or the field side. "Radio Loop" is a good name for a band though.