Grounding and Bonding a Production Line

So I should NOT connect the 'Protective Earth'/'shield' lines on the interconnect cables.

Thanks for the help.

Meaty 24Vdc loads such as relay coils are largely immune from ground loops, Mildred. I mean, we are not talking about microphone cables or magnetic cartridge pickup wiring for gramophone turntables, are we?

Well, Captain, from what I've read, ground loops are not just about keeping noise out of sensitive equipment, but also about protecting personnel from a current running through lines that should NOT be carrying current, hence my concern.

After all, to use a plumbing analogy, we don't want the water being pimped through the wrong poops, if you'll pardon my French.

(Sorry, couldn't resist, I've got a lot of excess energy this morning. Wonder if someone snuck an espresso shot into my morning coffee?)

• The principal use of an earth/ground connection is to carry the fault current for as long as it takes the circuit protective device to disconnect the fault. National wiring standards exist to enable simple selection of cable size and protective devices. Following standards will always alleviate concern, and that is what a qualified Electrician will do.
• Ground loops are a greater concern in high-sensitivity phono and microphone audio equipment, which manifest as unwanted "mains hum" at the speakers.

In both cases, star topology is the best way to go.

Thank you, I've already determined that the interconnect cables will have their shilds grounded at only one end, to prevent ground loops.

I'll freely admit when I'm in over my head, and with this setup, I'd say I was in up to my lower lip, and wanted some advice before I sank any deeper.

Sometimes it's hard to tell what are true safety issues that were not mentioned during formal training, and what are 'minor concerns' that weren't worth mentioing during formal training.

Until I had gotten into an actual job in my field, Arc Flash was an unknown term to me, and when I had first heard of it, from the descriptions and PPE, I had thought I had not heard of it because it was an issue just for people working on high tension lines, Then I find out it's for any time someone is working around live Low Voltage (100V) or greater. That means any homeowner adding a new branch circuit to their breaker panel should have the protective gear, but the term and the dangers are never mentioned outside of industrial areas.

I was not sure if Ground Loop was another concept to be as concerned about as Arc Flash, but thanks to you all, I am better educated now, and humbly grateful for all your assistance.

You can connect the shields to common ground, but only at the supply end, and don't earth the 24v. By keeping the 24v floating, it has no reference to earth and is less likely to be corrupted by signals from other parts of the control system.

If you ground the shield at both ends, you are introducing a ground loop between the interconnected machines, this could cause problems in the event of an earth fault on one machine causing larger than acceptable currents in the shield conductor which could then induce a spike in the control cable thus damaging the PLC.

Thanks for the tip.

I've already decided that the cable shields will be connected to chassis ground at only one end, still up in the air about which end, since it's connecting between two machines, both of which are grounded. Once I find my decision-making coin and settle on which end gets bonded, I'll be consistent with it all down the line.

"Bond to Upstream equipment, good, bond to Downstream equipment, good, bond to both at once, squish, just like grape."

(So borrowing Pat Morita's line from the (real) Karate Kid movie(1) doesn't quite fit, but the gist gets across.)

Notes:

1) Not even Jackie Chan could save that train wreck of a remake that Will Smith bought for his kid.

Are there any PLC's or SMC's involved and is there any communications on blue hose involved? Or RS485, 232, 422, 423, or USB btwn the processes?

Some of the equipment uses PLCs, some are simple enough that a few relays can function as switch contacts for the speed controllers.

The reason I'm using the try contacts is so that each machine only 'sees' the power it itself generated, so I'm not 'injecting' power into a PLC input that has no main power source. Some of these PLCs are old, and I'm not sure if they have the internal signal/data isolation that newer PLCs have. IF a machine is off, it is not sending the 24VDC 'read pulse' to the other machine's dry contacts, so there is no voltage returning down the cable, aside from the other machine sending its 24VDC 'read pulse to the off machine's dry contacts.

So with the Very Low Voltage 24VDC lines, and the DC floating, I don't need to worry about ground loops posing a safety hazard as long as each machine with Low Voltage (100V) or greater is properly grounded? Thanks for the help.

The 'interconnect sections' are planned to be in segregated areas on the enclosure's backplane, to make it easy to identify the sections where machines may be injecting voltage into each other.

Due to the large amount of 'legacy machines' with their own wiring color schemes, it's difficult to have a single standard across the board, but with newer machines we're trying to enforce a color code based on the voltage level (even though I would prefer an additional level of detail, so 'power lines' could be distinct from 'control/data lines,' but sometimes one voice cannot be heard above the crowd) and 'external power,' regardless of voltage is planned to be Orange, and it is the ONLY thing that uses Orange.

That's also why I am 'in love' with the idea of using 24VDC as machine intercommunication lines, it's a high enough level that we won't have induced spikes making false signals, and yet low enough that it should be harmless to humans in case of accidental contact. I believe that 24V across dry skin is even below the threshold of sensation. You can't even feel it unless you stick the wires in your mouth, and anyone who thinks that is a good idea probably has had too many recent beers to be reporting to work.

Speaking from an electrical utility perspective where we use DC for device control purposes, we never ground our DC systems. We have up to 250VDC systems in some of our substations, but most common is 125VDC. The DC is really a separate issue from your AC power supply to your equipment and its grounding.

What the floating system allows you to do is keep the control system working if you do happen to get a ground fault. Usually our battery chargers have ground fault detection built into them, and we alarm when a GF occurs so that it can be traced and reparied before another happens.

As for your signal wires using the 24VDC, it probably would make sense to use shielded twisted pair wire to prevent electrical noise in the industrial environment from causing spurious signals. The shield should definitely be grounded. There is disagreement out there if only one end or both should be grounded. In my firm, we ground at the "receiving device" end of the shield, such as at the PLC or SCADA device.

Thanks for the info.

Since these lines are for machine-to machine communication, both sides will be a 'receiving' end. I have to look at it as 'upstream' and 'downstream.' The cables are straight-through, male to female, so there are clear upstream and downstream ends and, like USB, you physically cannot plug the cable in the wrong way. The cables aren't that long, typically 2-3 feet of cable, conductors are 18 gauge, and there generally won't be high voltage cables running next to them, the worst case scenario would be a 110VAC line that runs a conveyor belt from upstream to downstream, and that would be if Maintenance were to 'zip tie' the cables together.

I'll make sure the cable shields are all connected the same way, bonding either all to upstream or all to downstream. Now where's my decision-making quarter? Hope I didn't spend it in the coffee machine last week...

I would definitely let the 24v DC float. Grounding one side will tie all of your ELV supplies together and introduce possible signal corruption.

You say that each machine uses a dry contact to achieve its stop/go signal from the next machine in the line, so there is no interchange of voltage/current between adjacent machine controls and therefore no chance of corruption if the ELV supplies are kept separate.

For wiring and earthing, treat each separate machine as just that although with a basically central control. Taking all earths to a common point will effectively eliminate ground loop concerns, but this would be unlikely to affect your control circuit in any case as it is effectively broken up into separate modules at the control interfaces.

You need to ensure that there is adequate protection against inadvertent starting due to what's happening elsewhere in the plant, ie. well placed lockable Estops for the entire line, and individual isolation for each machine for maintenance safety.

Lots of good info there, thanks.

The interconnect signals are simple "Hand Mode" (Set for working as 'standalone' equipment), "Auto Mode" (Set for working as a team), "Call for Parts" (A signal sent upstream to let the neighbor know it's clear to send more widgets), and "Cycle Count" (A signal sent downstream every time the machine completes one 'cycle,' however many parts that machine's 'cycle' processes. This will generally be used when there are specific 'parts count' accumulators such as bagging/boxing units, or holding towers that are supposed to hold so many parts per chamber, and 'count' the parts be 'watching' the upstream equipment)

All the equipment will have individual E-stops and safeties, and if a machine's PLC says the machine is not 'Ready to Run,' meaning all feeds are good and all guards and safeties are in place, AND the machine is set to 'Auto' mode, than it will ignore any 'Call for Parts,' no matter how desperate.

The main purpose of the automation is to reduce the amount of 'babysitting' needed on the line, so each piece of equipment can tell its neighbor upstream, "Wait a moment, I'm too full right now." This way we don't need operators moving back and forth along the line starting and stopping the equipment by hand, typically getting caught up working on one machine and either allowing a hopper to overflow with widgets(1) or letting a machine sit idle when it's got someplace to put its finished widgets, which causes ANOTHER backup further upstream.

Notes:

1) We're VERY fussy about our widgets, we have to be. Any widget that is touched by a human hand, or touches the floor, is dead to us, as dead as the older Corlione brother from the second Godfather movie. We HAVE to be this fussy about our widgets, they need to be "clean clean CLEAN!" according to the federal agencies who are effectively our gods. Every US company has one of more 'federal gods' who hold the power of life and death over them, even Comcast and AT&T, who thought that they could get away with letting their god, the FCC, that it had no juristiction over the way they wanted to run the internet backbones. I'm sure they're still reeling from that Title II slap across the face(2).

2) I know they're going to appeal, Comcast was already violating the spirit of Net Neutrality while they were claiming that they wouldn't. In certain areas, If you had the "HBO anywhere' subscription (which was a subscription with HBO directly, not with Comcast) you were supposed to be able to watch HD HBO shows streaming on any device. That is, unless your device was a PlayStation 3 connected to the internet through an Xfinity (Comcast) ISP. There the service was blocked, and the users were 'encouraged' to watch the shows through Xfinity On Demand instead. You could watch the shows on a computer, you could watch them on an Xbox, but the PS3 was specifically blocked, probably because Sony wouldn't pay the 'protection racket' money Comcast was demanding.

I would be concerned with allowing the 24 to truly float, although there is rarely the case where that happens. If it truly floats, it will pick up static or other energy from capacitace or otherwise and may move to a voltage that will cause the low insulation level of 24 volts to fail. The OP stated that much of the equipment is old and of various vintage. Modern equipment has been designed so that although no intentional grounding is done, it never truly floats. The resistance from each side of the 24 volts is balanced so that it stays at the same potential to ground. However if you measure the resistance to ground, it will be only in the area of a few meghoms. This is intentional. If the resistance is two high, static charges will not be bled off. When PLC and Distributed Controllers first came into being, many failed due to static electric charges and the failures were at random and throughout the systems and unpredictable. With the old equipment we solved a lot of this by basically tying megohm level resistors on each side of 24 volt supply to a good ground. The resistor must be high enough to not cause any action of the control equipment, but low enough to bleed off any static charges to prevent erroneous errors. The equipment can be "checked" to see if it is new enough to have the protection built in, i.e. look at the instruction books. If the books say that the 24 volt should be grounded, then it probably is not built in. I personally consider the addition of megohm level grounding resistors grounding each side of the 24 volt system. With them installed, the 24 volt supply will exactly "float" +/- 12 volts to ground and if one side becomes grounded it can be detected and measures taken to clear it with no interruption of operation. Newer quality equipment does not need this but there is still some junk out there that does.

Thanks for the tip.

The 24V will be supplied by a 'modern(1)' 24V power supply, since the 'vintage' machines will be inspected and any 'old' or 'unsafe-looking' DC supplies will be replaced.

The 24V system will be used to run the relays for the dry-contact intercommunication, power sensors and operator switches feeding the PLC, run solenoids and indicator lamps under PLC control, and power the PLC itself. Everything is running as a digital signal, sensors show either 'on' or 'off' Solenoids are ether energized or de-energized, etc. There are no analog signals and every signal is either 0V/+24V above the negative line or continuity/no continuity. Any new 24V lines being added to a legacy machine will be isolated from the legacy machine's 'old' power source. A number of these widget polishers still work off the 'master camshaft' control scheme, useful in the old days because you only needed one huge, powerful motor to run the whole thing, instead of having multiple huge, powerful motors you had to try and synchronize, but now with PLCs being so mainstream they're being considered 'quaint antiques' by the PAC early-adopters, you don't need one huge, powerful, expensive motor to run things when you can use several small, strong-enough, inexpensive motors, servos, steppers and solenoids all synchronized by a PLC to run things.

I think I'll stop here, I'm beginning to get on my 'Old camshaft-driven machines suck for so many reasons' soapbox, and I don't need to be there today.

Notes:

1) As in 'current models available through Grainger/McMaster/Whatever-Other-Industrial-Parts-Distributer-Is-Out-There.' It's not the experimental, cutting edge stuff, it's probably using 30+ year old tech, but it's still meeting today's safety requirements. As I'm apt to say in conversations, "My momma didn't raise no fools that lived(2)."

2) Hint: EVERYONE my mom raised is still alive, so the last two words are just bait for everyone who's heard that saying without them. They hear the phrase, immediately dismiss it as the old cliche line, then a moment later they get that "Wait, what?" look on their face that is so adorable(3).

3) Yes I do know I have a somewhat tweaked sense of humor, it keeps life from being boring.