I have a grounding question for any sparkies: Have a metal 2-gang switch box, switches are on different circuits. Should I:
– ground the box to one of the circuits
– ground the box to both circuits
– splice all the grounds together
I assume #3, since it’s effectively what you get with BX. In this case it’s MC, with the grounding wire.
Edit: There is one legacy BX cable.
Edited 1/11/2005 7:04 pm ET by Taylor
This thoughtfully designed home preserves the character of its neighborhood while maximizing space for a growing family through careful attention to scale, rooflines, and material choices.
"I have learned so much thanks to the searchable articles on the FHB website. I can confidently say that I expect to be a life-long subscriber." - M.K.
Dig into cutting-edge approaches and decades of proven solutions with total access to our experts and tradespeople.
Start Free Trial NowGet instant access to the latest developments in green building, research, and reports from the field.
Start Free Trial Now
Replies
All the grounds get connected.
If there is more than one ground conductor leaving the box and going back to the panel you will likely create a loop. There is some small chance of 'circulating current' that could cause problems with some delicate electronics.
Typically not a danger of equipment damage rather a higher error rate because of an unstable ground reference. In most cases in residential construction this is not a problem. A fairly rare problem even in computer rooms and technical locations.
> you will likely create a loop. There is some small chance of 'circulating current'
This is the protective fault clearing ground we're talking about, not the neutral, right? If so, a ground loop shouldn't be a problem. It's only a problem if the ground carries a current.
We used to think about this in big recording studio installations. The ground was either a "star" system with everything grounded to one point, or a "buss" system with one hulking bar that everything got grounded to. This is for signal ground, not case ground. Amplifier designers know to not make ground loops inside their equipment. Installers know that when there are multiple shielded inputs, you always lift the ground on the source end. Ground both ends, you get a little whine or whistle. Ground neither end, you get a big AC hum. Pull XLR plugs until the noise goes away, and you've found the mistake.
In England, they use a "ring" system for power, not the "home run" system that we use. Every circuit goes out from the panel in a big loop, and the far end comes back to the same panel. I've never heard of them having a problem from that.
-- J.S.
"This is the protective fault clearing ground we're talking about" ...Yup.having an issue with the ground conductor being in a loop in residential construction is pretty rare. off the top of my head i can't remember any cases. Just being extra careful.I have seen a case or two, ghosty problems like this can leave me unsure as to what the problem was and what exactly fixed it, where a looped ground causes a minor problem.Sometimes with multiple switching power supplies, long and complicated runs and electronics that use the ground as a reference the error rate can climb between machines. After a long search we discovered several ground loops. Reconfiguring these to eliminate the loops lowered the error rate to normal levels.Engineers wanted to have us install a complete isolated ground system, the popular solution of the day and a surefire techno-buzz word, for two stories in the large building. A major PITA, several days work and thousands of dollars. They were right. IMHO an isolated ground system would have corrected the problem by isolating the ground reference from the machine and chassis ground but remaking a few wire nut connections to eliminate the ground loops was a whole lot cheaper and easier. Had the company been really short of work the isolated ground system would have made us more money and, in the case it didn't work, would have come down on the EEs calling for it. I talked it over with my boss in private. Years later they were still happy with our simple solution.To this day I have only installed a couple of isolated ground systems that seemed justified. Computer rooms or manufacturing process controls they are a handy way of maintaining a clean reference ground.IMHO a regular single ground system well designed, constructed and maintained will work just fine in 99% of the cases the engineers call for isolated grounds. The EEs like to over specify. Makes them feel powerful. Sometimes I don't object too loudly. Makes the company a lot of extra money and I get some overtime. Double time Sundays ... Mmmm
> Engineers wanted to have us install a complete isolated ground system, ....
This has long been SOP for recording studios, where you're concerned about signals down into the microvolts.
Another thing they worry about is getting a good low impedance connection to true earth ground. A friend of mine used to sink ground rods for recording studios and TV transmitters. His method was to solder together as much as 60 ft. of type K water pipe, hammer the bottom end almost flat, and put a 3/4" npt fitting on the top. Stand this thing up alongside the tower, and just let the water run and erode its way down into the ground. For grins, I tried it for my new service ground, sinking 30 ft. of 1" K in four hours. It's really kinda fun. I have the stuff to do a second one on the opposite corner of the house.
-- J.S.
I'm going to add to the question.
If the switches have a grounding screw, does each switch need it's own grounding pigtail back to the point (wirenut) where all of the grounds are connected or can a single wire be used to daisy chain the switch grounds?
Mike
It's O.k. to think out of the box, Just don't walk off of the plank!
Edited 1/12/2005 5:40 pm ET by Mike S
I have often seen the grounds on switches in a row connected using a single ground wire. I don't have any major issues with it. It effectively grounds the straps on the switches.I prefer to use multiple ground wires pigtailed off the main ground connection. With more than three or four switches I sometimes pigtail one conductor off this connection and then add another connection where I can get all the lines I need to ground the switches. A bit longer it does fold a bit easier into the box.My only complaint with a single wire daisy-chained between the devices grounds is that it makes working on connections, some testing or replacing a single device more difficult. Essentially you have to remove all the switches from the box to do much work. More trouble but not a major issue. On a six-gang box it can be a PITA having to kit the box to to find out the trouble is not there.Of course if the box is metallic, certainly it would be grounded, and the a switch mounting screws has a contact clip, has to have that spring metal clip, on it the ground can be eliminated entirely as the mounting screw and clip makes this connection.Worked in one jurisdiction where the inspector questioned why I grounded the switch at all. I grounded them and code says to ground them. He complimented my diligence but said it wasn't necessary given the non-metallic switch plates. Different strokes. The AHJ always has the last call.
"Worked in one jurisdiction where the inspector questioned why I grounded the switch at all. I grounded them and code says to ground them. He complimented my diligence but said it wasn't necessary given the non-metallic switch plates. Different strokes. The AHJ always has the last call."He is right.I wonder my the code changed. The only thing that I can think o f is that there was concern that "you never know when some one will install a metal coverplate".That or bribes by the GSMA (green screw manufacuters assocation).As to the "ground loop" problem I wonder it the problem was not caused by the large armonics in the neutral inducing a voltage in the EGC. And if the EGC was connected in a loop them it would allow current to flow.Reminds me of a control system on a printing press that I was called out on. The system had a central console and individual controls at each press tower. Worked fine when the presses where off, but when they where run the communication between the central and individual units where lost.They claimmed that it was checkout more than once by the installation/maintace people for the equipment manufacture. And taht they had local electricans check out all fo the press wiring.But nobody could find the problem.Now I was only the software engineer on this and I worked for the company that make the boards for the equipment manufacture.But I am good at trouble shooting.The problem was the connections for the communications wires. In one box there was a couple of strands that did not get under the screw head and when the presses where running there was enough vibration that it was grounding out.
> said it wasn't necessary given the non-metallic switch plates...
But you can still touch the screw heads. I've been zapped by carpet static that way.
-- J.S.