Problem is that you can get a lot of nuisance trips. For example, our TV has a well-filtered power input, including sizeable Y-capacitors. If you have a super-sensitive GFCI, no ground and then touch a headphone jack or something ... click ... THWOK ... pheeeooouuu ... lights out. And worst case the DVD that was recording this great Western for the last hour is toast. Thunderstorm nuisance trips are also quite common.
And then there are these:
formatting link
The price is off, normally yo can get them for $1-2.
We generally bolt our PC board ground planes to the chassis or box, as many places as possible, and the chassis/box is generally grounded through a line cord or rack frame or some such. Given that, our RS232 ground pins are hard grounded to the PCB ground plane, and connector shell is both grounded to the plane and bolted to the side of the box. If somebody wants to push a couple of amps of ground loop current into all that, let them.
If you have EMI concerns, put a ferrite bead or EMI filter in the signal leads.
Maybe your building has some flakey grounds. I've seen several volts AC between grounds in nearby rooms in a building, enough to garble a parallel printer interface.
Have you tried a new (more recent) GFCI? They're supposed to cut down on these nuisance trips dramatically. AIUI, the new models cut down on nuisance trips to the point where the "freezer" exception has been removed from the latest code.
Sure, those things are supposed to be fastened behind the plate screw to pick up the ground from the box (where BX was used). Most of those I've seen in people's junk drawers have the ground tab cut off, too. ...and when the widget doesn't plug into the outlet, it's a lot more convenient to grab a pair of pliers and hack the pin off than find one of those.
I remember back in 76 while working for a subsidary of ITT we had a group working on a HV 20kHz switchmode PSU for a transatlantic cable. The health and safety guys went through their lab grounding everything for safety reasons. After that the team leader downed tools. His arguement was if he touched 10kV and wasn't grounded he would walk away uninjured. 10kV and grounded, well he wouldn't risk it.
Current flowing in the safety ground system (and therefore voltage differences) does not necessarily mean there is any fault or unsafe condition present (although it might). You should never expect the voltage at two different points in a properly installed safety ground system to be exactly the same, although you should expect no more than a few volts difference. It is absolutely impossible to eliminate all parasitic connections to safety ground. Where motors are present there exists a large parasitic capacitance from line to ground, for instance. and significant currents will unavoidably be induced in the safety ground. There are many other sources of ground currents too, such as mutual inductance between power and ground conductors. Typical industrial installations will have amps of ground current flowing at all times even when installed 100% to code; there is no possibility of elimination the unintentional parasitic circuits which produce these currents.
Those of us who have read either Grounding and Shielding by Ralph Morrison or Low Noise Design by Ott (both is better, the two books complement each other nicely) would never make the serious error of assuming that ground == 0 volts; it just isn't so in the real world.
The purpose of safety ground is NOT to provide a zero volt reference, it is to provide a low enough fault return impedance to allow the protective device to clear the fault (enough fault current to trip the breaker).
Thanks for all the replies and the interesting discussion.
Here's the common mode choke:
formatting link
The power outlets check out ok.
The serial port isn't on a DB25 or DE9 with a conventient metal shell; instead it uses an 8P8C modular jack ("RJ45") with the Cisco console connector pinout.
That wispy little part isn't even suitable for V28 voltage levels, and definitely not something you would put in the path of unfettered chassis circulation currents. The impedance is too low to use with RS-232 anyway.
Yeah, the new ones shouldn't have nearly the problems with false trips due to reactive loads as the older ones. The normal brands (Leviton and such) should be better.
You really have no clue about whether the situation is dangerous or not, do you?
Try this: Suppose he has 3 #22 ga wires strung between two systems 1000 ft apart. The resistance of those 3 wires is about 6 ohms. Suppose it takes .8A of current to ruin his choke. Can you calculate what the voltage drop between the two systems is? I'll give you a hint: it is less than 5V.
Stop being such a clueless ass. He said the problem was within his lab. If there is that much current flowing in the grounds in the same room, it is quite likely there is a safety issue.
That's a *BIG* lab, dummy. If he's blowing things up he *NEEDS* isolation and he HAD BETTER look into the problem, rather than taking advice from a moron, like you.
Completely irrelevant.
You're the one who needs a clue. You're passing Slowman and moving in on DimBulb.
ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here.
All logos and trade names are the property of their respective owners.