GFCI Query

You are confusing polarity with line and load. It really doesn't care about polarity although the code does. The thing you do have to do is be sure the line side is connected to the feed and the down stream load is on the load terminals. The proposed U/L spec is that they punish you ( Won't reset?) if they are swapped. I imagine they could make them work if they were wired either way but I bet they won't be $10 anymore.

There are many types of GFCI's from breakers to receptacles to pigtail plug-in. All sense a current imbalance using a single core CT. So I don't think the polarity matters, although I've never tired it. I did run across a problem in the field when a two pole GFCI breaker was installed on a 240 V circuit. The two pole plug-in breaker has provision for connecting two hot and a neutral (3 terminals). On the plug-in end is the two hot clips and a white pigtail. Since there was no neutral the electricians just taped the pigtail instead of attaching it to the neutral bar. Everything worked fine except that the test function failed. After replacing the breaker a couple of times they decided to get a second opinion. As it turns out the test button routs 5ma of current from the line side of the CT to the load side neutral (white pigtail) So even if you don't have a neutral conductor you still need to attach the neutral pigtail to complete the test circuit.

Thanks for this info.

Can I ask a supplementary along the same lines.

I've heard that a PC should never be used on a non-grounded socket even if it is protected by a GFCI. Is this true, and if so, why?

Cheers

Den

You lose a significant part of the RFshielding. There can also be surge protection issues.

It's a myth. Consider a laptop with a 2 prong plug for the AC adapter - no ground, works fine.

in article gqKKb.23656$R snipped-for-privacy@nwrddc03.gnilink.net, Den Murray at snipped-for-privacy@nul.nul wrote on 1/6/04 6:46 PM:

This is true. The reason is to protect the silicon devices from ESD, electrostatic discharge. If a sufficiently large potential difference is generated by friction or charge transfer, little energy is needed to get a permanent breakdown failure in silicon devices, particularly MOSFETS with no built in charge leakage. GFIC is to protect people, not mineral devices.

If the equipment has a ground connection, many circuits purposely provide leakage paths to ground to prevent the buildup. Without a ground, that does no good.

Bill

Primary reason is for safety - if the PC case is made of metal, then it must be bonded (connected) to safety earth ground to prevent possible electrocution in case any live loose wire inside the PC should touch the metal case.

PC power supply is very noisy, as well. It uses a noise filter which dumps high frequency power supply switching noise into earth ground (it could be as high as a few milliamps for real noisy power supplies). If earth ground is not there as a "dumping ground" for noise, then strange malfunctions (intermittant) might occur.

PC power supplies contain RFI supression components, which as a side effect leak a small current to ground (too small to trip a GFCI). The power supply therefore expects that it has a ground connection through which this current can leak. If it doesn't, the ungrounded casework will float up to mains voltage (typically it sits somewhere around half mains voltage, but that depends on the quality and tollerance of the RFI components).

There are a number of issues with having a piece of metalwork sitting at such a voltage. If the RFI components are in good condition, the current available should be less than a 1mA (at least in the EU -- I don't know what the US regs say). So risk of shock isn't high, but the RFI components are not designed failsafe and a shorted capacitor (which could easily be triggered by a mains voltage spike) could easily leave the case with a low impedance connection to the live supply.

Connecting peripherals to such a PC has issues. For a Class II (double insulated) peripheral, it will be carried to mains level potentials via its data/signal connection to the computer unless it contains full isolation on the interface. For a Class I (grounded) peripheral, if it also has no ground connection, then it will also be leaking a similar current, so the available shock current available is now doubled. If it is grounded, then the signal/data connection between the two devices is likely carrying the leakage current from the PC. Depending on the construction of the connectors and the electrical protection on the interface, there are circumstances (particularly live plugging) where this current and voltage level could damage one or both appliances. If a ground fault happens in the PC, then the interface will carry the fault current, in so far as it is capable of doing so without burning out.

Finally, relying on a GFCI for your protection is not a brilliant idea. They are themselve prone to non-failsafe failures. In the UK, we don't allow grounded appliances on non-grounded outlets at all.

So, if you do this, it will probably work, but there's plenty of rope available to hang yourself one way or the other.

Try: Fundamentals of Ground Fault Circuit Interrupters at:

Gerald

I good write up on how a GFCI works, however the statement that a "...GFCI opens the circuit and limits the current to no more then 6 ma..." is not quite accurate. There is nothing in the GFCI circuitry that limits physically limits current. Your write in the second paragraph and the TCC graph for the LM 1851 shows a fault current of 200ma @0.02 sec is possible before the device would trip. Maybe a minor point, but in the past I have had to caution workers that were using power tools at heights that the initial shock may cause them to lose balance or drop the tool both which can have serious consequences.

Mike

A few years ago,my PC was not properly grounded (poor contact of the receptacle)and touching it I got a shock.

-- Dimitris Tzortzakakis,Greece Visit our website-now with aircondition!

Ï Den Murray Ýãñáøå óôï ìÞíõìá óõæÞôçóçò: gqKKb.23656$R snipped-for-privacy@nwrddc03.gnilink.net...

The laptop has a plastic case.

-- Dimitris Tzortzakakis,Greece Visit our website-now with aircondition!

Ï Ýãñáøå óôï ìÞíõìá óõæÞôçóçò: snipped-for-privacy@bellatlantic.net...

A laptop IS a PC, which automatically makes the statement under discussion a myth: "a PC should never be used on a non-grounded socket even if it is protected by a GFCI."

But read on - I'll address both laptops (with plastic cases - irrelevant as will be shown) and PC's that are not laptops, both with and without plastic cases.

My Dell minitower has a plastic case, and a 3-prong plug. Ground has nothing to do with whether either my ThinkPad or my Dell work or don't work. The fact that a PC's electronics work just fine with no ground is proved millions upon millions of time every day by the work done on laptops.

The ESR stuff people mention is not well thought out. Many semiconductor chips are so sensitive that just handling them can kill them, due to static discharge into the chip. BUT - seal them inside a conductive UNGROUNDED material - that's how they are shipped - and they can be handled with no problem. The static charge is spread via the conductive material over a wide area such that it won't harm the chips, even though they are touching the material and thus very close to your hands and the source of the static electricity. A PC, in a metal case with NO ground has a MUCH MUCH greater area for the static charge to spread over, AND the chips are no where near in contact with the case. There is no discharge from the case into the chips. The case simply does not have to be grounded to be effective against static discharge. If you remove the chips from their conductive package, or if you remove the case of the computer to handle the electronics inside, then you need to use a grounding method to ground yourself before touching the parts.

Think about it. The PC arrives at your home, shipped with styrofoam packing. If it arrives at the time of year when there is extremely low humidity, it is obvious that there is an awful lot of static electricity. The PC cannot be grounded when it is inside the box, and you need to remove all the packing to get to the PC. You are handling an ungrounded PC extensively, more than most people will ever handle it again, in a high "static environment" - possibly the highest static environment it will ever be exposed to - and no static electricity damage results. Ever.

The ground is there for safety. In the event of some internal failure, it might be possible for the exposed metal parts to become energized. If that happens and the case is grounded, you won't get hurt. That is the purpose for the ground - safety. If the PC is plugged into an ungrounded GFCI receptacle, and the same failure occurs, the GFCI will trip, preventing you from being hurt.

Someone pointed out that a GFCI could fail. That's certainly possible, and so one can say, based on that, that it is

*better* to use your PC on a grounded receptacle versus and ungrounded GFCI receptacle. But that possibility does not support the myth that one should *never* use a PC on an ungrounded, GFCI protected circuit.