A quick safety question

[mashaffer]

The general practice is using a three prong cord with the safety ground tied firmly to the chassis. However if someone uses a cheater without hooking the ground strap on it to a real ground all protection is lost. In situations where this is likely to happen (old homes with 2 wire system) I wonder if it would be safer to just use a normal 2 prong plug with a built in ground fault interrupter in either the cord or first thing in the chassis.

Thoughts?

[r0b]

Well, I would think it's still pretty darn unsafe.

Imagine a fault occuring. Hot has contact to chassis. In a normal setup, chassis would leak to ground -> fuse would blow.

With gfi, no leak to ground, no fuse blowing, gfi does not notice because no current is flowing. And the potentially lethal part: Chassis is now hot.

Now I don't know how fast and safe the gfi would trip, but it better be damn fast, otherwise body gets hot when touching chassis.

I am not an electrician, and I would love to hear if my train of thought is correct on this one.

I had a cheated socket in my old home. you could see the line was laid from another outlet, but my master-slave multiple socket adapter said "Connection Fault" no matter which way it was plugged in (in Germany you can plug stuff in two ways). That was when I noticed there were only two lines. I even told the landlord, but he didn't care. If someone got hurt there, he'd be in deep ... trouble.

[Andrew Eckhardt]

r0b, far as I know GFI trips if any (small) current going out the hot line doesn't come back on the neutral. So a hot chassis with current leaking to anything other than the wired neutral would be safer with a GFI in the supply. I think the 15A receptacle types trip at 10s or couple hundred uA max, and yeah, pretty darn fast. Like you said, a GFI wont prevent a hot chassis but it might keep one from killing something.

[mashaffer]

That was my understanding of GFI. If anything other than neutral becomes the ground the difference would cause shut down before it could kill someone. That is IIRC why GFIs are required for outlets near water sources and why they are used on hair dryers.

[Andrew Eckhardt]

A quick google says the trip current is nominally 5mA, which is 5 times what I learned is necessary to fry yer gourd. I don't know where I got my uA current ideas. I thought I read that somewhere on a device I recently installed... Apparently YMMV.

[M Gregg]

mashaffer,

I am an electrical engineer, however I do not work on US systems. You must not rely on leakage current through a human body (reduced or not) as a form of automatic disconnection of supply! I am sure that a US elec will step in here!

Regards
M. Gregg

[gyro]

A very recent personal experience to relate:

I use a wall mounted GFI to power all the sockets on my bench at home - nice and convenient, 'Test' button to turn the bench off, 'On' button to turn on. Then the other day I pushed the test button and it didn't turn off. I though it must be the internal resistor on the test button had come loose so I wired a plug with a big 4k7 resistor Live to Ground pin (well over 30mA) and tried that - nope still no trip. So then I thought it must be the mains socket ground connection so spent half an hour unscrewing sockets, testing continuity etc, but nope that was fine. Finally tried the cheap plug-in GFI that I use with the lawnmower on the same socket with my 'resistor plug', that tripped instantly.

Plain and simple, it's just broken (still need to replace it!), I thought it would be a bit more failsafe than that... but I suppose that's why there's a test button - it just isn't. It was a good quality one too, MK I think (UK).

It's not welded contacts either, I always turn everything off before switching the bench off and the 'On' button mechanically opens the contacts when it's held in.

Just seemed a relevant place to share. Not a comforting feeling when you're playing around inside mains gear.

Chris

[SGregory]

A GFI will not discharge the power caps. It would cut the mains, but the energy stored in the PS caps would have to drain.

[mashaffer]

Originally Posted by SGregory

A GFI will not discharge the power caps. It would cut the mains, but the energy stored in the PS caps would have to drain.

A good point I had not considered. Of course a mains fuse wouldn't either but a B+ fuse after the caps would assuming that the short is after the fuse. This business of defining possible failure modes is tricky stuff isn't it.

Of course depending on the safety ground is sort of having all of ones eggs in the same basket and my concern is when that ground is defeated either intentionally (cheater or to eliminate ground loops) or unintentionally (faulty mains wiring).

[Michael Koster]

Originally Posted by M Gregg

mashaffer,

I am an electrical engineer, however I do not work on US systems. You must not rely on leakage current through a human body (reduced or not) as a form of automatic disconnection of supply! I am sure that a US elec will step in here!

Regards
M. Gregg

A GFI is not an approved substitute for a 3 wire cord and ground connection to chassis