Dangerous troubleshooting

[rsjsouza]

Apart from the low battery, the DMM he is using is not very reliable... I have the Triplett 9005, which is the OEM of the UT-51 and its rotary switch is very flaky.

Thus, both the VTVM, the UT51 and the self-made leads are unreliable... Oh well...

[vince53]

I'm setting up an "electronics lab" for experimenting, etc, since being reenergized by Dave's videos. I'm a retired master electrician and hold an Extra class ham license.

I have a Variac and isolation transformer in a rack mounted chassis that I built 40+ years ago and will get it set up on the bench.  My old analog scope has the ground pin cut off and interestingly I just bought an old B&K function generator with a missing ground pin.  As an electrician I'm well aware of potential shock hazards from improperly grounded equipment.

When I get my isolation transformer set up I will not connect the ground wire to the outlet receptacle but the hot and neutral from the secondary will be fed through a ground fault circuit interrupter (GFCI).  These are used many places in the home for personal protection and are approved for feeding and ungrounded circuit.

From memory, this is how they work.  The hot and the neutral wires run through a current doughnut (no ground connection).  As long as all the current flows through these two wires, the sensor sees zero amps.  When some of the current travels outside of the circuit (you have your hand on a cold water pipe and touched a hot chassis) the imbalance is sensed and the circuit is opened.  The trip current is between about 5 and 10 milliamps, not enough to kill you.

So no matter how badly things are messed up, you won't get killed.  Also, if you do have a leak to ground you will be forced to to find and repair it.  I only intend to use the isolation transformer for the DUT when appropriate and restore the grounds to my test equipment.

Vince

[Andy Watson]

From memory, this is how they work.  The hot and the neutral wires run through a current doughnut (no ground connection).  As long as all the current flows through these two wires, the sensor sees zero amps.

But ...

the hot and neutral from the secondary will be fed through a ground fault circuit interrupter (GFCI).

Since the secondary is isolated there will be no imbalance in the L and N currents - therefore the GFCI will not function.

[mos6502]

Using a UNI-Turd meter for troubleshooting is like using a circular saw for trimming your pubes. The first thing I would do is sell most of that old junk and buy a Fluke 87V.

[vince53]

Andy,

If isolating the secondary for the GFCI to trip a number of things to go wrong.  Such as: DUT has an the chassis hooked to one side of the line and and a non-isolated piece of test equipment has the ground clipped on to the chassis.  There is now wiring in the chassis that is 120vac above ground.  If you touch that wiring and came in contact with an earth grounded object, The GFCI would open and you'd still be alive.

GFCI's are for human protection and they would work on the primary or secondary of the isolation transformer.

The GFCI maintains shock protection while legally allowing you to not connect the equipment ground.

Vince

[Ian.M]

The GFCI will not trip.  The isolating transformer secondary winding is an insulated wire with no way for current to enter or leave it other than via the ends.  Therefore no matter what you do downstream of a GFCI connected to the secondary, including clipping an earthed scope ground clip to isolated Neutral, and a Presto Hot Dogger between isolated Live and the nearest water pipe, you cannot get a current imbalance through it, so it cant trip. (I am neglecting the leakage current via the inter-winding capacitance as that will be an order of magnitude or two below the minimum GFCI trip current.)

The ONLY way to get a GFCI to trip on an isolated supply is to strap isolated Neutral to isolated Ground before running Live and Neutral through the GFCI.   Then any faults to grounded chassis in the D.U.T. will trip the GFCI. However that doesn't improve your personal safety.  All it does is serve as a half-arsed substitute for properly inspecting and electrical safety testing the D.U.T. before first plugging it in, and if you habitually plug in suspected faulty devices you don't know the history of without a safety inspection you are a Darwin award candidate waiting to happen.

OTOH feeding the rest of your bench that isn't on an isolating transformer, from one or more GFCIs with lighting on a separate circuit is worth doing and provides a considerable safety improvement over having a non-GFCI feed to your bench.

[Shock]

Ian can you think of any more additional risky scenarios than listed below?

HV/EHT, HV/EHT arching over, introduced earth, live chassis or heatsinks, neutral tied chassis, live via bench, live via test equipment, live while touching neutral, non discharged caps.

[Ian.M]

There are plenty more workshop hazards, but if you restrict it purely to electric shock hazards, and exclude specific hazards like working on camera flash circuits as subsets of HV or non-discharged caps you've already listed, what about your physical reaction to a non-lethal shock? e.g. jerking away and slicing yourself your on a sharp chassis edge or pulling a heavy rack of test equipment down on top of you.

I aso emphasise the necessity of having sufficient lighting on a separate circuit to the RCD/GFCI.  If you trip the RCD and now have to extricate your hand from the D.U.T in total darkness without encountering anything sharp, charged or hot, there's a significant risk of injury.

[vince53]

Ian,

It seems from your response you won't be easy to convince.  I'll get a GFCI recept over the weekend and see if I can simulate it.

Vince

[FlyingHacker]

It's a shame that companies decided to call these modern devices Isolation Transformers, as they are not fully isolated. The output ground is still tied to the input ground (which in itself is not horrible), and worse, the output neutral is tied to ground.

You either have to buy a real bench isolation transformer, or go in and fix it yourself.

Only the DUT should be isolated. Everything else should be grounded.

[OilsFan]

SDG did some good videos on it too where he specifically mentions the ground.

[tggzzz]

It seems from your response you won't be easy to convince.  I'll get a GFCI recept over the weekend and see if I can simulate it.

Excellent! Anybody that relies on a safety mechanism without having tested it is at best, um, "optimistic". I have  a mains plug fitted with an internal "leakage" resistor, just to check the safety devices do trip as I expect.

Please let us know the results - they will benefit us all.

[vince53]

When I drew out my scenario I could see the GFCI would not trip and all the current would flow through the sensing doughnuts.  I've stripped down my transformer/Variac setup and will be installing the GFCI on the front panel.  I'll do some testing to see if it works on setups that keep the equipment ground intact.  I'm using a US class A GFCI which is rated to trip at 4 to 6 ma.

Vince

[Shock]

When I drew out my scenario I could see the GFCI would not trip and all the current would flow through the sensing doughnuts.  I've stripped down my transformer/Variac setup and will be installing the GFCI on the front panel.  I'll do some testing to see if it works on setups that keep the equipment ground intact.  I'm using a US class A GFCI which is rated to trip at 4 to 6 ma.

Be interesting to see your results.

[Shock]