Arc Flash Mystery

[Kiljoy]

OK, this is a strange one. We build industrial machinery which we test in our shop via temporary pigtails to a fused bus system. Our QC guy was reaching onto a 480V panel to read voltage across the distribution block. All of a sudden, there was an arc flash. The distribution block he was reaching for didn’t have a mark, but the fuse block for the hydraulic tank heater about a foot to the left had it’s fuses melted on the tops only. They did not clear. His multi-meter probes were also melted. The main 400A molded case breaker disconnect did not trip, nor did the buss-plug fuses. I took apart the meter and power obviously went through it. Our guy swears he has the meter on voltage, was no where near the melted block, and was reaching behind some temporary wires to read voltage. The temporary run is SO cord from a VFD to a motor we use for testing VFDs.

Here’s the only thing I can think of. Either our very qualified QC guy is lying, or there was some sort of harmonic arc between the VFD leads and his meter. He swears that he was no where near the melted fuse block and that he hadn’t even touched the probes to the distribution block yet. He was wearing a nice metal watch and a ring (yes, not the smartest thing), but other than a slight UV burn from the arc, not a scratch on him.

 

[Skogsgurra]

It would be very easy to inspect the instrument to see what the setting was when the high current occured. If he will not let you do that, you must assume he is not only less qualified, but also lying.

Gunnar Englund

http://www.gke.org

--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

 

[GTstartup]

I know of a case where dislodged dust cause an arc flash, not everything is as it seems.

 

[Kiljoy]

He says the probes were in the voltage terminals. By the time I got the meter, the leads were removed. The

 

[GTstartup]

After rereadng your post I would ask him to explain how his meter was damaged when he had not yet connected the probe to the distribution block. That statement alone puts everything else he says in doubt.

 

[Skogsgurra]

Open it then. If there was a current heavy enough to melt the probes, then it must also show inside.

Dust? Yes, perhaps. But not very likely in a test setup.

Reaching to take a measurement in high power equipment is a very bad idea. Easy to slip and cause shorts. That would cause exactly the damage you describe.

I am so tired of listening to 'mystical' explanations - often with assumed harmonics and resonances - from guys that have made a mistake. Much better to train not to make those mistakes. And never accept those fairy tales. Such things do not happen when lead or cable lengths are short (meaning 1 - 100 m).

Gunnar Englund

http://www.gke.org

--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

 

[Skogsgurra]

OK, see that you already opened it. Current from A terminal or V terminal?


Gunnar Englund

http://www.gke.org

--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

 

[Kiljoy]

I was trying to figure that out and I'm not sure. Negetive terminal trace is fried, but the A and V terminals don't looked damaged. It looks like it arced to the back of the case where the little foil beeper is. The meter was given to him by our maintenance man. It was an old one he had lying in a drawer.

Let's assume our guy is telling the truth. He's an old navy radio operator and has been HAM certified for 30+ years. He used to work for us as a service tech years ago. Could a faulty meter be the culprate? I'll upload some pictures in a few.

 

[Kiljoy]

Here are the pics. There are temporary wires between the two distribution blocks that will be removed.

 

[Kiljoy]

That was the meter. Here is the panel.

 

[Kiljoy]

The fuses...

 

[Skogsgurra]

Must be the meter then.

Navy, ham certificate, decades of experience, still alive.

I would trust such a guy.

But - he never touched anything? That's where the problem starts. And that's where my doubts creep in.

Do you calibrate your meters regularly? You can usually also have meters checked for debris and overall condition if you have them calibrated.

Gunnar Englund

http://www.gke.org

--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

 

[Kiljoy]

That's why we are so confused. The meter was an old junker. I suspected the leads were in the amperage terminals, but he swears not. The amp side is unfused. I just can't see 480V jumping thorugh free air like that. We even called in an outside electrician and he was stumped. Said it looked like something shorted across the top of the fuse block.

 

[stevenal]

Old junkers don't have cat ratings, and can be damaged by surges. Might want to replace with a cat III or IV.

 

[Kiljoy]

Took a closer look at the meter. My suspicion is that he tried to read voltage when the probes were plugged in for amperage. Being a non-fused meter exacerbated the problem.

 

[Compositepro]

Where was the meter positioned? Perhaps the meter was hanging so the screws on the back of the meter shorted the fuse block It looks like there is tracking between the screw holes. If the probe tips were dangling, they may have been grounded somewhere on the bottom of the panel, and that caused the damage to the probe tips.

 

[Kiljoy]

That I'm not sure we will ever know for sure. I know that people's memories tend to get foggy in situations like this. Where he was and where he thought he was may be two different things.

 

[oldfieldguy]

480 volts won't "jump". However, once an arc is started it will span considerable distances until the source gives up.

I'd lean towards "compositepro"'s idea, or the "Probes in the current jacks" idea. The tips obviously saw arcing and that is not usually the result of a voltage measurement.



old field guy

 

[Compositepro]

Check the spacing of the screws against the parts in the fuse block. Look for arc marking on the screw ends on the meter that would have been facing the fuse block. Look for arc markings on the bottom of the panel and elsewhere. There is a lot of physical evidence you have available. How do you troubleshoot any technical problem? You have to make knowledgeable hypotheses and then find evidence to support or eliminate the hypothesis.

 

[Kiljoy]

There were no exposed screws. The meter was in a protective rubber boot.

I just talked to the guy. He thinks the probes may have touched the block as he was holding them in his left hand. He still swears the leads were in the Voltage position.

After the incident, the meter still worked. I tested a 120V souce with the meter. It jumped around a bit, but was still working. Now that's a tough meter. I think someone may have switched the leads around when he wasn't using it.