Could the mounting surface be warped? Maybe when the spring clip is bolted down with the 4 bolts the holder is being distorted causing the side pressure on the fuse to be less than it should be. Is the long through bolt tightened to spec when the fuse is installed for a tight fit? Just a thought.

Mark

I have considerable experiences with fuses. There is the wire coming in, the wire termination, the clip, possibly a clip hinge, the clip friction connection to the fuse and the fuse element connection if of the renewable type.

Most problems are at the wire termination or the clip friction connection. Rarely the clip hinge. Replacing the fuse doesn't replace a bad fuse, it is the reseating that corrects the problem.

When current is flowing, the "same potential to ground" is not exactly the "same potential". The IR drop across a bad joint creates heat, whether arcing takes place or not. Remember, 400A is welding current! Arcing to ground usually leaves tracks and only occurs in an exposed environment, worst when a dirty exposed environment.

Tom Mason, Ralph Tyler Companies, formerly BP Coal Technical Liason

I know this is late and I don't want to beat a dead horse into the ground more, but I used to work for Bussmann in the prototype department.

The clip is definately the problem, when they heat up enough they "relax" and then make a poorer contact, resulting in greater heat. Usually this will blow the fuse as it draws more current.
But in this application, with the high voltage arcing can and does occur, and the fuses are rated pretty high.
I worked in the DC side not the AC side of the prototype department. But I tested fuses up to 3000 amps (6VDC)and automotive fuses at up to 48VDC and about 2000 amps.
But the big misnomer about fuses, especially when I first got there is "if a fuse is rated at 100A it will open when it see's 100A". No they are "Rated" at that, but don't open until at least 135% of rating. So a 100A fuse can carry 135 amps for about a minute. It will have to open in less than 5 seconds at 200A. And then the opening times rapidly decrease at 500% and 1000% into the microseconds. And of course when they open they are supposed to not catch anthing on fire, etc. UL tests with cheese cloth over fuses to see if it burns when they open. If it does it's a failure. Yada yada yada. Sorry, I still find fuses interesting and the testing of high voltage fuses scarry. One of the demonstrations we vaporize a 18" piece of 10 guage wire with something like 1 million amps and 600V to demonstrate how our fuses protect circuits. Actually I think it's only 300,000 amps, but who's counting when it gets that high?

Duane,
That's one of the best descriptions for this failure mode that I've seen. I've seen the relaxing of the fuse clips and other pressure/friction connections several times over my career.

Thanks for the contribution. I don't think its ever too late to share good troubleshooting experiences.

One experience I remember was on a chiller system where they employed separable two piece terminal blocks on the control wiring. The terminal blocks relied on a pressure connection between the two halves of the block. With current flow, the connection would (intermittently) relax and open up. By the time you got there to troubleshoot it, the connection would cool down and make again. It took awhile to find that one. Pretty hard on 4160v chillers to shut down like that.

J-

Thanks for the compliment. I wanted to explain a little. First, about the misnomer, "I" was under the impression if a fuse saw it's rated current it would open then. Second, at 135% I think the fuse "has to open with in 1 hour". I remember 200% being 5 seconds. So there is a very steep curve for a fuse to carry it's rating yet open fast over 200% of it's rating. Also my experience is mostly lab related. (Actuall ALL lab related). I remember some fuses getting so hot carrying 135% rated current that the insulation on the wires melted off before the fuse opened. That was on the EV fuse for the GM electric vehicle years ago. The fuses in those were 250A ceramic fuses from what I remember.
Well that is totally irrelevant to this conversation, just wanted to illistrate the difference between lab and field testing.

I think it sounds very relevant to this conversation.

It would help the field greatly if Thermographers were exposed to conversations like this. I put a lot of study into fuse curves and gained field experience on the same.

Ultimately I don't think the field and lab aren't that different as far as fuses are concerned, except that the field is more unpredictable, and no matter how you try to emulate what might happen in the field, you never can in the lab.

Great thread here. Thanks to all.

Metallurical data suggests annealing (or "softening" as it has been termed here) begins to be significant when the component reaches 200F for approximatenly a month. At that point the clip is no longer doing it's job and heating increases.

It is vital to remember that as heating increases, so does resistance...and, of course, as resistance increases so does heating and the component quickly hits a "runaway" condition.

How long? Difficult to predict! The only time it gets cooler is when load drops or it welds itself together. This is one of the major reasons I refuse to base prioritization of findings on temperature alone.