The images for this article are now posted. Thanks for your patience.

I have enabled popups and still can't see fig 1 & 2. Is there a HTML problem?

I'm not sure why the figures didn't get into the online version of the article. They appear in printed magazine okay. Anyway, this comment section doesn't allow web links to be posted, but if I spell out the link, you can go ahead and download the original document that included those figures. Point your browser to (www dot pps2 dot com slash figures) and you will be able to get the figures I reference in the paper.

Please send fig. 1 and 2

Great questions from Mr. Stankavich (and assume Mr. Miller too).
1) Regarding the instantaneous tripping issue, yes, if in fact both breakers were in the process of opening, then you would have a point-of-no-return and while the smaller breaker might open first, the upstream breaker would still trip eventually. But what I was talking about was not the breaker contacts opening, but rather the tripping mechanism within the breaker that tells the breaker when to trip. The larger breaker will have a larger, heavier tripping mechanism that must be accelerated with a force and/or distance greater than what would be required on a smaller breaker. When that smaller breaker tripping mechanism operates a few milliseconds before the big one, you add arc impedance to the circuit. And I agree about your testing comment. You couldn't guarantee that this would work in all cases. Another interesting reason that could explain why we don’t hear about selective coordination issues, at least from properly sized and set breakers, is this – the 2005 code mandates selective coordination on emergency systems. Emergency systems are fed from generators. Generators are not like utility sources in that a bolted fault on the output of a generator results in voltage collapse within a few milliseconds. For example, you will use a generator’s subtransient reactance (x’’d) to compute the peak current available, but this current decays as a result of the higher transient reactance (x’d) that appears in a few milliseconds. That looks, to the circuit at least, as if some impedance were suddenly switched into the system, just as if an arc had been added to the circuit. Interestingly, this is a well known issue, but overcurrent devices (fuses and breakers both) are only tested on sources that provide continuous, rated current. I have a personal theory that because of the ways fuses and breakers are tested versus the power systems to which they are applied, this could explain why we don’t see breaker selective coordination failure on emergency systems even though “theoretically” there should be problems.

2) Series Rated Breakers
Series rated breakers passing fault current that exceeds the rating of the downstream device will operate the upstream device first to clear the fault before the downstream device opens. This means that selective coordination is lost in series rated applications where the fault current exceeds the interrupting rating of the downstream device. Good point!

I get back to the issue that I find interesting, namely, we just don't hear of selective coordination failures in the field. It may be that nearly all faults are arcing faults and they don't generate enough current to cause problems. Even for those real bolted faults (typically happens when you have multiple cables per phase and the cables are terminated improperly), you have inserted that downstream cable impedance into the circuit and perhaps that drops the fault current below the instantaneous pickup of the upstream device.

I guess my biggest concern is to make sure engineers take a balanced approach in evaluating protection. I hear that some believe that installing a fuse will solve their selective coordination problems. We'll, the truth is that it does! However, it cause another set of problems that in my opinion should not be ignored. Look at the IEEE 1584 standard for example. 1584 provides calculations for how much arc flash energy will be produced from a fault. But look at how they calculate the fault current... by assuming zero ground impedance. I'm sure they were trying to be conservative (since zero ground impedance results in maximum fault current), but the problem is that on a current-limiting fuse-protected circuit, the arc flash hazard increases as the fault current decreases. The result is a lot of well meaning people are underestimating the danger associated with fuse protected systems. By only examining the performance at bolted fault current levels, the arc flash hazard is underestimated since after all unless you have superconductors, there is no such thing as zero impedance ground path back to the neutral of the transformer. For typical CL fuses, the incident energy increases by the square of the current reduction. That means if you cut the current in half, your incident energy would increase 4 times. This is counter-intuitive to many people. To make matters worse, while we can calculate the maximum fault current, we can't calculate the actual fault current. 1584 makes allowances for lower levels of current, but again, this is for phase-to-phase faults. Phase-to-ground faults are calculated based on zero impedance and that is not a real world reality.

Anyway, my point is that engineering is the art of compromise. The best design is one where the engineer has evaluated all the factors and made their decision. My goal in writing this article was to make sure some of these factors were recognized.

Thanks for the good questions!