I'm sure this has been addressed, but I could not find a thread.

Is there an expert consensus on fuses versus circuit breakers in a context of arc flash protection? I've heard fuse manufacturers state that fuses are best and CB manufacturers state that CB's are best. I lean toward fuses because the underlying assumption from CB manufacturers is that there will be maintenance and care on the CB's, which in my estimation, is universally lacking.

In a recent article I read, the author made a good arguement for CB's (vs fuses) in data centers by putting forth data that ground faults are 2.5-70X more likely than phase-to-phase faults (citing IEEE 493-2007, Table 10-32), and that ground faults typically flow through non-copper conductors, lowering the fault current. The author states: "Under these lower current conditions, the very steep time-overcurrent curve typical of a current limiting fuse works against it."

The author then goes on to outline the subsystems available on today's modern CB's that will allow actuation at lower fault current levels, which I agree are effective, but I remain concerned about maintenance.

I do not believe there is not a single 'always use' answer.

The last two devices I recommended were CB's because the calculated arcing fault was below the current limiting range of 350A RK1 fuses.

For the most part if the available bolted fault current is not greater than 10x the fuse size, I tend to go towards CB's. For circuits below 100A, it doesn't seem to matter.

Great question. Since my first job was for a breaker company, I should be waiving the breaker flag but being independent, there are merits of both.

Fuses

Pro

• Generally higher interrupting rating
• Most are current limiting
• Easier to coordinate by using selectivity tables that take into account I2T energies

Con

• Most faults involve only one phase - This means only one fuse interrupts causing single phasing - additional protection may be required
• Need replaced instead of reset
• Keeping them in stock i.e. don’t want a 500 kCM conductor as a replacement.

Breakers

Pro

• Interrupt all three phases (assuming 3 pole)
• Can be easily reset (after finding out why it tripped )
• No live parts to contend with if dead front equipment

Con

• Not always able to have 100% coordination
• Breakers are not usually current limiting but they are available
• Interrupting ratings are not always as high as fuses

The decision really depends on preference. A strong argument can be made for or against each type of device. Properly sized current limiting protection can reduce the incident energy but if the arcing current is low, the device may take a long time to operate and result in more problems.

Feel free to add to the list!

Fuses

Con

• Hard to prevent replacement with incorrect fuse. (different vendors usually negates 'table based' selectivity, RK5 in place of RK1 can impact incident energy)

Breakers

Con

• Difficult to guarantee adjustable settings are not 'magically' re-adjusted.

Great point! Back in my breaker days, tested series ratings with fuses and breakers were just developing (yep - a while ago A big concern was the difference in let thru energy between the RK5 and RK1 fuses. UL finally standardized on testing with RK5 only (more let thru) That way if someone uses RK1 or RK5 it would be OK.

However, I still wonder when the RK1 vs. RK5 issue will be addressed with arc flash. Right now it is up to the facility to make sure they keep the right fuses in the switch. I guess there is only so much you can do. Better than a piece of pipe or wire which I have seen on several occasions.

If there are any RK5 fuses in the 'tool crib', my suggestion is to use RK5's in my arc flash calculations. If the results are unsatisfactory, then I suggest additional labeling on the switch requiring RK1 only fuses.

As far as the UL series ratings, they employ a specific 'umbrella fuse' for the fuse-breaker combination testing. All fuses actually manufactured (since the 80's) must perform better than than these worst-case umbrella fuses.

And the Umbrella fuses have I2T and Ipeak let thru values much greater than the production fuses. As of a few weeks ago (from an IEEE 1584 subcommittee meeting that I was at) the next edition will likely have new fuse equations based on specific fuse data rather than umbrella values.

I have no problem using exact information for those locations with strong 70E maintenance practices.

However, I like the idea of a 'worst case' fuse as an option.
I can't remember how many times I have opened a switch and found 3 different fuse vendors, much less different fuse sizes.

There is an incorrect statement in this thread. There are current limiting circuit breakers available. They tend to be around the 400 A category in 480 V configurations.

Two points that haven't been made. For the instantaneous fault condition, fuses trip in 1/4 cycle. At higher end voltages/currents, the fastest breakers out there trip in 2-3 cycles, AFTER a 1/4 to 1 cycle delay (depending on the trip unit) triggers it. The reverse tends to be true in the current limiting mode...circuit breakers with electronic trip units can generally be set tighter than fuses. And this is the area where the manufacturers love to argue the merits of their systems.

Second point: chapter 2 (maintenance) is effectively nonexistant For fuses. There are no false negatives (fuse fails to trip on demand), and the maintenance requirements are minimal. Even if maintenance gets ignored, the fuse will still trip on demand. If you follow NFPA 70B or NETA MTS to the letter the worst you have to do is annual inspections for bad contact and/or cleaning the contacts between the fuse and fuse holder (and the compartment in general) on an infrequent basis.

One of the most interesting things out there is the S&C electronic power fuse. This is a fuse link with an expulsion type fuse (with a muffler), with a separate compact relay and a capacitive trip unit on the side. When the trip relay triggers the unit, the capacitor puts a high current on the fuse and melts the fuse link. Very nice because you get the accuracy and control of a circuit breaker with the higher ratings and speed of a fuse.

What they haven't done is to make the trip unit programmable and/or allow it to trip all 3 fuses, or otherwise make it a truly universal solution. If someone else were to copy the basic concept, it could be possible to have the best of both worlds.

Finally there are also self-resetting fuses. These contain a small amount of polymer that melts on overcurrent just like a fuse but then resolidifies once the current goes away. The downside is that these are only rated for up to 1-2 amps so they are used in some electronics including self-protected IO from Allen Bradley but haven't seen much use elsewhere yet because of the inherent limitations. So nothing yet big enough to support a power level in the "arc flash" range.

"...Breakers are not usually current limiting but they are available"

I mentioned C/L breakers in an earlier post.

Sounds like we are all pretty much in agreement, the decision depends on a lot of things and there are many "pros and cons" depending on the design, specific conditions, equipment etc. etc. etc. That's why I usually sit on the fence with this debate, the answer usually ends being "it depends".

Thanks Paul for all your great additional info and comments.

I've watched some of the advancements that various manufacturers have been making over the years and am usually impressed with how they keep moving the ball forward.

Not sure if the writer is discussing LV or MV CB/fuses, or both. However, from the perspective of LV circuit breakers, generally curves are stated (drawn) conservatively and include all sensing, logic, mechanical operation and arc clearing time. For the brand I am most familiar that time is 3 cycles under 600V bolted fault conditions using a slower instantaneous algorithm than is possible, to attain certain desirable selectivity related capabilities. When testing the device and associated trip under 480V arcing conditions clearing times were ~ 2/3 of that shown on published curves. When using a simpler, faster and less selective algorithm, clearing time is less. This covers Low Voltage Power Circuit breakers in the 2000A range. Of coursem this can vary by manufacturer, specific device, age and condition of the device.

Molded case CB with or without electronic trips typically have some point at which the contacts part due to internal mechanical forces, even if they are not listed as current limiting. 20X or so of the frame rating is a probably a good guess for this. Devices designed for current limitation will limit let-through energy by their current limiting threshold quite effectively because of this popping action, which is improved by design in a CL device, even if the NET total interrupting time is greater than a 1/4 cycle. Keep in mind CB timing is always stated for all three phases whereas fuse clearing times are stated for single device operating on a single phase fault at a defined Powe Factor. In CBs the stated clearing time is the worst possible at a range of conditions up to maximum current, maximum power factor and worse closing angle for all three phases... or should be!

Nevertheless, as far as current limitation if the fault is expected to be well above a fuse's Current Limiting threshold it is doubtful anything could be faster. As fuses get larger, however, that is harder to achieve.

As far as maintenance the CB definitely needs some... however so does the switch the fuse is probably inserted in. In my experience fuse and switch maintenance should not be ignored as I have seen statistics that AF incidents are, unluckily, associated with the care and feeding of both types of OCPDs. Access to proper spares, type, brand and size is probaly a consideration.

There is no single best solution... the most powerful solution is probably a combination of actions that best fit the characteristics of the system and how the system will be used and maintained.

I've seen outdoor medium voltage power fuses fail to open under fault conditions. The chemical that helps snuff the arc can swell with moisture, preventing the element from pulling apart. The tube will glow white hot while it continues to carry fault current. Maintenance provides the solution; air flow testing to detect the condition.

I "believe" that the current 1584 fuse equations are based on a single manufacturer's fuses, and not on umbrella fuses. I am not sure what would drive different performance from fuse to fuse. I would "suspect" that degree of current limitation is less important than the value of the current limiting threshold of the particular fuses under 3 phase arcing fault conditions. I also believe that ohter manufacturers have stated that the current equations in IEEE 1584 are fairly representative of the current indsutry offering. Regardless, I would suggest it is probably best practice to consult with the individual manufacturer one is considering and then stick to the brand, model and size of fuse originally selected and analyzed.