For Phil Corso or anyone with specific information, There are still some PWM drives on the market without intelligent programmable/adjustable motor protection features. For these apps, in-line fuses are still typically used in cases where the motor rating is less than the drive rating. Example: drive protection 10Arms (not adjustable) and 5Arms motor. We recently had a customer with an unusual motion control application in that the machine can frequently jam. Because they desired to try and fight through potential hard jams, the following error was set very loose and a jam could occur such that the fuses were the only motor protection. The glass type 250Vac fuses did their job but would sometimes blow violently and a few times even started an electrical fire. They tried ceramic sand filled fuses and these could also blow in pieces although (per fuse mfg) the sand filling acts as an arc extinguisher and fires are much less likely. Finally, the fuse mfg agreed with my original hunch that fuses designed for 60Hz or pure DC waveforms may not be suitable for PWM waveforms where instead of 60 peaks per second (60Hz) you get 4,000 - 20,000 peaks per second of Vac X 1.4 with very high dV/dt and possibly some spiking. The fuse mfg advised switching to 600Vac class electrical power fuses versus the 250Vac class electronic fuses. We assume the customer installed the 600V fuses as we haven't heard back since (about 3 weeks). 1. Has anyone else had the same experience [or info] re. fuse voltage rating for PWM drive? When I talk to fuse app engineers about PWM waveforms they act as if I've invented some new kind of waveform etc. 2. Has anyone ever heard of fusing only two of the three motor phases? I'm trying to convince an engineer that this is a no-no. I gave him all the reasons why. We found the NEC rule that says all "un-grounded" conductors must be current protected etc. Thanks, Tom Gianni

 

In response to Tom Gianni's query: Even with a zero-ohm short, the maximum current can not be greater than that which the source can supply. Unless there is some feedback energy from the "drive" the current can't be "amplified" by "switching" devices. Supply fuses must be capable of interrupting maximum peak current. If they are connected to the DC bus then it is unlikely an AC-rated fuse can do the job. You description of the the fuse cartridge or glass tube rupture is indicative of application error. The waveform you described is voltage. The high dV/dt and switching frequency certainly affect insulation integrity, but neither influences maximum current. It sounds like the interrupting capacity of the fuse is too low. Have you tried a current-limiting type? Regarding your "peak" frequency observation: are you sure about the 4k-20k strikes per second? Are these figures the actual switching frequencies? They result in a period of 25 and 50us, respectively. What in the electronics is providing such rapid triggering? And why? BTW, while the sand in sand-filled fuses (mostly found in Europe) improves quenching characteristic, the sand is not effective as a peak current-limiting medium. Regards, Phil Corso, PE (Epsicon, Inc)

 

Further to my earlier response to Tom Gianni's query(ies): a) Please note correction. The value, 25, in the penultimate paragraph should read 25 ms. Also, what are DC Bus voltages? b) The use of two fuses (for cost reduction?) should have served as a portend of events that followed. As you discovered fuses are always required in ungrounded leads off of a supply bus. If specified per NEC, or the "Electrical" Approval Authority for the installation requires compliance with the NEC, then, yes, 3 fuses are required. Also I presume the engineer you are trying to convince is non-electrical. If I'm correct try this safety argument: If A and B phases fuses are blown or "pulled", and C is intact, then, A-B, B-C, and C-A measurements will all show zero volts. Because of semiconductor leakage currents, Phase C voltage will be present at the other two phases, hence the "zero" readings. This could seriously jeopardize safety! Regards, Phil Corso, PE (Epsicon, Inc)