NickWhen my Dad worked for GE in Auburn, NY, he used to cut the prototype blanks for the sintered ceramic discs that made up the active elements of the Tranquell MOV stacks. I have one that was originally intended for a big railroad locomotive connected across the main breaker panel of my house. It's been working for 15 years, through many lightning storms, without an electronic failure in the entire house. It runs through its own private 100 amp double pole breaker. It stays slightly warm to the touch in normal operation. I'd bet it dissipates 5-10 watts because of leakage current, but my Dad said that this was normal behavior. It's enclosed in a unvented 12" x 12" x 4" PVC utility box adjacent to the breaker panel. The electrical inspector didn't seem to have any problem with that arrangement.
I got my unit new, but if you recycle previously used ones, they can be a bit of a gamble. Each time they absorb a transient, their threshold increases a bit. The bigger the bump, the more the threshold increases. Testing them was always a problem for GE, because every test changed the characteristics of the unit. Eventually, they become less effective as their characteristics approach a permanent open circuit.
The short circuit failure mode is much less common, but potentially more spectacular. However, the Bakelite disc and copper plate construction of the unit is pretty robust, and it's supposed to be able to contain the rated dissipated energy as long as you have a breaker or fuse to prevent continuous current passage after the transient is over.
I also had a pair of small MOVs that came attached to a 5 amp Corcom L-R-C EMI filter that I installed to protect my home security system power supply. They were about the size of a 6 amp axial lead rectifier. One of them failed shorted after a few months of operation, and the resulting transient burned out about 6 Leviton solid state computerized light dimmers on that leg of the house wiring, before tripping the 20 amp panel breaker. I spent a long and frustrating day trying to figure out where the short was in that branch circuit, a job made more difficult by the fact that the alarm system was hard wired into the panel without a plug/outlet. The MOV didn't explode or change visibly in appearance. When I found the unit at fault, I removed both of them permanently with a pair of diagonal cutters.
The MOV elements are just flat discs of compressed shiny black granular material that resembles a grinding wheel more than anything else.
Dave On 3/15/2010 7:06 PM, Nick Andrews wrote:That would be one of my main concerns. I think closed in a proper box, the fire issue is small (albeit still cause for concern), but a closed short could be problematic. It would be interesting to look inside one of the GE whole-house units to see how they are constructed.
Nick ADate: Mon, 15 Mar 2010 18:17:04 -0400 From: Dave@xxxxxxxxxxxxxxxx To: tesla@xxxxxxxxxx Subject: Re: [TCML] Idle musings about surge suppressors Nick, Be aware that MOVs often fail shorted. Be sure to put them on the outboard side of a big breaker. The breaker won't trip fast enough to limit the surge suppression capability of the MOV, but it will prevent a big problem if a MOV fails in shorted mode. I've had this happen on my own house. Dave On 3/15/2010 2:44 PM, Steve Ward wrote:It might be worth considering what happens to these things when they DO have to clamp a surge. Is your construction material going to catch fire from the heat generated by the MOV? Steve On Mon, Mar 15, 2010 at 12:26 PM, Reverend Fuzzy <cmayeux@xxxxxxxxxxxxxxxxx>wrote:Most of them are just that... MOV's When I found that out, I built my own with 3 heavy-duty MOV's on a piece of perfboard, then covered the whole board in a chunk of heat-shrink tubing... wound up with a module as big around as a small matchbox, with one red, one black, and one green wire coming out of the end... total cost, under $10. (black to phase A, red to phase B, green to neutral/ground bus bar.)
_______________________________________________ Tesla mailing list Tesla@xxxxxxxxxx http://www.pupman.com/mailman/listinfo/tesla