Interesting. I thought MOVs were still the only game in town, despite their problems. Something new to read about ;(
EDIT - Hmmm. I won't call total BS on the BrickWall products because there is some validity to what they are saying, but IMO there is a bit of BS mixed in with the good stuff.
They make a big stink about ground current with MOV based systems, implying that there is some strong technical reason why MOV protectors have to dump surge current to ground as well as neutral. As far as I know, MOVs are connected from hot to ground as well as hot to neutral for protection in the case where part of the failure involves neutral being disconnected, or for "common mode" surges where there is a spike on both hot and neutral.
Their point that a typical MOV protector will dump surge current equally into neutral and ground on a Hot-to-Neutral surge is only true if the two MOVs have the same characteristics, which AFAIK is true on the cheapest protectors but not on the better ones (the threshold on the MOV to ground is higher so it doesn't fire unless (a) neutral is floating or (b) the neutral MOV has already sacrificed itself to the surge.
I have to admit, though, that from BrickWall's graphs it does appear that the competing products they tested used the same MOV characteristics on both ground and neutral, or (more likely) that the test surge was so large that both MOVs kicked in at almost the same time anyways. It would be interesting to see the same graphs with a smaller, more typical surge which is where you care the most about ground current.
Moving past the BS, there is some good stuff in the BrickWall products as well.
A lot of the BrickWall protection comes from a big series inductor, which seems like it would help on a spike but not on the very common "oops, the 5kv wire touched the 110v wire" scenarios. If you look at their test graphs they are all running with a very short surge (50 microseconds) which the inductor can absorb easily. With a longer surge (more typical of a transformer failure) there is a much greater chance of something being sacrificed. BW talks about the IEEE guidelines but I thought there was also a time element as well as voltage/current. Maybe someone else can check; I have to vacuum ;(
The rest of their protection looks pretty good (active, SCR-driven current dumps instead of MOVs) but I'm hoping they have oversimplified the drawings they provided. SCRs only conduct in one direction so presumably each current dump in the diagram is actually a pair of cross-connected SCRs and associated circuitry.
Anyways, the interesting thing about BrickWall is that there are two forms of protection -- one operating exactly like an MOV but presumably more rugged and another (the series LCR circuitry) which AFAIK could be used just as easily with MOVs.
Actually, the blurb for Tripp-Lite ISOBAR surge protectors claims that they have big honkin' inductors as well. Not sure if they are just part of a noise filter or are big enough to perform the same kind of function that the inductor in the BrickWall provides. BrickWall certainly seems to think not
