This is basically off-topic for stompboxes, but the principles would apply to power amps based on PCBs.

I've got an output stage (non-audio) that will be a rugged NPN tranny, probably an MJE3055 (or 2N3055), working as a common-emitter amplifier, with a 12V supply, driving a 1.5 ohm load.   Fortunately, the duty cycle is a couple of milliseconds at about 1 Hz.  But the peak current is still likely to be about 8 amps.  

Will the small amount of heat generated while "on" for 5mS be easily dissipated while "off" for 995mS?    Or will my PCB tracks (about 0.1") go to hell? :cry:

thanks for any info, including guesses.

PS this thing is to drive an auto ignition coil, which is to energise an electric fence

It does, of course, depend as much on the thickness and length of the tracks as their width - but I have seen 0.1" tracks take 1A continuous current, so with this duty cycle, there shoudn't be a problem. However, you'll have to consider the integrity of the actual soldered joints - they'll have a higher resistance than the tracks themselves, and will probably be the weak point. Cleanliness and soldering quality will be paramount.

Consider the current ratings of ordinary copper wire, and compare the cross-sectional area with that of your tracks.

Quote from: brettWill the small amount of heat generated while "on" for 5mS be easily dissipated while "off" for 995mS

No problem averaging over second. (my guess). But here is the official word (half way down):
http://www.epanorama.net/documents/wiring/wire_resistance.html

i like to smear a huge amount of tin (solder) on such tracks (and be it only to lower their resistance...)