MIKE ROSS wrote in a message to Wayne Chirnside:

 WC> Anyway I'm not a fan of stressing solid state components by
 WC> continually exceeding their ratings, leads to reliablity issues.

 MR> You should appreciate that semiconductor ratings are usually meant
 MR> for steady continuous duty at a conservative level. However, "time"
 MR> is a factor that is not often given in the specs 

It often is for rectifier diodes and SCRs...

 MR> and "time" is very important when dealing with very brief pulses. 
 MR> Semiconductors can and do withstand really huge current surges if 
 MR> the timing is kept very brief. 

Looking in this handy Motorola Master Selection Guide under rectifiers, 
there's some parts that specify I(sub)O of 6 amps.  They also specify
I(sub)FSM of 75 amps.  So your estimate of 40-70A for a surge rating
_would_ fall within the device's specifications.

I scrapped out an old linear power supply board that was originally made by
IBM,  who you'd think wouldn't tend to skimp on their components.  They had
a crowbar circuit on there with an SCR in a TO-220 package that did _not_
use a heatsink.  That's pretty typical from what I've seen.

---