Quote:
Originally Posted by
flexinator 
I am a bit confused.
Well lets get that out of the way first. Speaker ratings are generally how much power it can absorb
continuously over some period of time without damage. So if I put a 50Hz sine wave at 350W into your driver, it should be able to take that for say 2 hours with no damage. Music and movies SFX is
not continuous and averaged out over several minutes, at best is about 1/3 of putting in a sine or pink noise. Short term over powering of a driver (delivering more power than it's rated to take) seldom does damage, unless the signal causes the driver the approach or exceed Xmech where physical damage to the suspension might occur.
How much power the amp delivers to the driver depends on the signal. A 1000W amp does not automatically deliver 1000W to the driver. It means it is
capable of doing so,
if you gave it enough signal input to do so. Give it 1/10th and it will deliver 100W to the driver, 1/20th and it will give the driver only 50W etc. I have a test mule amp I use in my electronics workshop rated at 400W/ch and it has driven a pair of vintage Celestions rated at 25W for years with no damage because I never turn it up much.
Quote:
Originally Posted by
flexinator 
that amp is rated as 2 x 750W into 4ohms. My subs are rated at 350W max rms. Wouldn't that be way too much? Are all these amps rated as max power not rms? How do you figure out the correct amount?
Max/peak power specs for amps are a bullsh** spec. Ignore them.
If you had clicked on the link you would have seen an actual EP1500 measured bridged at 800W max, or max 400W/driver.
Here is the same amp, 2ch mode driving a 4ohm load, measured, not marketeering specs. I suggested that model because it and it's larger brother are used here a lot with good results, and it's probably the cheapest way to get that much power.
Generally if possible pick an amp rated greater than that of your speakers: 1.5 - 2x is my rule of thumb depending on the cost to allow for short term transient signals to be reproduced cleanly.