I was taking issue with your statement that the power supply was the only risk. Whacked out power supplies may be more probable, but not the only component that could cause problems. Based on the symptoms provided by the OP, it's not really possible to isolate the cause to the power supply. Assuming it is the power supply, there's no way to know what cause the PSU to fail. If the DVD drive caused the power supply to fail, then reusing the DVD drive would be a bad idea.

Power supplies are designed with over- and under-voltage cutoffs. Both conditions are harmful to electronic devices. The trip point should be set such that the abnormal voltage condition doesn't persist long enough to cause damage to the load. Of course, this assumes the power supply is reasonably well designed.

What really kills the electronics is excessive power dissipation at PN junctions internal to the device. In DC systems, power is easily calculated as current*current*resistance (I^2R). Assuming the load resistance doesn't change (and why would it?), the current delivered by the PSU will remain constant unless the output voltage changes. But, if it changes too much, as discussed above, the PSU should shut itself off.

If a load is faulted (e.g., the DVD drive) such that it draws excessive current, the power supply could be damaged because it attempts to provide current beyond its capacity. Once again, a decently designed power supply will detect this and shut itself off or faults will be cleared by fuses or circuit breakers (not likely inside of a PC, but it could be done).

Now take a (4-bit) uP with a fault such that address/data output AD1 is stuck low (grounded). Internally the uP will place address 1011 in the register, but, due to the fault, what actually ends up on the AD bus will be 1001. In this case, both data and addresses will be corrupted. Hard to determine what sort of odd behavior will occur.

Because AD inputs are bidirectional, when acting as an input to the uP, this grounded AD1 pin will cause excessive current draw from the buffer in the decoder or tranceiver connected to the AD1 bus line. This may cause a failure of that register.

You mean like restarting itself over and over?

The BIOS can't check 100% functionality of the uP. It'll check basic (primarily) I/O functionality. It would take hours to boot your machine if it ran a complete set of test vectors.

Yes, it will. I also have been happy with the three Gigabyte motherboards I've bought in the past couple of years.

Well, I do believe you, and know that you are an expert on this and that I am not. But I have a few questions.

If the DVD drive happens to drain too much power, thus causing the PSU to overheat and/or shutdown, how could this damage the motherboard or other components ? I know that there's a separate power cable going directly from the PSU to each component, and two go to the motherboard, one for the processor (it's supply is separate from the mobo) and one for the board itself (that many components draw from excluding the video card if it is PCI-E), right ?

So in this case if the graphics card was PCI-E, then it really isn't possible for it to have damaged any other component than the PSU, and really neither could the DVD-RW drive as it too has it's own separate supply line. Only the processor could have. But, from what I know of PSUs, they will all shutdown if the components drain too much power from it, you will also get odd behavior right before it shuts down, probably because the components are not receiving enough voltage. I also know that the BIOS closely monitors the main voltage and especially the processor voltage and temperature. It is designed to prevent damage to components and to the board itself, and it will shut down the computer if anything like this happens (which could have been the case, from the looks of it).

So far, I think only the processor could have caused this, as a component. But my bet is on the motherboard. I've never had problems with CPUs, but I've had many problems with poorly manufactured motherboards.

What brand of motherboard was this anyway, out of curiosity ?

I was only providing examples. I didn't mean to imply that they were specific to the OP's problem.

Depends on where the fault was. If the fault was on the PCIe bus, then other components connected to the PCIe bus could be damaged. There are components on the motherboard that are connected to the PCIe bus. Thus, a fault on the graphics card could cause a failure on the motherboard.

As far as the DVD drive, if the fault damaged the PSU, then the PSU could be providing degraded voltages to other devices connected to the power rail. This might include hard drives or the motherboard.

Bear in mind that more plugs doesn't mean more power supplies. All of the +12VDC plugs are connected to the same +12VDC rail inside the power supply. Thus, any device plugged into the +12VDC rail is connected to every other device on the +12VDC rail. Same for the other voltages.

I've had a 3.5" Floppy drive go bad for no apparent reason (it was rarely if ever used) when it failed it took the entire system down and it was unable to power up.. the system couldn't POST. Through process of elimination, disconnecting anything and everything in the system, then connecting each component back one at a time until I located the culprit. I certainly never expected it to be the floppy drive, but it proves that any device that goes bad, can pull the entire PSU down.

Yes, I agree, I've had this happen too, but can this damage the motherboard or other components ?

I imagine it could damage the PSU if it were cheaply made.. Since the drive was effectively a direct short I think it just pulled down the output from the power supply so no power was being provided to anything else in the system.

The only way I could see it damaging other components is if something died/shorted and bridged the 12 volt rail to the 5v rail or a similar situation. Nothing is really impossible though.

I've pulled components out of one PC and put then in another, everything from defective CPU's, RAM, Video Cards, etc.. to test them in a known good system. All I've ever managed to do was identify the specific bad part, and have never damaged the system I was using to test with.

The best PSU failure I've had is when one of the transistors/mosfets detonated; bright flash, boom, everything. :) I didn't have to replace anything other than the PSU.