I had a buddy 2 years ago go through 4 drives in 1 year just like you - He swapped out his PSU and has never had a fail since - there was NOTHING measurably wrong with his psu's VOLTAGES - but clearly changing it fixed the problem.
The reason here is far more technical than voltages - see PSU's don't actually make nice clean voltages - they use rapid spikes of high voltages at various frequencies in a process known as (PWM) pulse width modulation - that allow the PSU to dynamically "generate" the power you need and ONLY the power you need and then adjust it as your needs change.
These spikes are then smoothed out by capacitors to make them LOOK like a constant output voltage. This is all very simplified, and its a lot more complex than that -but the point is that what it outputs can change DRASTICALLY as your load changes - and its possible that under an EXACTLY specific set of conditions some form of pulse resonance or other form of not directly observable interference can be happening - especially if the PSU has one or two "slightly faulty, but not faulty enough to initially show any signs", parts.
Pulse width resonance - a potential result of all sorts of component failure in a high frequency switching power supply can have all sorts of bizarre and unpredictable effects - and definitely cause the loss of hardware attached.
Swing and a miss. I don't exactly know where you studied power supplies, but you have a few fundamental errors.
The PSU (modern ones) uses a high frequency signal that they generate from the initial outlet power. Switching on and off at certain points in the cycle to only allow certain voltages through. This voltage is rectified (AC to DC), then smoothed by a capacitor array or linear regulator. PWM signals are used to switch the system during the initial harvesting of the high frequency power, but are not the source of it.
After being smoothed, the power is 100% DC. There is a ripple voltage, but it is so small that it is well within the acceptable tolerances of the device it is connected to.
Where power supplies randomly kill things is not the ripple, but high frequency noise on the lines due to inductance. This noise is not detectable to most DMMs because of response times, but it does damage electronics.
This is why voltage readings are misleading as far as a PSU's health.
Basic electronics concepts aside, inductance is dang near impossible to predict. Minor variations in construction can produce very different amounts of induction, which is why some cheap PSUs are great and other are crap. Large suppliers occasionally develop this issue, but generally do enough testing to minimize its influence on produced goods.
Replacing a PSU is definitely the most likely solution, given nothing else seemed to be dying. Motherboards generally have surge suppression hardware, which insulates them from high frequency noise. I would bet that anything connected to that power rail gets fried, whether it be optical drive or HDD.
Ok, it's faulty, but at least you got no worries about getting it replaced. You haven't made the mistake that many many others have...
I'll keep the thread subbed.
