Just like the hundreds of memory issues, the PSU issues continue to be suspects in a lot of the technical queries related to motherboardsand Graphics cards. It has actually become a very confusing market for the general user ever since high power PSUs became an issue due to the inreased power demands of the "high" end home PC, and when I say "high power" PSU, I mean more than 500W. Doesn't sound that much but when one talks about half a kilowatt, only then does one realise what little power stations our everyday computers need. So the manufacturers started the novel idea of seperate 12V rails for stability and to prevent spiking and surges. And this is where the normal customer who actually has limited knowledge, (hopefully) looks at the amps and sees e.g. 41 Amps or 500W total on 12 V. The fine print though will clearly show that possibly 18A per rail is the max, meaning that utilising 2 of those rails limits the PSU to 36A for the GPU alone if two 12V rails are used. Thus the user has 4 12V rails, each capable of 18A individually, but as soon as he maxes two rails, in reality he has 5Amps left for the other 2 rails.
A similar single rail PSU would probably provide 50Amp or more on a single 12V output. Yes its all shared, but at least the "Juice" doesn't run out as early.
Admitted, they both have their advantages and disadvantages, but when we get to power delivery, the one advantage of the total amps available on 12V in my opinion outweighs the others. This does not go without saying that the customer should still do his homework in terms of stability and efficiency.
I sometimes wonder how many system instabilities are really caused by a lack of power or even unstable power delivery. Yes, we all expect those wonderful power circuits on the boards to manage a stable power delivery, but we seem to forget that any component/circuit, even on the best boards have their input limits. Hopefully the system will shutdown when the input is outside the limits, but how may times do we expect these circuits to protect the rest of the components, before they themselves have had enough? Even worse, I wonder how many boards have been RMA'd because of damage done to them due to power supplies that are just not up to scratch, and the manufacturers assume responsibility for component failures caused by the PSU without having any proof.
I always read with interest the hardware reviews with regards to PSUs and especially those where the tester actually ensures that the unit is tested to its claimed maximums to establish the real values and stabilities. It is actually frightning how many of the so called "better" supplies fail or are just inside the margins. One has to feel sorry for people that actually buy pre-assembled units which are built into a case that already contain a "No Name" 250W or 350W PSU.I still remember my first Intel machine that I bought in similar fashion, and later I upgraded and found that the 12V rail would actually fluctuate from 11,2 up to 12,3V.
It's amazing how much money people are prepared to fork out for high end boards, CPUs and GPUs and then save a couple of dollars on the PSU. It is not dissimilar to someone buying a super car, e.g. Ferrari or Lamborghini and then filling it up with regular unleaded.
My only wish is that people would pay more attention to buying the correct driving force behind the high tech components that they envisage to power with whatever PSU they choose. But then again, I suppose most forums are reactionary instead of being a place where ordinary people would enquire before buying and mixing and matching.