Yeah, it can also be because of overheating. Forgot that procs will trip a safety shutdown when overheated..

3A vsb would be quite hard to find. I'm not sure if its because of the vsb, if the psu can turn on while the board is connected then its not likely to be because of the vsb. I've started a PC with a PSU with sufficient vsb (only 0.2A ) by jump starting the psu while connected to the board.

It depends. If the power supply is truly reliable, then even at its limit it will still mantain good voltage & ripple regulation even for long periods of time. The CM Realpower 550w that I use can handle a full 100% load for 24 hours with no problem, and even a 650W load for short durations. Is your Enhance 460w the one from 3431? Seem to be hearing that the voltage regulation is a bit weak on that PSU at the limit. But at least it doesn't blow up like those cap ayam/akucomel/power tak logik..

-extra
Some reviews of some screwed up cheap @$$ power supplies, just something for a bit of amusement reading. Especially like the comments on the 2nd one.
http://www.jonnyguru.com/review_details.php?id=134
http://www.jonnyguru.com/review_details.php?id=129

At stock it boots properly but only need to press F1 right (no hanging, no need to reset)? If that the case then it -might- be the PSU. But even then I suspect something else, possibly a motherboard voltage regulator not functioning as well as it should. If the voltage regulators are to blame, then you will have this problem using most processors on this board except for the lowest power ones.

Single rails are better than mutli railed PSUs for most purposes because you can't accidentally place an unbalanced load across the rails. But multi rail PSUs have their advantages too, a true multi rail PSU (which has completely seperate transformers and regulators for each rail) isolates voltage flutuations & EMI on each rail from other rails. Effectively two seperate PSUs in one, but designed to work together if necessary. Unfortunately, there are very few PSUs (even from highly reputable manufacturers) using a true multi rail design, most cheat and share regulators, and just place a current limiter on each rail, which has none of the advantages of a true multi rail. Hardwaresecrets does a good job of examining PSU circuits to determine how good a PSU is, though they lack the equipment to stress test the PSUs.

I'm really stumped now..by right the multi rail should be giving the problem, not the single rail. There is NO increase in actual amperes going to the motherboard regardless of which PSU you use, the number on the PSU is just the absolute max. In reality it will only supply what is asked for at that particular time.

I'm really starting to think this is not a PSU problem..probably the mobo..

Its normal, the monitoring chip that reads the Vcore is nowhere close to super accurate. Sometimes you can even get some really riduculous readings which are obviously wrong.

Not necessarily. AM2 and C2D systems are trickier to overclock than previous systems, hitting high FSB or HTT often requires tweaking a lot of settings. C2D systems are well known to have FSB 'holes' due to chipset strap settings and AM2 systems are notoriously sensitive to memory settings.

Well that question should be asked elsewhere..but I'll answer it anyway.
For an overclocked system, the lower the multiplier, fsb and temp, the more stable the system is. More voltage to the processor core will stabilise it more, but if there's too much voltage and it overheats, it will become unstable, perhaps even be damaged; heat + extra voltage is a very bad combination for a processor. More voltage means cooling must be better. Keep in mind processor voltage is not controlled by the PSU, but by the motherboard.

XS forum looks for nothing but the very best..all of them are out of our reach, either financially or availability..
I tell you, if those guys heard what PSUs we're using here,

The Gigabyte is a good PSU, but if you're buying it because of the sensor, you will be disappointed. The software to read the sensor is messed up, takes up a lot of resources, and hangs sometimes. All the PSUs you mentioned are good, with the only exception being the Enermax. In fact, most of them are already excessive for your system, and will easily last you 5 years or more.

Wah, sniper69..you got a lot of biased opinions there..
But you are still entitled to speak yours

I stand corrected

I guess you just have your own cable management style. Some people prefer modular so you can get rid of unused cables. Voltage loss depends on design, a good connector design will not cause much problems. Actually come to think of it, the only wires which are voltage fluctuation sensitive are the main ATX connector, which is usually not modular. The 12v lines going into the CPU and PCI-E are only feeding voltage regulators, which are are there to step down the voltage and re-regulate them anyway.

I have to agree with you about single rails though..too many PSU manufacturers are cheating by using a single 12v regulator and transformer, and then splitting out the 'rails' using current limiters. There is a certain advantage to genuine multi-rail PSUs, but I've only heard of a rare handful, mostly of 800W units and above. Those use completely seperate transformers and regulators, and the best of the lot can even withstand a heavy crossload or a accidental rail linking. Its like having 2 PSUs in one, and both are designed to work together. Which was the whole idea when Intel came out with the specification in the first place, but pretty much nobody followed.

There is only one rail design better than that..but it doesn't exist for desktop PCs.. One FAT (20-30A) rail each for every connector, with seperate transformers and voltage regulators, with allowance to combine rails anytime.. But the cost and the size of that PSU will also be FAT

You know what? If such a PSU existed, you could connect all the rails together and use it to do arc welding..it would be so power
But it doesn't exist, except for gigantic servers which use true modular PSUs..can combine 2 or more PSUs if not enough power..

You mean like best 'bang' to your pocket..

450W is more than enough for any 478 system. I own one of the highest end 478 systems, and it only takes up less than 300W with ram(DDR500 2.5-3-3-8), graphics (X1950, 640/1580) and processor (3.7Ghz) on full overclock, everything on full load.

Well it depends, by right active PFC PSUs are supposed to be able to tolerate quite a wide range of input voltages (100-260V). Also good PSUs have decent hold up time, and can even tolerate a (very) short interruption of power.

AVRs do not improve much over APFC PSUs, but its nice having a cheap filter to take the bullet when something really bad happens (like a really strong power surge, or minor lightning strike). Best if you lose RM50-70 on the AVR than RM200++ on the PSU and perhaps even more on a fried system.

That was before the days of active PFC PSUs, which can accept some pretty wide variations in input voltage. Older PSUs with passive PFC or none at all could only accept input voltage in a narrow range, usually 200-240V or 100-120v only

It would still be able to run as it would not take more than 250-300W, but problem is that enermax has a history of high random failures. Can't figure out why, I've heard of enermax failing on various PCs, each with different power requirements, so its not because of overloading. Could be a QC problem.

Its actually not a problem if you distribute the load evenly across the rails. Anyway, most of the PSUs are actually single rail PSUs, the label is lying. For example, an Enermax Liberty 500W I opened has both 'rails' connected together-how on earth can it be dual rail? The explanation is simple-PSUs manufacturers have been lying so that they can keep up to the latest ATX spec (which is several years old btw) which states that no rail in a PSU is allowed to output more than 240W (more or less, I didn't read the spec back to back) which results in a maximum of 20A for the 12V line.

The spec was so unsucessful that eventually Intel gave up on it (unofficially) and PSU manufacturers have started to manufacture single rail PSUs in more recent years.

The 12V line powers the CPU and often the GPU as well. CPUs typically take 60-130W, and GPUs can take 30-150W, which results in a current of 7.5-23.3A on the 12V line. Most PCs are towards the lower end of the range, so 18A is enough. The high end here is assuming a quad core and a sinlge 2900XT, no overclocking. No other graphic card takes anywhere that much power, and dual cores take significantly less than quad cores, so most pc won't even touch 23.3A.

There is no sure way to tell if a PSU is pure power or not without using a load tester. Reviews by tomshardware, jonnyguru and silentpcreview are done using load testers so you can check their reviews. I think xbit labs does too, but I haven't had time to check.

It should be OK, I've yet to see an Acbel fail a load test. PFC is not important unless you're using a UPS. APFC does have a little side bonus though, its more tolerant of voltage fluctuations.

Yeah, the read headache is finding out what rail powers each connector. Sometimes its not even mentioned in the manual.

APFC is not important unless you're using a UPS or your area's AC power fluctuates (which in that case, you should be getting a UPS anyway). A side bonus is that APFC psu usually have better power efficiency as well (do not confuse with power factor). Otherwise, APFC is not important, just count it as a bonus feature.

For multirailed PSU, keep in mind that most of the PSUs below 600W are actually single rail PSUs with 30A+, but the wires are labeled as seperate rail to keep with the stupid ATX 2.xx specification.

Simple reason is this-a true multi rail PSU usually uses seperate transformers for each rail (which are pretty big & expensive). Even if the manufacturer just shares one transformer (still ok, but less EMI protection), it can still really be considered multi rail if there are seperate voltage regulators for each rail. But the sad truth is that is simply a LOT cheaper to use fewer but better components to design a single rail PSU capable of outputing FAT amps, and just relabel wires to say its multi rail.

Generally speaking, its not likely that you'll find a true dual rail PSU for less than RM300, simply because it doesn't make economic sense for the manufacturer (its like selling Core 2 Extremes at less than RM600).

Personally, I don't trust voltage readings given by software since the monitoring chips of motherboards are not to say terribly accurate to begin with. Sure, they're good enough to warn you when the PSU is going south and voltages are flying all over the place, but if you want to compare between decent PSUs to see which is the best you're not going to get a consistent answer. To check how good is a PSU's voltages, a digital multimeter is needed.

ESR is equivalent series resistance, which is a capacitor spec, not a PSU spec. What you're talking about is ripple voltage, which are small, very fast fluctuations which can only be detected by using an osciloscope.

Actually, those fake 'multi-rail PSUs' are generally safer than single rail PSUs if the current limiters are properly done-the result will be that you will not be able to overload the PSU; doing so will cause the safety trip to go off. And both current limited 'multi-rail' PSUs and single rail PSUs have one weakness in common-if there is a massive fluctuation due to a rapidly changing load, all the lines will be affected. True multi-rail PSUs have each rail isolated from each other, so fluctuations on one line will not affect the others. Single rail PSUs have only one real advantage over multi-rail (whether true multi rail or current limited), that is you can load the single rail all the way to the limit before it trips. On multi-rail PSUs, overloading even a single rail will trip the PSU. But all said, loading a PSU to the limit is never a good idea, it greatly shortens the life of the PSU.

The PCP&C Silencer IS pretty quiet at low load, because very little heat is produced the fan doesn't need to spin much. But at heavy loads, then you're right-it really needs to be 'silenced' by a 120mm fan.

Its a good PSU, like OC4/3 said. The review by jonnyguru.com explains it in full. The only downside is maybe the software bundled with it, which as I've been told this problem has been fixed.

I don't recommend this at all, the voltage regulation is ok, but the ripple voltage at high load is off the chart. Under overclocked conditions, your hardware (especially mobo voltage regulators) could give way even though the voltage readings appears ok (its only obvious on a osciloscope). For details, read the review at jonnyguru, scroll though the list till you find it.

One glance and I know your PC doesn't even take 500W, much less 650W. Its probably around 450W only (you're not doing SLI right?)

Overclocked C2Q -200W max
8800 ultra superclocked -150W max
rest of system - 50-100W

The max values are as I said-max. If they actually took up that much power, even water cooling won't cool it down to comfortable temps (it'll be around 50-60C). Aircooling would be close to useless, even high performance 3rd party coolers will be running close to limits (70C+).