12V1 14a AND 12V2 13a Amps?

[Algere]

Originally posted by: HardWarrior

Originally posted by: Algere

Not every 2 rail PSU follows that method of power distribution.

Having two 12v rails, in general, means 1 rail to power the mb though.

Yep :thumbsup:

Originally posted by: Algere
Count only one for components powered by one 12V rail*.

*You'll have to find out which 12V rail powers what component(s). May wanna contact PSU maker on that one.

 

[HardWarrior]

Originally posted by: Algere
Count only one for components powered by one 12V rail*.

*You'll have to find out which 12V rail powers what component(s). May wanna contact PSU maker on that one.

[/quote]

In my case it's painfully obvious. There's only one wire loom with an ATX power connector and it has no molex's.

 

[Mark R]

That's one of a few reasons I've heard for the use of multi-rails.

Other reasons or added benefits I've read/heard about multi-rails.

1) Cheaper to make. Would a single rail 700W PSU cost more than a 700W multi-rail PSU???
2) Easier to implement. That or there's the challenge PSU makers would face in attempts to regulate high amperage on a single rail. I suppose reasons 1 & 2 intertwine with one another.
3) Cooler operation, assumingly more important now than ever before. PSUs act as a secondary exhaust for a PC plus add to the fact that the higher the wattage the PSU is outputting, the more heat is produced. Both factor against a PSUs ability to provide full power & as some know, more heat = less power.
4) Prevents CPU power load (and/or other?) from straining the rails. In this case as load increases, voltage goes down. Now my guess is that voltage on the other rail doesn't go down/follow as well.
5) Cleaner power, or at the very least dirty power is isolated to it's own rail.
6) Anything else I might've missed .

That's basically right. The main reason has already been given - and that is safety.

The more power that you have available on one circuit, the more damaging a fault on that circuit could be, and the less sensitive you can make your protection systems.

It probably also is cheaper to make - when you start to go above 20-25A your electronic components start to get more expensive, quickly. Each component will also get hotter. Using 2 rails allow you to use 2 seperate circuits built from relatively low cost parts, and spreads the heat out out at the same time.

You also get a little bit of seperation between the CPU power draw and the rest of the system. It's unlikely to be complete, as I'd be very surprised to see a seperate regulator for each 12V line.

Although I haven't examined a dual-rail supply, I'd guess they work the same as older ATX supplies - all the rails* (5 and 12) come off a single regulator (the average voltage is regulated - if you load down the 5V heavily, you'll find the 12V goes up in response). Just in this case, there are 2 12V rails - load one down hard, and you'll probably find the voltage on the other rises in response.

Note * - you can include the 3.3V here as well - the 3.3V usually comes from stepping down the already regulated 5V supply.

 

[VirtualLarry]

Originally posted by: Algere
If there's filtering on board, why is it important for PSU makers to list ripple/noise regulation specs? If there's filtering on board why would you need ripple/noise regulation on PSUs & since OC'ing moves your system out of factory specs, is this on board filtering good enough for OC'd systems?

The way that I look at it, being a bit of a gross analogy since I'm not an EE - think of an optical scientific microscope (most of us had to use one of these in HS). It has a "coarse" focus adjustment knob, and a "fine" focus adjustment knob. The PSU's voltage-regulator stages are the "coarse" regulation, whereas the individual regulator circuits on the mobo and your modern video card are the "fine" regulation.

Does that help make better sense now?

Part of the reasoning is due to physical proximity between the on-mobo VRM ciruits and the loads that they are powering, and part of it has to do with the physical size of the components needed to do the regulation, and the physical space needed between them for proper (and safe) isolation. Sure, you could pipe normal 120V AC right onto the mobo, and then regulate that down directly to the 2.5v that the DDR DRAMs needed, for example, but that would require components nearly as large as are in the PSU itself normally, soldered to the mobo, and they would have to be spaced out from the other components on the mobo for safety. It just wouldn't work out.

Originally posted by: Algere
Other reasons or added benefits I've read/heard about multi-rails.
1) Cheaper to make. Would a single rail 700W PSU cost more than a 700W multi-rail PSU???
2) Easier to implement. That or there's the challenge PSU makers would face in attempts to regulate high amperage on a single rail. I suppose reasons 1 & 2 intertwine with one another.
3) Cooler operation, assumingly more important now than ever before. PSUs act as a secondary exhaust for a PC plus add to the fact that the higher the wattage the PSU is outputting, the more heat is produced. Both factor against a PSUs ability to provide full power & as some know, more heat = less power.
4) Prevents CPU power load (and/or other?) from straining the rails. In this case as load increases, voltage goes down. Now my guess is that voltage on the other rail doesn't go down/follow as well.
5) Cleaner power, or at the very least dirty power is isolated to it's own rail.
6) Anything else I might've missed :frown:.

Those are all essentially right-on. It's a question of when you reach a limit (either actual, or practical, or cost-related) to what you can achieve with a singular component, and the current level needed on the +12V line right now has hit that effective limit, so they've started to provide two rails instead to do the work instead, and split the load in half. It's much like CPUs - both Intel and AMD have had great difficulting pushing single CPU cores any faster, so in the near future, they will be giving us two cores (perhaps individually slightly-slower) instead.

As I've gained a greater understanding of PSUs, I don't see anything inherently wrong with these new so-called "dual rail" supplies, although I wish that the mfgs were more clear on how independent those rails actually are. "Dual independent rails" will have their own transformers and regulators, whereas "dual shared rails" (my terms for them), will share a transformer, but still have their own regulators. (AFAIK) The maximum load that you can draw from any one rail, is the maximum that either the regulator or the transformer will allow, whichever amount is lower. The maximum cumulative load that you can draw as a combined total from all of the split rails, depends on whether they are shared or independent, and if shared, that likely depends on the max load/current capacity of the shared portion, which is the transformer, which may well be lower than the added maximums of the each rail. The only thing to keep in mind is not exceeding the total max load overall, and don't overload any one of the rails with an individual load that exceeds that rail's rating. That's all. Otherwise, dual-rail should be fine for most things.

I think that the real reason for the introduction of dual rails (shared), is to provide better upstream voltage-regulation (less "noisy") for the mobo's CPU VRM circuits. Noisy power there, can lead to CPU and thus system instability overall, so it's important that the +12V power loads used, for example, by your powerful video card, won't negatively affect the stability of your CPU as well. All "high amp single-rail" PSUs will put you at risk of this being an issue, depending more-or-less on the quality and capability of the mobo's CPU VRM circuits for dealing with incoming "noisy" power. Cheaper boards (MSI and ECS comes to mind here - sorry Peter) will have more issues with this than more expensive boards that go all-out on the CPU VRMs, like Gigabyte's six-phase or AOpen's four-phase power circuits. OTOH, MSI only uses two-phase. ECS uses three-phase, but usually comprising of minimum-spec components. If you remember the stories told about the K7S5A's "need" for high-powered or high-quality PSUs, this is the reason why - less tolerance for noise in the mobo's CPU VRM circuits, requiring better-quality regulation of that power upstream - in the PSU.

 

[Algere]

Originally posted by: VirtualLarry
The way that I look at it, being a bit of a gross analogy since I'm not an EE - think of an optical scientific microscope (most of us had to use one of these in HS). It has a "coarse" focus adjustment knob, and a "fine" focus adjustment knob. The PSU's voltage-regulator stages are the "coarse" regulation, whereas the individual regulator circuits on the mobo and your modern video card are the "fine" regulation.

Does that help make better sense now?

From what I gather, you're saying that the PSU limits ripple/noise within a small +/- (0 = no ripple/noise) range & once power arrives to the motherboard, ripple/noise is then regulated to a even smaller +/- range by the motherboard? If so, would that mean if a PSU had bad ripple/noise regulation, the motherboard's regulators might focus into the wrong area? Say more into the + range (within the +/- range) for instance? And where does the "CPUs are noisy components/cause noise" come into this? Since that's what I've heard remembering as one of the reasons of multi-rails & why PSU (ATX) makers like Antec decided to isolate the CPU from the rest of the system instead of splitting the +12V load evenly (systemwide).

Assuming CPUs do cause noise, the CPU is at the end of electricity's travel (based on your limited explanation?). How does the motherboard's regulators help there? Is there return power from the CPU back into the PSU? If there is & in that return, is it possible that's why they say multi-rail isolation has it's benefits? -- dirty return power affecting other components (incoming power?). I assume from my limited understanding that's a possibility & could explain the reason for PFC -- prevent dirty power from going back to the electric company.

 

[dheffer]

I'm looking at an enermax that provides 14V to one rail and 13V to another - how much power is this really?
Is it enough for an athlon 64 754, geforce 6800 ultra, 1gb ram, 1 hard drive, etc?

 

[dheffer]

anyone know? do the dual rails equate to a total of 27V or what? I know for what I have, I was told 20V would be enough with a quality power supply, so will this work?

 

[ssvegeta1010]

No, you do not add them up. Go for a 20V+ single rail PSU.

 

[HardWarrior]

Originally posted by: dheffer
anyone know? do the dual rails equate to a total of 27V or what? I know for what I have, I was told 20V would be enough with a quality power supply, so will this work?

It depends on the PSU and how 12v accessories you have. I think my 2x12v rails add up to 35 amps, which is the way you should look at it.

 

[dheffer]

All newer revisions of power supplies are dual rail, like the new atx spec specifies.
Does anyone has this power supply, or know if I can run this well?
Thanks.

 

[HardWarrior]

Which PSU are you speaking of?

 

[dheffer]

enermax 370 watt, i can get one for really really cheap so itd be great if it worked
i think my biggest constraint on that is the 6800gt, but i know that thing only requires 380 watts, so I mean, the enermax is a quality power supply, think it could handle it?

 

[HardWarrior]

The cheapass in me says yes, but depending on how loaded your box is it may run pretty hot. What sort of deal are you getting on it?

 

[sman83]

I started this thread about getting a PSU and I decided on the Enermax 370 the one U are talking about above and I am NOT DISAPPOINTED at all with my decison it runs my:

Asus P4P800d-e Deluxe
P43.0ghz@3.3ghz
1gb DDR 3200
120GB HD sata
120GB HD ide
SB Audigy 3
ATI Radeon 9550 250/200@400/200 256 MB DDR
DVD ROM
DVD RW
2x80mm fans
1x120mm fan
Floppy Drive

And it runs PERFECT

 

[dheffer]

Well actually I was going to buy one from my friend, but this is only a few dollars cheaper - http://www.newegg.com/app/viewproductde...TC-Froogle&description=N82E16817103444
46.99 shipped is quite a good price, possibly a hot deal? haha

 

[HardWarrior]

I'd say that was a fine deal, dheffer! :thumbsup:

 

[dheffer]

Yeah, its a great deal, im just not so sure that its going to be enough power

 

[DSYee]

Does the 6800GT come with its own power connector (i.e. directly from the board to the PSU)? If so, then I think it draws power from both rails (from the motherboard via the PCI-E slot, and from the cable plugged directly into the PSU).

 

[dheffer]

Yeah it does, I would just love to know how much amerage it drew because that other rail is just for the video card, hard drives, optical drives, and i only have one sata drive, and one dvd burner... nothing too demanding

 

[Algere]

Originally posted by: ssvegeta1010
No, you do not add them up. Go for a 20V+ single rail PSU.

It's addable, see third post (top to down) on this thread.

Originally posted by: DSYee
Does the 6800GT come with its own power connector (i.e. directly from the board to the PSU)? If so, then I think it draws power from both rails (from the motherboard via the PCI-E slot, and from the cable plugged directly into the PSU).

IIRC from the manual, the 370W PSU in question has one rail dedicated to the CPU and all else on the other rail, higher powered models have it as you've said - one rail via mobo, one rail via cable. Therefore [dheffer] since the s754 A64s uses around 7A on one rail, you'd have...

7A used on one rail by the CPU + 13-14A available to all else. Giving you 20-21A usable in total. IMO you'd have enough power (Enermax 370W) for the following:

athlon 64 754, geforce 6800 ultra, 1gb ram, 1 hard drive

From personal experience I've powered the same setup minus the 6800 Ultra (Radeon 9800 instead) with a cheap 13A Raidmax - all stable.