Is a 1000w power supply worth it in this context?

[seon123]

12 minutes ago, Drak3 said:

Safety wise, no.

Do you know what multi rail is? Do you also think having circuit breakers in your house is unnecessary, and that it doesn't improve the safety?

Here's a thing. 

https://www.overclock.net/forum/31-power-supplies/944707-why-single-rail-not-better-than-multi-rail.html#/topics/944707

[Stefan Payne]

4 minutes ago, Drak3 said:

The gist of it: Multi/single makes zero difference if you're buying a quality PSU.

No, it makes zero differences, if nothing GOES WRONG!

Multi Rail = Breakers in your Breaker Box.

They have the same job: Limit the max. Current to reduce potential damage.

 

What happens if a Motherboard MOSFET dies and you have a high power single Rail PSU; you can see here:

https://www.overclock.net/forum/31-power-supplies/944707-why-single-rail-not-better-than-multi-rail.html

 

THAT happens if you have 1500W+ without multiple OCP Channels.

Because its usually 120-140% until it switches off, so at worst 1500*1,4 = 2100W /12 -> 175Ampere

You know how thick the cables for 175A in real life are???

 

https://en.wikipedia.org/wiki/American_wire_gauge

 

AWG3 ~ 25mm²

 

Now go into a hardware store and look for AWG3 or 25mm² Cables!
Usually the biggest ones are 10mm²...

8 hours ago, Chevy_Monsenhor said:

After settling up the deal i went to look for my new power supply, i ended up finding some really good prices on 1000w models, about US$10 more expensive as comparable (in quality) 650w and 550w models, such as:

1- Cooler Master V1000

2- Corsair RM1000X

3- EVGA Supernova 1000 G3

 

In this context, is it a better ideia picking one of the above instead of taking a smaller model for almost the same price?

No, its useless.

If you don't need it, its just totally wasted.

 

The reason for them beeing cheap is full warehouses due to Mining craze, wich led to increased production - and now they can't get rid of it.

 

The Problem is that it can be dangerous due to the high current capacity and thus you need a really really low resistance short for them to switch off.

 

for 100A on 12V we are talking about just 0,12Ohm.

And that's the minimum amount before the OCP trips, the nominal is somewhere around 140% -> 115A - or around 0,1 Ohm...

 

The results of such things you see above...

That's a shot of a Motherboard with a shorted MOSFET, PSU didn't have reasonable Over Current Protection and was a Single Rail design, so there's no reason to switch off until its under 2kW or even more.

 

 

TL;DR: I'd stick with what you really need.

Higher Wattage units cause more damage if something goes wrong...

[Stefan Payne]

17 minutes ago, Drak3 said:

Safety wise, no.

So you're saying Breakers in the breakerbox are bullshit and can be omitted, right?

 

ANd that there is no difference between max ~50A and ~120A  max...

 

[Drak3]

3 minutes ago, seon123 said:

Here's a thing. 

https://www.overclock.net/forum/31-power-supplies/944707-why-single-rail-not-better-than-multi-rail.html#/topics/944707

 

Just now, Stefan Payne said:

What happens if a Motherboard MOSFET dies and you have a high power single Rail PSU; you can see here:

https://www.overclock.net/forum/31-power-supplies/944707-why-single-rail-not-better-than-multi-rail.html

A singular forum post, of a singular event, covering a singular PSU model, from 10 years ago, that's (worthless) conjecture.

 

Conjecture, that single rail systems only can have 1 protection, that's wrong.

 

 

 

http://www.jonnyguru.com/forums/showthread.php?3990-Single-vs-Multiple-12V-rails-The-splitting-of-the-12V-rail&s=0216f2b3b1909b211e65252b63b332ef&p=37485#post37485

 

[Drak3]

10 minutes ago, Stefan Payne said:

No, its useless.

Peak efficiency of many PSUs is roughly 50% load.

 

And higher end PSUs often don't require active cooling at or below 50% load.

 

Plus, expandability.

[Stefan Payne]

57 minutes ago, Drak3 said:

 

A singular forum post, of a singular event, covering a singular PSU model, from 10 years ago, that's (worthless) conjecture.

 

Conjecture, that single rail systems only can have 1 protection, that's wrong.

 

 

 

http://www.jonnyguru.com/forums/showthread.php?3990-Single-vs-Multiple-12V-rails-The-splitting-of-the-12V-rail&s=0216f2b3b1909b211e65252b63b332ef&p=37485#post37485

 

I know all that.

But you did seem to ignore this Part:

Quote

So why do they do they split up +12V rails?? 

Safety. It's done for the same reason that there's more than one circuit breaker in your house's distribution panel. The goal is to limit the current through each wire to what that wire can carry without getting dangerously hot.

Short circuit protection only works if there's minimal to no resistance in the short (like two wires touching or a hot lead touching a ground like the chassis wall, etc.) If the short occurs on a PCB, in a motor, etc. the resistance in this circuit will typically NOT trip short circuit protection. What does happen is the short essentially creates a load. Without an OCP the load just increases and increases until the wire heats up and the insulation melts off and there's a molten pile of flaming plastic at the bottom of the chassis. This is why rails are split up and "capped off" in most power supplies; there is a safety concern.

 

54 minutes ago, Drak3 said:

Peak efficiency of many PSUs is roughly 50% load.

...here we go again...

 

No, its not.

And I don't give you 10 or even 20% tolerance.

 

Its roughly the same between 30-60%, in some units even 30-70% load and doesn't really change by even 1%.

You can look at the Reviews on TweakPC.de for example. I did many postings to show its not true.

Especially in 230VAC the talk about "Peak Efficiency" is just totally overblown.

 

Because the peak point is something like 0,2% more at a certain load. Yeah. Totally worth it...

 

Quote

And higher end PSUs often don't require active cooling at or below 50% load.

And how should that work if they use the same design, the same layout, have roughly the same efficiency at the same load??

And with the same Fan Controller obviously...

 

 

 

Its especially obvious on the EVGA/Super FLower made PSU. With the higher wattage units looking like this:

https://www.tweakpc.de/hardware/tests/netzteile/superflower_leadex_ii/s08.php

Yeah, no. Starting with 1100rpm is not desired.

 

 

Semi Fanless also isn't necessarily an advantage or how some marketing guys of a certain company once said to me, unironically:

 

We don't need no good quality fan, we have Zero fan RPM mode!

 

While on good quality, lower wattage PSU it looks like this:

https://www.tweakpc.de/hardware/tests/netzteile/bitfenix_formula_gold_450/s09.php

 

Or why would I want a Semi Fanless mode, when I can have something like this:

https://www.tweakpc.de/hardware/tests/netzteile/be_quiet_straight_power_11/s09.php

 

Or that looks pretty useless:

https://www.tweakpc.de/hardware/tests/netzteile/thermaltake_toughpower_gf1/s09.php

 

 

Quote

Plus, expandability.

what does a 1000W unit offer me that a good quality 550W PSU doesn't??

I can use my VEGA64 without Problems on some of my 550W PSU without any Problems.

 

I can see that it might make sense to go for a 750W for some reasons (such as Connectors) but 1000W??

What reason is there for a 1000W PSU in a Single CPU and GPU System??

[Stefan Payne]

16 minutes ago, Drak3 said:

Peak efficiency of many PSUs is roughly 50% load.

https://www.tweakpc.de/hardware/tests/netzteile/be_quiet_straight_power_11/s03.php

https://www.tweakpc.de/hardware/tests/netzteile/thermaltake_toughpower_gf1/s03.php

https://www.tweakpc.de/hardware/tests/netzteile/corsair_vengeance_650m_750m/s03.php

 

 

As you can see, the efficiency of those PSU already fell by around 0,5% at the 50% mark.

 

Or lets look at this:

https://www.tweakpc.de/hardware/tests/netzteile/corsair_ax1000/s03.php

 

also peak at 30%, already fell by 0,5% at the 50% mark.

 

At 230VAC the efficeincy stays around 95% or higher between 20 and 60% load.

The difference between the Peak and max in 230VAC is a mere ~2%.

 

With the Bitfenix Formula, 450W, its evenless than 2% difference between Peak point and 100% Load:

https://www.tweakpc.de/hardware/tests/netzteile/bitfenix_formula_gold_450/s03.php - its 1,7% at 230VAC Load.

Its a bit more on 115VAC though for whatever reason.

 

But really, if you want higher efficiency, get a higher efficient PSU -> Platinum instead of Gold...

[Drak3]

@Chevy_Monsenhor I'd say the 1000W is worth it, for the reasons I've listed. The cost isn't much higher.

 

Power curves on most units pin peak efficiency at or just above 50%. That's by design.

Many high end units run cool enough under 50% load to not need to run the fan, or only run it at low speeds, cutting down on noise.

You'll be able to expand significantly, should you ever get a particularly powerful (and power hungry) GPU setup, multiple storage drives, and other peripherals. These things often last 10 years.

The actual design of decent, modern PSUs do not have the disadvantages that decade old units do, disingenuous marketing, trolling, and blatant fear mongering disregarded.

 

https://www.techpowerup.com/reviews/Zotac/GeForce_RTX_2070_AMP_Extreme/36.html

https://www.gamersnexus.net/hwreviews/2827-amd-r7-1700-review-amd-competes-with-its-1800x

 

It's easy to cross over the threshold of low wattage PSUs with basic upgrades.

[Stefan Payne]

3 minutes ago, Drak3 said:

Power curves on most units pin peak efficiency at or just above 50%. That's by design.

...

Pls look at the Links I've posted.

Its UNDER 50%...

 

 

3 minutes ago, Drak3 said:

Many high end units run cool enough under 50% load to not need to run the fan, or only run it at low speeds, cutting down on noise.

...

No, they don't.

They just put a faster fan inside or use the same fan and fan controller for the same lineup.

look at this:

https://www.hartware.de/2012/11/27/be-quiet-dark-power-pro-10-650w/8/

 

550W at 100% load has a lower fan RPM than any other unit in that range...

3 minutes ago, Drak3 said:

 These things often last 10 years.

They might or might not.

ITs not recommended to use a 10 Year old design with modern components though.

 

And its also highly likely that there will be changes to the Spec that render modern PSU useless.

 

3 minutes ago, Drak3 said:

The actual design of decent, modern PSUs do not have the disadvantages that decade old units do, 

If a 10 year old PSU didn't come with multiple OCP Channels and was high wattage it acts the same as a modern unit, if something goes wrong.

 

If it doesn't, its not Single Rail Simple as that.

 

The laws of physics didn't change in the last decade or millenium, they stayed the same....

A short with 0,1 Ohm is a short with 0,1 Ohm, a Multi Rail unit detects that, singler rail but low wattage one als should detect that, higher wattage Single Rail does not and still runs without switching off.

 

Here a more modern, Single Rail unit:

 

That is due to high current from a single source without anything to prevent that!

 

 

In a server/Workstation enviroment it might be safe enough due to constant high load on the components.

When something fails then, it might detect it and switch off.

In a consumer enviroment with idles wattage aroudn 40-60W with a GPU Setup, it looks different.

Also due to the design, the high wattage units are usually not very efficient at loads under 100W either, wich is where a PC usually sits when it doesn't do anything.

 

So no, even if it is 10€ or so, it doesn't really make sense as there are some disadvantages in high wattage PSUs that must not be overlooked.

Safety is just one part of it, noise is another...

The thing is that if you don't need the wattage, there isn't necessarily an advantage.

As for modern PSU, a 400W and 850W might only differentiate in the OCP and OTP Trip points, Transformer and MOSFETs, the rest can be the same between those two (Bitfenix Formula, Whisper M for example)

[Drak3]

2 minutes ago, Stefan Payne said:

look at this:

It's disingenuous to use German sources against those that don't speak German, especially when they're not corroborated by other sources.

[Automobili3XF]

4 minutes ago, Drak3 said:

It's disingenuous to use German sources against those that don't speak German, especially when they're not corroborated by other sources.

I agree with that, i can't read any of those articles

[Stefan Payne]

13 minutes ago, Chevy_Monsenhor said:

I agree with that, i can't read any of those articles

You don't need to read any of it.

You just need to look at the diagrams/graphs/Pictures.

Whats written is not important at all. The important stuff is in the pictures.

17 minutes ago, Drak3 said:

It's disingenuous to use German sources against those that don't speak German, especially when they're not corroborated by other sources.

The point is that you are just wrong with your claim about the Efficiency.

I could also use Aris Reviews, just pick 10 units at random and it looks the same.

 

And also you don't need no German to look at that graph/Diagram, do you?

Where's the Peak in that and does it matter?!

 

And if you have questions, I'd answer them as well.

 

But you didn't even look at the sources and continue to claim things I have disproven with a couple of actual measurements.

 

 

And what other sources are you actually talking about?!

 

[Spotty]

1 hour ago, Drak3 said:

Peak efficiency of many PSUs is roughly 50% load.

 

26 minutes ago, Drak3 said:

Power curves on most units pin peak efficiency at or just above 50%. That's by design.

I don't think efficiency curves for PSUs are the way you think they are.
As an example, this is the efficiency curve of one of the PSUs mentioned in the OP (RM1000x)...


https://www.cybenetics.com/code/pdf.php?id=j21

This whole "PSUs are most efficient at 50%" thing doesn't really apply. Maybe it did 10+ years ago when the 80+ specs were outlined, but not so much today. If anything most modern platform PSUs it's closer to the 30% mark, followed by a very slight decline in efficiency. So much of a slight decline that it's not even worth considering where its peak efficiency lies as the difference between 30% and 70% may be as little as less than 1% (depending on the unit).

Efficiency is usually the worst at low load levels, so if anything buying a unnecessarily high wattage unit could make your efficiency worse when you're using the system as most people do throughout the day with just a browser tab open not really doing much. As you can see below the RM550x (2018) has higher efficiency than the RM1000x at load loads.

Spoiler

RM550x (2018) Low Load efficiency https://www.cybenetics.com/code/pdf.php?id=enO




RM1000x Low Load efficiency https://www.cybenetics.com/code/pdf.php?id=j21

Realistically if you're talking about 80+ Gold units then efficiency is good enough that it doesn't really matter where your system falls in terms of 'peak efficiency'. If you really want a more efficient PSU, you shouldn't be looking at higher wattage units, you should be looking at units with higher efficiency such as 80+ Platinum or 80+ Titanium units!
 

 

44 minutes ago, Drak3 said:

Many high end units run cool enough under 50% load to not need to run the fan, or only run it at low speeds, cutting down on noise.

It depends on the unit, but it's not uncommon for higher wattage PSUs to be louder than lower wattage models, as the higher wattage units need to use higher RPM model fans or different fan profiles in order to compensate for the extra heat generated by the higher wattage.
If you want a quiet PSU, don't buy a higher wattage PSU, buy a quiet PSU! One with a decent fan profile/controller and good quality fan that will spin at low RPM without issues.
 

[Drak3]

Just now, Chevy_Monsenhor said:

I agree with that, i can't read any of those articles

I can read enough to get the gist of it.

 

And so far, 

https://www.cybenetics.com/d/cybenetics_DWi.pdf

750M, 230V efficiency peaks right around the 50% mark

750M, 115V efficiency between 30 and 50% is similar, but the peak is still just below 50%.

 

https://www.kitguru.net/components/power-supplies/zardon/corsair-ax1600i-digital-power-supply-review/5/

AX1600i, efficiency peak between 750W and 850W (around 50%).

 

https://www.tomshardware.com/reviews/evga-450-b3-psu,5160-5.html

EVGA 450W B3

230V, efficiency peak slightly above 50%

115V. Efficiency peak at slightly below 50%

 

https://www.techpowerup.com/reviews/SuperFlower/SF-500P14FG/5.html

Peak efficiency between 200 and 250W (40-50%)

 

 

They're not getting the same graphs over and over. And 50% is either (roughly) the peak or part of a peaking plateau.

[Drak3]

1 minute ago, Spotty said:

ttps://www.cybenetics.com/code/pdf.php?id=j21

Peak for 115V is 400W, 230v is in between 400W and 500W. It's not really going against the 50% guideline.

[Stefan Payne]

https://www.tomshardware.com/reviews/asus-rog-thor-1200w-psu,5984-2.html

https://www.tomshardware.com/reviews/silverstone-sx700-g-power-supply,5969-2.html

https://www.tomshardware.com/reviews/corsair-sf750-psu,5979-2.html

https://www.tomshardware.com/reviews/silverstone-st1200-pts-1200w-psu,6026-2.html

https://www.tomshardware.com/reviews/seasonic-focus-sgx-650w-sfx-l-psu,6045-2.html

 

all show peak efficiency at around 40% load.

Yes, they are some with peak at around 50%.

 

BUT:

a) the difference in general is not that big with modern PSU, they have a rather flat Efficiency Curve
b) its not always there, its all over the place. In general: THe difference between 30-60% is not that big. Often its less than 1%

c) could we pls stop talking about the "Peak Efficiency"??? Or at least say its between 30-60%??? and stop with that 50% that only happened because 80plus tests at 20/50/100% (at the time)...

 

 

In the real world it looks far flatter and its not really important. And if you want higher efficiency, get 80plus Platinum instead.

 

 

 

An Efficiency difference of 0,2% is irrelevant!
That can also vary due to production!

So one unit peaks with 92,2% but the other unit is at 92% and the third at 92,4%. That's totally plausible.

 

Or just use this because I just had it opened:

https://www.tomshardware.com/reviews/seasonic-focus-sgx-650w-sfx-l-psu,6045-2.html

 

89%-89,7%-89,9%-89,5%-89,1%-88,5%

 

Thats from 20% up to around 70% load. Is it really worth arguing over that small difference?!
Especially when the next unit can be 0,5% different???

 

22 minutes ago, Drak3 said:

especially when they're not corroborated by other sources.

they tested the AX1000, here a comparisation with the AX850, wich looks pretty much the same, doesn't it?

https://www.tomshardware.com/reviews/corsair-ax850-psu,5986-2.html

ALso Peak at 40% load.

 

20%: 93,8%

60%: 93,1%

max. Difference: 1,8%

 

Really? That is what we are arguing about? Less than 2% between 20 and 60% load??

 

 

Can't we just agree that the Peak is usually between 30 and 60% load and that the difference in that area is quite slim??

 

A while ago I looked at the difference between Peak and lowest Efficiency between 20-80% load and 80-100% load.

The difference I found was 1,5% or less, based on Techpowerup Reviews.

It fell another percentpoint between 80 and 100% load.

 

 

So a difference of 2,5% between 100% load and the peak point at the time, based on a couple of Techpowerup Reviews I looked at...

 

[Spotty]

11 minutes ago, Drak3 said:

AX1600i, efficiency peak between 750W and 850W (around 50%).

You're kidding, right? You're really going to try to argue that the AX1600i is most efficient at 50%? That is not only incorrect, but also a completely pointless thing to try to argue! The AX1600i 80+ Titanium has one of the flattest efficiency curves there is!
 

Quote

10% = 93.3%
20% = 95%
30% = 95.7%
40% = 95.4%
50% = 95.3%
60% = 95.2%
70% = 95%
80% = 94.7%
90% = 94.4%
100% = 93.9%

It's +/- 1% efficiency almost across the board. The efficiency at different levels is going to make bugger all real world difference. The difference between 50% and 90% is 0.9% difference in efficiency.
 

Spoiler

https://www.cybenetics.com/code/pdf.php?id=Vtr

 

 

16 minutes ago, Drak3 said:

https://www.techpowerup.com/reviews/SuperFlower/SF-500P14FG/5.html

Peak efficiency between 200 and 250W (40-50%) 

You know that's almost 10 year old design, right? That review you linked to is from 2011. You're really digging deep to try and support your claims.

Different units will have different characteristics. Discussing discontinued 10 year old units is completely pointless as it has absolutely nothing to do with the units OP is asking about.
Instead, please try to help OP by discussing the units that they are asking about, or by suggesting alternatives.

[awesomegamer919]

I really have no horse in this race, but I would like to point a couple things out.

 

Intel’s new DSG 1.4.2 requires 70% efficiency at 2% load, this is *very* difficult to do on mainline units so there will be a trend towards efficiency peaks at lower loads. For example, here’s the new Corsair RM850 (not RMx/I): https://cybenetics.com/d/cybenetics_din.pdf

 

As you can see, the efficiency peaks at around 250w or so which is 30% of load. The old myth that 50% is maximum efficiency hasn’t been relevant for a while due to LLC topologies creating a much flatter efficiency curve and now is going away entirely due to the new DSG 1.4.2 requirements.

[Drak3]

1 minute ago, Spotty said:

please try to help OP by discussing the units that they are asking about

What do you call this:

 

1 hour ago, Drak3 said:

@Chevy_Monsenhor I'd say the 1000W is worth it, for the reasons I've listed. The cost isn't much higher.

 

Power curves on most units pin peak efficiency at or just above 50%. That's by design.

Many high end units run cool enough under 50% load to not need to run the fan, or only run it at low speeds, cutting down on noise.

You'll be able to expand significantly, should you ever get a particularly powerful (and power hungry) GPU setup, multiple storage drives, and other peripherals. These things often last 10 years.

The actual design of decent, modern PSUs do not have the disadvantages that decade old units do, disingenuous marketing, trolling, and blatant fear mongering disregarded.

 

https://www.techpowerup.com/reviews/Zotac/GeForce_RTX_2070_AMP_Extreme/36.html

https://www.gamersnexus.net/hwreviews/2827-amd-r7-1700-review-amd-competes-with-its-1800x

 

It's easy to cross over the threshold of low wattage PSUs with basic upgrades.

 

2 hours ago, Drak3 said:

Safety wise, no. The differences would come down to warranty, available connectors, noise, and rated efficiency.

?

 

[Automobili3XF]

3 minutes ago, Spotty said:

Instead, please try to help OP by discussing the units that they are asking about, or by suggesting alternatives.

Thanks a lot Spotty, i'm completely ignorant when it comes to the engineering side of electronics and i'm just getting more confused by the minute

[Spotty]

Just now, Chevy_Monsenhor said:

Thanks a lot Spotty, i'm completely ignorant when it comes to the engineering side of electronics and i'm just getting more confused by the minute

I think we're all taking things a little off-topic and going on rants about things that won't necessarily benefit you. Just the direction the thread has turned unfortunately...

I'll try and summarise things for you best I can...

Your system with a Ryzen 1700 + RTX 2070 will only require about 300W-350W. You would be able to run that system with a decent 450w or 550w without any issue. You have absolutely no need for a 1000W PSU.

The 1000W units you posted aren't preferred as they use what is known as a "Single Rail" design, where all of the 12V power which runs things such as the GPU(s) and CPU are dealt with on a single, powerful rail that is capable of delivering (in the case of 1000W units) around 80+ amps. Single rail doesn't matter as much on lower wattage units since you're dealing with less amperage anyway. What is preferred, especially for high wattage units, is multi rail design which uses multiple, smaller, 12V rails which may be rated for 30A each.
Multi rail PSUs will typically feature OCP (Over Current Protection) on the 12V rails and safely shut down the system if there is a problem and it starts drawing more power than it should. Since it is divided up in to multiple smaller rails, if there's a fault in something connected to one of the rails the unit will trip protections to shut down much sooner.
Single rail PSUs will trip at a much higher point as the entire power of the unit can be delivered over the 12V rail, and in the case of higher wattage single rail units such as 1000W PSUs that point may be around the 1200W (100A) or higher before it will shut down to protect itself.

If you're interested in the difference between multi-rail and single-rail I'd recommend checking out this video in which Jon Gerow (@jonnyGURU) who is the director of PSU engineering at Corsair who explains the differences between the single rail vs multi rail. He obviously does a much better job of explaining it than I did.

So, instead of choosing something like the the RM1000x which is a single rail design, the HX1000 or RM1000i which uses a multi rail design would be preferred *if* you actually needed 1000W, which for your system you do not.

In the case of the EVGA G3 1000W you were considering, during a reviewers test of its protection features the unit being tested failed before OPP kicked in to protect the unit - which is why some users noted earlier that the G3 has issues with its protections.

If you are concerned about efficiency then we can discuss some more efficient units you could look at, though the units you have posted are 80+ Gold rated which is generally speaking pretty good, and really good enough for most people.
Likewise if you're concerned about noise then we can give some recommendations for some quiet units.

As mentioned earlier, the most likely reason that the 1000W units are similarly priced to the lower wattage units might be because the store purchased a large quantity of the higher wattage units during the bitcoin mining craze expecting to sell a lot of them, and then the price of crypto currencies crashed and now they're left with a bunch of high wattage PSUs that no-one is interested in buying that they have to try and clear.

For your use case and your system, buying a higher wattage unit will give you absolutely no benefit, so you may as well keep the $20 or so difference in your pocket and get the 550W version instead. Use it to buy yourself a nice lunch or something instead.

As far as actual recommendations of units go, for your system with a Ryzen 1700 + RTX2070 something like the RM550x or Bitfenix Whisper would be good choices to consider and you should check what local prices are like.

If you have any other questions let us know.

[awesomegamer919]

17 minutes ago, Chevy_Monsenhor said:

Thanks a lot Spotty, i'm completely ignorant when it comes to the engineering side of electronics and i'm just getting more confused by the minute

 

Let me give you a quick rundown of Multi-rail vs single rail, both have their advantages and disadvantages, though they tend to do more good than not nowadays.

 

Single rail is fairly straight forward, there is a chip inside the PSUs that monitors how much power is being supplied at once over all rails, if a certain threshold is reached (on a 1000w it’ll be around 1250-1400w) then the PSU automatically shuts down. A few single rail PSUs have a dedicated pin for the 12V which is a little more strict than the general overpower protection, but it’s uncommon.

 

Multi-rail, despite the name, still only has 1 physical 12V rail (there used to be true dual rail units but they haven’t been around in over half a decade), what differentiates it from single rail is that the chip that monitors the output power/current is more advanced and can measure multiple output wires at once, this means it can set much stricter over current protection levels (usually 25-40A instead of the 100A or so on a single rail PSU). This obviously means that it can detect overdraw on a single cable much quicker, potentially stopping a cable from melting, but in the past badly designed “rails” has caused issues with GPUs and CPUs being stuck on the same rail and causing overcurrent protection trips even under normal load (Most, but not all, Multi rail PSUs will have 2-4 rails split between the PCIe cables, CPU cable, ATX cable and SATA/MOLEX cables).

 

This issue has mostly disappeared in recent times, though AMD Vega cards may cause issue with a few PSUs such as 550w Bitfenix Formula or, to a lesser and more user fixable extent, the Bitfenix Whisper 550M. Some units, such as Corsairs -I series (RMi, HXi, AXi) have adjustable multi-rail so you can give rails different amounts of max power before they shit down, or switch to single-rail completely.

 

Overall, Multi-Rail is safer, but slightly more restrictive on GPU choice, single rail is less safe but works well with any GPU and configurable Multi-rail is the best of both worlds.

[Automobili3XF]

i appreciate all replies but specially @Spotty and @awesomegamer919 for clearing up the way and leading me to a satisfying conclusion.

I'll follow the advice and buy a lower powered, high quality unit such as the ones mentioned here.

About 80 Plus Platinum models, near my price range i can only find this Raidmax model, apparently built by Andyson, no idea if its good or not.

[Spotty]

2 minutes ago, Chevy_Monsenhor said:

About 80 Plus Platinum models, near my price range i can only find this Raidmax model, apparently built by Andyson, no idea if its good or not.

Not too familiar with it to be honest, I don't think Raidmax are available in Australia - or at least I've never seen one.

There's some trustworthy reviews of the unit here. (Don't have time to read through them atm, might give it a look a little later)
https://www.tomshardware.com/reviews/raidmax-monster-rx-700at-titanium-power-supply,4530.html

https://www.anandtech.com/show/10461/the-raidmax-rx700at-700w-80plus-titanium-psu-power-supply-review

 

Looking at the website you posted, the Corsair HX750/HX750i are the same price.
That unit is $855 Brazillian, or about $215 USD? That's a lot to spend on a PSU for your system. I don't think you will need to be spending that much on a PSU for your hardware.

[Stefan Payne]

32 minutes ago, awesomegamer919 said:

This issue has mostly disappeared in recent times, though AMD Vega cards may cause issue with a few PSUs such as 550w Bitfenix Formula or, to a lesser and more user fixable extent, the Bitfenix Whisper 550M. Some units, such as Corsairs -I series (RMi, HXi, AXi) have adjustable multi-rail so you can give rails different amounts of max power before they shit down, or switch to single-rail completely.

I can only talk about the HXi because I have tested it.

And it switches off with VEGA.

If you are brave enough to use just one cable with the GPU. With 2 Cables I didn't have any issues under any circumstance ever.

Quote

Overall, Multi-Rail is safer, but slightly more restrictive on GPU choice, single rail is less safe but works well with any GPU and configurable Multi-rail is the best of both worlds.

Actually, with modern High Transiet GPUs it can be the other way arounnd as you can have far higher multi rail OCP Current, if you combine multiple rails than you have on a Single Rail unit.

Just take a 650W Unit.

For it to be useful, you need something around 700W or so - something like 60A.

With Multi Rail, you can have far lower Current per rail but combine multiple rail to one.

So even with as low as 35A you have 70A and with 40A you have 80A.

Though that kinda goes back to the original Problem of Single Rail..

 

12 hours ago, Chevy_Monsenhor said:

comparable (in quality) 650w and 550w models, such as:

1- Cooler Master V1000

2- Corsair RM1000X

3- EVGA Supernova 1000 G3

Now the real evaluation:

1. very old Design, from 2012 or so, before the agressive power saving mechanism of Kepler and Haswell. You see units based on this design relatively often with random reboots. Personally I'd pass on that
2. Single Rail, otherwise somewhat OKish. Best of the three. Though, as @Spotty said, your system is around 300W or so. So totally wasted. Also the new RMx is around 160mm while the 1000W is 180mm. that's something to keep in mind
3. Very high OPP points at around 1500W, loud as hell as soon as the fan starts. Nothing really of interest.

The thing is: If you don't need 1kW you don't have any advantage but have to deal with the disadvantages. ANd that is higher damage if something goes wrong (Higher OCP/OPP Trip points but also more energy in that system as there is something called the hold up time wich has to deliver 17ms of power. ANd that power has to go somewhere -> through the Short)...

But most importantly: 1000W Units tend to be on the louder side.

I've linked to a TweakPC Article about the fan RPM to a similar unit, wich, as soon as the fan starts, let it rotate at 1100RPM .

and that's a 140mm one. So that's what people tell "at full RPM"...

 

1 hour ago, Chevy_Monsenhor said:

Thanks a lot Spotty, i'm completely ignorant when it comes to the engineering side of electronics and i'm just getting more confused by the minute

Single Rail <-> Multi Rail is similar as the Breaker Box in your house.

You have something coming in (usually 35-63A, here in Germany usually 3 Phases). 

That is basically the Power Supply.

 

That goes then to the Breaker Box, wich breaks the one big thingy into many smaller things and trips once an overload is detected. The Breaker is in German (translated) a "cable protection switch" (Leitungsschutzschalter). And does exactly that: Protect the wire connected.

With Multi Rail, its similar to the Breaker but a difference is there: It switches off the PSU completely and not just the one rail.

 

Basically you see it here:

 

The Yellow Wires are +12V.

They are soldered to some points. 

These solderpoints are are connected via a Resistor (per Rail).and are not connected on the other side either.

The big area on top of that, with the silver stripes is the +12V Source, where the three Resistors are connected to.

Then you can see the small traces coming from under the Resistors, those measure the voltage and if they measure a big enough difference, they switch off the PSU...

 

The chip to supervise that is on the PCB you can see in the lower left corner.

 

 

And that's basically what Multi Rail does:
YOu have one big(ger) source and split it in some smaller portions.

In this case its 3x with exactly 20A

 

In case of the 1000W units, you have 83A according to spec but switching off at somewhere between 120 and 150%.(so around 100-124,5Ampere.

That's a lot!

Multi Rail (a mediocre 4 Rail) would split that in something around 4x40A.

Though more Rails are better.

 

 

And lets not talk about weird as shit Multi Rail designs that the Engineers totally messed up...

 

The bad name of Multi Rail came because PSU used the EPS12V Spec: 2 Rails of 4 for CPU, one PCIe, one Drives/ATX.

And then GTX280 came wich switched off those PSU...

And people didn't understand what was happening and blamed the PSU...