Will it explode?

[Stefan Payne]

Re: Will it explode?

Stefan Payne mentioned bad quality/design. What is probability of explosion under low load?

That's not neccissarily the problem.
The Problem is more something like a screw comes loos and does its fair share of carnage.

And with every unit that has two 8pin PCIe connectors on one 8pin connector on the PSUside, chance of burning that connectors with higher end cards (or two on one cable) are there.

And as mentioned eariler, shitty build quality can lead to premature failure or other problems.
If your PC sometimes randomly freezes or reboots, you must not forget the PSU.

CWT (the OEM of these) is known for shoddy QA.

Well, no...

CWT is know for doing what YOU want. If yo want high build quality, they will do it. They might charge extra for it, but they will do it.
And THAT seems to be the problem:
Some manufacturers don't want better build quality or pay extra for it...

But you have to have some contacts in the industry to know about this stuff

And that's one of the things that makes CWT one of the desired companys.

Seasonic on the other hand is something some companys don't want to talk about...
They are just the worst of the worst in terms of contract manufacturing...

And that's also the reason some companys won't ever work with Seasonic ever again!!
Even though there was a time where many companys worked with them...

RMx is one of the best series on market ATM, what the heck is all this talk about?

No it's not!
It's more like a glorified mid range unit that's a bit overpriced...

And only server/industrial grade units deserved to be called 'one of the best'. Sadly those things are rather rare these days...

And also RMx lacks a bit of protection depending on the chip used there is no UVP on +12V, OCP isn't there either. That is useful because OCP acts way faster than OPP. So it might be the difference between components bites the dust or not...

Either way, it's NOT "one of the best on the market", it's just another random/standard PSU with some ups and also many downs. Nothing THAT special.

[Behemot]

Re: Will it explode?

RMx is one of the best series on market ATM, what the heck is all this talk about?

[dmill89]

Re: Will it explode?

some new CPU like Intel Kaby Lake/Coffee Lake or AMD Zen APU
Stefan Payne mentioned bad quality/design. What is probability of explosion under low load?

Virtually any PSU will run that, you should be fine. There is nothing inherently wrong with the design of the RM series, but CWT (the OEM of these) is known for shoddy QA. Though with that it doesn't matter what load you put on it, if there is a loose screw rolling around in the case or bad solder connection it will likely either be DOA or fail early in the warranty period (and while RMAs can be a pain, that's what the warranty is for after all), and unlike the cheap off-brand Chinese specials these have enough protection circuitry any damage should be limited to the PSU itself in the event of failure.

[Behemot]

Re: Will it explode?

RMx has DC-DC converters so it does mot matter what do you hook to it.

[gc4e7t2u]

Re: Will it explode?

What CPU?

some new CPU like Intel Kaby Lake/Coffee Lake or AMD Zen APU
Stefan Payne mentioned bad quality/design. What is probability of explosion under low load?

[dmill89]

Re: Will it explode?

I bought RM550X which isn't connected yet before reading this thread.
Is it safe to use it in PC with 65W non-overclocked CPU and integrated GPU which means peak load won't exceed 80W?

What CPU? If it is a relatively recent (Pentium 4/ Athlon 64 or newer, i.e. just about anything made in the last 12 years or so), the low(ish) 5V/3.3V combined rating will be a non-issue since newer CPUs draw their power from the 12V rail (which is why most newer PSUs have very robust 12V rails and comparatively weak 5V/3.3V rails).

[gc4e7t2u]

Re: Will it explode?

Overall build quality on those Corsair units is really bad, sometimes there is threadlocker on the screws for the modular PCB, sometimes there is not. They are semi fanless (the S12G is not), use just a rifle bearing fan...
And last but not least:
They use those stupid and almost dangerous PCIe-Y Cables with 2 8pin connectors per 8pin connector (8+6 pin would be barely OK by the way).

Just because the voltage regulation is a bit better and riple and noise is a bit better doesn't really mean that the unit overall is better at all...

Oh and those Corsair units also have a whole lot of solid caps on the modular PCB as well...


Well, yeah, they are a couple of years old, so one would think that seasonic did a completely new plattform in the same (price) range...

I bought RM550X which isn't connected yet before reading this thread.
Is it safe to use it in PC with 65W non-overclocked CPU and integrated GPU which means peak load won't exceed 80W?

[dmill89]

Re: Will it explode?

Is there any other choice but to get either an older PSU and recap it or get a PSU with DC-DC?

^This, and if you don't want to recap, older workstation/server grade PSUs which often have Japanese caps can be had pretty cheap (even NOS units), just about anything that only has a 20-pin ATX connector sells for next to nothing these days (since any remotely modern system needs a 24 pin).

If you have to have new go with a DC-DC converter PSU that drives all the minor rails off the 12V so cross-loading isn't an issue.

[Stefan Payne]

Re: Will it explode?

Thanks Momoka!

So if you had to power an ancient +5V·heavy Athlon Classic/XP system, and you had to choose from amongst brand new current (2016) PSUs, which ones do you think would tend to give less headaches for this application?

Puuh, that's hard...
Because you should have around 20Amps on the +5V rail. And very little 'normal group regulated' units have that...
So that leaves the DC-DC things.
But they are rather expansive...

Or perhaps a DC-DC regulated one like the recently released Corsair CX650M (25A on +5V) to prevent crossloading problems?

Is there any other choice but to get either an older PSU and recap it or get a PSU with DC-DC?

But there are also other options I'd take a look at...
Like cooler master G-Series, InWin GreenMe, Xilence Performance A+ and probably some others too.

But if you want higher currents on +5V, you'll run into problems with the price...

The be quiet Straight Power E10 with 500W or more seems like a decent choice for older systems with 24A on +5V and +3,3V.
But the price is rather on the high side...

[TELVM]

Re: Will it explode?

Thanks Momaka.

... older group-regulated PSUs made for 5V-based PCs are the other way around, so with a heavy 3.3V/5V load, the 12V rail is normal.

Yep the best option would be to get a decent old-style PSU with plenty of amps on +5V, then recap and recondition it.

But I'm thinking about people that like keeping retrocomps alive, but can't recap and will never tinker inside a PSU. Old PSUs with OEM aged dubious caps inside (like those Fuhjyyu-loaded old Antecs ) aren't a reliable option for them.


So if you had to power an ancient +5V·heavy Athlon Classic/XP system, and you had to choose from amongst brand new current (2016) PSUs, which ones do you think would tend to give less headaches for this application?

Perhaps something in the style of the Seasonic S12II 520W (group-regulated, 24A on +5V, 130W combined +5V/+3.3V)?

Or perhaps a DC-DC regulated one like the recently released Corsair CX650M (25A on +5V) to prevent crossloading problems?

[momaka]

Re: Will it explode?

Suppose this PSU, whose label says '19A max on +5V':
http://i.imgbox.com/Wv9DSMPr.png

But that as a matter of fact comes with one 30A rectifier on +5V, of this model:
http://i.imgbox.com/H5ZR0wwv.png

Then suppose I use it to power an old Athlon XP system without 'P4' connector on the mobo, meaning the CPU draws all its juice from +5V (as opposed to from +12V like modern stuff).

Thus total system draw is something like say 25A from +5V, 5A from +12V, and 5A from +3.3V.

Will that PSU cope with, or will it bust?

Ignoring the possibility of severe cross-loading, the PSU will more than likely be fine with that load on the 5V rail. That 30 A rectifier should allow you to pull up to 20-22 Amps on the 5V rail (but not 30 Amps, because this PSU uses a forward-converter topology, after all)

Realistically speaking, however, with that heavy 5V load, the voltage on the 12V rail might get out of spec (too high). So be prepared for that and don't attach your most valuable HDDs.

Newer group-regulated PSUs that are made for 12V-based PCs usually have an extra turn or two for the 12V rail in the main transformer in order to keep the 12V rail high even when it is under load with a light or no load on the 5V rail. Conversely, older group-regulated PSUs made for 5V-based PCs are the other way around, so with a heavy 3.3V/5V load, the 12V rail is normal.

Because 3V3 is often generated out of the +5V rail, so it should be at least 130/5 = 26A

Not exactly.
This PSU has a forward-converter topology. That means the 30 Amp rectifier on the 5V rail will NOT be able to supply 30 Amps of current, unlike in a full or half-bridge topology (or any other topology that has a center-tapped output on the main transformer).

The 19 Amps rating was probably given for what this PSU can do without having the 12V rail go sky-high due to cross-loading.

That said, 19 Amps x 5 Volts = 95 Watts, indeed as you stated. There were very few 5V-based systems back in the day that could exceed this limit. You would probably need a Pentium 4 motherboard that draws power from the 5V rail (i.e. one that doesn't have a 4-pin 12V CPU connector) and a "high-end" Pentium 4 Willamate or Northwood CPU. Or a high-TDP Athlon XP combined with a Radeon 9800 Pro/XT.

OK so (just to be sure) where does exactly this '130w max +3.3V & +5V combined rating' come from?

** Transformer tap wire thickness for the 5V rail.
** output inductor wire thickness for 5V rail
** size of filtering caps on the 3.3V and 5V outputs
** the maximum power the PSU can supply from the 5V tap before the 12V rail goes too high.

*edit*
Don't know why I wrote all of that above . Wester547's explanations in post #9 are spot-on.
https://www.badcaps.net/forum/showpo...59&postcount=9

[Stefan Payne]

Re: Will it explode?

I would not use S12G or G series units, they are not really that good. Corsair RMx or RMi is completely different league.

Overall build quality on those Corsair units is really bad, sometimes there is threadlocker on the screws for the modular PCB, sometimes there is not. They are semi fanless (the S12G is not), use just a rifle bearing fan...
And last but not least:
They use those stupid and almost dangerous PCIe-Y Cables with 2 8pin connectors per 8pin connector (8+6 pin would be barely OK by the way).

Just because the voltage regulation is a bit better and riple and noise is a bit better doesn't really mean that the unit overall is better at all...

Oh and those Corsair units also have a whole lot of solid caps on the modular PCB as well...

I also have the feeling all these series will be EOL'd very soon as Seasonic is launching new ones, hopefully much better.

Well, yeah, they are a couple of years old, so one would think that seasonic did a completely new plattform in the same (price) range...

[Behemot]

Re: Will it explode?

I would not use S12G or G series units, they are not really that good. Corsair RMx or RMi is completely different league.

I also have the feeling all these series will be EOL'd very soon as Seasonic is launching new ones, hopefully much better.

[Stefan Payne]

Re: Will it explode?

Well, depends on the buck converters.
Some do, others don't...


Some just switch off, others burn.

But what I don't understand with those DC-DC units is the minor rail combined voltage.
Why do they do that?

[TELVM]

Re: Will it explode?

^ Yep, old-style +5V-heavy crossloading would be a problem with group-regulated PSUs like the CX430:




But not with modern PSUs which use DC-DC conversion for +5V & +3.3V, like the Seasonic G-550:



^ In these the main transformer only generates +12V, then buck converters step down the minor rails directly from +12V.




I wonder if these PSUs with DC-DC generated minor rails would be perhaps more tolerant of a (moderate) overload on the +5V rail?

[joshnz]

Re: Will it explode?

Basically I'm testing the waters about the dangers of using modern PSUs, designed with large +12V loads and little +5V loads in mind, to power old comps of the Athlon Classic/XP pre- 'P4 connector' era, which draw relatively heavily from +5V and little from +12V.

main issue is regulation is done on 12v rail so a heavy load on 5v rail causes 5v rail to dip low.

[TELVM]

Re: Will it explode?

^ Thanks. Guts pic of PSU in cuestion, looks like +3.3 & +5V come from same transformer tap:

... Are you asking if the rectifiers would explode or if the main switchers would explode? ...

Basically I'm testing the waters about the dangers of using modern PSUs, designed with large +12V loads and little +5V loads in mind, to power old comps of the Athlon Classic/XP pre- 'P4 connector' era, which draw relatively heavily from +5V and little from +12V.

[Wester547]

Re: Will it explode?

I meant to ask what exactly phisycally limits the combined +3.3V & +5V output to 130W.

So it is the transformer and the drive circuit, and they'll blow if I ask say 140W from +3.3/+5V combined on this particular PSU? Even though +12V draw is very little, say just 24W/2A?

Just to be sure.

The combined rating suggests that +3.3V and +5V are taken from the same transformer tap unless +5V and +3.3V are stepped down from +12V through a buck DC-DC converter. So the combined limit is representative of the maximum current that particular transformer winding can handle or the buck regulator in any case, whichever method the PSU uses to get +5V and +3.3V. In some older PSUs, they use a cheap design that linear regulates +3.3V from +5V (very inefficient but results in an exceptionally clean output).

The maximum current on each output is determined by a number of things - the size of the toroid, the number of turns, the core material it uses (-52 or -26, etc), the width of the PCB traces, the transformer taps, how many layers the PCB is made of, the current rating of the rectifiers, the rating of the output MOSFETs used for DC-DC conversion, etc. Are you asking if the rectifiers would explode or if the main switchers would explode? The output rectifiers could potentially short if you exceed the current rating for too long or if you exceed the peak repetitive reverse voltage. Depending on the topology as well, a half wave rectified output (forward topology) will result in a lower current limit for the rectifiers than a full wave rectified output (half bridge), since in a half wave rectified output, one of the diodes has to be used in a double diode pack to discharge the output inductor (and complete the current path). But that can all change depending on how well the PSU is ventilated and cooled. Of course, this might work a bit differently if the whole power supply uses synchronous rectification instead.

The two +12V rails are usually split through shunts after the output capacitors, so the +12V rail is basically split into two. This means that the limit for both +12V rails is given by the rectifier(s) themselves, but as stated before will depend on some other factors as well.

[budm]

Re: Will it explode?

The drive circuit will have protection circuit that will detect the over current draw (also will have over Voltage protection) by the load so if the protection circuit is well designed it should stop the circuit before something blow up badly.
The confusion is on that 130W label. I think they are trying to indicate that for example, if you are using 100W on the 5V then you can get 30W from the 3.3V.
If you look at the +12V1, +12V2 rating, it shows 408W 34A MAX so you cannot draw 24A from +12V1 and 21A from 12V2 at the same time, you have to derate one or the other with combined 34A MAX.. That is the way I see it. Do they details info at the web site?

[TELVM]

Re: Will it explode?

I meant to ask what exactly phisycally limits the combined +3.3V & +5V output to 130W.

So it is the transformer and the drive circuit, and they'll blow if I ask say 140W from +3.3/+5V combined on this particular PSU? Even though +12V draw is very little, say just 24W/2A?

Just to be sure.