Author Topic: switching-PSUs: why can't you usually exceed the 50% of declared power?  (Read 11744 times)

0 Members and 1 Guest are viewing this topic.

Offline 0dbTopic starter

  • Frequent Contributor
  • **
  • Posts: 336
  • Country: zm
The Superflower Leadex is a great series of Switching-Mode PSUs, and if you buy a 650 Watt unit, you can be sure the unit will serve you up to 650 Watt without dropping voltage, or increasing the voltage ripple.

But let's talk about cheaper PSUs. I have recently seen a lot of 500 Watt units where you cannot exceed th 50% of the declared power, thereore if you buy a 500 Watt, it will serve you only up to 250 Watt, and beyound it might get damaged.

What I cannot understand is why. Which are the technical reasons for this? And what one should look into the circuit to have some enlightenments about the quality of the product concerning its behavior under load?

Thanks
 

Offline wraper

  • Supporter
  • ****
  • Posts: 17952
  • Country: lv
Unless you buy cheapest garbage usually you can run them at 100% or nearly 100% load for prolonged time.
Quote
What I cannot understand is why. Which are the technical reasons for this?
Really, is it that hard to figure it out? It's cheaper to build lower quality PSU with fake ratings. More power/ higher efficiency = more expensive to produce.
 

Offline 0dbTopic starter

  • Frequent Contributor
  • **
  • Posts: 336
  • Country: zm
Really, is it that hard to figure it out?

I'd like to know details: is it for capacitors? if so, which is the difference between "good" capacitors and "bad" capacitors? And is there any table around telling which are good ones (vendor/models/etc)?

Is it for the transformer? Filters? Protection chip? Oscillators? What does make a PSU a "bad" PSU?

It's too easy to say "oh, because it's cheap". Cheap means all and nothing.
 

Online ejeffrey

  • Super Contributor
  • ***
  • Posts: 4034
  • Country: us
Not enough capacitance, not enough core material, not enough copper, not enough silicon, and not enough cooling.  Too low temperature ratings and too much ripple current. Even if the PSU is designed correctly, if parts are substituted at manufacturer time that can cause any of the above issues.

There are lots of ways to cheap out, not just one.
 

Offline engrguy42

  • Frequent Contributor
  • **
  • Posts: 656
  • Country: us
The Superflower Leadex is a great series of Switching-Mode PSUs, and if you buy a 650 Watt unit, you can be sure the unit will serve you up to 650 Watt without dropping voltage, or increasing the voltage ripple.

But let's talk about cheaper PSUs. I have recently seen a lot of 500 Watt units where you cannot exceed th 50% of the declared power, thereore if you buy a 500 Watt, it will serve you only up to 250 Watt, and beyound it might get damaged.

What I cannot understand is why. Which are the technical reasons for this? And what one should look into the circuit to have some enlightenments about the quality of the product concerning its behavior under load?

Thanks

Where did you see this? Do you have an example of an ATX that can't operate over 50%? Are you sure you're not talking about efficiency dropping above 50% or something like that?
- The best engineers know enough to realize they don't know nuthin'...
- Those who agree with you can do no wrong. Those who disagree can do no right.
- I'm always amazed at how many people "already knew that" after you explain it to them in detail...
 

Offline wraper

  • Supporter
  • ****
  • Posts: 17952
  • Country: lv
Really, is it that hard to figure it out?

I'd like to know details: is it for capacitors? if so, which is the difference between "good" capacitors and "bad" capacitors? And is there any table around telling which are good ones (vendor/models/etc)?

Is it for the transformer? Filters? Protection chip? Oscillators? What does make a PSU a "bad" PSU?

It's too easy to say "oh, because it's cheap". Cheap means all and nothing.
It's everything, including smaller transformer.
 

Offline richard.cs

  • Super Contributor
  • ***
  • Posts: 1201
  • Country: gb
  • Electronics engineer from Southampton, UK.
    • Random stuff I've built (mostly non-electronic and fairly dated).
Think of it like taking a 250 W PSU, removing any overload protection, then slapping a 500 W sticker on it and selling it as 500 W. Fundamentally all the power components are undersized for 500 W, there isn't usually just one component where they cheaped-out.
 

Offline engrguy42

  • Frequent Contributor
  • **
  • Posts: 656
  • Country: us
Think of it like taking a 250 W PSU, removing any overload protection, then slapping a 500 W sticker on it and selling it as 500 W. Fundamentally all the power components are undersized for 500 W, there isn't usually just one component where they cheaped-out.

Do you know of any examples of this, or are you just speculating?
- The best engineers know enough to realize they don't know nuthin'...
- Those who agree with you can do no wrong. Those who disagree can do no right.
- I'm always amazed at how many people "already knew that" after you explain it to them in detail...
 

Offline 0dbTopic starter

  • Frequent Contributor
  • **
  • Posts: 336
  • Country: zm
Where did you see this?

I bought a couple of AEROCOOL VX PLUS 550W just to make a few tests.
They litteraly exploded at 300Watt of load. And there was no PFC inside.
 

Offline 0dbTopic starter

  • Frequent Contributor
  • **
  • Posts: 336
  • Country: zm
CORSAIR VS 650W inside a PC at work. Same bitter end.
It worked only up to 450 Watt, then it presented overvoltage at its lines, and the motherboard died.

 


Offline chriva

  • Regular Contributor
  • *
  • Posts: 102
  • Country: se
Where did you load it and how?

If you don't have proper equipment to measure actual load, you can't really say "it blew at this and that wattage".
Just because a GPU states it'll only draw 150 Watts doesn't mean that is actual peak power. They can easily push almost twice that with a slight overclock.
Processors are pretty much as bad. Design power is far under their boost power draw
 

Offline rsjsouza

  • Super Contributor
  • ***
  • Posts: 6106
  • Country: us
  • Eternally curious
    • Vbe - vídeo blog eletrônico
Where did you see this?

I bought a couple of AEROCOOL VX PLUS 550W just to make a few tests.
They litteraly exploded at 300Watt of load. And there was no PFC inside.
Others already mentioned what could be cheaper. Think about your own experience with the "explosion" and imagine the manufacturer used a cheaper 15A MOSFET trying to carry 25A or 30A on the 3.3 or 5V lines, or used a 30A MOSFET but reduced the size of the heatsink. Both scenarios are quite possible for a catastrophic scenario but rarely go alone: the use of lower quality PCBs with smaller current capacity, bad soldering job due to low salaries, use of components removed from other products, etc.

I have seen all these in a plethora of power supplies sold in the marketplace - the worst offenders were in the periphery markets such as Brasil and Argentina (the ones I know better), where the supplies were sold as "500W (300W real)"
Vbe - vídeo blog eletrônico http://videos.vbeletronico.com

Oh, the "whys" of the datasheets... The information is there not to be an axiomatic truth, but instead each speck of data must be slowly inhaled while carefully performing a deep search inside oneself to find the true metaphysical sense...
 
The following users thanked this post: 0db

Offline 0dbTopic starter

  • Frequent Contributor
  • **
  • Posts: 336
  • Country: zm
Where did you load it and how?

I attached several bulbs in parallel to simulate the load. I started with a few bulbs to simulate 100Watt, then I added more bulbs to simulate 200Watt, 300Watt, and finally 500Watt. During the test I observed the ripple with an oscilloscope.
 

Offline engrguy42

  • Frequent Contributor
  • **
  • Posts: 656
  • Country: us
Hold on here...some alarm bells are going off...

Some basic questions:

- Are these new or used units?

- What tests did you do?

The reason I ask is that decent ATX power supplies are required to provide some basic protection. They generally have supervisor circuits which monitor output current and voltage and shut down the switching MOSFETS quickly when out of range. They also have fuses. I've done a significant amount of testing on a decent ATX supply, and the supervisor circuit worked great. I overloaded it multiple times with a huge rheostat and it shut down instantly.   

Now, I would be fairly certain that a Corsair unit would have these protections. So something seems fishy here. "The motherboard died??".

If you pay for used or extremely cheap power supplies you shouldn't be asking the question. You get what you pay for.

If you're paying for a decent, ATX power supply, I suggest you take a look at the protection circuits it's supposed to have (look at the ATX spec) and see what supervisor circuit is inside and what it's supposed to do.

I'm not saying nothing bad can ever happen, I'm saying a string of failures sounds more like user error (or cheap/used junk) than bad units.
- The best engineers know enough to realize they don't know nuthin'...
- Those who agree with you can do no wrong. Those who disagree can do no right.
- I'm always amazed at how many people "already knew that" after you explain it to them in detail...
 

Offline 0dbTopic starter

  • Frequent Contributor
  • **
  • Posts: 336
  • Country: zm
- Are these new or used units?

I bought them from Amazon, declared "new".
 

Offline engrguy42

  • Frequent Contributor
  • **
  • Posts: 656
  • Country: us
For some perspective, here's something very similar to the supervisor circuit on my Dell ATX power supply. And this is from a Dell desktop that's like 5-10 years old. As you can see it's monitoring current AND voltage on 5 outputs (3.3, 5, & 12v). It checks overvoltage, overcurrent, and undervoltage.

And it has protection on the input (fuse, MOV's, PFC, etc.).

So blowing up really isn't a reasonable likelihood, nor is destroying the motherboard. Hence my skepticism. 

Of course that won't stop the anti-ATX fanboys from jumping to the assumption that these ATX supplies are just crap, but anyway... 
- The best engineers know enough to realize they don't know nuthin'...
- Those who agree with you can do no wrong. Those who disagree can do no right.
- I'm always amazed at how many people "already knew that" after you explain it to them in detail...
 
The following users thanked this post: 0db

Offline 0dbTopic starter

  • Frequent Contributor
  • **
  • Posts: 336
  • Country: zm
The reason I ask is that decent ATX power supplies are required to provide some basic protection.

OverCurrent Protection?

umm, my old Cooler Master MASTERWATT 550W had a slow OCP protection at 600 Watt.
The PSU didn't shutdown immediately, only after 1 sec later with a VDrop of 0.8V at 12V.

Paid 50 euro for it, not exactly "cheap", it went back to Amazon.
 

Offline engrguy42

  • Frequent Contributor
  • **
  • Posts: 656
  • Country: us
The reason I ask is that decent ATX power supplies are required to provide some basic protection.

OverCurrent Protection?

umm, my old Cooler Master MASTERWATT 550W had a slow OCP protection at 600 Watt.
The PSU didn't shutdown immediately, only after 1 sec later with a VDrop of 0.8V at 12V.

Paid 50 euro for it, not exactly "cheap", it went back to Amazon.

I suggest you take a look at the Intel ATX power supply design guide. It's extremely informational if you want to understand ATX supplies. Here's a snip of the allowable voltage regulation. Note it says +12V output can get down to 11.4 which is slightly higher than what you measured. I'm not sure how you measured and how accurate you were, but it seems reasonable. And if you were acutally measuring the -12VDC, it says it can get down to -10.8
- The best engineers know enough to realize they don't know nuthin'...
- Those who agree with you can do no wrong. Those who disagree can do no right.
- I'm always amazed at how many people "already knew that" after you explain it to them in detail...
 
The following users thanked this post: 0db

Offline 0dbTopic starter

  • Frequent Contributor
  • **
  • Posts: 336
  • Country: zm
+12V - 0.8V = 11.2V < 11.4V  ---> not acceptable! I initially measured it on the oscilloscope, then I repeated the experiment with a multimeter of 3 digits in parallel. Same result.

Anyway, OVC means over current protection. If the PSU has a limit of 550Watt, and you try to sink out 600Watt, it should immediately shutdown.  Not after 1 sec, but rather the soonest possible!
 

Offline engrguy42

  • Frequent Contributor
  • **
  • Posts: 656
  • Country: us
Again, look at the Intel ATX design guide. It tells you exactly what the requirements are. I'm not sure why you feel 1 second is too long. Is that what the spec says? If the spec doesn't say the operating time then there's not much you can do.

As far as your supply dipping to 11.2 rather than 11.4...why is that important? Yeah, it's outside the spec, but does that really matter?
- The best engineers know enough to realize they don't know nuthin'...
- Those who agree with you can do no wrong. Those who disagree can do no right.
- I'm always amazed at how many people "already knew that" after you explain it to them in detail...
 

Offline 0dbTopic starter

  • Frequent Contributor
  • **
  • Posts: 336
  • Country: zm
I'm not sure why you feel 1 second is too long. Is that what the spec says?

Because in 1 sec it was enough to overheat, and it started making bad smell. Parts were for sure stressed.

 

Offline engrguy42

  • Frequent Contributor
  • **
  • Posts: 656
  • Country: us
BTW, here's what it says about overcurrent protection. No more than 240VA on each rail, greater than that causes a trip. Test that at a minimum rate of 10A/second. So taking 1 second to trip seems pretty reasonable.
- The best engineers know enough to realize they don't know nuthin'...
- Those who agree with you can do no wrong. Those who disagree can do no right.
- I'm always amazed at how many people "already knew that" after you explain it to them in detail...
 

Offline engrguy42

  • Frequent Contributor
  • **
  • Posts: 656
  • Country: us
I'm not sure why you feel 1 second is too long. Is that what the spec says?

Because in 1 sec it was enough to overheat, and it started making bad smell. Parts were for sure stressed.

Come on, did you actually measure the temperatures? Were they high enough to damage anything? You're really worried about it taking 1 second to trip?
- The best engineers know enough to realize they don't know nuthin'...
- Those who agree with you can do no wrong. Those who disagree can do no right.
- I'm always amazed at how many people "already knew that" after you explain it to them in detail...
 

Offline engrguy42

  • Frequent Contributor
  • **
  • Posts: 656
  • Country: us
Hold on...the OCP section I posted was from an older version. Here's the 2018 version
- The best engineers know enough to realize they don't know nuthin'...
- Those who agree with you can do no wrong. Those who disagree can do no right.
- I'm always amazed at how many people "already knew that" after you explain it to them in detail...
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf