Author Topic: ATX Power Supply Load Resistors  (Read 13762 times)

0 Members and 1 Guest are viewing this topic.

Offline GrandmastersexsayTopic starter

  • Contributor
  • Posts: 11
ATX Power Supply Load Resistors
« on: November 19, 2014, 03:21:52 pm »
I have a new 500 W ATX power supply with a 30 A 5V rail that I am going to use to power 300 smart RGB LED Christmas lights for my tree this year. The power supply is an Antec Basiq BP500U.

I tied the load wire to ground, and everything works well. My question is, do modern ATX power supplies need a minimum load, or do they come with internal load resistors? It sounded like some of the older ATX supplies needed a rather large minimum load.  With how modern computers are so concerned about low power, with sleep and hibernation, ect. I can't imagine they would always be providing much of a load. Voltages were good on the 3.3V, 5V, and 12V rails with no external load. I ran it for a good hour testing a strand of 50 of the RGB LEDs without issue. There were extended periods of time where the only load on the supply was a little controller and the fan of the power supply.

Can I confidently use the supply as is for the Christmas season? Can anyone give me a good explanation for why some ATX power supplies need an external load and others don't?

Thanks!
 

Offline ConKbot

  • Super Contributor
  • ***
  • Posts: 1398
Re: ATX Power Supply Load Resistors
« Reply #1 on: November 19, 2014, 03:36:24 pm »
It has to do with what rail is regulated, and the power supply topology.  In your 'classic' ATX power supply, its a single transformer operating in flyback mode with a variety of windings.  The turns ratio of the windings is matched (I.e. a 12:5:3.3 ratio for the main 3 secondary windings) The output inductor also has multiple windings in the same ratio, which helps provide regulation between the rails.  However the SMPS control IC is only monitoring one of the outputs.  On older designs it could be the 5v or the 3.3v outputs. If you drew a bunch of current from a non monitored winding, with none on the monitored winding, then stuff could start to get strange, hence the load resistors.

Modern high end PSUs are high current mains->12v converters, and have a 12-5v and 12v-3.3v DC-DC converters to supply the other rails, since a majority of the load is on the 12v rail in a modern PC anyway. That also means that each rail is individually regulated too.
 

Offline wraper

  • Supporter
  • ****
  • Posts: 17581
  • Country: lv
Re: ATX Power Supply Load Resistors
« Reply #2 on: November 19, 2014, 03:54:02 pm »
It has to do with what rail is regulated, and the power supply topology.  In your 'classic' ATX power supply, its a single transformer operating in flyback mode with a variety of windings.
Not flyback but rather half bridge in most older ones transitioning to forward topology variations in more modern ones. Flyback topology usually is used only in <150W power supplies.
Quote
However the SMPS control IC is only monitoring one of the outputs.  On older designs it could be the 5v or the 3.3v outputs.
No, usually it is monitoring average of 12V and 5V rails while 3.3V rail is regulated separately (powered from 5V winding).
 

Offline zapta

  • Super Contributor
  • ***
  • Posts: 6289
  • Country: 00
Re: ATX Power Supply Load Resistors
« Reply #3 on: November 19, 2014, 04:25:41 pm »
Modern high end PSUs are high current mains->12v converters, and have a 12-5v and 12v-3.3v DC-DC converters to supply the other rails, since a majority of the load is on the 12v rail in a modern PC anyway. That also means that each rail is individually regulated too.

Do you know such a power supply that have separate regulation for 12V and 5V and doesn't require 5V load throughout the current range of the 12V?

I needed recently a 12V 16A and used an ATX PSU but had to use a 5V load to have it regulated (see https://www.eevblog.com/forum/reviews/is-this-power-supply-fake/45/).  My conclusion from this little exercise was that the capabilities of ATX PSU's are greatly hyped (e.g. claiming to have 2  x 12V rails where in practice they are both shorted to each other).
 

Offline ConKbot

  • Super Contributor
  • ***
  • Posts: 1398
Re: ATX Power Supply Load Resistors
« Reply #4 on: November 19, 2014, 06:59:29 pm »
http://www.hardwaresecrets.com/page/power has teardowns to the component level of power supplies, they usually spell it out on the "secondary analysis" page of the reviews. However youre not going to see much under 500W there for recent reviews.
 

Offline wraper

  • Supporter
  • ****
  • Posts: 17581
  • Country: lv
Re: ATX Power Supply Load Resistors
« Reply #5 on: November 19, 2014, 07:49:48 pm »
Modern high end PSUs are high current mains->12v converters, and have a 12-5v and 12v-3.3v DC-DC converters to supply the other rails, since a majority of the load is on the 12v rail in a modern PC anyway. That also means that each rail is individually regulated too.

Do you know such a power supply that have separate regulation for 12V and 5V and doesn't require 5V load throughout the current range of the 12V?

I needed recently a 12V 16A and used an ATX PSU but had to use a 5V load to have it regulated (see https://www.eevblog.com/forum/reviews/is-this-power-supply-fake/45/).  My conclusion from this little exercise was that the capabilities of ATX PSU's are greatly hyped (e.g. claiming to have 2  x 12V rails where in practice they are both shorted to each other).
I did a bit research after that tread and there appears to be a few models with less than 1KW which have separate rail regulation. Like this http://www.amazon.com/Cooler-Master-G550M-Compact-Mini-ITX/dp/B00MAZK3VE . But they cost almost 2x price of the decent PSU of the same power. And I still insist that you should have bought that cheap MeanWell instead of throwing your money in the air.
 

Offline zapta

  • Super Contributor
  • ***
  • Posts: 6289
  • Country: 00
Re: ATX Power Supply Load Resistors
« Reply #6 on: November 19, 2014, 09:33:25 pm »
...  And I still insist that you should have bought that cheap MeanWell instead of throwing your money in the air.

The main reason was the built in mains connector and switch of the ATX.  I ended up with an Antec, non modular ATX PSU. I removed all the wires I don't need (it was not easy, it's a jungle of wires within the case and added a 5ohm resistor for load. I looks quite nice now.
 

Offline Kjelt

  • Super Contributor
  • ***
  • Posts: 6572
  • Country: nl
Re: ATX Power Supply Load Resistors
« Reply #7 on: November 19, 2014, 09:41:16 pm »
The newer high efficiency psu all have 12V rails even multiple and have seperate dc-dc converters for the 5 and 3,3V supplies that need no load to work ok. Look for instance at the 80+% BeQuiet PSU,s.
There was/is somewhere this great website i think it was german that review psu's and show pictures of the inside and tell the way they are build and even quality electrolytes.
I use the bequiet 350W as 12V supply for my ledlighting indoors and the quality is way better than the meanwells but they are not cheap.
 

Offline GrandmastersexsayTopic starter

  • Contributor
  • Posts: 11
Re: ATX Power Supply Load Resistors
« Reply #8 on: November 20, 2014, 12:01:35 am »
I would have gone with a meanwell, but this is for my livingroom Christmas tree, and exposed 120 V terminals and toddlers don't mix well.  Using an ATX power supply saved me the hassle and added expense of sourcing an enclosure.
 

Offline Phaedrus

  • Frequent Contributor
  • **
  • Posts: 714
  • Country: us
Re: ATX Power Supply Load Resistors
« Reply #9 on: November 20, 2014, 12:48:21 am »
The Antec Basiq is a double forward unit with the +12V and +5V group regulated (averaged value) with the +3.3V using separate feedback. This is considered a barebones, somewhat dated topology by today's standards. It has a limitation in that the +12V and +5V must both be loaded by about the same % of their total rated capacity, or else the averaged regulation fails and you get out of spec voltage. For instance, with 20A on the +12V and 1A on the +5V you may see +11.65V and +5.25V outputs. With 20A on the +12V and 0.1A on the +5V you may see +11.45V and +5.35V outputs, which by the way are still inside that unit's OVP/UVP protection limits. I'm just guessing on those load/voltage figures, but they're typical for similar units. It's not a great PSU anyway, intended largely for the "my PSU just died and I need to buy a replacement at the store today" market.

Modern high-end units will use a double forward, LLC resonant half bridge, or ZVS half bridge or full bridge topology with a single +12V transformer output, with the +5V and +3.3V rails buck regulated from that output. These are generally much better suited for general purpose electronics work than older or cheaper units.
« Last Edit: November 20, 2014, 12:57:40 am by Phaedrus »
"More quotes have been misattributed to Albert Einstein than to any other famous person."
- Albert Einstein
 

Offline GrandmastersexsayTopic starter

  • Contributor
  • Posts: 11
Re: ATX Power Supply Load Resistors
« Reply #10 on: December 01, 2014, 10:20:18 pm »
I accidentally tied the load wire(green) to -12V(blue) instead of ground, and now I can't get it to work. Did I fry this thing, or can I reset it/fix it somehow? 
 

Offline Kjelt

  • Super Contributor
  • ***
  • Posts: 6572
  • Country: nl
Re: ATX Power Supply Load Resistors
« Reply #11 on: December 02, 2014, 12:32:16 pm »
Ouch!
Fixing it depends on the PSU implementation, look where the green wire leads to. If it is a standard logic ic , optocoupler or transistor you can try replacing that.
 

Offline GrandmastersexsayTopic starter

  • Contributor
  • Posts: 11
Re: ATX Power Supply Load Resistors
« Reply #12 on: December 02, 2014, 09:54:54 pm »
Ouch!
Fixing it depends on the PSU implementation, look where the green wire leads to. If it is a standard logic ic , optocoupler or transistor you can try replacing that.

I'm guessing this guy.

4-Channel Secondary Monitoring IC

http://www.siti.com.tw/product/spec/Power/PS223.pdf


 

Offline Kjelt

  • Super Contributor
  • ***
  • Posts: 6572
  • Country: nl
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf