Page 1 of 3 123 LastLast
Results 1 to 10 of 21

Thread: Designing a load machine for PSU testing

  1. #1
    Join Date
    Feb 2011
    Posts
    31
    Thanks Thanks Given 
    0
    Thanks Thanks Received 
    0
    Thanked in
    0 Posts

    Default Designing a load machine for PSU testing

    I want to build a load machine for PSU testing : Load it with resistive load, see how much it holds, test OCP OPP SCP, measure ripple an efficiency

    The PSUs which are to be tested for now are Antec VP450 and others of similar specs : http://www.antec.com/Believe_it/product.php?id=MjQwNw==



    I have a preliminary design as for construction, as shown in the pictures below. The design mostly is :

    Resistive loads built in an emptied enclosure of an old A/C, and use the fan of the A/C for cooling

    All control is manual, using standard 10A mains switches (of the type used in appliances) connecting the loads to each rail. They limit each load unit to 10A max

    The resistors are made of wire from heating elements (the type where the wire is wound on a ceramic sick and is easy to disassemble to take the wire). I will cut the wire to short segments so that it can be used at 3.3/5/12V (the entire heating element was meant to work on mains power)

    The higher power load units are made of more than one resistor in parallel. The resistors are connected between pairs of threaded rods, and secured to the rods with nuts

    The threaded rods (carrying current) are held to a construction of metallic drywall support members, and are isolated from the construction at each end with a piece of blank PCB. The earth (black) is isolated too to prevent mess-up in ripple measuring

    To prevent heat from conducting along the rods (and melting isolation on the wires that connect the rods), each rod has heatsink on the end - a few sheet metal plates tightened to it. It is supposed to make the end of the rod (wherre the wire connects) cooler




    The questions :

    Do you have any improvement suggestions ? Cost cutting suggestions ?

    The heating elements are meant to work on mains power, without fan. loaded at 750W / entire wire length, they get yellow-hot when working. If I load the wire at 1/4 that power (in W per length) , and cool it with a fan, how much hot it will be ?

    Can I load it more than 1/4 the power without running into : 1. the wire changing resistance too much when hot, 2. heating to the extent that the heatsinks cannot deal with the heat and I melt isolation on the wires, 3. A/C fan with plastic blades (which pull air out so it gets the hot air) melts ?

    Are those load levels good for testing the Antec VP450 and similar PSUs ?

    12V1 : 25 / 50 / 4x100 W (2 / 4 / 4x8 A)
    12V2 : 25 / 50 / 4x100 W (2 / 4 / 4x8 A)
    5V : 5 / 10 / 20 / 3x40 W (1 / 2 / 4 / 3x8 A)
    3.3V : 3.3 / 6.6 / 13.3 / 3x26.6 W (1 / 2 / 4 / 3x8 A)
    5VS : 5 / 10 / 20 / 40 W (1 / 2 / 4 / 8 A)
    -12V : 3 / 6 W (0.25 / 0.5 A)

    The high loads on 3.3V 5V and 5VS are to test OCP (I dont think an actual computer loads them that much)

    Why are the OCP levels in the Antec VP450 so much higher than the actual current ratings (like 5V 15A rating 50A OCP) ?

    A scenario I imagined, which I think is possible when there is no OCP on 5VS : Low quality mainboard has no fuse on USB power output. A mouse plugged in shorts out (due to stressed isolation where the cable enters the mouse) and takes short circuit current (which is not very high because of resistance of the thin wires). High current heats up the mouse's cable (and some mainboard tracks) and the melting cable (laying on papers etc) starts a fire. What you think ?

    I want to test SCP. What type of switch or contact can I use (which takes the blow of closing the short circuit) which won't die from arcing / welding ?

  2. #2
    Join Date
    Aug 2010
    Posts
    2,700
    Thanks Thanks Given 
    116
    Thanks Thanks Received 
    0
    Thanked in
    0 Posts

    Default

    The PSUs which are to be tested for now are Antec VP450 and others of similar specs : http://www.antec.com/Believe_it/product.php?id=MjQwNw==
    We'll have a comparison then

    One question:
    Does testing SCP even have a point if you test OCP, and from what I've learned here, OCP is likely to kick in quicker?

  3. #3
    Join Date
    Nov 2010
    Location
    Serbia
    Posts
    2,248
    Thanks Thanks Given 
    129
    Thanks Thanks Received 
    15
    Thanked in
    7 Posts

    Default

    It does if the OCP is not present, and OPP is a bit slow...
    Careful what you wish for... You just might get it.

  4. #4
    Join Date
    Feb 2011
    Posts
    31
    Thanks Thanks Given 
    0
    Thanks Thanks Received 
    0
    Thanked in
    0 Posts

    Default

    I want to have this option : For example if I test a PSU without OCP, or if the load is too low to kick in OCP

    About OCP : Does it kick immediately as you cross the limit, or there is a time/current curve (as in circuit breakers) ?

  5. #5
    Join Date
    Nov 2010
    Location
    Serbia
    Posts
    2,248
    Thanks Thanks Given 
    129
    Thanks Thanks Received 
    15
    Thanked in
    7 Posts

    Default

    It should cut immediately. But it's rarely set at what's actually stated on the sticker. Usually if it says 20A, the actual OCP trip point will be at least 23A, to allow for some tolerance/headroom...
    Careful what you wish for... You just might get it.

  6. #6
    Join Date
    Feb 2011
    Posts
    31
    Thanks Thanks Given 
    0
    Thanks Thanks Received 
    0
    Thanked in
    0 Posts

    Default

    The specs of that antec look like this :
    5V 15A / 50A OCP
    12V 18A / 27A OCP
    3.3V 24A / 50A OCP

    Not overkill of a headroom ? why is that so ?

  7. #7
    Join Date
    Nov 2010
    Location
    Serbia
    Posts
    2,248
    Thanks Thanks Given 
    129
    Thanks Thanks Received 
    15
    Thanked in
    7 Posts

    Default

    Probably to handle transients and spikes better. The silicon in that PSU is certainly capable of handling those kinds of loads... Would be a waste to make the OCP limits too tight. The minor rail OCP is a bit too relaxed, though...
    Careful what you wish for... You just might get it.

  8. #8
    Join Date
    Feb 2011
    Posts
    31
    Thanks Thanks Given 
    0
    Thanks Thanks Received 
    0
    Thanked in
    0 Posts

    Default

    Does this mean that the PSU can be destroyed if I run it long enough just below the OCP values ?

  9. #9
    Join Date
    Nov 2010
    Location
    Serbia
    Posts
    2,248
    Thanks Thanks Given 
    129
    Thanks Thanks Received 
    15
    Thanked in
    7 Posts

    Default

    I don't think so. The 12V rail is actually capable of putting out that many amps. The minor rails aren't, but I imagine OVP or OTP would trigger in time to save the PSU.
    I also sincerely doubt you'd find a realistic way of pulling > 15A on 3.3V and > 25A on 5V rails...
    Careful what you wish for... You just might get it.

  10. #10
    Join Date
    Feb 2011
    Posts
    31
    Thanks Thanks Given 
    0
    Thanks Thanks Received 
    0
    Thanked in
    0 Posts

    Default

    There is one : a "weak" short circuit, that has just enough resistance to keep tthe current under the OCP value

    You can get one by cutting wires with sharp edge of the case

Similar Threads

  1. 12V load for burn-in testing
    By ashiekh in forum PC Power Supply Discussion
    Replies: 4
    Last Post: 01-14-2018, 03:52 PM
  2. Ripple testing with which load ?
    By Ash in forum PC Power Supply Discussion
    Replies: 21
    Last Post: 02-20-2011, 07:06 PM
  3. NSC: Load Transient Testing Simplified
    By Travis in forum Testing Methodology Discussion
    Replies: 8
    Last Post: 09-16-2010, 11:37 PM
  4. Combine 12V rails for load testing?
    By ~*77*~ in forum Testing Methodology Discussion
    Replies: 10
    Last Post: 05-21-2010, 02:45 AM
  5. Load Testing
    By davidhammock200 in forum Testing Methodology Discussion
    Replies: 8
    Last Post: 04-08-2009, 04:30 AM

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •