Page 1 of 3 123 LastLast
Results 1 to 10 of 22

Thread: Ripple testing with which load ?

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

    Default Ripple testing with which load ?

    I have tested some PSUs for ripple with a PC doing various stuff. Found that there is more ripple when doing certain light tasks (e.g moving a window) than when doing heavier tasks (e.g running benchmarks). "lighter" and "heavier" are consistent with the power drawn on the AC measured with a watt meter

    This raises the question : The ripple in the ATX spec is measured with "switching" load like a computer moving a window (rapid spikes in CPU load) or with load like resistors ?

  2. #2
    Join Date
    Jan 2011
    Location
    WATERBURY, CT USA
    Posts
    545
    Thanks Thanks Given 
    0
    Thanks Thanks Received 
    0
    Thanked in
    0 Posts

    Default

    Hopefully the standard pertains to all conditions.
    You are correct testing is done under load at various "sample" (% of load) points.
    The people will want full details of your test procedures and equipment, please start with details. Thanks.
    Fact is though under all testing I've seen max ripple occurs at full load.
    Did you discover ripple exceeding that of supposed worse case scenario?
    Let the games begin..

  3. #3
    Join Date
    Aug 2010
    Posts
    2,700
    Thanks Thanks Given 
    116
    Thanks Thanks Received 
    0
    Thanked in
    0 Posts

    Default

    Measuring ripple when computer is a load is bound to go wrong, apparently you pick up an interference from (most likely) fans, etc. All in all, I had FSP400-60 GLN (80) having 20-30 mV of sinusoidal ripple on 5V and 12V, but also some nasty, aperiodic, several hundreds of mV high, spikes.
    Last edited by rafal_iB_PL; 02-19-2011 at 04:19 PM.

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

    Default

    I have done a short test. I did not have a scope so I plugged it into the "line in" audio port on a laptop, unplugged it from power / network / anything else to avoid interference, and recorded with Audacity

    In this test I measure +12V on a very bad generic power supply JNP ISO-500. The load is Pentium 4 + Geforce 6600 + 1 HDD running Windows XP

    0:00 - 0:10 : Sticking the probe into +3.3V to have a scale of how much V/div on the scope
    0:10 - 0:30 : Idle
    0:30 - 0:40 : 15W extra resistive load (lightbulb) on +12V
    0:40 - 0:50 : Moving (rapidly shaking) "My computer" window across the screen
    0:50 - 1:00 : Selecting (rapidly shaking the selection) icons in maximized folder window in "My computer"
    1:00 - 1:20 : ATITool starting and displaying 3D furry cube
    1:20 - 1:30 : ATITool doing artifacts scan
    1:30 - 1:40 : Paint opening 1280x1024 BMP file
    1:40 and on till the end : FLood fill of complex shape in the BMP, which pins CPU to 100%
    1:50 - 2:00 : +15W resistive load
    2:00 - 2:20 : +ATITool starting and displaying 3D furry cube
    2:20 - 2:30 : +ATITool doing artifacts scan
    2:30 - 2:40 : +ATITool doing artifacts scan +15W resistive load

    Watt meter on AC displays ~85 W idle, ~110 W moving the window, >130W in ATITool OR Paint, >190W in ATITool AND Paint

    jnp_iso-500.mp3 - 6.2 MB

  5. #5
    Join Date
    Nov 2010
    Location
    Serbia
    Posts
    2,245
    Thanks Thanks Given 
    129
    Thanks Thanks Received 
    12
    Thanked in
    6 Posts

    Default

    I am so remixing this.
    Careful what you wish for... You just might get it.

  6. #6
    Join Date
    Jan 2011
    Location
    WATERBURY, CT USA
    Posts
    545
    Thanks Thanks Given 
    0
    Thanks Thanks Received 
    0
    Thanked in
    0 Posts

    Default

    Ah McSteel "what's the buzz." Though I suspect generic psu's can indeed do weird.
    And some not so generic [Nexus 1.1K over 80 %].

  7. #7
    Join Date
    Jan 2007
    Location
    Lake Hartwell
    Posts
    375
    Thanks Thanks Given 
    10
    Thanks Thanks Received 
    0
    Thanked in
    0 Posts

    Default

    I was under the impression that ripple can only be tested/observed with an oscilloscope, as detailed in the ATX spec.
    A lot of people ask me if I were shipwrecked and could only have one book, what would it be? I always say 'How to Build a Boat.'
    - Stephen Wright

    "Land was created to provide a place for boats to visit."
    - Brooks Atkinson

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

    Default

    This is in fact an oscilloscope. Not the 20 MHz with AI trigger and fancy features, but you can see wave and measure it. I have tested some more PSUs and the results are very consistent with the PSU quality visible with the cover off

  9. #9
    Join Date
    Dec 2009
    Posts
    41
    Thanks Thanks Given 
    0
    Thanks Thanks Received 
    0
    Thanked in
    0 Posts

    Default

    Quote Originally Posted by C'DaleRider View Post
    I was under the impression that ripple can only be tested/observed with an oscilloscope, as detailed in the ATX spec.
    Most power supply reviews are written against oscilloscopes that aren't fast enough to meet ATX spec anyway. The USB Instruments Stingray is only good out to 250KHz, and all your better sound cards record at 192.

    Of course the sound card's readings are dimensionless and there may be a band pass filter that you don't know about.

  10. #10
    Join Date
    Jan 2011
    Location
    WATERBURY, CT USA
    Posts
    545
    Thanks Thanks Given 
    0
    Thanks Thanks Received 
    0
    Thanked in
    0 Posts

    Default

    Are comparative results indicative?
    Ash does seem to be saying values found decrease with quality. This would appear consistent.
    Please Note: ripple results on generic/low quality power supplies are difficult to come by. examples that fail often fail early in testing, plus interest and performance focuses, except for occasional "bad example" on psu's that appeal to enthusiasts. Or so it seems.
    Last edited by walterm; 02-19-2011 at 07:10 PM. Reason: spell

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. Has there been any ripple-testing with the in-cable caps removed?
    By schnopsnosn in forum PC Power Supply Discussion
    Replies: 4
    Last Post: 08-11-2016, 12:26 PM
  3. Designing a load machine for PSU testing
    By Ash in forum PC Power Supply Discussion
    Replies: 20
    Last Post: 03-28-2011, 09:01 PM
  4. Load Testing
    By davidhammock200 in forum Testing Methodology Discussion
    Replies: 8
    Last Post: 04-08-2009, 04:30 AM
  5. Harwaresecrets now into load testing [Sunmoon]
    By Super Nade in forum PC Power Supply Discussion
    Replies: 4
    Last Post: 12-15-2007, 11:13 PM

Posting Permissions

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