Extreem low ohms tester?

Discussion in 'General Electronics Chat' started by ErnieM, Dec 22, 2013.

  1. ErnieM

    Thread Starter AAC Fanatic!

    Apr 24, 2011
    7,386
    1,605
    I could really use a very low ohms tester to give some go/no-go indications for 2.5 milli ohms (0.0025 ohm) interfaces. It's a common measurement around here, and I would like to get the assembly guys something to pre-test an assembly before submitting to test.

    A 4 point (Kelvin) measurement is assumed, battery operation is desired but a wall wart is not out of question.

    Anyone have any ideas? It's a simple measurement but pulling out a voltage difference is difficult considering it's 1) near ground and B) small compared to an op amp offset voltage.
     
  2. #12

    Expert

    Nov 30, 2010
    16,257
    6,757
    There's a battery operated 1 amp supply in the finished projects section.
     
  3. Alec_t

    AAC Fanatic!

    Sep 17, 2013
    5,773
    1,103
    What is the limit for current injected through the DUT ?
     
  4. inwo

    Well-Known Member

    Nov 7, 2013
    2,433
    315
  5. Potato Pudding

    Well-Known Member

    Jun 11, 2010
    684
    92
    Seems like you would want a wheatstone bridge type of system.

    Or a small precison constant current source that is split between a device under test and a known good typical comparison value of resistance. An op amp to drive a small analog plus/minus/zero center coil meter and you are done with a homebrew tester.

    I agree with inwo that the test connections needing to be <<2.5mOhm are going to be a pain. They could easily be a deal breaker.

    If your product and requirements support it then an onboard self test might be the way to go except that adds another system that might need to be verified.
     
  6. #12

    Expert

    Nov 30, 2010
    16,257
    6,757
    The OP-07 has typical 60 uV input offset voltage and has trim terminals. It will run on 6 volts (9Vbattery) and you can use a D cell to throw am amp through the resistance to get a 2500 uv DC reading (41 times the typical input error with no trim pot).
     
    • OP07.pdf
      File size:
      870.2 KB
      Views:
      10
  7. MikeML

    AAC Fanatic!

    Oct 2, 2009
    5,450
    1,066
    I would do this by using a transformer with a low-voltage, high current winding on it to create the current through the junction being measured. I would amplify, rectify, and display the AC voltage developed across the junction being measured (effectively a sensitive AC voltmeter).

    The transformer can be driven either with the AC line, or some sort of power oscillator if the device needs to be battery-powered.
     
  8. GopherT

    AAC Fanatic!

    Nov 23, 2012
    5,991
    3,737
    I would track down a Leeds and Northrop precision potentiometer. They essentially have everything on board for the Wheatstone bridge spit out the low ohm values once the bridge is balanced. You will likely get one on eBay for $50. You can have mine for $50 if you want. It was used about 5 times back in the 80s when new. Perfect condition.
     
  9. ian field

    Distinguished Member

    Oct 27, 2012
    4,413
    782
    Measuring low-Ohms wasn't invented yesterday.

    Various magazines have published designs at one time or another.

    The usual method is to energise the resistance with a precise current limited source and measure the volt drop produced - you can make the test current as low as you have a millivolt/microvolt meter sensitive enough to measure the resulting voltage.
     
  10. MikeML

    AAC Fanatic!

    Oct 2, 2009
    5,450
    1,066
    A system based on amplifying an AC voltage is a lot better than trying to do it with DC, which might be contaminated by Galvanic action at the joint being measured...
     
  11. ian field

    Distinguished Member

    Oct 27, 2012
    4,413
    782
    Just cut the current and note the galvanic potential, then subtract it from the voltage measured when conducting.

    Most people do that all the time in respect of test-lead resistance when measuring resistances of a few Ohms.

    But hey - you can build Houston control in the kitchen sink if you want to.
     
  12. Dr.killjoy

    Well-Known Member

    Apr 28, 2013
    1,190
    156
  13. #12

    Expert

    Nov 30, 2010
    16,257
    6,757
    Last edited: Dec 22, 2013
  14. trader007

    Active Member

    Feb 27, 2010
    222
    19
  15. Dr.killjoy

    Well-Known Member

    Apr 28, 2013
    1,190
    156
  16. trader007

    Active Member

    Feb 27, 2010
    222
    19
    Hah i know that's why I said maybe. I should have been more clear though. I just thought maybe that guy would have a practical way to do what op wants.
     
  17. studiot

    AAC Fanatic!

    Nov 9, 2007
    5,005
    513
    I suggest a pulsed DC with a low mark to space ratio.
    This will significantly reduce the power effects of passing a high enough current through the test resistance.
     
  18. #12

    Expert

    Nov 30, 2010
    16,257
    6,757
    I can't quite imagine a power dissipation problem. Throwing an amp through .0025 ohms causes .0025 watts of heat.
     
  19. trader007

    Active Member

    Feb 27, 2010
    222
    19
    im getting a tan just reading that! :D
     
    #12 likes this.
  20. ErnieM

    Thread Starter AAC Fanatic!

    Apr 24, 2011
    7,386
    1,605
    Thanks for all the responses. Let's see if I can bang back all the questions:

    Current limit by specification is either 50 or 500 amps, in any case way above what test current I will inject.

    Commercial $200 meter is OK, I have some in mind (DT-5302), but if I could make a $20 go/no-go box 1) we could get 10 for the same price and B) we don't have to train assemblers to read a meter. (Similar for meter add-ons: why add on if I can just get a meter to do it directly, but go/no-go is a no brainer choice for production people.)

    There is nothing "on board" to add self-test functions. This is simply two pieces of metal being connected. Frequently switches on one side, though all start off as just two pieces of metal.

    No galvanic action as both sides are the same metal.

    It did occur to me (and someone here also mentioned) that by pulsing the test current I could measure the no current and full current potentials, with the difference being a continuous adjusted accurate measurement. Combine that with a current reading and the bonding reading is easy peazy.
     
Loading...