PXI

cancel
Showing results for 
Search instead for 
Did you mean: 

pxi 2575 - 2 wire resistance measurement

Does anyone know why the 2575 is not recommended for 2-wire resistance measurement?

 

thanks

0 Kudos
Message 1 of 6
(2,903 Views)

The armature relay used on the PXI-2575 requires a minimum switching current to ensure the mechanical contact surfaces remain clean.  If that minimum switching current (1 mA on the PXI-2575) isn't met, then the resistance will gradually increase over time, which will introduce measurement error into your resistance measurements.

 

If you want to perform 2-wire resistance measurements, you should use a reed relay instead.  I suspect that since your question is in regard to the high-density PXI-2575, that you're looking for high channel density.  In that case, you should look at the PXI-2530B (2-wire 64x1) instead.  Lower density, but uses reed relays, which are hermetically sealed and won't foul up over time with low switching current.

 

A more accurate option is to use 4-wire resistance measurements, which are immune to increased relay contact resistance.  Of course, this requires twice as many relays, which is why MANY solutions use 2-wire measurements.

 

-John Sullivan
Problem Solver
0 Kudos
Message 2 of 6
(2,883 Views)

Hello John,

Thank you for the reply. I was wondering, does it state that in the specs, " If that minimum switching current (1 mA on the PXI-2575) isn't met, then the resistance will gradually increase over time"?

 

Basically, I am measuring resistance on a DUT that isn't powered on.  The DUT has many pins/channels and I will need at minimum 3 modules, even with a high density module.

0 Kudos
Message 3 of 6
(2,865 Views)

The manual isn't as clear as you've suggested.  The specific wording is:

"Minimum switch load: 20 mV/ 1 mA, Note: The NI 2575 is not recommended for 2-wire resistance measurements."

 

Given enough cycles, any armature relay that isn't hermetically sealed with a noble gas will exhibit this issue.  The phenomenon isn't specific to NI products... NI just chooses to make customers aware that these laws of nature exist.  If you don't want the relay contacts to foul out, you either need to use reed relays (hermetically sealed with a noble gas), or you need to run enough power across the contacts to blast the otherwise accumulating junk off the contact surfaces.

 

FYI: The eventual failure mode if you don't meet the specified requirements is intermittent increased contact resistance.  Purely as an example, you might see the following resistance after a few hundred thousand cycles:

Cycle        Contact Resistance (mOhm)

204,458    524

204,459    801

204,460    25057

204,461    384

204,462    753

204,463    3854

 

In the above example, the four green contact resistances are all within spec, whereas the two red resistances are out of spec.  I'll repeat that the above data is fictitious, but representative of the behavior if minimum specifications aren't met.

 

If you're measuring relatively high resistance measurements - say over 1kOhm - then the eventual increased relay contact resistance might not be an issue.  If you do decide to go with the PXI-2575, you could occasionally run higher current through the relays to clean off the contacts.  That's not a surefire guarantee that the contact resistance will remain low every single cycle, but it does work in principal.

 

If I were you, I'd go with reed relays (e.g. PXI-2530B).  The density isn't as high, but the long reliability will be much higher.  I'd only go with the PXI-2575 if your measurements are well above 1 kOhm and you can deal with up to an additional 20-30 Ohms in your measurements over the long term. 

-John Sullivan
Problem Solver
0 Kudos
Message 4 of 6
(2,862 Views)

Hello John,

Thank you for the detailed response.

I would expect the PXI chassis to deliver the correct amount of current, yes?

 

thanks again,

Jose

0 Kudos
Message 5 of 6
(2,847 Views)

Does anyone know the manufacturers part number for the relays used on this ?

0 Kudos
Message 6 of 6
(2,678 Views)