Failing relays on PXI-2569

[jplotzke]

Hi All,

I have a PXI-2569 (100 SPST relays) module which has relays which are failing.  The relays that are failing show an exteremly high resistance (> 15 ohm) when they are first closed and then actually decay in resistance to a lower resistance (typ 4-8 ohm after about 30 seconds).  This is just measuring with a instrument connected directly to the pins on the terminal block.  I've replaced one of the relays according to the relay replacement guide -- That fixed the problem with that relay, but over the next few days, 3 other relays on the card started showing the same symptoms.  (and the newly failing relays aren't close to the one I replaced... so I don't think the heat from soldering caused a problem on others)

Has anyone seen this before with a SPST relay module?  Is there a minimum wetting current required for these relays (I could not find any specifications for this in neither the PXI-2569 specifications nor the IM42GR relay specifications)?  This is really bothering me on how this is happening.

Thanks for any suggestions!

Jeff

[Now_With_Underscores]

Hey Jeff,

Page 2 of the PXI-2569 specification sheet lists a minimum wetting current of 20mV/1mA.  The failure mode you're observing is typical if we do not meet these minimum values, as the sparking action forms a solid connection.  If we exceed the minimum wetting current, the resistance will come back down after a few cycles (after any surface debris is removed from the surface by the spark).  Obviously, if we then continue to use the module below the wetting specification, we'll see the contact resistance rise again.

Note that Voltage measurements and 4-wire resistance are not affected by this level of resistance increase.  2-wire resistance, current (below 1mA), and thermocouple measurements are affected.

Wetting current is a phenomenon of mechanical relays.  If you need to switch low voltage/low current signals, I'd recommend either reed, fet, or SSR relays (note that NI does not offer an equivalent 100 channel SPST module in reed, fet, or SSR). 

If you're not able to increase the Voltage/Current to above the minimum levels, we can discuss how to implement a cleaning algorithm that will attempt to reduce the resistance without manual user interaction.  Post up with your system parameters and we'll see if we have any options.

[jplotzke]

Thanks, John.

Thanks for pointing this out.  Some of these relays pull a logic line to GND, so they pull very little (in the order of uA of current), definitely below the 1mA wetting current spec.  My system configuration actually has the PXI-2569 going out to a generic interconnect adapter with test-specific ITAs being plugged into that.  Without making a "relay cleaner ITA", I'm going to have a hard time getting to the wetting current.  Do you have any suggestions on things to do if the wetting current isn't acheiveable?

Jeff

[Now_With_Underscores]

You can typically pull a logic line to ground with a 10k resistor (or more).  The resistance on these relays is going to be much lower than that regardless of how many times we switch, so this shouldn't be an issue unless you're pulling a relatively high load to ground.

If the actual resistance is an issue despite the above, we could also add an additional resistor to pull more current when the relay closes.  At 5V, we'd need to add 5k dummy load resistance to get to 1mA.

[MOBMOB]

We also have the PXI-2569 under the following conditions.

All relays are set with Zero load,  after relays have settled, we apply power.    So there is zero current,  is this condition ok and not cause a high resistance load?

Thanks Marty

[MOBMOB]

Sorry I ment to say :  "So there is zero current to start, current will flow when we turn on power (relays are already set).  "    Instead of saying: So there is zero current. 

[Now_With_Underscores]

Hey MobMob,

I assume you're also setting the current to zero before opening relays, too?

How much current are you switching?

What is your DUT?  Can you vary the current level at will?

Reed relays - NOT used on the 2569 - are sealed inside a glass enclosure filled with a noble gas that prevents oxidation.  Thus, reed relays don't require a minimum current to clean the contacts (there's no particulate to build up on the contacts).

Armature relays - used on the 2569 - can handle more current than reed relays, but aren't typically sealed in an inert environment.  Thus, particulate matter can build up on the contacts, and since there's oxygen, the contacts can corrode.  The minimum current spec listed on armature relays (e.g. 1mA/20mV) ensures there's a small spark across the contacts as they come into contact with each other.  This spark 'cleans' the relay surfact by literally blowing contaminates out of the way.  The relay contacts are coated with elements that do not have stable oxides (e.g. platinum, rhodium, palladium, etc), which prevents the contacts from oxidizing.  However, if your switching current is very high, the spark can blow through the non-oxidizing metals, at which point the metal will oxidize faster than the sparking action can 'clean'.

My recommendation is to set the switching current above the minimum switching specs (20mV/1mA for 2569, so set to maybe 10mA).  If you're not able to do this, then I'd go with 0mA while switching, since you can always go back and switch current through the contacts to clean them, whereas you can't go back and undo the damage caused by overcurrenting the contacts while switching.

Post up with more details and I can provide a more directed response.  Happy switching.