Solved! Go to Solution.
As relays connect and disconnect, they can over time build up a less conductive oxide across the contact terminals. This oxide is caused by many things, such as surface impurities, airborne impurities, and overcurrenting or overvolting the relay. For applications with voltages and currents over the minimum switching specifications, the spark that occurs each time the relay operates aides in preventing this oxide from building up to a state that the lowest path resistance is through only the oxide and not the metal surface. In essence, the spark acts as a broom to brush off the junk that accumulates on the relay contact. Too much spark will accelerate the failure mode of the relay; too little spark will prevent the switch from being able to 'clean' itself, and will add surface impurities each time the relay operates.
What you would see in an undercurrent/undervoltage situation is the gradual increase of path resistance as the device switches. If we were to occassionally place a larger test current through the relays and operate a few times, we'd burn off the contaminates and thus reduce the impedance.
So what can we do about this? The best option is to use a reed relay, which is a relay contained inside either a vacuum or a noble gas. Reed relays, therefore, are not exposed to the airborne impurities. To reduce the surface impurities (a property of the relay contact), most of our reed relays have contacts covered with rhodium, which has no stable oxides. Any oxide that forms from rhodium will sublimate the oxygen back out over time, thus preventing a metal oxide from building up on the contact.
The downside of using reed relays is their limited current handling ability. This doesn't apply in your application because the current you're using is very low, but it's good to keep this in mind if you ever decide to use reed relays in another application.
This doesn't mean that electromechanical relays won't work for DMM measurements. For example, when we measure voltage the DMM has an input impedance of say 10MOhm. So an increase in path resistance caused by a switch from the specified 1 ohm to 10 ohms isn't going to substantially affect the system. On the other hand, if we're using the same DMM for a current measurement, this path increase will introduce a sizeable error.
I would recommend looking at the 2530 for a similar topology that uses reed relays.
Hi! All! Sorry for long silence, it was time for holiday 🙂
John S, aNItaB, Than You for answer and recomendations, I will think about 2530 and Digital Multimeters (In fact I have to use one our DVM, but in future I will need some other DVM).
After your answers I was still confused, I have not enough understand - what is the minimum current? And I have started my own investigation: I have made some experiments: I have serial connected amperemeter, load (10 kOhm) and with help several (from 1 to 10) 2529 switches I have connected power supply. I change voltage of power supply and measured current. So there is no any problem to pass current 10uA, and smaller like 1 uA.
After I star investigate the AGQ210A4H (it is relay used in Matrix) datasheet (http://www.panasonic-electric-works.com/catalogues/downloads/relays/ds_61008_0000_en_gq.pdf) and some relay technical information from AGQ210A4H manufacturer (Panasonic) http://www.panasonic-electric-works.com/peweu/en/downloads/ds_x61_en_relay_technical_information.pdf. Because probably this minimum current parameter has been take from relay datasheet.
And it is said that AGQ210A4H have a Min. switching capacity: 10μA 10 mV DC. And there is some explain about Minimum switching capability This value is a guideline as to the lowest possible level at which it will be possible for a low level load to allow switching.
But this explanation only adds some confusion, because, what is the reason for minimum limitation of load? Because if we have a low load, the main idea, that its voltage can be enough big for compare with thermal EMF, and current will be limited with resistive of relays and resistors of connected circuits in case voltage source using is all scheme.
After that I try to find "Min. switching capacity relay" in Google. And I have found this: http://answers.google.com/answers/threadview/id/519904.html Most interesting answer (it is similar to John S answer) - "...The current required to remove the oxide layer and establish a metal to metal contact and hence an electrical connection is also described as a wetting current..."
So here is the questions:
Is it meaning that minimum current specified in 2529 specification it is “wetting current” or sealing current (like here: http://en.wikipedia.org/wiki/Wetting_current)?
And what is the consequence of long using relay for connect DMM like voltmeter (over minimum current pass through the switchers)?
Is it mean that after some years of such matrix using conductive oxide layer become enough big and I will lose the connection?
Is this you try to explain me in the first post?
To answer your questions, yes the minimum current is very much like the wetting current, it is the current that is needed to consistently knock the junk of the relay to make sure it keeps working properly. The NI Switches Help defines the minimum current as "the minimum current that can reliably flow through the switch". So, while you'll likely be able to measure a signal even if you're not meeting the minimum requirement, we can't guarantee your switch's behavior if you're operating out of specification. If you are operating below the minumum current specification for extended periods of time, the contact resistance of the relay will increase over time, and your relay will eventually become unusable. Essentially, by not meeting the specification, you are going to shorten the life of your relay.
As for your question, "Is it mean that after some years of such matrix using conductive oxide layer become enough big and I will lose the connection?" The answer is yes.
Hopefully that is a little more clear, let me know if you still have questions about it though. Have a good one!
I just want to reiterate that periodic higher currents will increase the life of the relay. For example, say you're running 10nA through a relay that is rated at 1uA minimum switching current. The 10nA won't provide enough of a spark to clean the gunk that accumulates on the relay contacts, which means over time the path impedance will increase.
If we run 10nA through the relay most of the time, but every 100th cycle or so we run 10mA through the switch, then we can reduce the accumulation of gunk that remains on the relay, which will increase the life of the relay.