Silver contacts lifetime - relay switch 3

[kr7]

Hi, I am trying to figure out the reliable lifetime of the silver contacts in a pressure switch. I am trying to use this switch in a logic circuit as relay control. Since the switch serves a failsafe function in the device (it senses over pressure in the system), it should never close, and thus I am worried about the silver contacts tarnishing over time. The device is a medical device that is going to be used in controlled environments (such as hospital, physician's office, etc), and we're aiming for a lifetime of 7-10 years.

If the switch were to close, it will be driven by 5V and carry a load of 10mA. The maximum resistance in the switch my circuit will tolerate is ~80 ohms.


would anybody have opinions on whether silver contact would be reliable for this application for 10 years, or know where I may be able to find information that could help me figure it out?


thanks much,
Ethan

 

[Skogsgurra]

You should reverse the action. A contact that is supposed to close and then activate something like a relay is not a fail-safe design and against good practice in safety related designs.

Gunnar Englund

http://www.gke.org

 

[kr7]

hey Gunnar thanks for the reply. yea we thought about doing that but the device has already been designed and sold that way for years, changing the design now would cost a lot, as well as create compatibility and manufacturing book keeping headaches... so I guess we just gotta figure out how to make things work the way it does now...


Ethan

 

[Skogsgurra]

OK,

There is an old book (my is from 1986) named "Modern Relay Technology" by Hans Sauer. It discusses "your" problem and has many design data. Try to get hold of it. Or talk to relay manufacturers. They know a lot about such things.

Gunnar Englund

http://www.gke.org

 

[itsmoked]

Just remember if your logic circuit is less than 9V DO NOT use anything but gold contacts. All other common contacts build an oxide layer that insulates the contacts in a short time. The voltages above about 9V are high enough to punch thru this oxide layer.

 

[kr7]

thanks Gunnar I'll try that book if I can find it.

itsmoked I think silver may be an exception to the 9V rule. I've just tried a sample switch that I've used in a lifetime test for a couple weeks in an equivalent 5V circuit, and it does form a reliable electrical connection.

the book "Silver: Economics, Metallurgy, and Use" by Allison Butts says that silver, unlike most other metals, actually does not oxidize. although it does tarnish by reacting with sulphor in the atmosphere and form silver sulphide.

silver sulphide, however, is also conductive. it's 10,000x less conductive than copper, but a thin film of it would still not have much resistance.

from a purely resistance point of view, it looks like silver would be absolutely fine for my application. (the silver sulphide deposit would need to be one foot long and thin as a small gauge wire to give 80 ohms of resistance)

but I'm not sure what other factors need to be taken into consideration. There must be a reason why people say you need to use gold contacts for low voltage/current circuits.

I am not sure what "low voltage/current" is defined, and I don't know if my circuit (5V, 10mA) would be considered "low voltage/current" in this sense.


Ethan

 

[itsmoked]

"There must be a reason why people say you need to use gold contacts for low voltage/current circuits."

YES! It doesn't work otherwise!!!

Hi kr7,
I can guarantee, you WILL have reliability problems with non-gold solutions in five volt circuits. I thought the same thing you do.[several times] I was wrong. As are you, if you think it is going to work.

Here's how it will work. You will put it all together. It will work fine. You will try lots of tests.. They will all work. Then you will ship. A year later the problems will start creeping in growing to epidemic levels.

If for some reason you must use silver or any non-gold contacts you need to put in a 5V capacitor doubler and run 10V preferably 12V through the contacts. You must! Period!

If this is a "fail safe" you have to do this with something that is reliable. Silver is not.

Don't forget you have already decided to use the wrong sense for fail safe. Just wait for the contacts to do nothing when they finally close.. ["Jury. What say you?]

 

[Skogsgurra]

Smoked is right.

Your example with a thin "wire" made from silver sulphide is pure nonsense. It is not about resistance of the material - it is about breaking isolating films on the surface.

The following table gives voltage ranges for reliable operation with different contact materials in dry circuits:

Ag >12 V
AgCu >12 V
AuAg >1 micro V
AuNi >100 milli V
AgPd >1 V
AuAgPd >1 milli V

These data are from the book I mentioned.

So, if you are stuck with a large base of installed silver contact units which cannot be replaced with gold contacts, you have a problem.

One way of solving it would be to increase the voltage over the contact. I assume that the contact is "potential-free" (available to you on two wires with insulation to rest of circuit). A voltage doubler will almost do the job, but if you have a 12 or 24 V source available in the design, I think that you should use that. Then see to it that you load the circuit when the contact closes. A resistor that pulls about 50 mA works. You can use about 10 mA if you combine the resistor with a parallel capacitor that causes a small surge to clean the contact point, but make sure that the contacts do not weld. Usually not a problem unless you are using a reed element.

The other solution - and that's the one I would recommend - is to change the design. Offer the clients a "safety upgrade". Tell them that the equipment can be a lot safer if they change to this new fail-safe supervision. Make the price for the upgrade attractive and get rid of this situation once and for all.

If your superiors tell you that "changing the design now would cost a lot, as well as create compatibility and manufacturing book keeping headaches" then tell them that the first time the device fails will "create court and lialibility headaches".

Good luck!




Gunnar Englund

http://www.gke.org

 

[Noway2]

This Jury Member agrees. I have encountered plenty of cases where the relays appear to work throughout testing and early into the product life cycle, only to fail after time.

You also mentioned in your post that this is a medical application. While I am not up on the medical requirements, I understand that they are typically pretty stringent when it comes to safety issues. There may be some specifications or other 'legal' documents you can reference that will answer the question for you.

 

[kr7]

Thanks guys, it's so great to hear opinions from other professionals. I think we will now be seriously considering reversing the design to have the switch normally closed as a failsafe.

Would silver then be an acceptable contact material in that case? Or has any of you experienced many "false alarms" later into the product life due to poor electrical contact?


thanks,
Ethan

 

[itsmoked]

many, Many, MANY! times I have experienced silver/logic problems and they are the worst kind of flakey works-most-of-the-time but not always problems.

 

[Skogsgurra]

It will, of course, be a safe solution and I am glad that you are thinking about a safer way of doing it.

The problem is - as Smoked pointed out - that you will have false alarms if you use silver in a dry circuit. There are ways of dealing with intermittent false alarms (like time delays) but it still is a problem.

Gold is recommended if you cannot work with a higher voltage and good contact loading.

Gunnar Englund

http://www.gke.org