Intermittent "short" to case of rotary switch

I'd say it is, going by

-- john, KE5FX

I don't know Elma, we do use some NKK switches I haven't gotten to the point of searching for a different vendor. I've really liked grayhill, they were helpful in the past when we had a problem. I sorta thought they would be all over wanting to know if there was a potential problem... but the feeling I got was that they would be doing me a favor by looking at the failed switch. It's not all that hard to understand the attitude. I get calls all the time from people who can't make things work and think something is broken... I'd say that less than half the time it's a real failure.

George H.

Yeah I should check NKK. Getting all the foot prints and panel heights right is problematic. Is whisker growth moisture related?

George H.

John, thanks for the very nice article... (much more than I want to know.) I was looking at our inventory program, and we've put about 1,000 of this style switch into field in the last 3-4 years. (with probably equal numbers of their two pole switch.) I was mostly thinking last night about how my customers could "clear" this fault, without having to replace or return the switch. Blowing with compressed air? Flooding with IPA? Zapping (over current melting) with some voltage. (The last idea came from your above article.)

It would be nice not to have any returns, which are expensive. So I may not get the chance to take another picture... but if I do I'll be sure to post it.

George H.

Nobody seems to have a good handle on that. Dendrite growth and creep definitely require chemistry - ionic exchange reactions involving moisture.

I've seen whiskers growing on plated tin over copper over 260 brass which bubbled when heated which I assumed meant that a lot of hydrogen was in the plating. The parts that whiskered were in an area that could see some condensation during use - same parts in dry areas didn't whisker. I grumbled to the plating company and they added a reflow step which is supposed to eliminate stress and drive out the hydrogen - didn't bubble anymore and I never saw whiskers on those parts after that.

I suspect that moisture is not always required, because some of the whisker failures have been in satellites, specifically in parts under vacuum where the whisker can initiate a metal vapor arc: (3)

Looks like some progress is being made:

Just published in SMT Magazine

Well here's another piece of data, the two failures I've had were the same switch on the same relatively new instrument. There's a gain of 20 after the switch so in normal operation not that much voltage. But it's a student instrument... There's a gain of 5-50-500 in front of the switch so no problem to rail it. (Heck, I think at some point in the manual, I invite them to plug a function generator in and measure how well it works.) The switch is just taps on a string of 1k resistors to ground.

I could set some prototype up on my shelf and bang it around forever, I wouldn't be surprised if nothing happened.

(that sounds like too much work... and my shelf is a mess.)

George H.

The quickest thing I can think of is to saw open the existing failed switch and see if there are whiskers inside using a microscope. At least it should be the same material that caused the failure, and the right age, possibly unlike some other switch from a different batch.

Chris

Well I took apart a different one to see how it was made. There's not any potential for contact between the pins and the outside housing... It's all shielded with plastic bits. I did however look at the bad one under a microscope and I can see a a bunch of little whiskers... I don't have anyway to take a photo though.

George H.

I had a bit of an exchange with Xilinx a couple years ago. I was making some boards that had a variety of analog stuff on them, and one Xilinx FPGA. I started finding shorting slivers on the boards, and at first though these were somehow left over from the lead frame manufacturing process, and were released when the solder was reflowed. This might not have been what was happening, as the slivers were sometimes seen to run under SEVERAL leads. I would think a released sliver couldn't do that.

But, anyway, at a later date, clear Tin whisker formation was seen at the places where the leadframe was bent in the gull-wing shape. Xilinx asked about the reflow profile, and I told them I only went to about 230C as I was using a tin-lead process. They said this was the problem, I was not getting the chips hot enough to relieve stress built into the lead frame when it was bent. They said if I heated it so the CHIPS went up to 240 C, then the problem would disappear.

I wonder if the crimping of the switch body puts stress on the plating of the housing, and later causes whisker growth.

And, yes, whisker growth is associated with humid conditions.

Jon

I have had pretty good luck holding a camera up to the eyepiece of a microscope. Some sort of adapter to ensure alignment would make it easier, but I haven't got around to making that, so I just take a lot of photos and pick the best one.

Chris

Both failures from the same customer? If so, what environment does the switch live in?

I agree with Grizzly. That "fuzz" could conceivably be due to exposure at the customers and could easily be conductive.

JohnS