Reed switch failure modes

We're seeing some failures (about 0.1%) of reed switches we use (Hamlin

59045-T). They seem so get stuck in closed position in some cases. The switches are used to activate the device from outside the waterproof case with a keyfob magnet.

The circuit has a 1M||100nF pulldown+filter and reed switch from there to +2.8VDC.

I'm now wondering what is going on. Some ideas are:

-Too strong magnets used for activating the devices

-Welding of contacts (with 100nF 2.8V ?!)

-Moisture destroying something (supposed to be hermetic glass!)

Sometimes the switch can be opened by knocking it, sometimes not. The reed switch is moulded in plastic, so opening it up typically causes enough mechanical vibration to destroy the fault before getting there.

Any ideas what could be the culprit ?

--
mikko OH2HVJ
Reply to
Mikko OH2HVJ
Loading thread data ...

A microscope will tell you if the contacts welded, even after decapping breaks them apart.

Cheers

Phil Hobbs

Reply to
Phil Hobbs

100nF at 2.8V is 0.4uJ nominal. Hamlin claims that below 1uJ welding should not be a problem, others (such as NI) urge caution with cable capacitance of hundreds of pF. Personally, I would definitely use a series resistor.

I guess you could test the remanence theory by measuring the pull-in voltage before and after exposure to a massive rare earth magnet.

--sp

--
Best regards,  
Spehro Pefhany 
 Click to see the full signature
Reply to
Spehro Pefhany

I am suprised your failure rate is so low! Reeds love to dry-weld. I have experienced much worse rates with much smaller capacitors and now routinely add series resistors - in the tens of ohms typically, enough to limit peak current to 100mA or less.

Somewhere there is an HP Journal article which mentions they added 22 ohm series to reed contacts switching ATE backplanes because even the

10s to 100s of pF wiring capacitance would weld reeds!

piglet

Reply to
piglet

If there is some double digit uH of stray series inductance it may be enough to save the contacts, but it will ring away for a while with a good capacitor. A series resistor ends all that nonsense.

--sp

--
Best regards,  
Spehro Pefhany 
 Click to see the full signature
Reply to
Spehro Pefhany

Yes. Just 10 ohms should make it safe.

Reply to
John S

Reeds are terrible. They can weld from arcing, weld from just being closed too long, make high-resistance contacts, or just plain forget to open or close. Mag fields, like from other reed coils, can keep them closed. And they are big, expensive, and have terrible thermoelectrics and reed twang noise. Bad news all around.

Can you use a more conventional relay, or better yet and SSR?

--
John Larkin         Highland Technology, Inc 

lunatic fringe electronics
 Click to see the full signature
Reply to
John Larkin

How would you suggest he implement that in his described environment

Reply to
John S

Hall effect sensor maybe. There are halls that act like open-collector switches, all in one chip, tiny and probably cheaper than a Hamlin reed. Certainly more reliable.

--
John Larkin         Highland Technology, Inc 

lunatic fringe electronics
 Click to see the full signature
Reply to
John Larkin

Be sure to extract some "good" ones to compare, as well as some brand- spanking new ones.

Might be repeated damage going on that's not showing up until the uh-oh moment.

--
Tim Wescott 
Wescott Design Services 
 Click to see the full signature
Reply to
Tim Wescott

There was a big problem with CP Clare reed relays about 30 years ago, due to contamination before the contacts were put in the glass. (A friend of mine did the electron microprobe study, and found, I think, Sodium deposits, which had no business being inside the switch.)

So, you have a 100nF cap directly across the contacts? What current do you get when the contacts are closed? Could be tens of amps for a few us, certainly enough to weld contacts, assuming minimal resistance in the path. Do you actually need that HUGE a capacitor? Or, you could put a 100 Ohm resistor in series with the contacts, so you eliminate the current pulse.

Jon

Reply to
Jon Elson

Another failure mode for your list is shock damage. The internal contacts are tiny cantilevered beams, and it doesn't take a big hit to eliminate the few mil spacing of the contacts.

You can do a rough test of relative sensitivity of specific parts by recording the switching distance from a weak magnet using a caliper setup. The switching distance will be inversely proportional to the contact gap. If the problem is shock damage, you should see a range of sensitivities and the most sensitive ones may be on their way to failing closed. Of course you need repeatability of reed location inside the plastic package, which you likely don't have.

Maybe you could replace the plastic packaged reeds with unpackaged reeds. At least then you would be able to see the problem.

--
Best Regards, 

ChesterW 
 Click to see the full signature
Reply to
ChesterW

We went from reed switches to Hall effect devices for shaft speed measurement ( low speeds, 100 - 5000 rpm ) about 30 years ago, for exactly the same reasons. But its true that a few ohms in series with a shunt capacitor can vastly improve the reed life.

The one advantage of a reed is that it does not draw power in the off state, so using it as an on-off trigger is much more power efficient than a Hall device for battery operated systems.

Reply to
Adrian Jansen

Some Hall effect switches operate on low single digit uA.

ChesterW

Reply to
ChesterW

The Bell System always ran a tiny DC sealing current to clean the contacts, this was one of the reasons why. They invented the reed, they should know why. You might want to read up on them in BSTj.

STEVE

Reply to
sroberts6328

Mikko, I've seen a paralleled 100nF edm a hole in a snap/dome switch! I shudder to think what it would do to a reed.

Count me in the series resistor camp , definitely.

Cheers, James Arthur

Reply to
dagmargoodboat

Welding from the capacitor, as others have said, plus the slow removal of the field causing nearby metals to remain magnetized. Together it's getting stuck at a much lower threshold than you'd think.

--
I will not see posts from astraweb, theremailer, dizum, or google 
because they host Usenet flooders.
Reply to
Kevin McMurtrie

This is a bad case, as the reed gets cycled thousands of times per minute, in most cases, and can use up the rated life in a week or two. So, there, a Hall sensor makes very good sense.

Jon

Reply to
Jon Elson

True, forgot about using more than 20X Mantis!

--
mikko
Reply to
Mikko OH2HVJ

Sounds like your experience matches our case. I measured over 5A peak currents through the switch. Probably the original designer thought that a tiny 100nF cap could not do any harm. Neither would have I.

Next PCB revision will have series resistor, I'll have some statistical data by the end of the year and will make a calendar note to report back here!

--
mikko OH2HVJ
Reply to
Mikko OH2HVJ

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.