Switch Types

The important part to take away from the quotes I provided was that they can operate a million times without failure. The point is the makers claim they are reliable and when I ask you for info that refutes this you don't offer anything at all. "Reed switch failures are a long known issue." Do you have any supporting info?

--

Rick C 

Viewed the eclipse at Wintercrest Farms, 
on the centerline of totality since 1998

[...]

I have several hundred Elfein and Pickering SPST reeds with more than 250M commutations behind them and not a single one ever failed. That's pretty reliable for a mechanical gadget.

I'm not so happy with those tiny DPDT Omron relays. Those *do* fail much sooner than I like.

Jeroen Belleman

I have not opened the flashlight , but am confident in has a reed switch. I did verify that a magnet would operate the flashlight.

sed in aircraft? I do not know the answer to that, but can say that they are not used in Submarine launched ICBM's.

Reed switches can be very reliable, but they are not a drop in replacement for conventional switches. If they are used in a circuit that has much cur rent , they are prone to having the contacts weld together. Reed switches do not have much mechanical force to open the contacts. I would not use re ed switches where I could not control what they switched.

Dan

k,

When you say "Omron" relays you mean the ones on boards sold on eBay for very low prices? Yes, I expect they are bottom of the barrel for quality and again, not representative of quality in relays in general.

250 million operations is high reliability for any mechanical device that isn't MEMS. I saw one reed switch that is MEMS. They tout a few advantages which are a bit specialized, but they also have a wider variation in field strength between operate and release. I would need to have as little motion as possible to operate the switch so I would likely not use such a MEMS device. I see 14 mm reed switches for very low prices on eBay, but they are probably not the highest quality. Heck they mostly have photos of switches with bent pins. You have to hold the pin next to the glass body to straighten it and the images show them bent coming out of the glass! Good devices to avoid I think.

--

Rick C 

Viewed the eclipse at Wintercrest Farms, 
on the centerline of totality since 1998

he minimal pressure & lack of wiping. We can't use mercury wetted switches nowadays.

cious or

lope and

d 0 years

their

re

reliable

. Small elastic amounts of metal bending do not make switches reliable! It would be silly to suggest they did. Of course one must stay within their ra ther small specs on v, i and arc energy, and observe wetting current. But t hey still can and do fail, as does every switch on the planet. IRL reeds ar e not one of the best. To suggest an airplane falls out of the sky if a ree dswitch fails is also a little removed from reality.

icient

liable

rse

a switch.

on't bend a lot they must be ultra-reliable is a total nonsequitur.

many years ago, know from experience that reliability is just not one of t heir strengths. And if you understand the basic requirements for a relative ly reliable switch, it is no surprise.

can

hey

r

Your quoted claims I don't buy. I explained why before. It's your issue not mine, and up to you whether you want to go digging. All I can tell you is I did run into reliability problems and I was far from the only one.

NT

Ok, I'll take your comments for what they are worth.

--

Rick C 

Viewed the eclipse at Wintercrest Farms, 
on the centerline of totality since 1998

The failing relays are Omron G6K-2G-Y. No, I don't buy on eBay. These relays would have been very nice to switch RF, if only they would last. Very low parasitics. I have about 200 in one system here but they keep failing with dodgy and intermittent contacts. I've been wondering if they'd been abused somehow during production.

Jeroen Belleman

Oh dear, I just was doing BOM for a board that has a few G6K-2P-Y relays. Have you tried the ones JL likes? (Fujitsu I think... FTR-B3 or -B4??)

George H. No, I don't buy on eBay.

No, but they seem to be pretty much identical.

I thought I could gain some space, since one of these little G6K-2G-Y relays did the job of four SPST reeds, but I won't make that error again.

I've used Elfein and Pickering SPST reeds in the past. Never a problem, even after more than 250M operations.

Jeroen Belleman

The Western Electric #5 ESS used "ferreed" switches, that were basically polarized reed contacts in a big memory core. These were the active switching element of the telephone switching system. The coincident current scheme, just as used in the core memories of old, allowed them to select one out of thousands of reed switches with a number of drivers that was 2X the square root of the number of switches. When these switches (reed) numbered in the tens of thousands, saving on drivers was desireable. As far as I know, these reed switches generally lasted the life of the ESS system, except for possibly some damaged due to lightning events and such.

Jon

Another thing that ran reed switches hard was a tachometer made for the mucsle car generation. They had a coil that was connected to the ignition points and a SPDT reed that charged a cap and then connected the cap to an RC filter and the meter movement. So, the reed switched every time one spark plug fired. At 3600 RPM, a V8 would do 4 sparks/ revolution, so that is 240 times a second. The reed would take this for thosands of hours of engine operation (although not all at 3600 RPM).

Jon

Many years ago CP Clare had a problem with reed contacts failing. They had made a "minor" process change, and been chugging along making the reed switches when they started coming back from the customers. My friend ran an electron microprobe, and they brought some sold contacts and new ones by to see if he could see a difference. He quickly found the new ones had some element showing up that was not on the old ones. I don't recall now whteher it was Sodium or Phosphorus, but something like that. As soon as he told them what the element was, they knew it was a cleaner they had introduced. Apparently, either washign that residue off or goign to a different cleaner solved the problem. But, by that time, they had made many millions of switches that were going to fail early.

jon

Oh Dear, what did they do? The minimum IMHO is price of relay or a replacement. I think if you tell people, "Look I/we f'ed up, here's the replacement and you have to put it in." and then pic's or a video. Then customers will appreciate it, as long as it's once in a while.

We've sent out circuit repair 'updates', mostly diodes or zeners with those resistive fuse things. I think we get about 25-30% of our customers asking for the 'update'. (Others may put it in themselves, or our emails are to the wrong person, or it's still in the box.)

George H.

Well it's good they told you, less good when you have to find it.

George H.

snipped-for-privacy@gmail.com wrote on 11/19/2017 9:12 PM:

There seem to be a number of users of reed switches responding here who are very happy with the reliability. Obviously your experience or whatever you are basing your comments on differs from theirs.

--

Rick C 

Viewed the eclipse at Wintercrest Farms, 
on the centerline of totality since 1998

Jon Elson wrote on 11/20/2017 4:09 PM:

I had a bad experience with CP Clare where they made a pair of chips for a POTS connection circuit where one chip was designed so it would not work properly from day one and the other had 0 dB of PSRR. The PSRR problem could be worked around if you were aware of it (they didn't spec this in the data sheet) and the other chip basically was not usable. After that CP Clare was black listed at my company.

--

Rick C 

Viewed the eclipse at Wintercrest Farms, 
on the centerline of totality since 1998

The problem with magnetics, reed or otherwise, is hysteresis. That may not be an issue for your actual application, but I think it would be for your example of a Morse key.

Cheers

--
Clive

Why is that?

--

Rick C 

Viewed the eclipse at Wintercrest Farms, 
on the centerline of totality since 1998

Really? I can't imagine using a conventional Morse key with hysteresis.

Cheers

--
Clive

For a Morse key user, the mechanical feedback from the contact is essential for good code input. A switch without mechanical touch (e.g. magnet + reed) is useless.

--

Tauno Voipio (OH2UG since 1960)