2
\$\begingroup\$

I'm doing a project with an Arduino pro mini 3.3v that uses internal pull-up to detect changes on a reed switch sensor. I'm buying the reed sensors from AliExpress, and most of the time they work OK without problems, but in some random cases they suddenly stick together... if I tap on them, they open again and work properly... which is worse because it is hard to test if they will work for long term.

I cannot allow this to happen since what this sensor measures is the main purpose of the device, and I cannot sell a product knowing there is a good chance this can randomly happen, making the device unusable.

For now I'm using 2 reeds in series, and doing some failure detection in code. But I would like to ask if someone already had this problem, and ask for a good reed switch brand that is more reliable and maybe spare me this problem for good. I looked for good brands in the internet, but I haven't found enough information. I was thinking maybe also buying plastic ones since the glass ones break so easily.

FYI, the magnet is a small cylindrical 3mmx3mm neodymium magnet, which I guess is not too strong for a 14mm glass reed switch, at a distance of 2-5mm.

Thank you for your help!

\$\endgroup\$
6
  • 1
    \$\begingroup\$ if you have capacitors in parallel with the switches can can cause the contacts to weld. \$\endgroup\$ Commented Apr 21, 2021 at 4:51
  • 1
    \$\begingroup\$ My only extended experience with reed relays in commercial products in large numbers was with keyboards that used to be built "back in the day." (Early and mid 1970's when electronic keyboards were real electronic keyboards.) There, we used cylindrical magnets, too. But they were NOT the fancy, over-powered neodymiums of today. When the key was pressed down, the magnet centered over the reed body and and the reed switch closed. When released, the magnet moved out of position and the reed opened up. They were extremely consistent devices. I don't recall anything like what you describe. \$\endgroup\$
    – jonk
    Commented Apr 21, 2021 at 5:12
  • 1
    \$\begingroup\$ The first thought that comes to mind is that your source isn't supplying quality devices. Another thought that comes to mind has to do with a question: are these mercury-wetted? And finally, I think that you should consider a minimally powerful magnet for the purposes. Just enough to be absolutely sure of consistent behavior, but no more than that. On that last point, can you consider a lower powered magnet? At least for experimental purposes to see if it improves the situation? (The distances I experienced were in the "few mm" range, much as you describe, by the way.) \$\endgroup\$
    – jonk
    Commented Apr 21, 2021 at 5:14
  • 1
    \$\begingroup\$ What voltage? What load (ALL attached components on both sides of switch.) I have used hundreds of thousands of Reed switched with good reliability. \$\endgroup\$
    – Russell McMahon
    Commented Apr 21, 2021 at 5:28
  • 1
    \$\begingroup\$ "the magnet is a small cylindrical 3mmx3mm neodymium magnet, which I guess is not too strong for a 14mm glass reed switch, at a distance of 2-5mm." - how close does the magnet have to be to close the reed, and and how far away to open it? \$\endgroup\$ Commented Apr 21, 2021 at 5:41

1 Answer 1

2
\$\begingroup\$

Reed switches are very simple devices and usually have only a few failure modes. Here's some advice on how to troubleshoot:

The reeds themselves have to be magnetically very "soft," and if you are purchasing them from a source that does not do a good job of material selection and annealing, the reeds themselves can become slightly magnetized by an external magnetic field and "stick." You can determine if this is your problem if your reed switches fail in this fashion when using a meter to measure the contact pins, (or if they are failing in a circuit where the current is low). If this is a critical application, I would suggest purchasing through a reputable electronics distributor.

If your reeds don't stick with a meter and are only sticking in your device, it may be possible that your energizing magnet is still too close when you want to open. When the magnet moves closer to the switch, there is a gap between the reeds causing a higher reluctance, so it takes a larger field to magnetize the two reeds until they are attracted together. Once in contact, the reluctance is lower so the field strength to keep them magnetized is less than that required to pull them together. This means that the magnet position at which the reed switch opens must be farther away from the reeds than the position when the switch pulls in. Your problem could be this, and could be made worse if the required field strength varies from switch to switch. Once again, I would recommend buying from a reputable supplier; good reed switch manufacturers usually specify a narrow range of field strengths.

Bending of leads can change a reed switch's performance if not done properly. Leads must be supported during bending so that the bending stress is not experienced by the glass capsule. I recommend the glass over the plastic because the contacts are in a hermetically sealed atmosphere for long life.

Finally, as @Jasen correctly points out, if you are switching any capacitive loads you can be introducing instantaneous currents capable of welding the contact surfaces. This will happen regardless of the supplier in a poorly designed circuit. For sensing applications, make sure you have series resistance to limit the instantaneous current through the reeds.

Good luck!

\$\endgroup\$
3
  • \$\begingroup\$ Thank you for your response! I will try with a good manufacturer, do more research about the magnet strength and reed sensibility. Any good reputable supplier to recommend? I'm new to electronics :) \$\endgroup\$
    – ernewston
    Commented Apr 21, 2021 at 15:57
  • 1
    \$\begingroup\$ I use Littelfuse (formerly Hamlin). Try Mouser or Digikey for distribution - they will have a good selection. \$\endgroup\$ Commented Apr 21, 2021 at 17:01
  • \$\begingroup\$ Awesome, thank you!! \$\endgroup\$
    – ernewston
    Commented Apr 21, 2021 at 17:12

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.