Fail Of The Week — Accidental Demagnetization

There’s a trick in the world of plastic enclosures. The threaded insert is a small cylinder of metal with threads on the inside and a rough edge on the outside. To make a plastic part with a hole for securely connecting bolts that can be repeatedly screwed without destroying the plastic, you take the threaded insert and press it (usually with the help of a soldering iron to heat the insert)  into a hole that’s slightly smaller than the insert. The heat melts the plastic a little bit and allows for the insert to go inside. Then when it cools the insert is snugly inside the plastic, and you can attach circuit boards or other plastic parts using a bolt without stripping the screw or the insert. We’ve seen Hackaday’s [Joshua Vasquez] installing threaded inserts with an iron, as well as in a few other projects.

This trick is neat. And I’ve now proven that it does not work with neodymium magnets.

It happened while working on a new product. I’m using a plastic part as a rotating cover. In one position it covers two holes. In the other position, a magnet embedded in the plastic moves over a magnetic reed switch, supplying power to the microcontroller and turning the device on. It’s a slick way to turn on/off the device without a visible mechanical switch. Unfortunately, the magnets kept failing to trigger the switch. Eventually I discovered that the magnets were losing their magnetism when I was trying to press fit them into the plastic with the aid of a soldering iron. This was not a simple problem to troubleshoot.

We covered the basics of magnets nearly two years ago, and there is a specific property of magnets that tripped me up. Once heated up, magnets can lose their strength. For some metals these temperatures are pretty high. For neodymium, it happens to be very low.

Working Temp, Irreversible Loss, and Curie Temp

The max working temperature of a magnet is the temperature below which there should be no loss of strength, and for neodymium magnets this varies by grade of neodymium. The temperature for N-type comes in at 80°C, while AH-type is workable up to 230°C.  Since ABS extrusion happens around 230°C, the temperature needed to insert a magnet will almost certainly be above the max for the magnet. Fortunately, the max working temperature is only for reversible loss. Once the magnet cools again, it should be at roughly the same strength, though prolonged heat can permanently reduce the strength.

The next level of damage is irreversible loss, which happens above the max working temperature but below the Curie temperature. In this region of heat, the magnet loses its strength, and that loss is maintained even after it is cooled. The only way to repair the magnet is to put it in a strong enough external magnetic field, but who wants to add an extra step of recovering the magnet after inserting it?

The Curie temperature is where the permanent damage takes place. Above this temperature, the magnet is done. For a neodymium magnet, the Curie temperature is 310°C for N-type, ranging up to 350°C for AH-type.

Magnet type Max. working temperature Curie temperature
Neodymium N 80°C 310°C
Neodymium M 100°C 340°C
Neodymium H 120°C 340°C
Neodymium SH 150°C 340°C
Neodymium UH 180°C 350°C
Neodymium EH 200°C 350°C
Neodymium AH 230°C 350°C
Nickel 354°C
Iron 770°C
Cobalt 1127°C

In theory, then, it may be possible to heat up a magnet to snuggle into some ABS and only suffer some reversible loss of strength, but if you aren’t careful and have your iron set to max, you’ll destroy the magnet permanently. In the future I’m not going to risk it. I’ll be 3D printing my holes slightly larger than the magnet and using an adhesive.

This lesson only cost me a buck in lost magnets and some time. May my shame and failure bring you success.

41 thoughts on “Fail Of The Week — Accidental Demagnetization

    1. My Weller soldering iron has a magnetic element for temperature control. When the tip reaches a certain temperature, the magnet loses its magnetism and the heating element disengages. when the tip cools down a bit the magnetism returns, and so does the heating element. You can hear it clicking on and off while the iron is setting in its stand.
      IIRC, this is called the Maxwell temperature.

      1. The magnet does not lose its magnetism.

        There is a slug of iron alloy in the back end of the tip. The alloy chosen has a Curie temperature at the chosen tip temperature, 600F, 700F, or 800F. The magnet itself is placed just behind the back end of the tip and has a much higher temperature of demagnetization. It has a weak spring pulling it away, the magnetic attraction to the iron on the back end of the tip overcomes this spring and closes a switch. The tip reaches the Curie temperature of the iron slug in the back end of the tip, the magnet is no longer attracted to it, the spring pulls it away and the switch is opened.

        Weller’s document on one of this type of soldering iron:
        https://neurophysics.ucsd.edu/Manuals/Weller%20Tech/Weller%20Tech%20Sheet%20WTCPR.pdf

        A cutaway of the handle and tip:
        http://codeandlife.com/wp-content/uploads/2012/03/magnastat-medium.png

  1. I found it best to use a threaded neodymium magnet.
    After a while the glue fails and the magnet sticks to the metal surface and not the plastic part.
    Maybe cyanoacrylate is the wrong adhesive for this ?

      1. It’s good for gluing skin lacerations (commercial name: Dermabond) though the formulation is a bit different than the stuff you buy at the hardware store in terms of side chains on the molecular structure.

      2. Beef says: “ca is the wrong adhesive for anything”

        Then you’re doing it wrong. Boeing uses it to hold together parts of passenger airliners. Seems to be working out OK for them.

          1. Explains this then…

            “”Boeing 747F experienced a number of engine falloffs:
            On Dec 1991, a China Airlines’ flight 358, a 747-200F lost an engine near Taiwan and crashed.
            On October 1992, a El Al 747-200F crashed after takeoff due to engine seperation, at Amsterdam, The Netherlands.
            Another 747F from Evergreen Airlines lost an engine over Anchorage, Alaska soon after.
            On October 2004, a Boeing 747-132SF of Kalitta Air lost an engine while climbing and landed without further incident.
            Engine seperation has been reported on 737s too:
            In November 2007, Flight CE723, a Nationwide 737-200 lost an engine during takeoff; the aircraft was landed without further incident.
            On December 1987, USAIR FLT 224, B737, lost an engine during climb; the aircraft was landed succesfully.””

            That was for the cheap joke, but in reality engines are designed to fall off… yah really… because if they get unbalanced or catch fire or something it’s better they fall off than rip the wing apart or set fire to the rest of the plane.

    1. Shape the cavity to be trapezoidal in cross-section, with the wide end *inside* the part. Make the cavity deep enough to have a millimeter or so clearance above the magnet when it is inserted. Use epoxy, and make sure there’s enough to cover the magnet surface when it’s pushed down into the opening. It will be permanently locked in place, and the magnet will still be close enough to the surface to activate any reed or hall-effect switch.

    2. Too brittle for some purposes. The shock of the metal smacking into the magnet tends to crack brittle superglue. It may also slowly weaken your magnet, as physical shock has that effect on magnets.

      Ever notice those magnetic catches, the magnet itself doesn’t hit anything? There are steel plates stuck to the magnet, the steel plates hit the catch plate.

  2. You could make a cavity in the part for the magnet to sit in, pause the print just before covering it up to insert the magnet, then resume the print to close the magnet inside. No magnet heating or glue needed.

      1. Hotends are usually pretty rigid, so I’d say the bigger worry wouldn’t be the hotend itself being affected, but rather the magnet moving from its hole sticking to something. Still probably unlikely if your hotend is mostly aluminum and brass near where the plastic comes out.

  3. I go over friction-fitting magnets in holes in one of my .IO projects:

    https://github.com/OpticsBench/laser-cut-optics-bench/wiki/Before-You-Cut

    Basically, if you make the hole 0.05mm smaller than the object, you can press the item in with a strong friction fit that will hold the item for most purposes. The project specifically describes doing this with magnets.

    If using a laser cutter, you need to also compensate for the kerf of the beam, and the text explains how to do that. For example the laser at my hackerspace: a 6mm magnet needs a 5.87mm hole.

    I don’t know how you would get that level of accuracy in a 3d print, but you could drill or ream the hole to the correct size after printing.

    Friction fit: Hole 0.05mm smaller
    Snug fit: Hole exact size of object
    Rattle fit: Hole 0.2mm larger

    “Friction fit” means the object won’t come out without being pressed.

    “Snug fit” means the object might fall out under vibration.

    “Rattle Fit” is for when you want a person putting things into the hole; for example, when lasercutting a holder for your socket-wrench sockets. Each hole should be 0.2mm larger than the socket for ergonomic ease of use.

      1. I think he means a washer, which would be a ring the magnet sits on, not a ring that the magnet goes into. The latter would indeed shunt some of the magnetic field; a washer on one end, only slightly larger than the magnet, will have hardly any effect on the field on the opposite side of the magnet.

  4. Small neo magnets make nice multi-size battery connectors – solder wire to some copper foil tape, place battery on adhesive side of tape and wrap it around the battery (ideally with a bit of fold-over). With conventional Alkaline and 18650 type LiIon cells, the magnet has attraction to the battery contacts, holding the copper sufficiently for all but high current draws. There’s some really cheap alkaline cells out there it doesn’t work so well with.

    Anyway, when fabbing these, it is really important to keep the magnets away from the workarea — they’ll happily leap to their doom like lemmings to the passing soldering iron as you’re soldering the wire, and then it’s game over. Needless to say, one also doesn’t wrap the copper foil on the magnet before soldering the wire either.

    I do laser cut undersized holes in acrylic and press magnets in, which make great hold downs and repositionable devices. Note if you’re doing a through hole with the laser and not precisely adjusting the focal length to the centre of the medium, one face of your material will have a slightly larger opening than the other, and you should press the magnet into the larger face down flush to the smaller one (which should face the direction it should latch towards), since it’ll less likely pull through the smaller one. If your focal point is set right at the middle of your workpiece, that narrow point of the cut is right in the middle (think of the shape of an hourglass), so you might be best off setting your focus high (esp. if also etching).

    If incorporated into the design (at least with laser cut things), you can affix another layer of material, thus encapsulating the magnet in the final assembly – no opening to the outside world means it can’t pop out even if the fit in the opening at the layer where the magnet lives is a bit large or it relaxes with time.

    Wood will swell and contract with humidity – different species and grains will be impacted to different degrees. You can get magnets with countersunk holes that you can drive a retaining screw through, or certainly at least drive a metal screw into the backside of the opening where the magnet will live to give it something to attract to – in the latter case, this would work fine for a magnet interacting with a reed switch, but possibly not with a metal catch with more mass than the screw on the backside of the magnet.

    Also with wood, you can bore the hole deeper and glue in a wooden plug that you cut flush and sand smooth, which can be more aesthetic than an exposed magnet.

  5. While 80C is the max temp before having irreversible losses, you can have irreversible loss even at lower temperatures. It depends on the permeance coefficient which is the height to diameter ratio. So the more cylindrical the high working temperature, the more disc like shape the lower working temperature. The magnet I was working with was a thin disc magnet (n grade neodymium) and it’s max working temperature was around 40C. I found that out the hard way and it took a long time to determine root cause.

    1. How would the shape of the magnet affect the working temperature? I’m not saying it isn’t true, but I’ve never heard of that, and I can’t really think of a plausible reason why that would be true. Do you have any references, or know what mechanism would cause this?

      Frankly, I find it a little hard to believe.

  6. What helped me in the past working with magnets is to tie a thin metal wire string across the magnet and then put a small stick at the end and wedge it in between whatever you’re working with. If you use plastic the little stick will be molten into the surface and the magnets force won’t drive it out.

Leave a Reply to RobotCancel reply

Please be kind and respectful to help make the comments section excellent. (Comment Policy)

This site uses Akismet to reduce spam. Learn how your comment data is processed.