Neodymium Magnets. Self Center?

[Falcon69]

I have 8mm Diameter Magnets at 3mm thick right now.

My question is:

If I purchase some 4mm Diameter at 3mm thick, when the two flat sides of the magnet stick together, will the smaller diameter self center onto to the bigger one?

I have a project I am thinking of doing, and if they do self center, it will simplify things.

So, anyone who just happens to know, or does have two magnets of different diameter, could you check please and take a snap shot? 

Thanks in advance.

[Psi]

I think it will depend on the friction of the two surfaces.

They probably will want to find dead center but i suspect once the two surfaces are touching the force trying to move an off-center pairing to a centered one will be quite small and likely not enough to overcome friction.

Maybe with some polished magnets and a coating of oil it is has a good chance of finding dead center.

[electr_peter]

Short answer - no, magnets will not self-center completely*. There will be some play + rotation instability. It depends on magnets and what you want to do - it may be good enough for your application.
*Centering is only meaningful with very light items and/or small lateral forces acting on magnets. Small lateral force can move magnets a bit.

Few reasons - lateral pull is not that strong; magnets themselves are not completely centered in terms of magnetics vs dimensions; longitudinal pull creates too much friction (if magnets are touching anything).
 
If you care about exact placement of two pieces (assuming magnets will have to hold some item exactly in place), you should add additional mechanical means of centering/guiding.
I have experience with boxes/latches containing neodymium magnets on both sides - such system has strong pull force, but not enough lateral strength.

[Falcon69]

Well, the thing is, the magnet sits 1mm of the PCB.  The problem is, I am super gluing a 1/4"1/16" nylon spacer to the PCB and then to the Magnet.  The super glue is not working very well.  It's not holding and they are coming apart.  I need it to hold long enough to cast these things in resin.

My thought was if I got the smaller magnet, made the cut out in the circuit board for it with a 4mm diameter plated hole, solder that magnet in place (which would then be centered, no guess work trying to center both the nylon spacer and magnet), and then let the 8mm diameter magnet self center on top of the 4mm diameter one.

But, if it's not going to center as you say, I need to find another way of doing this.  I tried hot melt glue, that isn't sticking to anything.  I am using Gorilla Super Glue.

I know I can have magnets made that are called TopHats. They appear to have smaller diameter magnet under a larger one, but That will be expensive.

[IanB]

Well, the thing is, the magnet sits 1mm of the PCB.  The problem is, I am super gluing a 1/4"1/16" nylon spacer to the PCB and then to the Magnet.  The super glue is not working very well.  It's not holding and they are coming apart.  I need it to hold long enough to cast these things in resin.

How is your superglue process? Are you thoroughly cleaning both surfaces with a degreasing solvent like tric or perc before gluing? Are you also slightly roughening the nylon surface to aid adhesion?

[Falcon69]

cleaning them with IPA.  The surface of the nylon is fairly rough to begin with

[Paul Moir]

The nickle plate on the magnets is very smooth which prevents getting a good bond.  First completely degrease the magnets (solvent wash like acetone is probably easiest) and then abraid the surface with sandpaper.  You'll get a good bond with superglue that way.  Epoxy is much more permanent, but it takes a while to cure.

Soldering won't work for at least two reasons.  The heat will destroy the magnet and soldering nickle can be tricky.

Your nylon bond will be much improved if you roughen it and activate the surface.  Loctite 770 is the activator I've used, I think it's mostly heptane.  I have seen someone activate aluminum with a propane torch and it worked surprisingly well!  Google the loctite plastic bonding guide for details. 

EDIT:  interesting, loctite says not to roughen Nylon if using 770. 

[SeanB]

You might need to use an activator on the nylon . Does the bond fail on the magnet or the plastic? Which part holds to the glue will tell you which adhesion is failing.

[Falcon69]

Well, I tried the soldering in the reflow oven (I had an old circuit board with a voltage regulator footprint big enough for the magnet!).

It soldered just fine, did not discolor it, and didn't seem to lose the magnetic properties (maybe just alittle, but it was actually alittle too powerful to begin with).

SO, this method may work.  I might explore that possibility.

It would certainly reduce my labor time of making these things. What do you think?  Did I just get lucky on the heat NOT destroying it?

[SeanB]

They will start to degrade over 176F ( 80C) depending on the grade.

https://www.kjmagnetics.com/neomaginfo.asp

Curie points for the various grades are listed here as well.

https://www.kjmagnetics.com/specs.asp

You can see the top grade will operate till 200C, so it will survive SnPb soldering without losing strength. All though have a curie point of 350C where they become demagnetised.

[Falcon69]

thank you sean

It looks like I need at least a NUH (180°c), instead of a N brand. I can only guess what that would cost.  Way too much I suspect.

I read that it will lose partial of it's magnetic properties, but if temp reaches its curie temp (350°c), it loses all of them.

So, with that said, the initial baking in the reflow oven will drop the magnetic properties a little. But, will it continue to drop, even in ambient temperature? Or it will stay at it's reduce rate forever (or unless super heated again somehow, of course)?

[Falcon69]

Or, maybe I can just buy the standard N42, once it is reflowed, it would probably lose enough magnetic properties to equal a N35?  What do you think?

I also read that they can catch on fire if soldered.  I'm guessing that is only if it exceeds the Curie temp?  I'd hate to have these things catch fire and explode in my reflow toaster oven. 

[SeanB]

It loses strength over 80C and that is permanent. Over curie point it becomes plain metal as the magnetic alignment is destroyed. It will retain any external imposed magnetic field present as it drops below the curie point though. As to catching fire I do not know the ignition point, but that is likely at around 600C. Using the stronger one and weakening it by soldering to a lower level is going to be very variable, depending on how hot it got and for how long, but if your use is able to work with a large variation from the weakest field to the strongest in the original it will work.

You might have problems though with the coating peeling off the magnet, probably it will be best to place the magnet in the hole and pot it with some epoxy to seal it fully top and bottom. I have some that have turned into magnetic powder from corrosion through a pinhole in the case, and one in a breaker which peeled off the one face where it was stuck to a steel pole piece.

[Falcon69]

i am not worried about peeling or corrosion.

Once these come out of the oven and cool, they go right into the mold and resin cast over them.

so, once it reaches the temp and goes over (but doesn't exceed the curie temp), it will lose some magnetic properties, and that new magnetic strength will then become it's permanent value.  i.e. N42 then becomes a permanent N37.94 for example (if that's what it ends up being)

I understand that each magnet will vary, depending on the milli-seconds it is left in the oven, etc. over the previous magnet that was in the oven.   There is adjustment on the end product that will account for that.  I just don't want it to be a drastic change, like N42 becomes N15 on one, and N41 on another, etc.

I'm just using numbers as an example. I am unsure of the real values.

[SeanB]

As it is not a controlled part of the processing in making them you will likely have batch variation on the magnets, but there should be consistency in a batch. Only way is to try and see how it goes, just placing some in a reflow oven on some plain copper board with solder paste dots and reflow and see how it performs. Worst is they all fail, and if you just use your regular profile it will show roughly how it will perform on your production run. Likely most will still have a good part of the original field in them.

[Falcon69]

I am going to order up some bare copper clad boards, and get some samples in and test.  I think I will just order the N42 ones.

I'll keep this thread posted.  It will be some time, as I need to get the samples from the manufacture in China.

Any regular copper clad board work that I can get quick off eB (*cough*) crap store (*cough*) ay?

[Falcon69]

Aha!

These should work for that, just need one lot (10 pcs), I can test with these.

http://www.ebay.com/itm/10pcs-lot-Star-Aluminium-PCB-Heatsink-for-High-Power-LED-emitter-20mm-White-/111147111069?pt=LH_DefaultDomain_0&hash=item19e0e2529d

[SeanB]

I also ordered some, I have some LED units that I can use with them. Just have to place them on the graphite plate and use hot air gun to tin them before putting the LED units on. I have done them by hand though the board is very uncomfortable to hold.

[Falcon69]

i was just going to place the magnets onto the circle pads, and then shove them in the oven.  Do 10 different reflows for 10 magnets, then test and see if they all trigger the sensor okay.  If they do, then I think I'm safe to go ahead with the plan.  If it triggers the sensor all at or near the same distance away, that will tell me if the consistancy of the magnets are near the same for that reflow curve