Dismantling and repairing the motor :o

[bob2.0]

So, over all a good research job, that'll help one or the other of us in the future!

I will continue and finish the job. I will see if there is more to figure out and go over any pointers and tips that might help.

FWIW I tried measuring the height of the magnets, and there was nothing to measure. I was hoping it was a way to tell which ones need to be changed, but, no, you cannot tell from this. You have to actually remove them to see. They can be clean and shiny, but rusty underneath.

They do not need to be standing high. Just that adhesion/cohesion became so weak that the magnets must have been moving as the motor rotated.

They touched because the magnetic forces were stronger than the rust and they rippled as the motor rotated.

The glue I am using, AA326, is a polyurethane methacrylate (slightly flexible so does not fracture with stress and heat) which cures by anerobic process, single component, and a secondary cure using an activator which comes as a 30% acetone solution. So I basically used the curing accelerant as a cleaner, spraying it on then rubbing the surfaces down with a coarse sponge and spraying a further coat. They seem glued on very robust. I suspect if I wanted to take these off then they will have to be chipped off and broken. Many times stronger than the rust!

I have also experimented with cutting out polycarbonate filler to go between the magnets (as they are smaller). I don't think it is needed now but I have already had this problem twice so why risk it? I can make this filler, I have the materials, and this glue is also good to bond polycarbonate to aluminum, so the instructions say.

WhatsApp Image 2025-04-03 at 20.20.17.jpeg (153.52 KiB) 498 mal betrachtet

[jeff-jordan]

Don't know whether polycarbonate is a good stuff as filler or not. It resists only to temperatures up to 120°C (150° for a short time).
Maybe just a little strand of high temperature silicone will do the job as a filler? Or just use some remaining glue .

@dominik: From your experience, how hot will the magnets get inside the motor?

[bob2.0]

I think the 'replacement' magnets I bought are only 'H' (120C) so if they were to get that hot then they are probably already doomed.

QS motors were unhelpful in any way towards telling me even what the magnets are, let alone offering to sell me any replacements. But I suspect they probably only used SH (150C).

I tried this out myself before, running at full speed for 30 minutes to see if the hub warmed up, and it was still luke warm. I did not find any temperature in them.

This is another reason for a better layer of (more) glue. Cured glue is not an idea thermal material, but it is better than a layer of air!

[dominik]

I have had the windings in my QS60V5000W heated up to 160C (with >20kw peak) and the outside of the rim was still under 90C in hottest summer days.
The magnets are somewhere in between, and the better they are glued, the more heat can be transfered to the rim.

They should stay way under 120C to not loose their magnetic strenght.
I do not think they used high temperature magnets with up to 150C working temperature.

[jeff-jordan]

Ok, if the windings of the stator will be at about 160°C max. the magnets and the rim will certainly remain below (or max.) 100°C.
That's fine for the magnets (and the polycarbonate).

...
and the better they are glued, the more heat can be transfered to the rim.
...

Of course, a bubble of air underneath the magnets will be a good isolator.

[Amp_Sauger]

Hallo Jungs

Wehre es nicht eine gut Idee die Magnete nach dem einkleben nochmal mit Klarlack einzupinseln damit keine Feuchtigkeit mehr darunter kriechen kann.

[bob2.0]

Hallo Jungs

Wehre es nicht eine gut Idee die Magnete nach dem einkleben nochmal mit Klarlack einzupinseln damit keine Feuchtigkeit mehr darunter kriechen kann.

I have gone over and over this in my mind, thinking about what benefits and problems this might have.

The problem is that adding a further layer between the rim and the magnet means the connection is only as strong as the weakest link.

The AA326 'literally' leaves a layer of acrylic plastic adhered to the back of the magnet. As long as it is a 100% coverage, it should be better than a further layer of lacquer.

If it was lacquered before gluing then it would only be stuck as strong as the lacquer. The glue is structural adhesive. The lacquer is not.

So long as it is 100% coverage then it will prevent ingress of condensate, therefore it should prevent rusting.

If only these magnets had 100% coverage of glue.

You can see that if there is full glue coverage, there is no rust.

Whether you need to lacquer the stator side of the magnets, and also the edges of the magnets, I agree this is more thought.

I decided to smear a fair bit of ATF (Dexron III) onto the magnets when I reassembled, and I think this actually saved the magnets from worse when the one moved, it did actually add lubrication. All the dust has bound to the oil. There is basically no oil nor dust that was left, they mixed and spread around.

My concern is the back of the magnets. I don't think the tops are an issue.

The edges are a concern. If they start corroding, can it work its way into the back of the magnets? I have seen no evidence of this, and some of the magnets have lost their plating at the edges but did not rust. It was only on the backs. Long term? I don't know.

[MEroller]

The idea is to cover the sides and UPPER surface with clear coat, not the glue sides

[bob2.0]

I think that is OK to do that coating.

I have looked more at these magnets, so there is yet MORE cost-down.

Not content with being cheap on glue, these are Zn-Ni plating only.

I decided to wet-sand a test magnet with 800 grit and light acid (citric acid) to remove the rust chemically, and see what came of it. Too much sanding would remove protective plating.

We see here clearly this is a zinc second (inner) layer.

Most (quality) magnets are two layers of nickel (a single layer can be too porous), with a copper interlayer to improve the binding of the nickel layers.

This is simply zinc on the magnet, then nickel, which is the cheapest option.

The zinc does bond well to the sintered magnet, but once it is exhausted, the iron has then rusted underneath (now black after reduction to ferrous citrate).

In time it should oxidise to ferric citrate which is more soluble in water, maybe it'll wash off if I leave it a day or two.

Then I can make sure the glue has covered the bear NdFeB pitted areas, and it should be sealed up.

I thought the edges (some of the magnet edges show this grey patina) might be exposed NdFeB but I can now see it is the zinc underlayer. So, it would be good to think about some sort of coating option for the edges where it is exposed.

[bob2.0]

Yesterday I began the slow and tedious process of refurbishing the magnets.

It is so slow and dull. Cleaning them up so they are clean enough to lacquer. Sigh. Perhaps 2 hours on just 8 magnets, one I will probably reject.

I did spend time trying to perfect an efficient process, which I will share my tips with further improvement.

For now I have another question.

I put 3 coats of lacquer, and cannot really tell if it is covered! Because it is clear and everything is shiny, I wonder if using coloured primer might be better? Then I can ensure a coating.

Of course, the primer might contain metallic particles in its colour, but it would still be less metal than the magnet plating. So, should I just use grey primer instead?

You can just see here on the edge that there is a feature of the lacquer.

There is none on the lower (corroded) face, to paint them I stuck them to cardboard with masking tape. I could therefore coat the edges but the underside is free of coating. This will receive a full coating of glue.

To paint, or to lacquer?