Can epoxy withstand the torque of a hub motor axle?

[shenzhen_ex]

For my first eBike I decided to go with a 2WD from components. So far I have snapped 1 rear and 2 front axles, all within a few 100 km. I have the problem with the rear axle solved. Now I need to design a solution for the front 20mm axle-fork.

Follow the various views of design below you will see that I am considering adding epoxy to keep my axle from snapping.

I used the design pictured below (without epoxy) and my axle snapped.

NOTE: Both adapters have internal threads and are made of 420 stainless.

My questions are:
1) Will the epoxy hold up to the torque? DrBASS say that epoxy is similar to weld, right?
2) Do you think that adding epoxy (as shown below) to the above design will prevent the axle from snapping again?
3) Will I be able to remove the adapter on the non-wire side with heat, without melting the phase/hall wires?

Epoxy shown in BLUE:
The below adapter screws on tighter with motor torque.I think this screw-on effect caused the snap.

The below adapter un-screws itself with motor torque which makes the TA on this side useless. This is the main reason why I think epoxy may prevent the axle from snapping. Will the epoxy hold up to the torque?

NOTE: I cant get a nut in there due to the short length of the axle on this side.

Image of snapped axle. The break point is at the red arrow:

Here is how the adapters look off the bike:

Here is how the adapters look on the bike:

 

[teklektik]

I think you may be going down a Bad Path here - epoxying this together will introduce some serious service issues. It seems you are taking the original poor flatted axle design and going to something worse in an attempt to get the TA functionality you need combined with the axle adapters.

Assuming getting a new fork is off the table:

I'm not sure about the clearances you have there, but I would look at a redesign to separate the axle adapter and TA functionality. Apologies if you already considered this, but how about flatting the axle up to the side covers (or as appropriate) and using TAs inside the dropouts on the newly flatted portion of the shaft? They would be held in place against the small remaining axle shoulder when the adapters are tightened down and the adapters in turn would be clamped up in the forks without torque issues. The TAs would be more or less conventional in configuration with the slot length appropriate for the axle diameter next to the side cover.

At your power levels if you can squeeze a single 1/4" TA on one side (or maybe 1/8" on each side) then you should be fine. (I use a single 1/4" TA on my front BMC and have had no issues at all in 7000+ miles regularly running 1600W - 2000W.) It sort of looks like you could put a single thick TA on the drive side which would make the familiar flatted shaft design more effective because of the added axle diameter. If you want to put a TA on the non-drive side, there may already be enough exposed flatted shaft to use the axle without modification - just switch to the inside TA...

Just a thought....

 

[shenzhen_ex]

...... I would look at a redesign to separate the axle adapter and TA functionality. ....

That is probably the best answer. All of my broken axles are likely related to the fact that I am using the adapter threads as TA's.
Like you said, I probably should design a TA that is separate from the adapter on the non-disk side but that will require that I mill flats into the axle (see OP images).
The disk side only has a few mm's clearance. I might still go with the epoxy on that side.

I am also considering building adapters that are keyed (no threads, rather flats to match the axle). But then I will defiantly need epoxy on the disk side.


BELOW: Non-disk side

BELOW: Disk side (not much room for a TA)

 

[teklektik]

I'm guessing that a large part of the problem may have been subjecting the long axle stubs to torque - which due to the threading was applied even though the width of the fork clamps. This was probably particularly bad on the drive side because the drilling reduced the strength. If you can relocate the TA as described above so that it is adjacent to the side cover on the drive side where you have a beefy and very short axle shoulder very close to the stator - the protruding axle stubs should experience minimal (if any) torque and adapter assembly should be reduced to bearing the load of the bike.

I don't think that epoxy in that situation seems to provide material benefit - the shear strength of the epoxy to prevent rotation does not appear to be serving a purpose. At best a little blue thread lock might be considered.

Anyhow - looking forward to reading of your revisions. Very interesting build

 

[shenzhen_ex]

...If you can relocate the TA as described above so that it is adjacent to the side cover on the drive side where you have a beefy and very short axle shoulder very close to the stator - the protruding axle stubs should experience minimal (if any) torque and adapter assembly should be reduced to bearing the load of the bike....

Is the drive side = to the disk side? Sorry for my total newb question:

It must be, because it ts the short axle side. So why is that called the drive side?

 

[teklektik]

Is the drive side = to the disk side?

Apologies for not being clear:
drive side = chain drive side = driver right = non-disk side
I think from your photos that this is the side with most clearance to install a TA...

Anyhow, regardless of terminology, the idea is to get the TA close to the motor and perhaps where there is a larger diameter axle. You may be able to re-purpose your existing shaft adapters by turning down the TA discs and machining two flats in the disc edges for a wrench.