Hub motor axle bolt tightening torque - How much is enough?

[mrbill]

Folks:

I've got an ongoing problem with loosening axle nuts.

I'm using a medium-power Nine Continents M3006RC as sold (but now discontinued) by Grin:
http://www.ebikes.ca/shop/discontinued/m3006rc.html

and on both left and right dropouts, Doctorbass's hardened steel torque arms:
https://endless-sphere.com/forums/viewtopic.php?f=31&t=29129&start=300#p698328

The motor axle is 12mm with 10mm flats. Thread is M12x1.25. The bolts are serrated flange nuts, probably made of mild steel by the look of it. Although I don't have vast experience with a wide variety of hub motor axles, I suspect this is pretty standard for motors of this size and rating.

The dropout slot in the torque arm is just a tad over 10mm to facilitate insertion and removal of the axle, so there is a tiny bit of rotational play. This play can be observed as an audible "clunk" when torque is reversed from regen to driving mode and vice-versa. When the axle bolt on one side or the other comes finger-loose, the "clunk" noise takes on a sharper "tink" or "tap" noise.

Various sources recommend maximum torque for M12 nuts (usually 1.5mm threads) ranging from around 80 to 130 Nm, depending on the hardness of the nut. I've tightened these to 60Nm, and that feels quite tight to me for a nut/thread this size, especially considering that the nut is only engaging threads around half the axle (the other half being the flat). I don't want to damage the threads.

But after a hundred miles of riding the nuts are loose again to the point of clunking. After a few hundred more miles of riding, the axle nuts are finger-loose. I now carry an 18mm offset wrench with me on the bike.

I'm suspicious that the serrated flange nut being softer metal compared to the hard surface of the torque arm can't get any bite into the surface of the torque arm, so it slides easily and given repeated axle torque reversals, works its way loose.

What axle nut torque is sufficient to prevent nut loosening?

Would there be any benefit to roughening the nut flange-facing surface of the torque arm?

Thanks.

 

[amberwolf]

If the axles have play, it's possible that no amount of tigthening (up to t eh point of stripping threads) will stop it from eventually loosening.

But you could try Nordlock washers instead ofd the regular ones. That worked on CXrazybike2.

FWIW, there's a thread by Justin_LE on ES about testing dropouts taht has some info avbout axle nut torque levels. I can't copy paste the URL on this stupid borrowed tablet though; it clears all data f rom everythingg whenever I switch tabs in any browser. (or pretty much anythihg else I do)

 

[teklektik]

I'm suspicious that the serrated flange nut being softer metal compared to the hard surface of the torque arm can't get any bite into the surface of the torque arm, so it slides easily and given repeated axle torque reversals, works its way loose.

Exactly right.
Unfortunately, that same serration concept is what makes NordLocks work - they rely entirely on the two washer components remaining fixed with relation to the adjacent nut and contact surface. Nordlocks will be equally ineffective with the hard steel of Doc's torque plates since the lower washer will not be able to get any bite and will simply rotate against the torque plate - actually deforming the serrations instead of digging into the contact surface (ask me how I know this...).

This pretty much tells the tale:


I would try auto lug nuts for a long thread surface and blue Loktite before resorting to other rotation-stopping construction.

For an "alternative NorLock" solution, I might try putting two thin finished stainless washers from the hardware store under the nut with a little lubrication and again, Loktite the nut. Use tight-fitting washers (likely SAE) to get full face contact. Here we are trying to encourage the washer faces to slip against one another and the TA face instead of the nut unthreading and relieving tension. With sufficient torque, there should be minimal or only incremental axle motion and the nuts should stay tight and not work loose. (This is the same idea as lubricating NordLock cam faces so they can slide to absorb and relieve rotation instead of the serrations breaking loose.) This last bit is just a thought, but cheap to try...

 

[mrbill]

For an "alternative NorLock" solution, I might try putting two thin finished stainless washers from the hardware store under the nut with a little lubrication and again, Loktite the nut. Use tight-fitting washers (likely SAE) to get full face contact. Here we are trying to encourage the washer faces to slip against one another and the TA face instead of the nut unthreading and relieving tension. With sufficient torque, there should be minimal or only incremental axle motion and the nuts should stay tight and not work loose. (This is the same idea as lubricating NordLock cam faces so they can slide to absorb and relieve rotation instead of the serrations breaking loose.) This last bit is just a thought, but cheap to try...

This I might be able to try without buying more hardware. Thanks for the suggestion.

With all this talk of Loctite and other semi-permanent means of attaching the rear wheel, I'm glad I switched to using Gaadi innertubes to facilitate easy flat repair.
https://www.bike24.com/1.php?content=8;product=58538;menu=1000,2,103,106

 

[teklektik]

Blue Loktite just cranks off w/o problem, so no issue w/flat repair.
That said, in the field I'm a big fan of "leave the wheel in place, break the bead, yank out the tube but still trapped by the axle, and patch in situ".

I carry one of these Park Wrenches nowadays just in case I need to do a tube replacement and they break the Loktite free w/o trouble.

 

[teklektik]

Can't help but wander OT - how do those Gaadi innertubes work out?
Using them for long?
Thick, thin, in between?
They are very weird but very intriguing...

 

[e-beach]

AW is probably right about the movement of your axle loosing your nut. Hard to know about Dr. Bass's hard steel vs the washer issue. I guess you have to be there.

I watched the Grin video about torque a couple of years ago and have been tightening my bolts to 55lbs ever since...about 75nm. But I am torquing into aluminum alloy forks. You may have to go tighter .....or epoxy something a bit softer washer to your Dr. B. torque arms and torque your bolt into that.

If you want a couple of nord-lock washers PM me as I have a few sets left that I can mail to you for $5.00 per set of 2 (that's one set on each side), including postage.

 

[mrbill]

Blue Loktite just cranks off w/o problem, so no issue w/flat repair.
That said, in the field I'm a big fan of "leave the wheel in place, break the bead, yank out the tube but still trapped by the axle, and patch in situ".

I used to be a fan of that technique until I got tired of fixing leaky patches. Patches need a day or so to properly set. I have had better luck patching at home, pressing the patched section of tube in a vise for several hours, then letting it sit for a day before inflating and riding on it.

http://sheldonbrown.com/brandt/patching.html

I carry one of these Park Wrenches nowadays just in case I need to do a tube replacement and they break the Loktite free w/o trouble.

My axle nuts have too much hardware on the dropouts to use a flat wrench with a broad head like that.

https://endless-sphere.com/forums/viewtopic.php?f=31&t=29129&start=500#p983592

I need a ratchet wrench with a deep socket or an offset wrench. 18mm for the M12x1.25 nuts that came with the motor.

 

[mrbill]

Can't help but wander OT - how do those Gaadi innertubes work out?
Using them for long?
Thick, thin, in between?
They are very weird but very intriguing...

The Gaadi tubes are pretty thick, almost too thick for the 1.5" tires I use. But, that's ok, as I prefer thick tubes. They puncture less easily.

I've been using a Gaadi tube in my rear hub motor wheel for only the last few months, not really long enough to come to a conclusion. They greatly ease replacing a tube on the road, but I have had a couple of problems, all on one tube.

1) The presta valve is 47mm, long enough that it makes inserting and removing the tube difficult. On my first tube the head of the valve got bent as I pried the stem over the rim wall. Then yesterday when I removed the tube from the tire, the valve head hit the rim wall, breaking the thin neck of the valve head. A better technique is to push the tire aside when inserting the valve stem into the rim, then pry the tire over the tube at the valve. Live and learn.

2) The failure of the tube that prompted me to replace it was a small cut on one of the flat ends of the tube, an odd location for a puncture, and would have been a difficult spot to apply a patch. I could find no debris in the tire or corresponding cut in the tire casing that could explain this, nor could I find anything on the rim bed such as an exposed spoke hole or other sharp edge. The cut appears along a circular pattern of roughness on the tube end. It's possible the tube was defective, or that I installed it carelessly so that the material stretched abnormally in that area when it was inflated, or perhaps the material at the end of the tube is thinner than elsewhere and herniates when the tube is inflated. Other unused tubes in my collection did not exhibit this marking.

 

[mrbill]

AW is probably right about the movement of your axle loosing your nut. Hard to know about Dr. Bass's hard steel vs the washer issue. I guess you have to be there.

I watched the Grin video about torque a couple of years ago and have been tightening my bolts to 55lbs ever since...about 75nm. But I am torquing into aluminum alloy forks. You may have to go tighter .....or epoxy something a bit softer washer to your Dr. B. torque arms and torque your bolt into that.

If you want a couple of nord-lock washers PM me as I have a few sets left that I can mail to you for $5.00 per set of 2 (that's one set on each side), including postage.

The torque arms are harder than the axle. I can see no marking or other evidence of material flow on the inside face of the torque arm slot, but the axle does show signs of repeated "bashing" (due to the frequent torque reversals upon regeneration), which is somewhat worrisome because the axle is hitting the torque arm at the cut ends of the threads. It's these cut ends that are bearing the brunt. Already I have noticed that the axle nuts do not without difficulty spin past these threads. Fortunately, the nuts don't need to travel that far down the axle when I install the wheel on my bike. But, if this "bashing" continues, I can see that the axle could eventually be compromised.

Thanks for your offer. Let me try teklektik's "poor man's" Nordlock, but if that fails to hold, I will contact you. Might take me a few weeks before I know.

 

[e-beach]

........ Might take me a few weeks before I know.

I will put a set aside for you until the end of fall. :wink:

Other then that, I say tighten them up. If 80nm is not crushing any spacers then go for a tight torque.

 

[Punx0r]

The average hubmotor axle does not take much torque to strip the threads due to the poor material quality and partial thread. Ideally you want clamping torque arms. Otherwise, I would consider disabling regen braking. That frequently torque reversal will round-off or snap the axle eventually since it has a little movement in the drop out.

 

[mrbill]

Just wanted to update this thread with some of my recent findings.

For the last few months I've been using two different direct-drive hub motors, Nine Continents M3006RC and Edge1500.

The Nine Continents motor is rated for 500-1000 watts, although I've driving it a bit harder than that in testing. It has a 30mm stator, uses a 12mm axle, and has a Kv of 8.9.

The Edge1500 motor is rated for 1500 watts, 2500 watts peak, and I've driven it that hard. It has a 35mm stator, uses a 14mm axle, and has a Kv of 11.5.

Both motors are installed on bikes using Doc's "Ultimate Torque Arms"

https://endless-sphere.com/forums/viewtopic.php?f=31&t=29129

also using Nord-lock wedge-locking washers with anti-seize applied to the wedge contacts (but not the serrations) and to the axle threads to assure that the nuts are installed to the measured torque.

I'm driving both motors with an ASI BAC2000 controller and am allowing regeneration power up to 1500 watts, although I have seen peaks up to 2000 watts in regen.

I summarized some recent findings in Doc's thread, starting with this post:

https://endless-sphere.com/forums/viewtopic.php?f=31&t=29129&p=1242646#p1240704

Although I am still testing for the long-term, I have found that given my installations I must tighten the axle nuts when installed with Nord-Lock washers under both axle nuts to at least the following installation torques:

12mm axle - 60 Nm
14mm axle - 50 Nm

These figures may change if I discover loosening after long-term use, but they are a good starting point for others planning to use regeneration up to 2000 watts. If the 14mm nuts remain tight at 50 Nm installation torque, I may try 40 or 45 Nm, my goal being to find a lower-bound installation torque that holds tight for the long-term. I'm seeking the lower-bound installation torque because I have found that tightening to over 60Nm deforms the axle threads slightly on these motors.

 

[cal3thousand]

Thanks for posting the follow up. I think it is an excellent point to figure out the lower bound of torque required.

A couple thoughts:

The lower bound torque number should be different for the varying applications we have depending on the axle material, fastener material, lubrication, and use of any washers/Nord-locks. However different, I do not know. Each person should test out to find the right spec for their bike, using these numbers above for a starting range.

Using a material that is harder than the axle material for the fastener (most materials are harder than the cheese-like axles on some of these motors) means excessive tightening will damage the threads of the harder-to-replace axle and provide no appreciable security from loosening.

 

[mrbill]

Although I am still testing for the long-term, I have found that given my installations I must tighten the axle nuts when installed with Nord-Lock washers under both axle nuts to at least the following installation torques:

12mm axle - 60 Nm
14mm axle - 50 Nm

These figures may change if I discover loosening after long-term use, but they are a good starting point for others planning to use regeneration up to 2000 watts. If the 14mm nuts remain tight at 50 Nm installation torque, I may try 40 or 45 Nm, my goal being to find a lower-bound installation torque that holds tight for the long-term. I'm seeking the lower-bound installation torque because I have found that tightening to over 60Nm deforms the axle threads slightly on these motors.

After riding my bike that uses a 14mm axle hub motor over hilly terrain for a few hundred kilometers, I discovered the drive-side nut, secured with the aid of Nord-lock washers, had loosened slightly. I re-tightened it to 50Nm and will monitor it again. The non-drive side nut, secured in the same manner as the drive-side nut, had not loosened from its original 50Nm torque installation.

50Nm may be borderline too low for my installation. If I discover that the either nut works loose at 50Nm, I will re-tighten them both to 55 Nm.

 

[MadRhino]

Axle nuts should not need to be very tight. I mean, the torque plates or whatever mean that you have to hold the axle torque have to be built strong enough in the first place. Then the axle nuts are only holding lateral placement of the motor between the dropouts and that doesn't require much strength.

If the nuts are getting loose because of back and forth kick of acceleration and regen combination, I suggest to use a simple locking screw as commonly used by gunsmith, locksmith, etc... You just need to drill and tap for a small screw just aside a flat of the axle nut after it is reasonably tight. This screw will prevent the axle nut to get loose, as it will need to be removed to free the nut.

Getting into a fight against motor torque kicks by increasing the torque on the axle nuts, is a game that you can't win. Sooner or later this will result in stripping the axle thread.

 

[waynebergman]

One idea for helping with the axle rocking back and forth loosening your nuts is to fill in the drop out opening better so its a super tight fit to the hub axle. Check out the start of page 6 of this thread https://endless-sphere.com/forums/viewtopic.php?f=28&t=74706&start=125 to see if this is a solution that may work for you. In my case the spacers where efective enough as I am using a mid drive and its not such an issue as hub drive but the opening on my bike did not fit the axle thickness at all so I had to do something. In your case the brass stips I used may be too soft to work over time? Also one thing that may help the norlock washers key into the Dr Bass drop outs better is to epoxy just ONE of the norlocks in place to the Dr Bass drop outs on each of both the drive and non drive sides of the bike. The nordlock that touches the tightning nuts would not get the epoxy treatment and hopefully keys into the nut like its supposed to. Not sure if this is a good idea or not but just thinking out loud here. Good luck.........wayne

One more option is to get a wrench (spanner) that is a tiny bit too small to fit the axle flats. File this wrench opening so you can just barely hammer it into place around the axle flats being used as kind of a washer between the drop outs and the nuts. Then either shorten the wrench and Drill a hole through the wrench and the dr bass drop out and rivet it into place or if the wrench can be cut and welding so its orientation falls in line with chain stays or seat stays then hose clamp it in place leaving it nice and long like a big lever. A total hack job for sure but it may work even if it was just done on one side of the bike. I did this kind of thing on my giant build here https://endless-sphere.com/forums/viewtopic.php?uid=20134&f=3&t=63283&start=0 and it worked pretty good.

 

[drew12345]

When I got hs3540 from grin couple years ago, they told me 40nm. So I been using that on all my hubs (now mxus 2) since and never had one come loose. Maybe because I'm not using regen.


Sent from my iPhone using Tapatalk

 

[mrbill]

Thanks, guys, for your suggestions.

I have a few comments:

When I got hs3540 from grin couple years ago, they told me 40nm. So I been using that on all my hubs (now mxus 2) since and never had one come loose. Maybe because I'm not using regen.

If you're not using regen, then you aren't going to be seeing this problem that only occurs when torque alternates frequently between forward and reverse with similar magnitudes. Axle nut loosening arises when substantial torque from regeneration/plugging is used. Even the regeneration available on the typical Infineon-style controllers (Lyen, Xie Chang, Grinfineon, et al) offers only about 1/4 the braking force that I'm discussing.

One idea for helping with the axle rocking back and forth loosening your nuts is to fill in the drop out opening better so its a super tight fit to the hub axle. Check out the start of page 6 of this thread https://endless-sphere.com/forums/viewtopic.php?f=28&t=74706&start=125 to see if this is a solution that may work for you. In my case the spacers where efective enough as I am using a mid drive and its not such an issue as hub drive but the opening on my bike did not fit the axle thickness at all so I had to do something. In your case the brass stips I used may be too soft to work over time?

Using spacers may help with a 12mm axle that has more space to rock back and forth. The 14mm axles don't have any detectable play when the axle is installed to the bottom of the slot. While the brass is malleable, it may be too soft for long-term use. The leading edge of the axle flat exerts tremendous pressure against the dropout wall and is probably enough to flow the brass, allowing the gap to open over time.

Also one thing that may help the norlock washers key into the Dr Bass drop outs better is to epoxy just ONE of the norlocks in place to the Dr Bass drop outs on each of both the drive and non drive sides of the bike. The nordlock that touches the tightning nuts would not get the epoxy treatment and hopefully keys into the nut like its supposed to. Not sure if this is a good idea or not but just thinking out loud here.

Using epoxy on the torque arm side of the Nord-locks will prevent wheel removal, unless you're proposing that the epoxy be broken and re-applied upon each occasion of wheel removal and installation. Besides that, I think the Nord-locks are just hard enough to get some bite into the torque arm, so I don't think this is necessary.

One more option is to get a wrench (spanner) that is a tiny bit too small to fit the axle flats. File this wrench opening so you can just barely hammer it into place around the axle flats being used as kind of a washer between the drop outs and the nuts. Then either shorten the wrench and Drill a hole through the wrench and the dr bass drop out and rivet it into place or if the wrench can be cut and welding so its orientation falls in line with chain stays or seat stays then hose clamp it in place leaving it nice and long like a big lever. A total hack job for sure but it may work even if it was just done on one side of the bike. I did this kind of thing on my giant build here https://endless-sphere.com/forums/viewtopic.php?uid=20134&f=3&t=63283&start=0 and it worked pretty good.

That's a novel approach. Before I start welding or riveting hardware to my dropouts, I'll fabricate custom replaceable clamping dropouts. But, I'll only consider that approach if all else fails as this would involve major surgery and heat-treatment of my aluminum frame.

If the nuts are getting loose because of back and forth kick of acceleration and regen combination, I suggest to use a simple locking screw as commonly used by gunsmith, locksmith, etc... You just need to drill and tap for a small screw just aside a flat of the axle nut after it is reasonably tight. This screw will prevent the axle nut to get loose, as it will need to be removed to free the nut.

Interesting idea. But, locking screws, cotter's pins, etc. are designed for low-torque installations. High torque installation is required to prevent axle rotation during torque reversals, which is my goal. If the axle is allowed to rotate enough for the flats to strike the dropout wall, then wear on the axle flats and threads will occur and lead to axle fatigue. The problem using a locking screw is that it requires the nut and axle to be in perfect alignment at the required high pre-load. Due to the slight axle thread deformation upon each re-installation, the positions of the locking screw holes will drift apart.

Although I'm willing to live with the current situation that may require axle nut re-tightening every 300 km, I'd like to find the minimum pre-load torque for long-term use, where "long-term" is longer than 300km. That figure appears to be around 50-55Nm for a 14mm axle installed with Nord-lock washers under both axle nuts, where regenerative or plugging braking force is similar to that of a friction brake (e.g. over 2000 watts).

 

[Punx0r]

IMO the axle does not need any initial play in the drop out to cause damage (rounding) of the axle. It's soft enough that it will occur anyway, just take a little longer. The simplest, most effective solution is clamping torque arms.

 

[MadRhino]

IMO the axle does not need any initial play in the drop out to cause damage (rounding) of the axle. It's soft enough that it will occur anyway, just take a little longer. The simplest, most effective solution is clamping torque arms.

It does need to have some play (friction) to wear. That is why tight fit through axle torque plates or clamping dropouts are best, because they are not giving the axle any play.

The dropouts or torque plates need to stop rotational friction. Higher torque on axle nuts will not, and even when there isn't any rotational play the nut will get loose by the hammer effect of motor kicks. For that a lock washer, lock screw or even a nylock or thread lock is good, but none are good means to stop friction caused by rotational play.

 

[Punx0r]

The axle is soft enough that it will deform even in a very tightly-fitting drouput and create its own play, then wear accelerates. I've had this using hammer-tight torque plates (8mm thick). The axles are soft and the forces applied to the corners of the flats simply exceeds the yield strength of the material. It's like using a spanner on a soft bolt head.

 

[fechter]

I have the nuts tightened to about 70Nm on my A2B Metro motor. That's slightly over the factory recommendation. I'm running about 2.5kW and use regen. So far they have not loosened significantly but I do see some bash marks on the torque arms, indicating the axle moved at some point. After fixing a flat, I did have one side loosen a bit after the first ride. Re-torqued and it has held since.

 

[mr.electric]

I find it is general good practice to check ebike axle bolts whenever tuning/ repairing the bike. I frequently find one side loose. The expensive norlock washers as on Stromer bikes seems to stop the nuts from loosening. Using a longer nut like a truck lug nut has been discussed here on the forum before.


Sent from my iPhone using Tapatalk