The ultimate torque arm (For <2kW and standard axles)?

Go get a plate of Steel a 4in grinder with cutoff wheel. Make a template get some chalk or a soapstone and a vise. Weld on two nuts and drill one out.
 
Chalo said:
Buk___ said:
In the thin axis, these springs are capable of being wound tight, and stretched flat and back again many millions of times. Each infinitesimal section of metal can be flexed though 20°. 30°, 90° or 180° without harm.

Flat springs are thin throughout their lengths. Your TA is "thin" (not in absolute terms, but relative to the whole) in a single spot on each side. You'll get material failure there as soon as static friction is overcome and the thing is called upon to do something.

Hm. Have you heard of Hooke's Law?

What you are suggesting is that if I take a piece of spring steel and apply a force to it as the green arrow below left, it will bend and return as any spring, but if I apply the same force to the same material in the direction of the red arrow, it's suddenly going to become fragile and break.

Sorry, but your intuition is wrong on this occasion. Steel is isotropic. That is, its properties are the same in all directions.

So if the material can withstand the force (green) across its thin section, it can certainly withstand it across the thicker section (red).

Of course, unsupported, it might twist (distort) away from the force, but in my design, I have many layers acting together and constrained either side so that they cannot twist. Per the green arrow on the right.

junk56.jpg

Maybe this will be less likely to throw your intuition out of kilter.

If the design looked like the dark grey area below, I don't think you'd have any problem seeing that when the axle tries to force the jaws apart, the force paths are clean and uninterrupted with smooth transitions.
junk57.jpg
Now imagine nothing changed except I put back the material in light grey. Does the part suddenly become weaker because of extra material?

(You might find this, particularly the sections on tensional stiffness, spring energy and isotropic materials instructive.)
 
999zip999 said:
Go get a plate of Steel a 4in grinder with cutoff wheel. Make a template get some chalk or a soapstone and a vise. Weld on two nuts and drill one out.

Some context as to what this is meant to address would be useful?
 
Your plate is excellent for trying this approach, which I used for its superior hold, yet it was stupidly simple (perfect for me), and dirt cheap. It is though, a bit of a pain to undo the two bolts when you need to change a tire. An open notch will do fine if its thick enough, and fits tight. Usually not so tight fit is how they fail.

mongoose pinch dropout.JPG
 
Buk___ said:
Chalo said:
Flat springs are thin throughout their lengths. Your TA is "thin" (not in absolute terms, but relative to the whole) in a single spot on each side. You'll get material failure there as soon as static friction is overcome and the thing is called upon to do something.

What you are suggesting is that if I take a piece of spring steel and apply a force to it as the green arrow below left, it will bend and return as any spring, but if I apply the same force to the same material in the direction of the red arrow, it's suddenly going to become fragile and break.

What I'm pointing out has nothing to do with materials properties but everything to do with your part shape. You can't put a narrow neck in the middle of a piece that's loaded in bending without causing stress to concentrate there. If you put a much thinner spot in the middle of a flat spring, it too would permanently bend or break at the thin spot.
 
Chalo said:
What I'm pointing out has nothing to do with materials properties but everything to do with your part shape. You can't put a narrow neck in the middle of a piece that's loaded in bending without causing stress to concentrate there. If you put a much thinner spot in the middle of a flat spring, it too would permanently bend or break at the thin spot.

I'm sure you've seen if not handled a pair of old-fashioned sheep shears:
Sheep_Shears.jpg


That thin section is where the stresses concentrate, because that is exactly where they want them to concentrate.

Ditto with my design.

And it doesn't matter if stresses concentrate somewhere, if they are well below the yield strength of the material, and these are.

You cannot ignore the material properties or the reason for my choice of material.
 
If you can't tell the functional difference between an entire loop of flat spring added as a feature, and a localized thin spot created by cutting a slit partway across a plate, I'm afraid I can't help you.

Make one and test it to failure with a torque wrench. See for yourself what you can't seem to understand otherwise.
 
Chalo said:
If you can't tell the functional difference between an entire loop of flat spring added as a feature, and a localized thin spot created by cutting a slit partway across a plate, I'm afraid I can't help you.

Make one and test it to failure with a torque wrench. See for yourself what you can't seem to understand otherwise.

Would you still be so concerned (about the stresses involved) if the part looked like this?

junk58.jpg

Because if you can bring yourself to admit that you wouldn't be, then look at this:junk59.jpg

And realise that all I did was remove the light grey areas.

But the key point is that adding back the removed material does not make the part weaker!

The stresses -- both the maximum values and the areas of concentration remain identical.

And even if only one layer were used, the stresses involved in the axle pushing the wings apart by 0.25mm or 0.5mm or even 1.0mm would never cause the part to fail, no matter how many times it happened. They are so far below the material's yield stress, they do not inflict any damage on it. And as the wings are in constant contact, there is no impact stress, nor shear stress; just good old-fashioned tension and compression.

Of course, with just one layer, there would not be enough counter pressure to serve the purpose of denying the axle the room to accelerate, so it wouldn't be fit for purpose.

So, I added more (9 or 11) layers, which increases the counter pressure by an order of magnitude, whilst leaving the stresses per component the same.

That's why design is different from guesswork.
 
You either fail to understand or else fail to acknowledge that in stiffening up that stressed underside arch by adding material, you cause a huge stress riser at the relief slit. That isn't guesswork; it's plain to anybody with mechanical engineering chops. The hard inside corners of the "ears" would have a similar effect even if the connecting arch were left with a slender snap-ring-like profile.

You can't just go putting notches and sharp corners on parts that are stressed in tension. When you do that, you're telling it exactly where to fail, and to do it quickly.

Make it and test it. Then you'll see what I'm talking about.

plane_stress_notch_syy_602DF947-F23B-4D34-F07C68352FC1FFFB.jpg
 
Buk___ said:
The design I posted is such that the jaws are the same diameter as the axle nut flange -- based on those supplied with my motor -- so if the flanged nut fits inside any lawyer lips (my bike doesn't have them), so would the jaws.


Many motor axle nuts dont have any flanges, or any kind of ribbing/etc on any side of them (completley smooth otehr than the threads inside), though they often come supplied with flat washers, which may or may not be small enough diameter to fit within any lawyer lips on dropouts front or rear.

Doesnt matter if this isnt intended to be a (semi-) universal solution, but if it depends on a certain diameter / type of axle nut (im not technical enough to be sure from what Ive read here so far), then that nut would need to be supplied with it to ensure proper operation.
 
Buk___ said:
In your last post you mentioned that the hardness of the materials that the Nord-Lock was in contact with [paraphrase] was critical, and went on to list those as "the torque arms and axle".
Just to be clear (since Im always trying to learn much more about mechanical engineering types of stuff and so follow all of these types of threads, but I get lost sometimes in them, requiring many re-reads) are you referring to this post?

Alan B said:
The rocking back and forth of acceleration and regen seems to defeat star (and most other locking) washers and requires wedglock type washer (such as NordLock) to prevent loosening. These require the surfaces they engage (nut and torque arm in this case) to be softer than they are to work properly.

If so, then Im not sure if one of you said nut and meant axle, or the other said axle and meant nut, or if I really missed the point and its something completely different.
 
You can't just go putting notches and sharp corners on parts that are stressed in tension. When you do that, you're telling it exactly where to fail, and to do it quickly

Chalo, you sound like you may have some legitimate concerns, could you throw together a quick and dirty drawing of what you consider to be a useful shape, one that is close to what the OP is trying to use?

Using riveted stacks of sheet steel is a common way to manufacture padlocks, and laser-cut sheet steel might be a useful method for a TA...

M15XDLF.jpg
 
Chalo said:
it's plain to anybody with mechanical engineering chops.

Sorry pal, but you do not have any mechanical engineering chops. You're a spanner monkey with ideas above his station.

When you start posting something more than your hyper-inflated opinion, of stuff you so obviously have little more than a passing understanding, of -- you know, some thing like a calculation, an appropriate formula, even just logical reasoning why you think adding material makes something weaker -- then come back.
 
amberwolf said:
Buk___ said:
In your last post you mentioned that the hardness of the materials that the Nord-Lock was in contact with [paraphrase] was critical, and went on to list those as "the torque arms and axle".
Just to be clear (since Im always trying to learn much more about mechanical engineering types of stuff and so follow all of these types of threads, but I get lost sometimes in them, requiring many re-reads) are you referring to this post?

Alan B said:
The rocking back and forth of acceleration and regen seems to defeat star (and most other locking) washers and requires wedglock type washer (such as NordLock) to prevent loosening. These require the surfaces they engage (nut and torque arm in this case) to be softer than they are to work properly.

If so, then Im not sure if one of you said nut and meant axle, or the other said axle and meant nut, or if I really missed the point and its something completely different.

I also thought -- and mentioned obliquely -- that Alan_B may have misspoke when he suggested that the axle needed to be softer than the nordleck washer for it to work. I'm pretty sure that instead of "the torque arms and axle" he should have said: ""the torque arms and axle nut".

In order for the Nod-Lock washer to be effective, the ridged outer faces (as opposed to the ramped inner faces) of the washer pairs:
junk60.jpg
need to be able to 'bite' into the the two clamping surfaces -- the dropout and axle nut, or torque plate/arm and axle nut -- in order to operate.

They need to be able to grip well enough that when undoing the nut -- which causes the washer in contact with that nut to slide and rise over the ramp in the other washer -- they do so without either washer slipping against the other material it is in contact with. The design relies on them staying put against that other material whilst being subjected to huge** torque.

That means those surface materials should be less hard than the washers in order to obtain the required bite.

**A standard 'high tensile' (8.8 grade) M12 bolt is normally recommended to be tighten to a maximum torque of 85Nm; Nord-lock's recommendation for M12 8.8 when used with their washers is 107Nm!

Personally, the more I learnt about them, the less likely it became that I would ever use them. They are warrantied for one use only. I really wouldn't want to subject my ungraded motor axle to that level of strain.
 
amberwolf said:
Buk___ said:
The design I posted is such that the jaws are the same diameter as the axle nut flange -- based on those supplied with my motor -- so if the flanged nut fits inside any lawyer lips (my bike doesn't have them), so would the jaws.


Many motor axle nuts dont have any flanges, or any kind of ribbing/etc on any side of them (completley smooth otehr than the threads inside), though they often come supplied with flat washers, which may or may not be small enough diameter to fit within any lawyer lips on dropouts front or rear.

Doesnt matter if this isnt intended to be a (semi-) universal solution, but if it depends on a certain diameter / type of axle nut (im not technical enough to be sure from what Ive read here so far), then that nut would need to be supplied with it to ensure proper operation.

My bike doesn't have lawyer lips, and I've never owned one that does, so I'm the wrong man to address that issue.

The standard bike axle is M10, and they have a maximum (across-corners) dimension of a gnat's under 18mm.

Hub motors come with M12 or M14 axles, with across-corners dimensions of 20mm and 23.5mm respectively; so they would not fit inside the lawyer lips intended for M10 nuts.

An M10 flanged nut has a flange diameter of just under 22mm, so if the lawyer lips were designed for use with M10 flanged nuts, then they might get away with using an non-flanged M12 nut; but whether that combination of circumstances ever arises I have no idea.

In any case, anyone fitting a hub motor to a frame that has lawyer lips, will need to deal with that issue regardless of whether they use TAs. How they chose to deal with it is their choice and not my province.
 
I've seen all kinds of torque arm and dropout failures over the years. With a powerful motor, the axle could exert enough torque to make the bike do a wheelie. Some fairly simple math can give you a ballpark about how many Nm we are talking about. Just the friction between the axle flange and the dropout contributes a significant amount of torque resistance as long as the nuts stay tight.

Any torque arm that is C-shaped will have a strong tendency to spread apart. A full circle around the axle is much more likely to survive.
I like the concept of clamping the axle by tightening the axle nut so there is no free play. Especially if you have regen the free play will tend to work the nuts loose.
 
spinningmagnets said:
You can't just go putting notches and sharp corners on parts that are stressed in tension. When you do that, you're telling it exactly where to fail, and to do it quickly

Chalo, you sound like you may have some legitimate concerns, could you throw together a quick and dirty drawing of what you consider to be a useful shape, one that is close to what the OP is trying to use?
junk61.jpg
His concern is that the undercuts shown in red are a fatal flaw in the design; despite that it is obvious that provided that the sections of material roughly indicated by the green arrows are sufficient to withstand the forces the design requires them to withstand, which they are, with a safety factor at least 3, then the presence of those slots has no bearing on the utility of the design.

Indeed, without those carefully sized slots, the section would be too thick to act as a spring as intended; the spring rate would be too high.

The best explanation is this image:
file.php
.

If you apply a given force to a single flat steel spring 0.5mm x 3.6mm (the minimum section at the back of the slot), in the direction of the left green arrow, and it caused it to wiggle back & forth by 0.5mm, nobody would have any concerns about there being any chance of fatique of any kind.

I've turned that (same) section of that same material through 90° to the same applied force, but the section (sorry to repeat myself) hasn't changed. So the stress calculations remain identical. The presence of extraneuos material above and below that section, does not weaken the component, nor change the calculation.

I've then added a bunch more layers meaning that either: the individual components will experience 1/10 the stresses; or: the combined layers can provide 10 times as much energy absorption, so resist 10 times the torque.

But enough pointless repetitions of perfectly clear arguments with the only counter response: I'z finks its gonna be busted.
 
fechter said:
Any torque arm that is C-shaped will have a strong tendency to spread apart.

That is exactly what it is designed to do. Absorb the energy rather than fight it, thus preventing the axle flats from impacting the dropout faces.

It's novel, but not that hard to understand.

fechter said:
A full circle around the axle is much more likely to survive.

no. A full circle will, of necessity, have to have clearance, which means the axle will accelerate and the impact it. Sure, you can make it thicker and thicker to cope, but it will still be fatigued. You've just moved the impact from one component to another.

fechter said:
I like the concept of clamping the axle by tightening the axle nut so there is no free play. Especially if you have regen the free play will tend to work the nuts loose.

Exactly. If you prevent the axle from twisting -- and this mostly does -- you prevent it from imparting torque to the nut to loosen it.

The axle will tend to spread the jaws slightly, but like a rubber ball, the more energy it imparts into the springs, the more they impart back.

Most of that energy will be dissipated as heat from friction, and due to the combination of the fulcrum effect of the axle and the C shape, causing the forces to oppose each other -- one side will try to rise up as the other is pushed down; and vice versa -- very little of the energy will get as far as the retaining bolt, but with the 4:1 ratio of torque moment, it will easily handle whatever residual force makes it that far.

It's like the karate kid, rather than meeting force with force, it turns that force back against itself. Literally turns it via the C shape.
 
I went back and checked, and I was clear - the wedge locking NordLock washers engage the nut and the torque arm. To do so the washer must be harder than either. The axle is not involved in this unless it fails under the tension of the nut, which is a different problem.

We have seen thick solid chrome moly steel dropouts with no relief slots spread by 12mm hubmotor recoil when they were slightly less than fully seated. I think this is a somewhat similar "force multiplier" to the slots in this proposed design.

Thus I wonder about the viability of this approach.

The "C" type torque arms I have seen that did not spread and fail were:

1) the same thick chrome moly steel dropouts, but with axle fully seated, and no weaknesses like slots, so the axle was in essentially full contact with steel on both flats and one curved side,

2) with open end wrenches employed as torque arms, these are hardened and quite thick,

3) commercial ebikes that generally have lower current limits and lower power, generally in steel frames.

Limiting the motor current controls torque, but this thread states a power limit of 2kW, which doesn't limit the torque to a modest level. At low speed with no back EMF the motor current at 2kW can be quite high. If the motor had 100 milliohms resistance the motor current could be 140 amps at 2kW.

There's not much point in arguing about it, only testing can prove anything. Build some and fail them with a torque wrench, as Justin has done so many times.

Calculations might be correct, or might not be. We have no way of validating them adequately here.
 
Alan B said:
There's not much point in arguing about it, only testing can prove anything. Build some and fail them with a torque wrench, as Justin has done so many times.

What I'm sayin'. I don't have to prove the design is bunk; he can just try it out and see for himself.
 
Chalo said:
Alan B said:
There's not much point in arguing about it, only testing can prove anything. Build some and fail them with a torque wrench, as Justin has done so many times.

What I'm sayin'. I don't have to prove the design is bunk; he can just try it out and see for himself.

YOU can't prove the design is bunk; you don't have the know how.

I will test it; but having done the math, I have a high degree of confidence in it.
 
Buk___ said:
I also thought -- and mentioned obliquely -- that Alan_B may have misspoke when he suggested that the axle needed to be softer than the nordleck washer for it to work. I'm pretty sure that instead of "the torque arms and axle" he should have said: ""the torque arms and axle nut".
AFAICT, searching his posts for the word axle and rereading this thread, tahts what he did say (though he did not use the word axle, only the word nut), and that was why I asked the question because I was confused since he appears to have actually said what you think he should have said.

Now I am even more confused. :?

I left my original post quoted below, since it contains the references Im confused about, but Ive bolded the phrase in question.

amberwolf said:
Buk___ said:
In your last post you mentioned that the hardness of the materials that the Nord-Lock was in contact with [paraphrase] was critical, and went on to list those as "the torque arms and axle".
Just to be clear (since Im always trying to learn much more about mechanical engineering types of stuff and so follow all of these types of threads, but I get lost sometimes in them, requiring many re-reads) are you referring to this post?

Alan B said:
The rocking back and forth of acceleration and regen seems to defeat star (and most other locking) washers and requires wedglock type washer (such as NordLock) to prevent loosening. These require the surfaces they engage (nut and torque arm in this case) to be softer than they are to work properly.

If so, then Im not sure if one of you said nut and meant axle, or the other said axle and meant nut, or if I really missed the point and its something completely different.
 
amberwolf said:

I guess I was mistaken. Strange though that I would quote it, and then misread it.

Even stranger that you would come along and make such a song & dance about such an off topic point so long after the fact.
 
Wedge Locking washers like the NordLock must engage the two faces on either side to work. One side faces the torque arm, the other the axle nut. Teeth on the outside of the washer pair engage these surfaces. Wedges between the inner surfaces of the washer pair cause the washer set to become thicker when the nut is rotated in the removal direction, so to release the nut takes more torque than for it to stay where it is, so the nut is caught in a minimum "valley" and tension increases for it to move in either direction. So once it is settled in (which requires a couple of re-torquing events) it takes effort to remove it. Vibration or small motions of axle vs torque arm can't loosen it unless they exceed the wedge rotation angle (or the teeth fail to engage the surfaces).

Look for utube videos on how they work if there is any further question.

They will almost certainly work whether we understand them or not.

In my experience they do take a couple of rides and a few re-tightening events to settle in, after that they don't tighten more and don't move. When I go to release the nut I can feel the tension increase before the wedge angle is exceeded, and the nut "pops", after that it is loose. It is a great solution for keeping nuts tight without wires or cotter pins that are a pain to take out and reinstall, or troublesome glues.
 
Back
Top