MG-Cars.net

Welcome to our resource for MG Car Information.

Recommendations

Parts

MG parts spares and accessories are available for MG T Series (TA, MG TB, MG TC, MG TD, MG TF), Magnette, MGA, Twin cam, MGB, MGBGT, MGC, MGC GT, MG Midget, Sprite and other MG models from British car spares company LBCarCo.

MG Midget and Sprite Technical - wheel bearings article finally finished!

Howdy,

I finally finished the article on midget front wheel bearings, combining everything together into one place, with clear technical background behind each point. Hopefully, successfully.

http://www.mgexperience.net/article/mg-midget-wheel-bearings.html

I welcome your input, especially if you find any typos!


thanks to everyone who contributed to this effort,
Norm "boy, are my fingers tired" Kerr



Norm Kerr

Norm
Thats great work, thanks for the dedication to the cause. I will file a reference copy in my workshop manual.
Ian
Ian Webb '73 GAN5

Norm,

Good work by the look of it, I'll have more of a look later. The section modulus study is interesting but I think the analysis is too simplistic in that it fails to take into account that the bearing race-spacer-bearing race assembly can't support tensile loads only compressive. I think a true analysis would involve working out the effective section modulus of a composite of the stub axle and half the race-spacer-race assembly which would support the compressive loads along with some part of the stub axle.
David Billington

I totaly agree with you David.
R Mcknight

Nice work Norm, thanks for all your time and energy you put into the research on this problem. I'll never figure out the spacer purpose entirely I guess. I can see the need to not compress the ball type bearings too much so as to lose their operating clearance, but I still can't see where it adds any significant structure to the spindle. I would understand if it was actually under compression, torqued to a spec , but as I understand it is torqued and then backed off to the first position on the castle nut where a cotter pin will go in thus removing most if not all of the compressive force. May be only a few ten thousandths of an inch or so, but that still allows the spindle to deflect ever so slightly in relation to the upright portion.
The bit about the original spindle haveing no radius at the spindle to upright joint doesn't match my 73 which I am pretty sure is original and has a small filet there. But any bearing, ball or roller type has to have a matching filet to clear or they won't seat properly as you said.
Bill Young

Hi Bill,

you don't back off the nut, you tighten it to the next castellation (I think you are thinking of a '60s or '70s Big Three wheel bearing!).

And the spindle fillet radius is 1.5 ~ 2mm. Please don't get mixed up with the hub radius that was introduced by some careless hub maker recently. I was really worried about that issue causing confusion unnecessarily!

Norm
Norm Kerr

Hi David,

I would be interested to see what you can come up with. From studying the cut section of the assembly, it seemed to me that BMC intended for the "pipe" made of the inner races/spacer "stack" to be in compression. The spindle itself, is in tension, holding that larger, outer "pipe" in compression, by the hub nut.

Since materials are typically stronger in compression than they are in tension, it is interesting that BMC designed it this way. If I am understanding the design correctly, the inner spindle sees as little of the suspension loading as possible. Mostly, it just sees the load input into it by the hub nut (a 40 ~ 70 ft-lb, well defined, and controlled load). The larger (1.7x section modulus) structure the "pipe", instead, takes all of the suspension inputs, and transfers them into the base of the stub axle where it joins the king pin.

It is always tricky to "reverse engineer" an existing design when the original engineers are not around to question about it, and this is especially true then that design is rather unusual, compared with what much of the rest of the industry has done. But, I think it does make good sense, what they did, from a machine design point of view.


Norm
Norm Kerr

Norm,

The assembly is in many respect the same as a concrete structure with pre-tensioned reinforcing, concrete being poor in tension similar to cast iron. The race-spacer-race stack can't take tension and is under initial compression due to the tension in the stub axle itself.

<<Since materials are typically stronger in compression than they are in tension>>
A very broad and in many ways incorrect statement. The only common engineering materials I can think of where that would apply would be concrete and cast iron. IIRC cast iron having about a 1/4 or a 1/3 the strength in tension as in compression. Steel and other common materials are the same either way.
David Billington

good analogy! reinforced concrete, that I think you've hit it right on the head.

And I think you are right: it seems like they did it backward, putting the cast iron stub axle in tension, and the steel "pipe" in compression.

That fact has bothered me too, but can you think of any other way to interpret what they did?

The interesting thing is that the other option, of leaving the spacer out, would then put all of the road input loading into that (relatively) weaker cast iron, instead of some (most? all?) of the input being taken by the steel "pipe".


Probably what the designers did, was they calculated the tension in the cast iron, from the 40 ~ 70 ft lbs of torque, and deemed that to be within the safe working limits for it. If that is true, then, even though cast iron is weaker in tension than it is in compression, if the loading was within its working limits, it would still be a good design, wouldn't it?

Also, if it works like that, and the "pipe" takes most or all of the road input loading, then the, say, 70ft lbs max tension in the cast iron would be almost totally axial, compared with the (unknown) road input loading which would almost certainly include bending stresses into the cast iron. And bending would put some localized portions of the cast iron into very high, and harder to predict, tensile loads.

Norm "the suspension is killing me!" Kerr
":oD


This conversation is great. I think that I should remove that far too general comment you noted in your post, out of my article, and replace it with what we are saying here.
Norm Kerr

Norm,

I think the spacer is cast iron and the stub axle forged steel. Fairly easy to test using the old spark shower check and compare against some known samples. Any one got a stub axle to contribute to the cause for testing.
David Billington

I think you are concerning yourself too much with the action of the pipe. The "pipes" primary function (and perhaps it's only function is to space the bearing apart in such away as to make them a double ball bearing.

The modern vehicle front wheel bearings are exactly the same as this set up except the "pipe" is alraedy built into the bearing when purhased

Here is a typical modern front wheel bearing with double row balls.

http://is.gd/ifpsG

Difficult sometimes to see the Pipe, but this is the same as the Midget bearing but already assembled.

It is necessary to have 2 rows of balls (as in the Midget) to avoid rocking of the wheel, just not sure why it was necessary for the space to be 1.5 inches apart as shown in the diagrams I sent you.
Robert (Bob) Midget Turbo

LOL,

I probably have been over thinking things.

David and Bob, I agree with you, the spacer's very important function, and all that anyone really need concern themselves with is that it prevents the bearings from being compressed, and keeps them the necessary, few thousands apart, for their maximum service life.

The effect that the spacer's presence has on the section properties, and on fatigue life of the structure are, possibly, of secondary importance to that.

A modern double row ball bearing has this structure contained within it, preventing any of these questions from coming up!


Norm
Norm Kerr

It is good to see you guys sorting this out. Once you have it worked out I should tell you about something else I figured out about this project. Something I wish I had realized a year and a half ago. Doesn't change what you are working on now though.
tomshobby

what, then you don't tell us??

Tom, are you going to keep us in suspense?


Norm
":oD
Norm Kerr

I don't want to just throw this out because it might be taking the lid off the can of worms. I think it is better for a few to consider it first.
Norm, I will send you an e-mail.
Dave and Bob, send me an e-mail and I will return it with what I am thinking. tomatsmithtr6.com
tomshobby

wow. excellent article. thanks.
Rebecca
R Harvey

Guys,

I can confirm that the only reason for the spacer is to allow the correct clearance between the balls and the race. A spacer is needed because the races are angular contact and therefore the bearings have no inbuilt retainer as a standard ball bearing does, i.e the ball to race clearance is inbuilt in a standard type and cannot change or come apart.

Norm,
Your comment about the unusual design compared to the rest of the industry, I don't fully concur with. Most wheel bearings of this era and later are similar and of similar spacing, the only difference being that the majority are taper rollers and therefore do not need a spacer to keep the inner races apart as this function is fulfilled by the outer races and the hub spacing.

I don't believe that 1.5 inches has any particular relevence other than being a convienient length, but you require a reasonable distance to spread the load across the whole stub axle.

Trev

T Mason

Norm-
An excellent contribution to clarifying an often-misunderstood yet vital subject. Keep it up!
Stephen Strange

David and Norm,
I disagree with the concrete under tension analogy.
The reason is this. when the concrete is held under tension by a wire in the center it is the concrete on the outside of the wire that is the target of the increased strength.
But with the spacer on the stub shaft is held under tension it is the stub on the inside of the spacer that is the target of the increased strength.
ie - inside of one vs outside of the other is the target of the effect.

Here are a few things to consider. (as if we need more)
Cast iron has resistance to deformation. It has more ability to experience shock force (wheel hitting pothole, or wheel in hard corner) with out bending or breaking than many steel alloys.

Another thing to consider is that alloy steel, cold or hot rolled, can have a great amount of tensile strength (resistance to stretch failure) and still bend relatively easy.

And this, non-cast steel will be much more ductile or easily bend and stay bent, less resistance to deformation than cast.

We also do not know the toughness of the stud material. That being resistance to break from flexing.

So here is what I think, especially with the cone shape of the spacer, think triangle, and that it is used on end where it would be incredibly strong.
I think that the spacer sill add a great deal of resistance to fatigue of the stub axle by providing resistance to the flexing of the stub axle that does occur while driving. And the stub does flex, if it could not flex it would be brittle and break. And there is a point where flexing will by itself cause brittleness and fatigue.

ie - the purpose of the spacer is to keep the correct bearing spacing AND to reduce the flexing of the stub shaft. It is not necessarily there to add strength to the stub axle. I see the strength issue is an entirely different matter.

Some trivia, did you know? That cast iron gets is strength from it's "skin". And it gets it's resistance to shock from it's softer insides.
tomshobby

Tom,

The concrete is not held in tension but in compression. As concrete is poor in tension it is subjected to a compressive load by the tensioned steel members so that in use when subjected to bending the area subjected to tensile stress stays in compression but at a reduced level compared to the unloaded state. I still see the analogy as a good one.

Regarding <<Cast iron has resistance to deformation>>, if you're referring to stiffness then cast iron is between 1/3 to 2/3 less stiff than steel depending on grade so for the same load a cast iron part of the same dimension will deflect significantly more than a steel part. Steel is also stronger than cast iron, especially in tension where CI is poor, so will take higher loads before taking a permanent set or failing. So steel is stiffer than cast iron and significantly stronger so would stand up better in all respects compared to cast iron. Cast iron does have good uses though. I suspect the spacer to be cast iron, other possibilities are malleable iron and nodular iron and they have different characteristics again.

Cast iron having a skin all depends on the casting process and whether the skin is chilled and cooled rapidly compared to the core, it isn't necessarily the case that all cast iron castings have a skin. Sounds more like you're referring to a case hardened piece.
David Billington

Hi David,
I agree with your statement but that is not what I was saying.
With the spacer the attempt is to strengthen or provide support for the stub or center piece that is under tension.
With the concrete the attempt is to strengthen or provide support for the concrete or outside part.

My original purpose of working oh the bearing problem was twofold. The first importance was to find a correctly fitting bearing which has been accomplished.
The second purpose was to find a replacement bearing that would be readily available in case of bearing failure while touring. That has not been found.

I believe the drawings I e-mailed you might lead to solving that second problem. And I am working on that idea.

I also believe the spacer is required for bearing spacing. I also believe the spacer is important for limiting the amount of stub flexing.
I am less convinced that it is of much value for added strength.
tomshobby

regarding the spacer, and its role for front hub, the author of this MGA website raised an interesting point about bearing loading:

(excerpt from http://mgaguru.com/mgtech/index.htm )

"...be aware that the inner race spacer is a structural part of the front bearing spindle assembly. With the spacer in place both ball bearings will equally share the axial load when you are taking a corner at speed. With the spacer removed one bearing will have to take all of the axial load, effectively doubling the load on that one bearing, which could lead to premature bearing failure. "

and this, about the spacer's role in the hub strength/durability. I think he said better / more clearly than I:

"When the spindle nut is tightened the spacer and inner races of the bearings become a structural part strengthening the spindle."


I think we are all in agreement that the spacer is important. Whether its primary purpose is bearing spacing, stub axle strength, or bearing durability is probably less important than that it is doing all of those things to some degree, and that any one of them is important enough to justify using it.


Norm
":o)
Norm Kerr

Norm,

I don't see how having the spacer in there can make both bearings share the axial load as each bearing can only take load in one direction and the directions are opposed. In cornering the inner bearing of the outside wheel will take the axial load.
David Billington

David B - unless the inside wheel is actually airborne surely it takes some of the axial load too?
David Smith

David Smith,

The comment I was responding to was about the spacer making both bearings in one wheel share the axial load not the split in axial load between the inside and outside wheel.
David Billington

ah, see what you mean - agreed :-)
David Smith

The quote refers to the MGA design, which uses two Conrad deep groove bearings, not angular contacts like the Spridget. This may or may not result in both bearings sharing axial load, depending on actual tolerances and clearances on all the parts involved. Axial load is a red herring anyway, since total axial is relatively small and transient compared to bearing capacity which is a constant rating over a long time period. Similar for pure radial loadings.

For either type bearing, the big one is overturning moment, which on an outside wheel will be in and up on the inner bearing and down and out on the outer. In effect, there is no true axial load since this is applied off-axis as a moment. These forces will be on the order of 6-8 times the axial loads, and act on each bearing at the same magnitude but opposite direction. This is the same force set that the bearing spacer works against and why the spacer is important for spindle rigidity.

The moment loads are also applied only to one side segment of each bearing race, not distributed around the entire race circumference as true axial loads are, so they are again far higher in effect.

Roughly, for a 2000 lb car with equal weights on all 4 wheels, capable of a 1g downhill turn (entire car weight on that wheel), the maximum loads are:
Radial 2000lb divided by two bearings = 1000 lb upward per bearing
"Axial" 2000lb theoretical net lateral both at the spindle and at the tire/road (to turn the car at 1g), split up and in on inner, down and out on outer;
Inner now has 3000lb up/in, outer has 1000 down/out.
These forces must balance to this point, ie, the car is stable relative to the world.
BUT, for loads in the spindle and bearings, this must be multiplied by the ratio of tire radius/bearing separation, or about 12/1.75 = 6.86, so:
Inner load is 3000x6.86 = 20580lb up/in;
Outer load is 1000x6.86 = 6860 down/out

These forces clearly do not balance, and the balancing force comes from the CG of the car acting through stresses in the spindle and bearings. The local forces in the spindle & bearing set are likely much greater even than shown here, since they act on the bearing races at various unknown effective angles, giving "wedge" effects. This may be a reason for taper rollers being better for heavy loading, since the angles of contact are pretty well stable and known; effective contact angles might result in lower tensile stress in the spindle.

Gets quite muddy, but the point is that this is no simple system, and the forces can be extremely high.

FRM
FR Millmore

I am not an engineer and really don't understand the maths behind discussion of the spacer. But conceptually I always thought it was there to prevent or at least reduce any tendency for the stub axle to bend at its root. The increased diameter afforded by the spacer is surely going to do that when the assembly is clamped up, even allowing for the running clearance of the bearings.
A bit like fixing your Christmas tree into a slice of log to stop it from tipping over when the cat climbs up it.
Guy Oneandahalf Sprites

This thread was discussed between 03/12/2010 and 22/12/2010

MG Midget and Sprite Technical index

This thread is from the archive. The Live MG Midget and Sprite Technical BBS is active now.