34 replies to this topic

[Spider]

I have a non official listing which states 0.001" Preload.

 

I set them with 0.003" preload. IMO, 0.001" is not enough, they'll settle by that much in the first mile no matter how well they are done.

[MontpellierVanMan]

+1 for Moke and "settling" of the bearings - my car is in bits after 1.000 miles and the preload on the diff has changed considerably.

 

Furthermore, I have three good-looking and reliably-sourced diff gaskets with THREE quite different thicknesses - so the idea of calculating anything with without measuring is indeed a nonsense.

 

Has anyone used the "original" and more expensive gaskets from MS (and others ?) that are nominally 0.3mm / 12 thou and have lines of blue RTV-type material between the holes - on just one side ?

 

I get the idea that this raises the clamping force between the bolted zones but struggle to believe that these eliminate leaks if used dry.

 

I would normally use Three-Bond grey gasket paste very sparingly on both sides of a gasket like this (made by Loctite and re-badged for Ducati and Yamaha at least, personally preferred over other Loctite products since it softens with acetone) and am loathe to try the dry-approach.

 

Regarding the "deep-groove" bearings, am I right in deducing that there is nothing new about these at all - all standard bearings have always been referred to as "deep-groove", to differentiate them from angular-contact bearings where indeed there is a deliberate asymmetry in the construction and the contact patch.

 

Indeed, when you read the contradictory nonsense written, relating to axial and radial loads, and the pidgin English used, in the bearing description quoted here that has been lifted off the web-seller's pitch, you wonder whether it shouldn't come with a warning that "it's not because you read it on the Web that it's true".

Edited by MontpellierVanMan, 24 June 2016 - 02:23 AM.

[Spider]

MontellierVanMan - I'm devastated. 

 

I was dead set sure that what we all read on the internet is not just true but gospel.

 

 

As with anything, we can't make assumptions, like the gasket thicknesses and bearing pre-loads.

 

Another misnomer, which I think was started before the days of the web, (I think by our good friend Mr. Vizard in fact) is that incorrect biasing or pre-loading of the diff will cause the car to pull one side or the other.       Hmmmmm,,,,,,, 

Edited by Moke Spider, 24 June 2016 - 03:53 AM.

[MontpellierVanMan]

No, No Moke - the best Mini one is surely that torque steer is caused by the unequal tortional twisting of drive-shafts that are not the same length.

Edited by MontpellierVanMan, 24 June 2016 - 05:41 AM.

[Dusky]

Tortional twist anyone?

[Spider]

Dusty, I've done that quite a few times, easy

 

But, I'd rather do that then break something on the inside

[MontpellierVanMan]

What's the principle at play regarding the use of shims on one side only ? The whole thing looks satisfactorily symmetrical and I would have thought centering the diff gears was the only thing at stake.

Anyone tried splitting the shim pile in two and putting half on each end ?

[Spider]

No reason why shims can't be fitted either end or both. I have done this from time to time.

 

As you've mentioned, it's best to try to get the Crown Wheel Central on the Pinion. This is less of an issue on a Mini and especially one that doesn't make much power.

 

On the Big Wheel Mokes, with the huge torque multiplication through firstly the final drive (4.2:1) and then the Big Wheel (610 mm rolling diameter) This is very important of the Pinion will split in very short time. I actually machine down the LH Diff Side cover to move them over another 0.050 to 0.060". Every bit counts.

Edited by Moke Spider, 24 June 2016 - 11:20 AM.

[Ethel]

No, No Moke - the best Mini one is surely that torque steer is caused by the unequal tortional twisting of drive-shafts that are not the same length.

 I've wondered about this, not so much the shafts twisting as the angles in the joints. 

[MontpellierVanMan]

We're all apparently wide off the mark .................... hp oil drilling is what's needed ; I read it on the Internet -

 

"I've seen FIA race minis that have the middle of the shorter drive shaft bore drilled (high pressure oil drilling) by a certain amount allowing it to twist the same amount as the longer one..this completly removes the torque steer on the car if the calculations are correct!"

[nicklouse]

Or what about that dif side plate that allowed two short shafts to be run?

[Spider]

 

No, No Moke - the best Mini one is surely that torque steer is caused by the unequal tortional twisting of drive-shafts that are not the same length.

 I've wondered about this, not so much the shafts twisting as the angles in the joints. 

 

 

If your pumping lots of HPs, that would be closer to the mark - in regards to the drive train, but IMO, 80% of 'torque steer' comes from the inaccuracies of the front suspension.

 

If there was that much drag / friction through the CVs (inboard and outboard), I think you'd have a considerably harder time steering them at all than what we do.

 

With the drive shafts we all run in our Minis, any twist from torque in them would be some small as to be academic.

 

 

We're all apparently wide off the mark .................... hp oil drilling is what's needed ; I read it on the Internet -

 

"I've seen FIA race minis that have the middle of the shorter drive shaft bore drilled (high pressure oil drilling) by a certain amount allowing it to twist the same amount as the longer one..this completly removes the torque steer on the car if the calculations are correct!"

 

Oh well, there we have it   another internet break through!

Edited by Moke Spider, 24 June 2016 - 09:15 PM.

[tiger99]

These bearings are indeed the deep groove type, and a preload of a few thou would be about right. I would be very nervous about going higher than 4 thou. I think there is a taper roller conversion, and these are not supposed to be preloaded at all.

 

As to the gaskets, are there not small notches where you can put a feeler gauge to measure the clearance between side cover and diff case, and therefore gasket thickness? Or am I imagining this? Don't you do that with the transfer case gasket too, to be sure that the idler gear end float is correct? Or have I imagined this ever since doing my first diff swap in the car (just possible, not recommended) and the second one on the bench. I would have been using genuine Rover gaskets in those days.

 

As to torque steer, it was thoroughly debunked in a magazine many years ago. I forget which, but it is long gone. I will try to find the article. Basically torque steer is caused by the angle the wheel axis (CV shaft) makes with the drive shaft. The torque vectors add in 3-D such that there is a small torque trying to turn the hub fore and aft, which just stresses the ball joints a bit, and a larger torque acting around the steering axis. This is determined by the ANGLE between the CV shaft and the drive shaft. If both shafts are sloping at the same angle, and the steering is straight ahead, the torques are equal and opposite, and cancel at the steering rack. They just try to stretch or compress the rack, not produce a residual torque on the pinion. Now if anything unbalances the shaft to CV angles, the component of the torque vectors around the steering axis become unequal and we have torque steer. The steering axis geometry has to be such that the component of torque steer due to steering angle tends to centre the steering, otherwise the car would be dangerously unstable.

 

The effect of unequal shaft lengths is simply to cause the angles to be different even when steering straight ahead. If the diff is exactly level with the wheel centres, again there is equal torque, actually zero, about each hub's steering axis. Move too far away from that condition by changing the ride height (slammers beware!) and the torque steer gets larger. It is all about angles, not about lengths, and the lengths only come into it because when the drive line is deflected from its ideal position the lengths affect the angles.

 

There used to be a lot of utter guff about the drive shaft stiffness on Wikipedia, not surprisingly contributed by someone in the USA. I fixed that a while back, and last time I looked, it was still fixed. The stiffness is such that the torsional resonant frequency of shaft plus wheel is so high that any effects of shaft wind-up are over in a millisecond or so, and will not be perceptible. Some cars with very different driveshaft lengths have a damping device on the long shaft, which is in fact a torsional damper to discourage audible vibration and fatigue of the shaft, and does not and can not reduce or eliminate torque steer, which is a low frequency or static phenomenon. By static I mean that if you drive in a steady state condition, a circle, with torque applied, the torque steer will be there, and constant. The shaft wind-up will have completed very quickly, and will not be changing.

 

If you must be a perfectionist, use the proprietary device which extends one side of the diff by 2 inches, so you can use 2 short shafts. But you will be disappointed by the results.

 

This sort of thing is of far more concern to those with fine, worthwhile projects which involve a "foreign" engine, often of much greater power than the A series. There may be a problem with shaft lengths and angles, worsened by having to shift the powerplant to one side for clearance. You may like to take inspiration from the Talbot Express van (had one, dreadful thing!) and certain Peugeot models, probably many more, which run a long shaft from the diff to a bearing mounted in a bracket low down on the engine block, where there is an inner pot joint, so the shaft from there to the wheel is exactly the same length as the shaft direct from diff to wheel on the other side. The long side will be nowhere near as stiff in torsion as the short side, but there isn't any torque steer while steering straight ahead.

[Spider]

These bearings are indeed the deep groove type, and a preload of a few thou would be about right. I would be very nervous about going higher than 4 thou. I think there is a taper roller conversion, and these are not supposed to be preloaded at all.

 

 I will try to find the article. Basically torque steer is caused by the angle the wheel axis (CV shaft) makes with the drive shaft. The torque vectors add in 3-D such that there is a small torque trying to turn the hub fore and aft, which just stresses the ball joints a bit, and a larger torque acting around the steering axis. This is determined by the ANGLE between the CV shaft and the drive shaft. If both shafts are sloping at the same angle, and the steering is straight ahead, the torques are equal and opposite,

 

Tapered Rollers can withstand a much higher pre-load than Ball Rollers, not to say that's the advice given with the Tapered Roller conversion you speak of here. I looked at fitting Tapered Rollers to the 4WD Gearbox I did a few years back, but the case and flanges simply aren't up to any benefit tapered Rollers can offer. They also take up more Bearing Width real estate than a Ball Roller and have a high rolling friction.

 

<EDIT: Preloading bearings where a need to maintain alignment, in things like Wheel Hubs and Diff, is a necessity. By Preloading, the bearings not only hold alignment, but also last much longer as, done correctly, it removes spot loading of the bearing and transfers the load and heat across the bearing.>

 

 

Tiger, I know you're on relaying what you've read (and that's appreciated), however in regards to the Torque Steer debate, this is exactly what CVs are fitted ie, they do not have these effects. The early Land Roves used Double Universals on the front wheels and they did have some of these effects.

 

Another de-bunk on the Torque Steer discussion is that at standard ride hight, the drive shafts are straight, so the angle through the CVs is Zero, yet this 'torque steer' is still present. Taken further, even with suspension changes, the angles are so small.

 

On the other hand, what the Factory Data on the steering geometry on a Mini?   Camber is + 2⁰ with a tolerance of +/- 1⁰ and Caster is 3⁰ with a tolerance of +/- 1⁰.

 

So, 50% of the Camber angle can be lost or gained and 33% of the Caster before the Subframe is fitted to the car.

 

        How consistent are the rubbers?

        And Bump-Steer?

        The Road?   (and camber thereof)

        Tyre Grip to the Road?

 

I've driven a Lot of Minis with 'Torque Steer', which way should the steering pull on 'Torque Steer'?  I've had some go left and other go right? I've even had some even go straight ahead (my Moke for one does).

 

IMO, Torque Steer from the Transmission (Drive Shafts, CV,s etc) alone is another urban myth.

Edited by Moke Spider, 26 June 2016 - 10:10 AM.

[tiger99]

The torque steer is nothing to do with the joints being CVs. That only affects cyclical modulation of rotational speed if they are not. The torque vectors must sum in 3D to zero if the joint has zero friction and is centred on the steering axis, which passes through the ball joint centres. If there is any angular displacement of the shaft, there must be a torque on the third axis. It is sort of like the well-known triangle of forces except that it is torques in 3D and the analysis is awkward.

 

The shaft angles are small, but the drive torque is large in comparison to the ideal steering forces.

 

It is true that I did read about it, but only as confirmation of what I as a physicist (lapsed) already knew. In a perfect car it is the only mechanism for causing torque steer when the wheels are straight ahead. As I said, while cornering it is unavoidable, especially as there is always some Ackermann.

 

You are of course totally correct about the many other possible assymetries which will have significant effect. The worst is possibly a worn tie rod bush on one side, followed by a steering rack that is not properly centred, and that one effectively happens at random when a rubber mounted subframe arbitrarily shifts on its mounts..

 

The Land Rover Tracta joint which perhaps was not in the originals is a true CV of a rather primitive but robust type. Its friction may be the problem.

 

As for the tapered roller bearing, they are certainly more robust. However, every single car that I know of, which uses tapered rollers in the hubs, has specific instructions about avoiding preload and adjusting for a couple of thou end-float. There is a reason for that. It may not apply everywhere, in fact the diff in a rear axle is commonly quite heavily preloaded, but that is to keep nasty hypoid bevel gears in correct engagement. It causes a lot of friction and wastes power. The Mini diff is driven by parallel helical gears which do generate some side thrust and so any running clearance should be kept to the bare minimum. What I would like to know is why taper roller hub bearings must not be preloaded while in a diff they can be.

OK, you don't want overheating and a hub fire, as I have seen happen on several nearly new Mercs some years back (front wheel, either side, lubricant and/or brake fluid ignited followed by fire spreading under bonnet, and if the Fire Brigade were not quickly on the scene, bye-bye Merc...), but my idea of preload would just be finger tight.

 

You may then wonder, as I do, why, considering the interminable problems with Mini rear hub bearings of dubious origin, why no-one has introduced an adjustable taper roller conversion. The stub axle would need to be machined with a groove for a reliable anti-rotation washer to engage in, but what is the fundamental difference between a Mini rear hub and the front hub of any small late 1950s car? Ford and Triumph, amongst others, used adjustable taper rollers, but apparently BMC did not, at least in the A35 and derivatives including Sprite, and the Moggy, which are all angular contact ball bearings with fixed spacer. Little or none, bar the reliable anti-rotation washer behind the nut. The adjustable bearings used to do 100k miles at least, or the life of the car, if looked after. All of this is only relevant to understanding why a particular bearing type may have been chosen, and even then there seems to be no sensible choice.

 

But the problem initially was about set6ting bearing preload and gasket thicknesses, and I think we have not succeeded in satisfactorily solving that one yet.