Ball joint shims are 1.5" od and 1.3" id. 

I made up 7 bearing spacers from .35 up to .40 wide, and wouldn't you know it, the thick OEM spacer I installed last night gave about .001 preload once I sanded off the small compression lips on each side.  I removed a bit more for .002 preload.  I would have had no way of knowing without the wider spacer as a base measurement of movement.  

Now I have a bunch of spacers that are probably too wide for any Timken bearings, but I can sand them down and have a selection to choose from. 

Im hopeful that I might actually get a decade out of the right hub this time. Now on to the left side!

As I have mentioned before I save the spacers from all sets I replace. I also match sets as they come out if the look very good. I'm amazed at how many different sizes they are. Making your own is a good thing and being able to is a part of the hobby. I'm like you in the regard I like to find one with play and custom fit to the required preload. I'm not at the shop so I'll ask what is the inner and outer of a ball joint shim?

I was confused, I thought we were working with new bearing each time. Why ball bearings, I don't know we use them on the race cars to reduce drag and heat transfer. We always carry a spare set and have never had a failure with them front or rear. Steve (CTR) 

McMaster has no shims that are 1 1/2" OD and 1 1/4" ID surprisingly.  Otherwise, that would be the way I would rather go with this.  I could not find such a shim on Amazon or anywhere else.

The front wheel bearings on an MGB have their endfloat set with shims.  The spacer is (intentionally) about 0.040" too thin, and then you shim achieve your desired endfloat, and torque the stub axle to about 60 ft/lbs.  

With all these reports of spacers being too thick and too thin (I also have a Timken spacer that is too thickand has anyone found a shim (McMaster?) that is the right OD and ID to allow for adjustment, while using a thin spacer?  Having to thin down spacers, or machining your own thicker one to within a couple of thou is not a very desireable position to be in. 

Interesting, when ROO went down with a blown front inside tapered roller wheel bearing.  We replaced with new Ball Bearing set (and CV, as inner race was stuck on the CV) at the roadside as the new bits were in our parts bin.  Torqued up with "largeish man" on end of a 12" long wrench.

When ROO got serviced, Ken in Ipswich installed new tapered roller bearings because "it is easier to get the proper preload" with them than with the ball bearings in the Metro hub.

So the ball bearing set went back into spares and the CV with inner race removed too.  We did not need either on the almost 19,000 miles of combined driving to/from Tasmania.

I suspect that what is going on is that my hubs have never been setup correctly until the last time I did this when I machined a spacer of the correct thickness.  By this time, my bearings had been overloaded, but had not yet shown signs of failing, although they were well on their way. Over the last few thousand miles, the bearing (on the inside - but I replaced both inner and outer) finally failed.  

The new bearings are further outside as would be expected, so the same spacer (even a thicker one) does not give adequate pre-load, although without a thicker one I really don't know what the spacer to bearing clearance is. 

Today I picked up a 1 foot long steel tube of the same ID and OD (thankfully it is a standard size), and I am going to part off a bunch of pieces in .002" increments that are wider than my widest one.  I'll heat treat them and then I'll have a nice collection from which to choose when I re-do my left front, which has a very slight amount of play.  

Why ball bearings?  I didn't think they would hold up to the side-load that taper roller bearings would.

As we know, somthing is going on and I'm asking the same questions I'd ask myself. Have you tried ball bearings? Steve (CTR)

Thanks, I should have said pre load for my final spacer. (had that in mind, but fingers typed something different)  I wanted to find a thicker one initially so I could have some end float to measure and then subtract from that to get .001 - .002 Preload.  I already packed my bearings and I've been doing the measurements on the car with an indicator.  I may clean everything up tonight and dry assemble on the bench as suggested, or since I've found a setup with my indicator that works, I might keep it on the car.     

To answer some previously asked questions:

I am using a large washer with no cone washer installed for all of this setup work.  Hubs look fine, and the old races had not spun.   I've not measured the center hub ring for thickness, but it should be the same as it was on day 1 as the races have never spun in these hubs.  

Im torquing to 150 ft lbs as my axle has two holes and torque wrench is a Husky (Home Depot) 50-250 ft lb.  Seems like a good quality tool, and gets really good reviews on several car and tool forums I've visited.  Certainly not Snap-On, but better than what Harbor Freight or Northern Tool carry.  

You don't want endfloat (clearance), you want .001-.002" preload.

I made some thicker spacers, if needed because my others are too thin, I just try one in there for size, measure endfloat then face it to the length needed.

[edit] If you trial assemble on the bench with both the seals out and no grease, you can pretty much feel when preload is correct. There should be slight drag when turned by hand, but it should turn smoothly, no lumpyness.

Ok, upon further diagnosis, my inner bearing on the right front is shot.   I have a spare set of timkin bearings and races, and I'm headed for the parts store now before it closes to buy some quality grease.  The lip formed on the spacer is very small (a sliver) and there are no lips on the hub or cv bearing contact surfaces.  The hub was newly replaced under 2000 miles ago at the last time I had this apart.

once I get the new bearings and races in and tighten the upper and lower ball joints a flat of the nut (they are not loose, but could be a smidge tighter), I will check end float again.   

Edit:  the spacer I had in the right hub was .32 thick. There was no end float.  My thickest spacer is .35 and still no end float.   I don't have any shims of the correct ID/OD to figure out how much wider a spacer I need, so will have to find some or just machine up a few wider spacers of varying thickness to get .002 end float.   

Steve, I only have done this on each side of my car, reading from the book, and learning from people here like you.

will be happy to meet you when you come up for the other car . Glad my pics helped.   - email sent.  Harvey

yes- false positive torque reading sort of explains it -  Good medical term

Thank you Harv, as we all know I ramble on and have been trying to shorten up. As my wife is fond of saying " Don't you realize half the time you are the only one who knows what you are talking about?" 

I'm coming up for the green car want to meet for BBQ? Steve (CTR)

OK am I correct then to say that:

1. A primary suspect in all this is the split-collar washer.

2. It is supposed to sandwich the inner races between the CV and the castle nut.

3. The CV/axle is actually held longitudinally in place by the hub's machined spacer between the two bearings.

4. The split-collar washer can bind as lankford says, or break in my case.

5. This allows endplay in the CV/axle.  Torque is lost and the bearings are chewed.

6. The flat washer technique allows you to ensure #2 and #3.

??

Craig (Isleblue - fabulous car) asked: What could cause play when the bearings are being held tightly in their races by the drive flange and cv joint and the axle nut is torqued tightly? 

So assuming the above, the answer would be a 'false positive' successful torque reading from a binding/defective split-collar washer.

Without going back and reading this long thread.....

see standard procedure in Haynes manual page 10.4 Suspension and steering section - for disc brake models/ my manual is in metric

PROBLEM: the split-collar washer (cone) may get seated/squeezed onto the shaft first, preventing tightening "to home" on the bearings

SOLUTION:

make a flat washer from mild steel - 25mm hole,  50 mm diameter, 6.5mm thickness

insert driveshaft into hub but first put on flat washer, then nut, tighten to torque spec to pull everything in place. 

remove nut and washer, then install split-collar cone washer, then nut ,  and tighten to torque spec.

Make sure the split-collar cone washer is good, not crimped crooked or chewed

Can you clarify this comment?  I'm not familiar with the heavy washer. 

I don't know what's wearing from your description nothing. That's why I ask you to mike the shoulder in the center of the hub. Is it both sides or just one side? With my driver I'd be doing wheel bearing every week. I cover 8,000 miles a year. My left front hub has a bearing issue every 40,000 and I'm considering buying new. Do you have the big heavy washer to pre torque before adding the taper collar? Quite often when I rebuild hubs people want new bearings I tie wrap the races bearings and spacers together on the old ones when they show no signs of distress. I flush and use these on my car ( left front ) as required. What are you using for a torque wrench and torque value? I'm trying to help. Steve (CTR)

Having been involved in testing some bearings, the Number One rule os never ever use the grease sachet they come with.

Dark grey grease sounds more like Mob-Special used in CVs and pot joints than an LM type bearing grease.

This is of course an excellent thread but I am still trying to figure out what the upshot is.  It seems there is a procedure here waiting to be made into something for dumb guys to use.  Last week:

1. 8 month-old S brakes/hubs/flanges; CV's 1 year
2. Clunking sounds on the left.  Castle nut lost torque.
3. Conical washer was broken.  ( The cheap sintered one that comes with the brake kit )
4. Put in Timkens and a hardened conical washer.  Used the included grease.
5. The drive flange was down maybe .010 - if that - based on the little edge it left around the inner worn area.  
6. I put the flange back in, not quite remembering how much is a problem.
7. Torqued to 180 with the little bit extra to line up the cotter pin hole.
8. So far no noise from the left, but now the right is of course protesting.

As a result I am waiting to do it all over again in 6 mos, preferring the luxury of ignorance.  And the .002% chance that it will all just be fine.

A standard Sta-Lube dark grey wheel bearing grease.  

The bearings are coated in grease, and it is fresh looking after only 1.5 years and a few thousand miles. 

I just replaced the Timkens in the RH side of mine, they had developed about .003" endfloat and the hub nut was still tight.
They have done well over 75,000 miles, grease used was Mobil Supergrease. I have used Castrol HTB in hubs before but find it tends to dry out.

Swivel hubs were not new, I just hand select a spacer from my collection to give me .001-.002" preload.
I will do the bearings in the LH hub soon, not that they need it....