Gearbox locked

[Bjorn Nilson]

The 350kg is indeed just to seat the bearing in the case, and not involved in the measurement (or the preload itself), so it's an upper limit, and doesn't involve shims.

That is what I think as well. The 350 kg will only seat the bearing cone flat on top of bearing rollers with a small pre-load. But pressing more than 350 kg would probably cause a too big pre-load as the cone won't spring back in the case. If the cone is seated correctly (350kg), the shim will add correct/final pre-load. So seating the bearing at first step with correct pressure is important.
Using a dial gauge to obtain correct thickness on the shim must be the best method.

 

[Dr.Jeff]

That is what I think as well. The 350 kg will only seat the bearing cone flat on top of bearing rollers with a small pre-load. But pressing more than 350 kg would probably cause a too big pre-load as the cone won't spring back in the case. If the cone is seated correctly (350kg), the shim will add correct/final pre-load. So seating the bearing at first step with correct pressure is important.
Using a dial gauge to obtain correct thickness on the shim must be the best method.

Hower the manual does say you can do either method. So I doubt there really is much to be gained by the "pressure" approach.

I'm interested to figure this out because I'll be doing two of them in the near future.

 

[Kendall]

I think there's possibly something being missed in this discussion.

When measuring the preload with feeler gauges you are checking in this location, between the bearing retainer and the body. This gap must be .003" to .005". This is done with the assembly put together and following their steps to rotate the differential to seat the bearings.
When using the other method you are instructed to apply 350kg to the bearing to seat it in the case. Then you measure from the outer face of the case to the outer lip of the bearing. Next you flip the bearing retainer over and measure from the face that contacts the bearing to the flange that sits above the case. The difference between these measurements plus the .003" to .005" gap is the required shim thickness.

At the end of the day you are measuring the same dimensions. I would argue that the first method will be more accurate as you're measuring the parts as they will be used. Have I helped or just confused things more?

 

[Bjorn Nilson]

I think there's possibly something being missed in this discussion.

No, nothing is missing. I agree that any measuring method would work, but I prefer the dial gauge because it is more accurate.
However, seating the bearing with correct pressure is essential (regardless of measuring method). So I think a feeler gauge is useless except for obtaining a correct shim.
If you push in the bearing cone too far it would create a preload because the bearing cone won't spring back when you remove the applied weight on it. Rotating the diff after pushing it or knocking it in to place may seat the cone flat on the bearing rollers, but the preload could still be too high. And then when adding shims, the preload would increase even more, to an unacceptable level.
The instructions in 5-Speed 101 is excellent and helped me a lot. It covers all steps to successfully reassembly the gearbox but unfortunately this info is missing.

 

[Steve Hoelscher]

I think there's possibly something being missed in this discussion. View attachment 55971
When measuring the preload with feeler gauges you are checking in this location, between the bearing retainer and the body. This gap must be .003" to .005". This is done with the assembly put together and following their steps to rotate the differential to seat the bearings.

At the end of the day you are measuring the same dimensions. I would argue that the first method will be more accurate as you're measuring the parts as they will be used. Have I helped or just confused things more?

This is the method I use. There are several problems associated with it. And as been noted already, how to do you accurately seat the bearings? I use a zip gun and a bit I made, at low pressure while spinning the diff, to ensure they are seated. A couple of typical problems: The seal carrier can bend and give a false reading. The bearing race can cock in the bore and not seat correctly. Thus I use rolling torque to ensure I have the preload correct. But there is no factory spec for rolling torque so its purely experience here.

 

[Bjorn Nilson]

I don't have the knowledge and experience like Steve Hoelcher, therefore I find the method risky for a single reason: If you push the bearing cone to far in, there's no way to bring it back again, except by splitting the box and then push the cone back, or in fact pushing it completely all way through the case, and start over from outside again (I actually did that when I decided to reset the preload).
If I cannot find a "firm" press (350 kg) to seat the bearing I will go for Steve's method; Knock in the bearing to seat it meanwhile turning the diff checking for friction. But with my lack of experience in this matter, I would prefer numbers instead of trusting my finger tip sensation.

 

[Dr.Jeff]

As I mentioned earlier I haven't reassembled any of mine yet, so perhaps I'm misunderstanding it. But how can you seat it too far? It has a ledge that the race seats against, and can't go any further.....

Or did I misread your comment Bjorn?

 

[Bjorn Nilson]

Or did I misread your comment Bjorn?

Yes you didn't get my point fully right.
The bearing it self cannot be pushed too far in as you say. -The bearing race will seat at the ledge (blue arrow) and that is absolutely ok.
But if you push the bearing cone (arrow 4) too far in, you will create a too much preload and there is no way (I think) to pull it back to adjust it. The cone is tight fitted in the case so it won't spring back.
Per my understanding you have only one try to set correct preload. If you fail, you must once again open the gearbox to remove the bearing cone, reassembly the box and then press in the bearing cone again from outside.
To avoid opening the box, maybe some kind of puller or very small breaker bar or other tool can help to release the preload if you mistakenly pushed the bearing cone too far in?

 

[fiatfactory]

@ Bjorn Nilson

you don't need a press to seat the bearing cup, you can use mechanical advantage / fulcrum and a torque wrench.

I made a tool similar to this (rough) drawing many years ago, but I loaned it out and it never came back.

I got an old bearing cup and welded a bar / rod with a hole at it's midpoint across an old bearing cup, then I used a slitting disc on a grinder to cut the bearing cup at two points (so it wouldn't get stuck in the gearbox housing)

The fulcrum / pivot point to the left was (mig) welded to a piece of metal that mounted on the four bolts used by the cast iron cover plate, directly above the centreline of the bearing is the pushrod providing the downward force, the mechanical advantage is provided by the length to the right of the downward pushrod, and the torque wrench is used to measure the required load.

all this does is "settle" the bearing cup in the gearbox housing, then you set the actual preload with the correct thickness shim

SteveC

 

[fiatfactory]

As I mentioned earlier I haven't reassembled any of mine yet, so perhaps I'm misunderstanding it. But how can you seat it too far? It has a ledge that the race seats against, and can't go any further.....

View attachment 56010

Or did I misread your comment Bjorn?

The ledge you have arrowed is the step on the differential carrier, this is where the bearing CONE sits, the shims that provide the preload seat against the back of the bearing CUP (number 4 in the picture)

It's the bearing CUP you need to settle in the casing at the correct depth, so you can measure the clearance (X) to determine the required thickness of shim.

SteveC

 

[Dr.Jeff]

But if you push the bearing cone (arrow 4)

It's the bearing CUP you need to settle in the casing at the correct depth

OK, now I get what you are saying. Yes, I misunderstood which portion of the bearing you were referring to.


This is where I feel using the shims and feeler gauge method is best (for a home assembly). Because....

the shims that provide the preload seat against the back of the bearing CUP (number 4 in the picture)

Therefore as Steve C says, that preload will be properly set by the shims when the bearing cap (#2) is torqued correctly. By measuring the gap between the bearing cap (#2) and the case (#3), you will know when the proper shim thickness has been achieved. Which is the method Steve H described earlier, which has been done successfully countless times by him. To me that seems as accurate as making a tool to try and replicate the factory one (with a dial gauge), or converting the forces to a lever/fulcrum torque approach. After all it is what the factory manual says to do when you don't have the proper factory tool (one of the two options given). But I'm sure any of those approaches will be sufficient.

 

[Kendall]

Assuming the top cup of the bearing is installed squarely, I don't think it's possible to over preload it, so to speak. If you put so much force on it that you damaged the races or bearings that's one thing but I think that's harder to do than would be thought.
How about this as a method: get a collection of shims that will certainly create a gap between the bearing retainer and the housing. Install the bearing and the over size shims and slowly tighten the retainer nuts and rotate the diff until you feel the bearing is just seated into the case. From there start taking measurements for what shims you'll need and test fit until you have the right gap. I believe that if you end up too tight at one point, loosening the bearing retainer and spinning the diff should allow the upper race to back off. If it's not a press fit it should be free to be pushed back up in the case by the bearings themselves. The amount of force required to deform the races or bearings would likely ruin them, but working with this you should not be in that realm at all. Everything is rubber until you move into plastic deformation.

 

[Bjorn Nilson]

I believe that if you end up too tight at one point, loosening the bearing retainer and spinning the diff should allow the upper race to back off.

I also thought about using that method but I am not sure the upper race will back off if seated too low because it sits really tight. This means you have only one chance to get it right.

 

[Dr.Jeff]

the upper race .... sits really tight.

Just curious. Is it possible your bearing is a little tighter than normal? Either the outer race OD or the case bore ID might be slightly off spec? My thought is in order for the bearing the seat at the correct "preload" setting it seems the outer race should be allowed to move a bit more than it sounds like your does. If it can't "slide" at all then it is subject to 'jumping' (microscopically speaking) as it moves down, which may not allow a proper preload. The top cover/cap (#2) holds the proper tension (preload) once everything is set (hince the shims). So the outer bearing race does not necessarily need to be a 'press fit' to stay in place once set and secured by the cap. Only a thought.

 

[Kendall]

There are many items that can be broken down here, and lots of good information in each of these comments.

One nomenclature item, it might be clearer to say interference fit instead of press fit (the distinction being that an interference fit has a physically smaller ID than the OD of the mating part). Something can seem a tight tolerance "press" fit while not being a true interference fit. Because this bearing is a taper seat bearing each race may be an interference fit to the case. Typically the races of a bearing need to be stationary to allow the balls or rollers to do the work of rotating. If possible, could Bjorn closely measure his case and outer bearing race? .0005" to .001" difference one way or the other could make a major difference in this application. It is possible that the preload on the bearing does keep the races stationary but I'm thinking now that isn't necessarily the case. For a tapered bearing you cannot apply too much axial load through it or it will lock up. On a normal ball bearing you could typically load it axially on the races to keep them from spinning.

I think as Steve mentioned a rolling torque would be the correct way to build this, but unfortunately there is no specification. And I think it stands that the most accurate way to set this preload is by doing it with the parts assembled to the case. That way the end result is the same as it will be in use. Always measure things in the way their going to be used. That's a major piece of my understanding of metrology and kinematics. Same goes for manufacturing/machining.

I think, based on the manual description of setting the preload, there should be no axial forces on the bearing without the retainer in place. Then with the retainer in place and the fasteners torqued, including the right gap, the clamping force from the fasteners should set the preload correctly.

This is one piece of the puzzle that I don't understand: why do they specify a gap measurement for the preload? When the retainer is on and fasteners torqued correctly, would not the clamping force on the bearing be controlled by the fastener torque? Wouldn't any gap allow for the right preload? I haven't taken one of these transmissions apart, so I don't know what the whole assembly looks like.

 

[ng_randolph]

This is one piece of the puzzle that I don't understand: why do they specify a gap measurement for the preload? When the retainer is on and fasteners torqued correctly, would not the clamping force on the bearing be controlled by the fastener torque? Wouldn't any gap allow for the right preload? I haven't taken one of these transmissions apart, so I don't know what the whole assembly looks like.

After you torque down the cover, there is no gap. The specified torque of 24.5 Nm each on 4 M8 (?) studs / nuts will give over 60 kN (13,500 lbf) of clamping force (cover to case).

A crude way to apply the 350 kg suggested in the FSM to settle the bearings (the way I interpret the FSM, the 350 kg is not the preload, and you measure after removing the 350 kg load) would be to look up, (e.g. on omnicalculator.com) what torque it would take to get to 350 kg / 3430 N of load, remembering there are four studs/nuts. Torque all four nuts to this torque (which would be much lower than the factory spec. for torque for final assembly), then turn the differential a few rotations before undoing the nuts. But again, I have never opened up an X1/9 transaxle.

 

[Dr.Jeff]

There are many items that can be broken down here, and lots of good information in each of these comments.

One nomenclature item, it might be clearer to say interference fit instead of press fit (the distinction being that an interference fit has a physically smaller ID than the OD of the mating part). Something can seem a tight tolerance "press" fit while not being a true interference fit. Because this bearing is a taper seat bearing each race may be an interference fit to the case. Typically the races of a bearing need to be stationary to allow the balls or rollers to do the work of rotating. If possible, could Bjorn closely measure his case and outer bearing race? .0005" to .001" difference one way or the other could make a major difference in this application. It is possible that the preload on the bearing does keep the races stationary but I'm thinking now that isn't necessarily the case. For a tapered bearing you cannot apply too much axial load through it or it will lock up. On a normal ball bearing you could typically load it axially on the races to keep them from spinning.

I think as Steve mentioned a rolling torque would be the correct way to build this, but unfortunately there is no specification. And I think it stands that the most accurate way to set this preload is by doing it with the parts assembled to the case. That way the end result is the same as it will be in use. Always measure things in the way their going to be used. That's a major piece of my understanding of metrology and kinematics. Same goes for manufacturing/machining.

I think, based on the manual description of setting the preload, there should be no axial forces on the bearing without the retainer in place. Then with the retainer in place and the fasteners torqued, including the right gap, the clamping force from the fasteners should set the preload correctly.

This is one piece of the puzzle that I don't understand: why do they specify a gap measurement for the preload? When the retainer is on and fasteners torqued correctly, would not the clamping force on the bearing be controlled by the fastener torque? Wouldn't any gap allow for the right preload? I haven't taken one of these transmissions apart, so I don't know what the whole assembly looks like.

First, thanks for clarifying the terminology. It offers a much better description of what I was trying to say regarding the fitment of the bearing races in the case.

Second, although you did not come out and say it, I believe you implied that the factory manual is somewhat vague here. And I fully agree, here and in many places as far as I'm concerned. I assume some of that might be due to language translation?

Third, regarding a better method to determine the preload by a rotational torque measurement (such as Steve H and I have suggested). Could that be utilized by determining a nominal torque value based on the bearing size, load, etc? Even if some assumptions are required, how much less accurate can that be than the other options we have so far? However I really do not know if that is possible, so I am asking. Or is what I just said the same as what Bjorn (@ng_randolph) just posted (I was writing this as your post came up)?

Fourth, regarding the "gap measurement" method described in the manual. I imagine they determined that given all of the parts are relatively consistent in dimensions, that gap value was found to equate to a target load value for the bearing (preload). In other words it is more of a approximation based on distance that a true reading of load force. Perhaps not a great analogy, but something along the lines of measuring bolt stretch to determine clamping force rather than torque values. Just a wild guess, I really have no idea. For that matter, is their approach of using a dial gauge any different? Itis still based on distance and not torque or other load measurement.

 

[Dr.Jeff]

After you torque down the cover, there is no gap. The specified torque of 24.5 Nm each on 4 M8 (?) studs / nuts would give over 60 kN (13,500 lbf) of clamping force.

A crude way to apply the 350 kg suggested in the FSM to settle the bearings (the way I interpret the FSM, the 350 kg is not the preload, and you measure after removing the 350 kg load) would be to look up, (e.g. on omnicalculator.com) what torque it would take to get to 350 kg / 3430 N, remembering there are four studs/nuts. Then turn the differential a few rotations before undoing the nuts. But again, I have never opened up an X1/9 transaxle.

Bjorn, your post and my last one crossed paths while writing/posting. So I'll add this thought here.

I may have it all wrong, but isn't there still a gap after the bearing cover is torqued down? I thought that was what the feeler gauge measurement was all about. Maybe I need to go back and reread the manual's description once again. I keep interpreting it differently every time based on our conversations. Sure wish they explained things better.

 

[fiatfactory]

There is no gap (for oil to leak out of) when the case is properly assembled and the correct shim thickness is used.

The gap is ONLY there for measurement after you fit a group of shims that are too thick, so you can then measure the gap (x) and subtract this value from the combined thickness of the nominal group of shims that were fitted.

once you have a number for X, you subtract the X value - less 3 to 5 thou - from the measured thickness of the "test" set of shims, , this calculation will provide you with the required (combined as there is often more than one) shim thickness for the case

You assemble / measure and fit a new set of shims that give the calculated thickness, and that shim group when the cover is torqued down gives the taper bearings their correct preload.

the 350 kg load is not applied during any measurements, it is simply to "settle" the bearings and make sure they are seated squarely and the "floating" side CUP is at the required depth

SteveC

 

[GregS]

As I see it, the key to Bjorn's problem is the fit of the cup (outer race) in the trans case. It has to be free to slide after the initial settling of the bearings (350kg load), so there is no extra preload, but it has to be a tight enough fit in the case so it doesn't spin in the case once everything reaches full operating temperature.