Someone explain Ford 9" bearings to me...

Working on swapping the 9 inch in my truck for another housing. The new housing already has bearing races in it which look to be in good shape, and the bearings on my axles are very low miles. So I am wondering if I can mix n match and use the existing bearings on the axles with the bearing races in the housing without problem. I don't know if, for example, different bearing manufacturers use a different angle on the race or something. Thoughts?

OK now on to the part I need explained to me.

On a typical front axle tapered bearing situation, the load of the bearing is set by tightening the spindle nut. It is an adjustable arrangement. However, on the rear end, there does not seem to be any adjustability. From the outside in, the "stack" of stuff looks like this: End plate (trapped between bearing and axle flange), then the brake backing plate, then the axle housing, which contains the bearing. No adjustable parts there, that I can see. When you torque the 4 nuts that you get to through the hole in the axle flange, it's going to put the bearing into the housing end by x amount. But suppose the bearing race is sticking out by x + .005". Wouldn't torqueing the 4 nuts ruin the bearing by crushing it? Either something has to give/slip, or the bearing will get ruined in this situation, I would think.

So, does something give/slip? Or is literally everything in the stack machined with such precision that the bearing will land in the bearing race as a perfect fit every time? This seems very unlikely to me.

Somebody enlighten me please, and thank you.

are the bearings tapered? I know that in the 8.8 axle they are not. I have one apart getting ready for a rebuild and I'm pretty sure they are not tapered. it is very common for it to be necessary to use a slide hammer to pull the axle due to the bearings being non tapered. one major difference from 9 to 8.8 the bearings press into the housing for an 8.8 and the axle slips through and captured by a c clip in the diff. but the 9 has a bearing pressed onto the axle and has a collar behind to capture it. then its held by the flange

The bearings in a Ford 9" are not tapered.  You get all the right parts, press everything all the way on and bolt it up tight.  That should give you the correct positioning.  There will be more end-play that you get with correctly torqued tapered bearings, so it's not like the tolerances precise enough for that.

As far as mixing and matching bearings and races, generally I wouldn't, but I don't know that it would generally be a problem.

However what you do need to be careful of if you're thinking of mixing and matching is that you have the same bearings.  As I understand there are 6 possible combinations.  5 of them were available in some stock Ford vehicle and my Bronco has the 6th.

There are "small bearing" and "large bearing" housings.  That's a bearing OD of 2.875" or 3.15".  My Bronco has the small bearing 2.875".

And there are three different shaft diameters: 1.378", 1.531" and 1.562".  I upgraded my Bronco to 31 spline axles which came with a shaft OD of 1.562".  That's the combo that Ford never offered, and I had trouble getting the right bearing when I needed to replace it last summer.  (I guess they do offer 31 spline axles with the 1.531" OD, which is easy to get bearings for).

I was thinking about this more, and trying to remember better what my bearings look like.  It was almost 5 months ago that I had mine apart, so it was digging back quite a bit .

But at least my bearings don't have separate inner and outer parts.  Mine were a single assembly, so mixing and matching parts wouldn't be a thing.  And mine are ball bearings, not rollers.  This post from my Bronco's build thread has a picture of my bearing after I cut it off (you can go back a bit to see the entire saga of replacing that bearing).

But again, there were several possibilities for sizes, and it wouldn't shock me if there were different possibilities for bearing design either.

... and thinking about it more...

In a full-floating setup (like in a front axle) there are a pair of tapered roller bearings on each side.  Each bearing supports force in any radial direction, but each one only supports force in one axial direction.  So you need a PAIR of bearings to support the wheel in all directions.  If you only had one bearing there would be nothing to keep the wheel from coming off.

There are semi-float axles that have something like that going on.  They have a C-clip inside the differential to keep the wheel from coming off.  But the Ford 9" (and Dana 44) isn't a C-clip axle, so the (single) wheel bearing needs to support axial forces in both directions.

In the 9" in my Bronco this is done with a ball bearing.  As I said above, I don't know if all Ford 9" axles use this same style, but whatever they use, the bearing itself needs to support axial loads in both directions (in and out).

Whether all Ford 9"s use the same bearing type or not, I'm pretty sure that they all retain the bearings in much the same way, because I don't see different types of retainers.  So going back to your "stack of stuff from the outside in," as you describe above the retainer plate traps the backing plate tightly against the end of the axle housing (which of course it needs to do) but only closes off the end of the housing so the bearing outer race can't come out, with the tolerances giving the freeplay.  But that isn't the case.  The outer race stands proud of the axle housing and the retainer plate plate holds it tightly as well.  Sitting at my computer I'm not 100% sure how it works that it holds two things tight at the same time, but I'm thinking the retainer plate must tighten on the bearing race first, and then deflect to tighten up on the backing plate.

So the end play ends up being controlled in the wheel bearing, not by the tolerance stackup of all of the parts you mention.  And that makes me feel like it's probably important to keep the parts together, regardless of what the bearing design is, and not using bearings and races that didn't come together.  But keeping in mind that my experience is with a 9" that has ball bearings, that's extrapolating a bit, and I might be missing something.

My axle, and every one I have seen, is a tapered roller bearing setup. The taper is about 15 degrees, but is pretty obvious.

I sure don't understand the statement about the "end play ends up being controlled by the bearing". How does that work? If the end play is too tight the bearing is going to end up crushed. If the end play is too loose, the bearing isn't going to hold anything up, it will act like a bad wheel bearing and allow slop.

The only thing I can think of is that either the bearing itself, or the race it goes into, is designed to slip a bit as you tighten the retaining plate bolts. Then you would end up with the correct fit. But it also seems that you would not want to be putting the bearing under that kind of strain while you were actually tightening the bolts. So I'm not sure that's it either.

On the Ford 9" most have ball bearings which will handle both radial and axial loading. Some 9", notably the early F150s had roller bearings for the increase in weight capacity. These were built to handle the axial load in addition to the radial load.

This design is called a semi floating axle.

There is an in between design designated a 3/4 floating axle where there is an inner bearing for the hub and the axle with it's bearing provides the outer support. I have never actually seen one of these.

Most 3/4 ton and higher trucks have full floating axles, where the weight is carried by a pair of roller bearings on a tube extending through the brake backing plate or the disc brake assembly. On this design the axle shaft carries no weight at all, it only turns the hub.

Back to the semi floating, for years there were two basic designs, ball bearing where the axle is retained with a pressed on bearing and retainer or tapered roller bearing where the two axles ran against a thrust block inside the differential and required an adjustment for correct loading. Chrysler built axles were this way for years. Spicer axles used a simpler system, a plain straight roller bearing riding on a hardened area of the axle and the axle retained with a C clip captured in the side gear. This design depends on the axle lubricant being kept high enough to lube the bearings. If it gets too low, the rollers will wear right through the hardened surface and destroy the axle.

OK I watched this video:

https://www.youtube.com/watch?v=iPvCpb8HsUI

He removes what are clearly roller bearings, and replaces them with TIMKEN SET-20 which he never actually shows but I assume are roller bearings as well. He keeps the bearing race on the bearing at all times so you never actually see the roller bearings. But I think that's the answer to my question - the bearing race slips into the axle housing the correct amount on installation. That sets the bearing race in the proper location - for that particular bearing set.

So the right way for me to proceed would be to remove the races from the old housing, put them on the old bearings on the old axle, and tap them in to the new housing to set the bearing race depth.

The wrong way would be to just insert the old axles and bearings into the new housing and bearing races. That would produce the unknown clearance that I am trying to avoid.

Thanks for your thoughts and advice, everyone, I appreciate it!

 I just watched the video. I remember seeing it a few years ago too. I think that you will find some differences between the currie aftermarket 9" axle that he was working on and the as issued f series axle in your truck.  

Pete Whitstone wrote

My axle, and every one I have seen, is a tapered roller bearing setup. The taper is about 15 degrees, but is pretty obvious.

I sure don't understand the statement about the "end play ends up being controlled by the bearing". How does that work? If the end play is too tight the bearing is going to end up crushed. If the end play is too loose, the bearing isn't going to hold anything up, it will act like a bad wheel bearing and allow slop.

The only thing I can think of is that either the bearing itself, or the race it goes into, is designed to slip a bit as you tighten the retaining plate bolts. Then you would end up with the correct fit. But it also seems that you would not want to be putting the bearing under that kind of strain while you were actually tightening the bolts. So I'm not sure that's it either.

As I said, I'm only relatively familiar with the ball bearings in my axle.  I didn't know before, but obviously do now, that there are also Ford 9" with tapered roller bearings.  I still don't know anything about them though so I won't say much about them.

As to the end play being controlled by the bearing, in my axle the outer race of the ball bearing is held tightly in the housing.  The outer race bottoms in the counterbore in the end of the housing, and the retainer plate holds it tight.  So there is no end play associated with the fit of the outer race since there is no freeplay in the outer race.

And the inner race is pressed tightly on the axle shaft, with the heavy retaining collar being pressed on as well to ensure that the inner race doesn't move on the axle shaft.

With no movement between the outer race and housing, or between the axle shaft and the inner race, all of the movement is in the bearing itself.

Again, this is with the ball bearing setup in my axle.  I don't know how end play is managed with the tapered roller bearing.

Pete Whitstone wrote

....  He removes what are clearly roller bearings, and replaces them with TIMKEN SET-20 which he never actually shows but I assume are roller bearings as well. He keeps the bearing race on the bearing at all times so you never actually see the roller bearings....

Here's an ad for Timken SET 20 bearings where you can see the ends of the rollers.  Again, I'm no expert on this, but overall the bearing set looks very similar to the ball bearings in my axle, with the inner race, outer race and rolling elements (balls or rollers) all contained in one assembly that isn't intended to be disassembled.  In the picture it looks like the visible end of the rollers is captured by flanges on both the inner race and the outer race.  If that's true on the other side too (and I suspect it is), then this bearing would also control the end play.

But that's just off my seeing the picture, so no guarantees.

I didn't really look up the Timken 20 product, but yeah you are right, it does appear to be a non-tapered roller bearing set. In which case it would not matter a great deal exactly where in the housing it sits. It could be flush with the stop at the back of the housing, or it could be more towards the retaining plate side and I don't see that it matters. That would absolutely not be the case with tapered roller bearings.

I'd consider switching to non-tapered roller bearings like these, if it didn't mean having to break down the old ones already on the axle, which are super-low miles.

mat in tn wrote

I just watched the video. I remember seeing it a few years ago too. I think that you will find some differences between the currie aftermarket 9" axle that he was working on and the as issued f series axle in your truck.

I don't really see any obvious differences. What exactly are you referring to?

Pete Whitstone wrote

But I think that's the answer to my question - the bearing race slips into the axle housing the correct amount on installation. That sets the bearing race in the proper location - for that particular bearing set.

Well, I'm wrong again. The REAL way this works is that the bearing itself sets the clearance. The bearing "carrier" (with the roller bearings) and the bearing race meet each other at their edge. So when you put the two halves together, the proper clearance is established by the two mating edges of the two halves.

So the bearing DOES get driven all the way into the housing up to the stop. And the reason it's safe to drive the bearing in is, while you're pushing it into the housing, you are not putting pressure on the roller bearings, the carrier is acting as an installation tool for the bearing race.

Duh. I feel kinda stupid now.

you figured it out! there is absolutely no reason to feel stupid.  one of the differences is the retainer plate and another is that you should have non tapered rollers. the timken set 20 is a tapered set. one of ways this gets confusing is the fact that the ford 9" has become the standard design for years as a hot rod and racing conversion. so, there are a few companies making ford nines that are NOT ford nines. custom housings for 28, 31, 35 splines etc. the one in that video was one of those and not a ford truck.in the truck the depth is set by the bearing in the housing captured by the retainer plate on the outside and the locking collar on the inside. the axle can carry weight evenly on the strait rollers better this way. the splined ends just spline in and are carried along with the carrier in the carrier bearings.

mat in tn wrote

 one of the differences is the retainer plate and another is that you should have non tapered rollers.

"Should have" is kind of a maybe in this case, I'm not working with the trucks original 9". I took it out so I could put another 9" housing in, the one I've been adapting the Dodge brakes to. While the bearings from the old axles went into the housing just fine, I found that the retainer plate has a different bolt pattern on it. I was aware that there were different patterns but I always heard them associated to the bearing size (big bearing is one pattern, small bearing another). After measuring, I find the housing with the Dodge brakes on them to be a "new style" or "Torino" style. So I have to change the retaining plate.

OK next question. In the stack of stuff on the axle, you have this from the inside out:

Retaining collar
Bearing race
Roller bearing
Seal
Retainer plate

So is the seal pinned in place by the retainer plate and unable to rotate, making the point of "friction" (point of movement) between the diameter of the axle shaft and the outer (smaller) part of the seal?

Or does the seal rotate with the roller bearing, basically fixed in relation to the rotating axle, which would make the point of friction the other end of the seal?

I hope that question makes sense. Thanks!

i kind of look at it from the axle in. axle, retainer plate, seal facing the bearing, bearing, collar. the seal sits on the recess of the bearing and the larger part gets sandwiched "held" in place once the four bolts are tight and the wear/slip point is at the axle surface.

mat in tn wrote

 the wear/slip point is at the axle surface.

Thanks, that's kind of what I figured. The reasoning being, if the entire seal was able to turn with the axle, then there would have to be some lateral (in and out) play in the axle to allow that, and what happens when you go around a corner and the seal is forced into the retainer plate?

withe locking collar there is little to no walking . on an 8.8 the axles are held by a c clip in a recess with the "walking" end play held against a main center axle pin. the 9"is locked in place at the collar.