so in my latest obx install, I got the diff torn down & rebuilt as usual. But when I got the ring gear on it and the assembly in the case, it was visually obvious that the ring gear was running out quite a bit. I had seen Bone's and a couple of other guys' threads and vids regarding the runout at the stem, but I wanted to get measurements at the ring gear flange as well. So I got ahold of a junk case and drilled and tapped a hole for my dial indicator

here's what I came up with...


1st test: Post-rebuild, case torqued in typical criss-cross pattern. Runout at ring gear flange measured 0.25mm (0.0098") yowza.

http://www.youtube.com/watch?v=V3hJ24Fu2cM


2nd test: Same as test #1, in closed and torqued case. Runout at ring gear flange measured 0.11mm (0.0043"), but diff felt tight due to the bearings binding from the case runout

http://www.youtube.com/watch?v=3kDH6XeIWOE


3rd test: Diff case bolts were removed, reinstalled 2 bolts loose enough to allow for adjustment of case alignment but tight enough to hold it in place after adjusting. Runout at ring gear flange measured 0.07mm (0.0028")

http://www.youtube.com/watch?v=pfl1cZEeVqI


4th test: Same as test #3, in closed & torqued case. Runout at ring gear flange measured 0.03mm (0.00118"), and diff spins way more freely in the case now that the radial misalignment in the bearings has been mostly alleviated.

http://www.youtube.com/watch?v=izPITr1qQlg


it seems like the bottom line here is where the case halves mate together isn't exactly a tight fit and allows for some adjustment. Luckily it was enough in this case to get it within acceptable tolerances

knowing that the monkeys at the obx factory don't care much about stripping bolts, belleville spring orientation, cases being torqued down evenly, etc etc, I would assume they don't bother making sure each unit isn't running out either. I am making this runout test and adjustments part of my rebuilds from now on


thanks to Zack at Black Forest Motorsports for letting me borrow a couple taps that I didn't have

 

Runout at the ring gear flange doesn't matter. Proper way to check runout is on v-blocks. I've had enough experience with crap runout on OBX's to tell you that if you can see the wobble with the diff in the clutch housing, it is trash.

 

Not sure what you are accomplishing by shifting the cases around, but the only thing that does is tip the rotational axis. it doesn't fix run-out.

 

I didn't rotate the cases on each other, my point was that there is enough play in the machining that it is easy to misalign the two halves. All I did was shift the cases and retorque the bolt

and aqua, I don't see how you figure that the ring gear running out doesn't matter, considering that gear tooth contact is critical in handling load on the gear. This one was running out enough that it most likely would have bottomed out in the countershaft and broken one gear or the other, or both, and/or the case as well, or the entire trans

 

The OD of the ring gear flange isn't precision machined like the bearing flanges and actual mounting surface of the RG flange are, not that anything on the OBX is precision. If you assemble the trans correctly there is no adjustment of the case, put the dowel pins back in.

 

again, shifting the cases just tilts the axis, runout is not changed. Don't know how you can shift the case at all with the dowel pins in place.

 

are you guys talking about the dowel pins in the trans case? The adjustments I'm talking about are to the diff case itself ie when you do the obx teardown & rebuild, NOT shifting the outer trans cases around. You know that lip where the two diff halves go together? THAT's what I'm talking about- it's less of a tight fit than you think and allows for some minor adjustment when torquing the 9 housing bolts, allowing you to alleviate the runout. If you have any obxs laying around, disassemble one and reassemble it without the belleville springs so that there is no preload on the diff case. Then play with the radial alignment between the two diff halves. It's alot easier to see what I mean with the washers out so you're not fighting with the preload. I'll get some more shots to illustrate what I'm talking about next time I'm at the shop but if you can see the markings on the diff in the vids it should give a better idea

 

Runout is not caused by the two case halves being misaligned, its caused from the production process. The diff case has a dowel pin as well.

 

You are correct, although aligning the diff halves could be a (bad) way to fix (band-aid) the runout problem. However you should not measure it the way you are doing it. As good of an idea as it seems, what happens if the flange you are measuring is not perfectly concentric with the diff ends? then you are only compensating for the difference between where the diff sits in the bearing and the flange that the ring gear bolts to. This will still put an increased amount of stress (and therefore strain) on the diff and its components since the alignment of pretty much everything could still be out, although technically the ring gear would be very well aligned.

Maybe i miss understand something though.

 

^Which is why RG OD flange runout is useless.

 

:sigh: if you really want to get into this...

you guys are missing the point. You have to remember that these diffs are not made in one piece. They are machined as two seperate halves which then are fit together. The problem is there is some amount of slop when you bolt the halves together. That is what everyone is referring to as "runout". This is not technically runout. Runout is when something that is one piece is true at one end, but spins off-axis at some point- middle, other end, or wherever. Or somthing ends up with a different id or od at some point where it should be uniform. Like say something that is machined from one angle or one end but you have to change its position in the machine or transfer it to another machine and work it from another angle or from the other end. It is very common for the working centerline to change. This is possibly the most common cause of runout, other than working with worn equipment or heat fatigue from heat treating or welding.

Now if you are measuring runout on the same half of the case that is mounted in your fixture/bearing/etc, then yes that is true runout, be it radial or axial. If you are measuring at the opposite half of something that is comprised of two halves, then it is more likely a misalignment issue between the halves, be it radial or axial. If it is axial, then you definitely have a problem unless the off chance that you have a situation like you have two case halves each faced at a 1* angle. That way if you bolt them together perfectly opposing each other, they will zero out and the entire unit should run true when assembled. Any other orientation of the two halves will result in both halves leaning off of the centerline from each other to varying degrees of between anywhere greater than 0* and 2*. In the case of these diffs, correcting this would be difficult due to there being the 9 offset case bolt pattern with the 1 dowel pin. I'm not going to get into how you could go about correcting that

A radial misalignment is when the two halves don't line up along the axle centerline, whether it be from machining error OR just slop in the machining which allows for the halves to be bolted together off-center from each other, which was the problem in my case. Simply taking extra care to ensure that the two halves are mated together exactly on the same centerline means it spins true, no matter where you measure it from- stem to stem, stem to middle, etc.

my point is I think you guys may be mistaking axial misalignment which means the diff is mostly fked, for radial misalignment which means you just need to bolt it together more accurately

like I said earlier if anyone has one laying around, take the belleville washers out, better yet take all the guts out and put just the two case halves together and see how much play there is in sliding them around on each other. Its not a tight fit

 

Yeah time is always the real important factor eh.

How much power and what sort of setting is this being used for?

 

to dumb it down a bit, your measuring process is useless.

Two V blocks, ring gear mounted, now measure the runout AT the ring gear.

BTW I've never seen a Honda OEM setup be perfect either.

Using your methods measure the runout AT the bolt circle, where it matters.

 

I agree with Mr Bone. ring gear tooth contact surface run out is what matters. not the tip of the teeth(major OD). The OD of the ring gear is not necessarily concentric with the tooth contact surfaces, which is the important part. A dowel pin or steel ball between each tooth might work as a measuring point for the indicator, but If I were building a lot of diffs, I woud make up a fixture that allowed a drive gear with a radially adjustable axis position to mesh with the ring gear, adjust the lash to zero at the tightest point on the ring gear, then rotate the gears to the loosest point and check the lash with feeler gauges. That would give a better indication of radial run out.

 

so here in hawaii, we dont have good machine shops and im not that inclined in machining language.

so what exaxtlly is the runout? where are you measuring it? are you shaving where the bearings goes on to make it spin straight?

 

you just have to make sure the case halves are aligned. That way the diff will spin straight, you will have no ring gear runout and the bearings wont bind in the case

 

would you care to elaborate, st motorsports

if aquafina or anyone one else care could jump in an help me

 

this may help illustrate what I mean

there is enough slop in the bolt holes, dowel pin hole, and the lip where the case halves go together that it makes getting the case alignment dead on very difficult

again, I'm not saying this will fix all runout issues with these diffs. This is just how I got mine straightened out.

 

great picture! It explains the misalignment very well.

 

there should be no slop at the dowell pins, that is what they are there to prevent. If there is slop, the machining is poor. In that case align with your indicator, tighten bolts, and re drill/ream dowel holes to larger size or drill new ones. The ring and both halves should have a register(shoulder) on the inner diameter to prevent misalignment any way, if it doesn't, it is poorly designed or machined. ie. junk.