bellhousing alignment question

[Belgian1979]

I've been measuring, remeasuring my bellhousing to crank alignment, and just want to check whether my reasoning here is correct.
So I set the dial indicator into the bore of the housing and zero'd it at the 12 o'clock position. I get the following readings :
- 9 o'clock position : +0.002"
- 6 o'clock position : +0.010"
- 3 o'clock position : +0.011".

So the 3 o'clock position is slightly out of tolerance. Worth correcting ?

[PackardV8]

My machinist and I get in fights about this all the time. He says let it run. I spend way too much time trying to get it within .005".

Here's a quick easy way to see if it can be brought in closer. Leave the dowel in the 9:00 position and install an eccentric dowel in the 3:00 position. Rotate that one until 12:00 and 6:00 are equal. The 3:00 will then usually be closer as well.

[Belgian1979]

Thanks. This fiddling with the bellhousing is making me go crazy.

[Belgian1979]

My machinist and I get in fights about this all the time. He says let it run. I spend way too much time trying to get it within .005".

Here's a quick easy way to see if it can be brought in closer. Leave the dowel in the 9:00 position and install an eccentric dowel in the 3:00 position. Rotate that one until 12:00 and 6:00 are equal. The 3:00 will then usually be closer as well.

I took your advice on this. I had to clock the offset dowel to 45° up from the 3 o'clock position. Like that :
- 6 and 12 o'clock were both 0.
- 9 and 3 o'clock were respectively +7 and -8.

I would say this centers the damn thing.

You saved my life here. I was about to kill myself.

[tenxal]

As an aside, these offset dowels from RobbMc makes the process a bunch easier.

http://www.robbmcperformance.com/products/dowels.html

[ClassAct]

I've run them over .060 out and it didn't matter. Once I get them fairly close I let them go. You have to be way off before it starts killing pilot bushings or breaking front bearing retainers.

Had a customer who wasn't capable of dialing his can in. Had run it for years and never checked so I helped him hoping he could learn how to do it (and yes it ain't that hard). When we got the indicator on it it was .125 out. SMH. He said it's working, should we move it? I let him make the decision. He ran it another 5-6 years before he died and his son, AFAIK, is still running it like that.

[PackardV8]

I've run them over .060 out and it didn't matter. Once I get them fairly close I let them go. You have to be way off before it starts killing pilot bushings or breaking front bearing retainers.

Had a customer who wasn't capable of dialing his can in. Had run it for years and never checked so I helped him hoping he could learn how to do it (and yes it ain't that hard). When we got the indicator on it it was .125 out. SMH. He said it's working, should we move it? I let him make the decision. He ran it another 5-6 years before he died and his son, AFAIK, is still running it like that.

Your experiences vary, but yes, most owner/builders are fortunate drivetrains are so forgiving. We've seen vibrations blamed on the driveshaft, clutch dragging, input shaft dragging in neutral, shortened synchro life and a host of other problems go away when the bell housing was dialed in.

Many aftermarket high dollar bells are worse than the factory. If one wants to really make trouble for himself, dial in the center hole, then sweep the mounting face. That's also often out of square.

[exhausted]

While I can understand the problem and how to remedy the tolerance, it all makes me wonder about all the automatic transmissions? Is it not to worry??

[peejay]

If you are zero at 12 o'clock and .002 at 9 o'clock, and .010-.011 at 6 and 3, then it sounds like the actual high spot is actually around 10:30. Move the bellhousing .005 towards 4:30 and recheck. I think. If I'm interpreting your numbers correctly.

Do this AFTER checking for parallel. When you shim the bellhousing for parallel, it will affect the centering.

It didn't make so much of a difference on old 3 speeds and 4 speeds where the tolerances inside were very sloppy, but if you are using a modern trans like a T56 then you need to get the numbers closer.

I have measured old Chevy 4 speed bellhousings to be .010" off center and .040" out of parallel!

[peejay]

Many aftermarket high dollar bells are worse than the factory. If one wants to really make trouble for himself, dial in the center hole, then sweep the mounting face. That's also often out of square.

I have not yet seen a certain bell manufacturer's T56 units that had faces that were even flat, let alone parallel. I can understand that the interrupted face required for the hydraulics means that the face can warp when it is welded up, but if we can fix it with a (relatively) quick facing in a mill, then they can too.

And yes, doing this did fix a clutch disengagement issue.

[Belgian1979]

If you are zero at 12 o'clock and .002 at 9 o'clock, and .010-.011 at 6 and 3, then it sounds like the actual high spot is actually around 10:30. Move the bellhousing .005 towards 4:30 and recheck. I think. If I'm interpreting your numbers correctly.

Do this AFTER checking for parallel. When you shim the bellhousing for parallel, it will affect the centering.

It didn't make so much of a difference on old 3 speeds and 4 speeds where the tolerances inside were very sloppy, but if you are using a modern trans like a T56 then you need to get the numbers closer.

I have measured old Chevy 4 speed bellhousings to be .010" off center and .040" out of parallel!

I haven't checked it for being parallel. I guess I first need to remove the power coating on it.

[af2]

I haven't checked it for being parallel. I guess I first need to remove the power coating on it.

That could be .020" thick in 1 place and .005" in another.....

[Belgian1979]

Well I check the paralellism of the mounting pad today and it seems to stay within .002".

However, when mocking up I found another issue that went unnoticed priorly. The Corvette C3 clutch arm is so short that it touches the sides of the scattershield (Lakewood). Now, I had to open up the window for the arm before, but apparently not enough causing the arm not to be able to draw the release bearing complety against the bearing retainer of the trans. I assume this plays a part in the whole scenario as well.
Had to open up the bellhousing some more and then grind some sleeves into the arm so I could bend it a little more backward. After a lot of mocking up I seem to have that working ok.

[ClassAct]

:

While I can understand the problem and how to remedy the tolerance, it all makes me wonder about all the automatic transmissions? Is it not to worry??

Exactly. The bell housing has to be off a good bit before you start having issues.

[BigBlocksOnTop2]

First things first! The first thing to do with set up is to mock it up on something ( in a vice) that you can pick up and turn up side down and see what the numbers look like. You will have sag from the weight of the dial indicator mounting.... the bars, the weight of the indicator. etc. The only way to know ''dial indicator sag'' is with a bench set up...exactly how you would set it up on the bell housing and see how much difference there is with it set up on the bench and turning it up side down and seeing the numbers....half way between is zero. Go from there....