Clutch / Torque Converter: Bellhousing runout and correction

[jpepper]

It is commonly accepted that it is necessary to dial indicate non-factory engine/bellhousing pairings to insure proper alignment. I am posting this to alert our membership to dial indicate ALL pairings, even if it is factory original. I have measured and aligned numerous bellhousings. Some factory installations are misaligned as bad as a random selection. The pattern I have recognized is that quite often, the rear bellhousing flange on the block is out of square (90°) to the crankshaft centerline. Every error I have found is the top of the flange surface is leaning toward the front of the engine. The ones I have measured have always been accurate side to side. Shimming the bellhousing so the face is perpendicular to the crank usually brings the bore in alignment as well. Evidently Studebaker had an error in their fixturing used for machining the blocks and it was never corrected. Assembly just did the best they could and lived with the error. The most recent example I just worked on was a 259 which was out .012" from the base of the flange (crank centerline) to the top. I have found several others that were out .020". If you project this measured error to a theoretical full circle, the total indicator reading error would be .040 for a .020 measured error. If the engine is to be rebuilt, check this feature and have the machine shop correct it if possible. If the engine is together, your only option is to shim the bellhousing. I add enough shim at the top to establish face alignment. I snug the bellhousing using 3 bolts (the two at the bottom and one in the center at the top with the shim). You then measure the gap at the other bolt hole locations and make appropriate shims for those locations. I buy shim stock assortments from McMaster Carr (.001,.002, .003, .005, and .010 increments). It is easily cut with a snips or scissors and punched with a hole punch on a block of hardwood.
One last note, when checking face runout, always put some light thrust pressure on the crankshaft so it does not float back and forth giving an inaccurate reading.

Jim

[firestoper 25]

Engine assy. must do data alignment

[QUOTE=jpepper;1131820]It is commonly accepted that it is necessary to dial indicate non-factory engine/bellhousing pairings to insure proper alignment. I am posting this to alert our membership to dial indicate ALL pairings, even if it is factory original.

Thanks for the refresh on alignments Jim. When I put the bell housing & flywheel I bought from you this spring on my engine run-in stand I will be sure to follow your data. Sherm / Green Bay

[PackardV8]

The pattern I have recognized is that quite often, the rear bellhousing flange on the block is out of square (90°) to the crankshaft centerline. Every error I have found is the top of the flange surface is leaning toward the front of the engine.

Hi, Jim,

After all these years of working radially, learning the block may be out of square is very bad news. So you've found it's the rear face of the block which is out of square to the crankshaft centerline? Usually tilted in at the top?

But thinking about it, not surprising, as most blocks benefit from being square decked. When we make the block decks square to each other and to the crank/cam centerline, we locate off a 2" round bar and locating rings which ride in the main bore. We'll have to make an adapter and an arm to hold a dial indicator to sweep the front and back faces of the block. That would eliminate crankshaft end play as a variable. Have you ever checked the front face of the block? If the front face is found to be square, it could be clamped on the bed of the surfacer and the back milled square to that. There goes another $150 in machine time, but it would be subtracted from the shop time of dial indicating the bell housing.

Evidently Studebaker had an error in their fixturing used for machining the blocks and it was never corrected. Assembly just did the best they could and lived with the error.

That would explain why the Studebaker Shop Manual is the only one I've ever read which contains the section on dial indicating a changed bell housing. I always said it should be done on any brand, but it's embarrassing to learn Studebaker probably knew they were putting out improperly machined blocks and never corrected it.

jack vines

[altair]

Crankshaft alignment is all over the board I think they set these blocks on the boring machines by hand and eyeballed them. I once had two or three engines and disassembled them and mistakenly did not keep the main bearing caps in order. When I tried to assemble one engine with the mismatched main caps it just bound up tight, they weren't even close. However with a handful of shims I managed to align the bearing caps to make it work. So yes they are out substantially. That is why some engines will leak and some won't, some will wear out bearings quicker than others, shifting gears may grind, clutches may howl all because of this misalignment and sometimes from the factory.

[jrlemke]

If the back of the block is out, couldn't you machine the bell housing to correct? -Jim

[PackardV8]

Crankshaft alignment is all over the board I think they set these blocks on the boring machines by hand and eyeballed them. I once had two or three engines and disassembled them and mistakenly did not keep the main bearing caps in order. When I tried to assemble one engine with the mismatched main caps it just bound up tight, they weren't even close. However with a handful of shims I managed to align the bearing caps to make it work. So yes they are out substantially. That is why some engines will leak and some won't, some will wear out bearings quicker than others, shifting gears may grind, clutches may howl all because of this misalignment and sometimes from the factory.

A thousand times no! Main caps must remain in their assigned position in their original block. The main line is always machined with the numbered caps in place. They don't fit correctly anywhere else.

Never shim main caps; that can only cause the bores to become oval and can't correct for lateral misalignment. If it becomes necessary to replace a main cap, the main bores should be line-honed back to round and square.

If the back of the block is out, couldn't you machine the bell housing to correct? -Jim

Chasing an out-of-square block rear face to end up with an out-of-square bellhousing is bad science. The correct way is to make the rear of the block perpendicular to the main bores. The field expedient is as Jim recommended; shim/dial indicate the bell housing into alignment with the main bores.

jack vines

[Dan Timberlake]

Hi Jack,

I look under a lot rocks on line, and in antique/junque stores.
I think I've seen clutch bell housing alignment checks in a few old factory manuals.


1949 Pontiac for sure. Both bore and face.

http://pontiac.oldcarmanualproject.com/manuals/1949/1949%20Pontiac%20Shop%20Manual/12-Transmission%20and%20Gearshift%20Control/image15.html

http://pontiac.oldcarmanualproject.com/manuals/1949/1949%20Pontiac%20Shop%20Manual/12-Transmission%20and%20Gearshift%20Control/image16.html

[TWChamp]

What PackardV8 said. It's absolutely essential that all rod and main caps be numbered and remain in the correct position.
I knew a guy that worked on a Buick straight 8 and never marked the rod caps. Naturally when he put it back together, the engine was locked. Now, to figure the number of possible arrangements for 8 rod caps, you need the factorial for 8, which means 1 x 2 x 3 x 4 x 5 x 6 x 7 x 8, which equals 40,320. That could keep you busy trying to find the correct placement.

[altair]

It is conceivable that the blocks and bellhousings are for the most part square it is the crankshaft that is off, left or right or up or down. It is easy to see if a square block is sitting off square in the jig by one or two thou and the boring jig is off a few thou the final result could be off multiple thousands. The average life of a vehicle was maybe 75 - 80 thousand miles. With a few leaks, vibrations, chattering clutch, grinding and sticking gears was the norm for the average life of the vehicle. The model T Ford differential was assembled with no provision for any adjustment and was assembled for the average climate and driver with up to .040 clearance between the crown and pinion. Today people try to assemble the rear ends with .003 - .005. To make up the difference a #2 grease was used in lieu of oil, today people use oil and there are several opinions as to the proper weight and grade of oil.

It would be very difficult to perfectly align the crankshaft with the transmission shaft as the crankshaft could be out in any direction or in more than one direction. The front main bearing in the transmission would have to take all the thrust, there is some limited allowance. The factory would set these up as close as need-be however they may not have been perfect because it was not necessary. Today I think we try too hard to make it perfect when a reasonable tolerance is acceptable. With a misaligned crankshaft as most are and a transmission that fits with out binding should be ok. The manual has a reasonably tight tolerance for the bellhousing but as reported above anything up to .020 appears to work. The flywheel and pressure plate are aligned with the crankshaft and the disk is aligned with the transmission. Excess misalignment could cause a chattering clutch and/or a shortened life of the front bearing and difficult sticky shifting.

[altair]

I do agree that the mains must/should remain in their respective factory locations, I was left with no option and had to make good with what I had. If the foreign mains were shimmed to align with the crankshaft why would that cause them to wear oval? I used plastigage on each main and assured they were perfect in all directions. The engine in question runs quiet and perfect with good oil pressure, I fail to understand what is so evil about the process. ie "a thousand times no".

[PackardV8]

It is conceivable that the blocks and bellhousings are for the most part square it is the crankshaft that is off, left or right or up or down. . . . Today I think we try too hard to make it perfect when a reasonable tolerance is acceptable.

Agree, good enough to get by is good enough at times. Some cars/trucks and some situations don't need or cost-justify a precision rebuild.

There is also those who want and can afford as good as an engine can be built. Since the labor to R&R, disassemble, clean and reassemble is a given, some just won't do anything less than as good as it can be done.

We begin by line honing the mains to within .0005". A crankshaft laid into a line honed block spins noticeably smoother than the factory alignment after 100,000 miles of use.

A bar is inserted into the main bores and another into the camshaft bores and fitted into a precision jig to insure the decks of the block are resurfaced perpendicular to each other and parallel to the crankshaft and camshaft bores. There are varying degrees of factory tolerances to be corrected. When done, the block is closer to the blueprint. We just completed these operations on a block, so next week, I'll dial indicate the rear block surface to find if it confirms to those Jim Pepper has seen.

Your opinions and results may vary.

jack vines

[jpepper]

Yes you can introduce error in the bell housing face but doing that will cost as much as correcting the block.

[bezhawk]

I don't know how common the out of square was, but I do have a Studebaker marked tool that checks runout, as well as parallelism. It has three dial indicators on one common shaft, that tightly registers into the pilot bearing.
It is a production line tool and not a service tool so there is no Kent-Moore markings anywhere on it.

[jpepper]

Hi, Jim,

After all these years of working radially, learning the block may be out of square is very bad news. So you've found it's the rear face of the block which is out of square to the crankshaft centerline? Usually tilted in at the top?

But thinking about it, not surprising, as most blocks benefit from being square decked. When we make the block decks square to each other and to the crank/cam centerline, we locate off a 2" round bar and locating rings which ride in the main bore. We'll have to make an adapter and an arm to hold a dial indicator to sweep the front and back faces of the block. That would eliminate crankshaft end play as a variable. Have you ever checked the front face of the block? If the front face is found to be square, it could be clamped on the bed of the surfacer and the back milled square to that. There goes another $150 in machine time, but it would be subtracted from the shop time of dial indicating the bell housing.

That would explain why the Studebaker Shop Manual is the only one I've ever read which contains the section on dial indicating a changed bell housing. I always said it should be done on any brand, but it's embarrassing to learn Studebaker probably knew they were putting out improperly machined blocks and never corrected it.

jack vines

I have never checked the front face so I do not know if it is perpendicular to the main bores or if it follows the rear flange. I do know that I have checked 2 60's, 3 62's and a 63 all with the same issue. A 63 R1 I had was within .002". Go figure. My guess is the operator who was mounting the block on the fixture left some dirt/chips in place and the main did not sit down tight all the time.

[PackardV8]

I don't know how common the out of square was, but I do have a Studebaker marked tool that checks runout, as well as parallelism. It has three dial indicators on one common shaft, that tightly registers into the pilot bearing.
It is a production line tool and not a service tool so there is no Kent-Moore markings anywhere on it.

For the good of the order, would you share a photo of this tool?

jack vines