Ask our Experts, we're here to help!
On any quiet evening you might hear it. That distinctive pull in Third gear as the engine accelerates and the listener waits for that lift of the throttle and a quick, clean manual trans gear change into Fourth. But instead, you hear that instant rpm gain that telegraphs to all within earshot that the driver missed the shift and the engine revs to the moon or bangs up against the rev limiter.
Universal blame is usually reserved for the driver’s lack of hand-foot coordination. But often we can trace these missed shifts to a more mechanical source. Blame then usually shifts to the bellhousing but good blueprinting technique will point out that it’s ultimately more an issue of misalignment that can come from a combination of problems, something the engineers call tolerance stack-up.
Muncie four-speeds (foreground) use a ball bearing input shaft that can withstand a large amount of misalignment. Later Tremec five-speeds like this 3550 (background) use a tapered roller bearing for the input that do not function properly when faced with misalignment well. This is why late model transmissions demand that the bellhousing be at or within the 0.005-inch spec
To perform this alignment test correctly requires a magnetic base and a dial indicator. We prefer to remove the spark plugs from the engine and turn it slowly by hand using a ratchet on the crankshaft bolt. Never attempt to spin the engine with the starter motor.
It’s this stack-up that can cause the misalignment between the crank centerline and the center of the transmission’s input shaft. When this occurs, the input shaft is forced into a diverging line compared to the crankshaft centerline. Because the input shaft is supported by the pilot bushing or bearing, this misalignment forces the input shaft into a bind. Think of engine rpm as a force multiplier that aggravates an already bad situation.
In the past, this wasn’t a huge concern. Early manual transmissions like the GM Muncie four-speed, the Super T-10, Ford Top Loader, and Chrysler A-833 all employed a ball bearing to support the input shaft. These ball bearings will accept a significant amount of misalignment. The overall spec for manual transmission alignment is 0.005-inch. So if the input shaft was off by perhaps as much as 0.010 to 0.015-inch, these older transmissions could accommodate that mismatch and allowed the driver to complete the shift.
Paint, dirt, debris, and high spots can all contribute to a parallelism problem so it’s best to completely clean the bellhousing. This one had thick paint on the transmission mounting surface that had to be removed. Take either a sanding block with 150 grit sandpaper or a whetstone to clean both mounting surfaces on your bellhousing to ensure there are no high spots that may cause a misalignment. This is also a good idea even for brand new bellhousings.
The new modern manual transmissions like the Tremec TKO / TKX five-speeds and the T-56 and Magnum version six-speeds employ a tapered roller bearing to support the input shaft. These bearings offer superior load capacity but are less tolerant of input shaft misalignment. That 0.005-inch spec is pretty much the limit especially for high rpm applications. That’s why it’s essential for late model transmissions that the bellhousing be checked to ensure that the input shaft and the crankshaft centerlines are within spec.
While previous stories merely focus on concentricity, QuickTime bellhousing’s Ross McCombs points out that the first dimension to check is the parallelism between the bellhousing flange on the engine and the bellhousing’s transmission mounting flange. McCombs emphasized that if those two surfaces are not within the parallel spec of 0.005-inch, this will also affect the concentricity test. In other words you will not be able to dial in the concentricity if the parallelism is out of spec.
The easiest way to check bellhousing parallelism is with a dial indicator placed 90-degrees to the trans mounting face. Then rotate the crank slowly and take readings every 90 degrees. We write the specs directly on the face with a Sharpie. A typical dial indicator will measure one full inch of travel with each hash mark representing 0.001-inch. Preload the dial indicator so it can travel in either direction. Watch the indicator needle move as you turn the engine slowly. If the needle indicates 0.090-inch, that means the bellhousing opening is moving inward 0.010-inch. If the indicator reads 0.010-inch, this is outward movement of the same distance. We prefer to zero the indicator at the 12 o’clock position and then check concentricity every 90 degrees. We also use an arrow to indicate the direction the bellhousing opening is moving either side to side or up and down. This new Lakewood bellhousing (left) checked within 0.002-inch using the stock factory dowel pins so it did not need to be relocated. This particular aftermarket reproduction bellhousing on the same big-block (right) tested extremely poorly with 0.055-inch of run-out. This may be brought into spec with 0.021-inch offset dowels placed pointed toward the 5 o’clock position. But if not, the bellhousing would have to be modified by a competent machinist. It might be easier and less expensive just to buy a new Lakewood bellhousing.
Before testing for parallelism, McCombs says its best to use a sanding block or whetstone to run over the engine block and both ends of the bellhousing to remove any dirt, debris, or burrs that may cause a misalignment. Even a small burr can easily multiply over the distance to the trans mounting face. It’s also a good idea to run a large straight edge over the block mounting surface to make sure nothing protrudes that could affect bellhousing alignment. We noticed on LS engines that the top portion of the rear cover plate will hit the inside radius of some steel small-block scattershields that will prevent the bell from sitting flush with the block.
In rare cases, the block mounting flange may not be square to the crankshaft. We set up our dial indicator on this old 327 block and found the block was high on one side by 0.011-inch.
The best way to measure parallelism is with a magnetic base placed on the crankshaft with a dial indicator positioned so that the unit is placed 90 degrees to the bellhousing transmission mounting face. We prefer to zero the dial indicator at the 12 o’clock position and take readings at the 3, 6, and 9 o’clock positions as viewed from the rear of the engine. If the reading is more than 0.005-inch, this will require modifications to the bellhousing in order to bring the bell into spec.
After measuring dozens of bellhousings, one situation caught our attention. The application was an iron-block 6.0L LS engine in which we had installed a cast aluminum engine swap oil pan. When we installed the pan, it ended up slightly rearward on the engine block. In a small- or big-block Chevy for example, pan placement would not be a problem. But LS engines use the oil pan’s rear face as a structural member that is bolted to an automatic transmission bellhousing.
On LS engines, make sure the oil pan does not contact the bottom portion of the bellhousing and create a problem in parallel. We filed our oil pan until we could measure a minimum of 0.005-inch clearance between the bottom of the bellhousing and the oil pan. That solved our parallel problem.
In our case, the oil pan moved the manual transmission bellhousing rearward over 0.020-inch, which is four times the allowable spec. To rectify this situation we had a choice to either relocate the oil pan or file the thick oil pan flange. We decided to file the oil pan because we didn’t want to risk an oil leak by moving it. We filed the pan until we created a 0.008-inch clearance between the lower portion of the bellhousing and the LS oil pan. Once this was corrected, the bellhousing was well within spec with only 0.002-inch of vertical top to bottom.
If there’s still an issue, the next procedure is to check to see if the block or the bellhousing is at fault. Borrowing another bellhousing to check is one way to determine where the issue lies. If a second bellhousing creates a similar error, then you can assume that the block may be the source of the problem.
Another option is to check the original bellhousing on a different engine to see how it compares. This may be more difficult based on access to another engine of the same family. If the bellhousing also measures out of spec on a different engine, then you can assume (yes, we’re aware this is not an ideal check) that the bellhousing is at fault. A final check would be to use the dial indicator to check the block bellhousing flange at multiple positions to measure its relative location. We checked an old 327 Chevy block and discovered it was angled relative to the crank by as much as 0.012-inch.
If the bellhousing is at fault, it’s possible to take it to a machine shop to face the transmission side to square it up relative to the engine. This also means that once the bellhousing is machined, it can only be used on that specific engine as it will likely be way out of spec on a different engine.
Assuming the vertical check is within spec, we can now move on to measuring the concentricity for the input shaft. Again, we’ll use the magnetic base and dial indicator but now the indicator will measure off the inside diameter of the input shaft opening in the bellhousing. Again, we prefer to start at the 12 o’clock position and zero the gauge and then take measurements at the 3, 6, and 9 o’clock positions.
Let’s assume for this discussion that the bellhousing is good in parallel but is off center in concentricity with 0.015-inch down at the 6 o’clock position. This will often also reveal that the 3 and 9 o’clock positions will show inward movement. What we’re looking for is the largest amount of movement on the dial indicator and then correct for that maximum offset. In this particular case, it is at the 6 o’clock position with 0.015-inch of run-out. Since the maximum allowed is 0.005-inch, an offset dowel pin that can relocate the bellhousing upward around 0.015-inch is needed. Keep in mind that when measuring concentricity, the dial indicator is displaying total indicator movement – what the engineers call total indicator run-out (TIR).
Lakewood makes offset dowel pins in 0.007-, 0.014-, and 0.21-inch offsets. The largest offset dowel will compensate for as much as 0.042-inch of offset in the bellhousing. With offset dowel pins, we like to mark the end to indicate the direction of the offset to ensure both dowels are installed consistently in the desired position. If they are not pointed in the same direction, the bellhousing won’t fit and you will have to start over.
But when determining the amount to correct the error, we will use an offset dowel pin rated at half that dimension. Offset dowel pins come in 0.007-, 0.014-, and 0.021-inch offsets. Each offset dowel will compensate for twice its offset. So a 0.014-inch dowel will relocate the bellhousing 0.028-inch. The same doubling effect is true for the other two dowels. This means total bellhousing relocation can be as much as 0.042-inch.
In our situation where the bellhousing checks 0.015-inch down relative to crankshaft centerline, we can use a 0.007-inch offset dowel to move the bellhousing upward by double its value – 0.014-inch. Theoretically, this should place the bellhousing with 0.001-inch of offset which is near perfect.
Lakewood, Moroso, and others offer offset dowel pins. Before installing these offset dowel pins, we like to mark the face of the dowel with a Sharpie to indicate the position of the maximum amount of offset. Then, the non-offset portion can be driven into the block while positioning the offset in the desired direction. In this particular case, we placed the bushings with the maximum offset at the 12 o’clock position. It’s important to make sure that both dowels are both positioned exactly the same or the bellhousing will not fit over the dowels.
It’s always best to recheck your work to ensure that the dowels were placed correctly as it’s easy to get them backwards. In our case, the re-test revealed our bellhousing now within 0.002-inch of perfect alignment.
With most engines you have a choice of either a pilot bushing or bearing. If you choose a bearing, this one uses a seal located just inside the tapered entry. Always place the seal pointed toward the transmission. This retains the lube in the bearing and prevents damage to the bearings from outside dirt and debris.
As we mentioned earlier, the key to checking alignment is to start by ensuring the bellhousing is parallel. If it is not, this creates a hole that essentially becomes an oval. No amount of changes with offset bushings will correct this. The bellhousing must first be as parallel to the block and perpendicular to the crank within the 0.005-inch spec.
It’s also worth mentioning that creating this proper alignment does wonders for creating much smoother clutch operation because now the input shaft is not placed in a bind relative to the crankshaft. Proper bellhousing alignment will also extend the life of a pilot bushing or bearing. We know mechanics that do not like pilot bearings because they sometimes seize on the input shaft. This is most often caused by a bellhousing misalignment. When all the components are in near perfect alignment, the clutch and transmission will last much longer and perform as intended.
Armed with this updated information, it should be relatively easy to accurately determine if your bellhousing is within spec. Once that’s achieved, that will make your 21st Century manual trans shift as smooth as silk. And then you can be that driver who executes that perfectly-timed high rpm fourth gear shift to the acknowledgement of all your friends.