Here is the new Atlas-4SP 32-spline input shaft.

Part number 340332.

Side view of the new 32-spline input shaft. Note the helical cut gears that mesh with the planetary gears in the 4-speed's planetary housing.

The internal spline length is 2-1/8", and from the inside bottom of the input shaft to outside end measures 2-5/16".

The new input shaft ships from Advance Adapters with the two required needle bearings already installed.

The bearings are part numbers 340361 (first one) and 340362 (second one).

You can clearly see them inside the shaft in this picture.

Before we explore the process and our options involved with mating the Atlas-4SP to the back of the TH400, the first job is to swap out the input shaft.

Before we tear into the Atlas, first drain the oil.

This is the planetary reduction assembly of the four-speed Atlas with my original 23-spline input shaft.

The first step is to remove the six socket-head cap screws securing the planetary housing to the transfer case.

"socket-head cap screw" is the technical term for an Allen-head bolt.

The screws require a 5/16" Allen-key or hex-drive socket.

With the six cap screws removed, strike the planetary housing with a plastic or rubber mallet or dead-blow hammer, to separate it from the transfer case.

Hold the input shaft as you do so so that the planetary housing doesn't loosen suddenly and fall away uncontrolled, smashing to the concrete floor with a bang (don't ask!)

Once you have a gap established between the two, you can use a plastic or wooden tool to gently ease the separation of the two components.

DO NOT use a metal tool such as a screwdriver as you will gouge the aluminum (don't ask!)

The next step is to remove the four socket head cap screws that secure the input seal retainer to the planetary housing.

Screws removed from the input seal retainer.

Note the small relief in the face of the planetary housing just at the edge of the seal retainer, at about the 7 'o' clock position in this picture.

Back side of the input seal retainer, showing the seal.

If the seal is in good condition, it need not be removed and can be reused.

If the seal is worn, now is a good time to replace it.

Remove the snap-ring from the input shaft and discard it.

This snap-ring secures the input shaft and entire 6-gear planetary gear assembly in the planetary housing.

Place the snap-ring to one side for reuse and pull the input shaft from the 6-gear planetary assembly.

Note the lock-ring on the input shaft, located just above the helical gear. It looks like a large washer with a tab on it.

Underneath this lock ring there is a composite fibre washer. They will both be transferred to the new input shaft prior to it being installed in the 6-gear planetary housing.

It is now time to reassemble the case.

For the sealing surfaces, make sure you use Locktite 518 Anaerobic Gasket Maker.

Do not use Silicone or other RTV sealants.

This is because standard RTV is difficult to get even, consistent, and sufficiently thin when mating precision components. RTV is unsuitable for use between two precision parts where the parallel between the two parts is critical. In addition, RTV begins to cure in air as soon as it is applied, making it time sensitive. In contrast, an anaerobic sealant like Loctite 518 only cures when the two parts are mated and oxygen is excluded from the joint (the word anaerobic means "in the absence of oxygen"). This makes it non-time sensitive, which allows for easy precision alignment without the worry of the sealant beginning to cure prematurely, as can be the case with RTV. 

Lay a bead of 518 around the input shaft bearing, inside the input seal retainer bolt holes.

The other reason for the Loctite 518 is that its viscosity is classified as "thixotropic". This allows it stay put during application but also to flow well when being compressed between the mated components, resulting in a very uniform film thickness. It isn't until this happens that the product begins to cure. Regular RTV sealants simply do not exhibit these properties and can cause problems if used where an anaerobic sealant such as Loctite 518 is called for.

Carefully install the planetary housing onto the main case, making sure the dowel pin and all the bolt holes line up.

Prior to reinstalling them, each of the long socket-head cap screws should have a bead of RTV run around their circumference at a point about half the way up the non-threaded area of the shank. This is because these 6 screws thread into holes that go through the front wall of the case and into the oil cavity. Also, each one should have "blue" thread locking compound added to the threads prior to installation. (Loctite 243 is better than 242 for this application, as it has improved oil tolerance).

Reinstall the 6 long socket-head cap screws...

Congratulations!

You have successfully swapped the Atlas-4SP input shaft.

Before installing the Atlas in the vehicle, you must refill it with the proper lubricant and ensure oil is provided to the planetary housing.

First add 2.5 quarts of the required oil.

Stand the case on its front so the rear output shaft points straight up as shown in the pic, and block it in place.

Disconnect the upper end of the oil sight tube, and add the oil through the fitting in the case by using a funnel and a short length of rubber hose.

Once the case has been filled, re-attach and tighten the sight-tube hose-end. Be sure to hold the adapter fitting with a wrench while tightening the hose end, as shown.

Filling the case while tipped up on its input shaft provides oil to the planetary housing, which is critical before the unit is operated.

Verify the correct oil level in the sight tube.

If you haven't done so already, when the case is installed in the vehicle at the desired rotation (clocking angle) a permanent marker or small zip-tie can be used to mark the "full" level on the sight tube.

The TH400 adapter is part number AS-6401.

This is the side that mates to the back of the GM TH400 4-speed automatic transmission.

And this is the side to which the Atlas-4SP will bolt.

The countersunk holes are the ones used to bolt the adapter to the transmission, using the supplied socket head cap screws.

The other bolt pattern is the one to which the Atlas-4SP bolts using the supplied studs and nuts.

Part number AS-0404 is a one-inch thick spacer. You may or may not need this part when mating the Atlas-4SP to a TH400, as I'll explain momentarily.

This is the side that mates to the back of the adapter (AS-6401).

For a combined thickness of 2.25".

Here's the output of the 4WD TH400, this one out of a one-ton Chevy 4x4 truck.

The stock transfer case adapter has been removed.

The output shaft has 32 splines, and the spline length is 2-1/4".

Inside, this is how the Atlas input mates to the TH400 output shaft.

Depending on the exact model of 4WD TH400 in question, there are two 32-spline versions suitable for mating an Atlas to, distinguished by the different "stickout lengths" of the output shaft.

The shorter, found in TH400s that came from the factory with an NP203 behind them, sticks out about 1.25 inches.

The longer one, shown here, sticks out about 2.5 inches. It is found in:

• 1981+ TH400s that came from the factory with NP208 or NP241 transfer cases; or
• 1985+ TH400s that came from the factory with an NP205 transfer case.

There are two other 4WD TH400's, but niether is ideal for use in front of an Atlas. One is found in 1980 - 1984 TH400s that came from the factory with an NP205 transfer case. It has a stickout length of about 4-1/2". The other is the 10-spline "Jeep" TH400 that has a very long stickout somewhere in the region of 8+ inches.

The stickout of the transmission output shaft is very important when mounting an Atlas transfer case.

As shown in the diagram at left, there must not be ANY stickout of the shaft beyond the face of the adapter.

If there is, even a sixteenth of an inch, and you force the Atlas onto the adapter, you will preload the bearings in the case. If you run the case in this condition, the bearings will be destroyed in very short order - followed by the rest of the internals. A common mistake is to use the mounting nuts to pull the transfer case and adapter together - this is a serious no-no!

To summarize - when mounting an Atlas, ANY transmission shaft stickout = very, very bad!

This is where the 1" thick spacer (AS-0404) comes into play. If you have the longer stickout (~ 2.5"), to mount the Atlas you will need to use the spacer as well as the TH400 adapter.

However, even with the adapter and the spacer combined, you will likely still have to trim a little off the output shaft. Let's do the math:

My TH400 output shaft has a stickout of 2.33".

The adapter (AS-6401) and spacer (AS-0404) combined have a thickness of 2.25", leaving a stickout of 0.08" or about 5/64" (between 1/16" and 1/8").

Granted, with adapter and spacer, there would be the thickness of three gaskets to consider as well, but I wouldn't trust that alone, as the consequences are too severe.

Instead, no matter whether you have a TH400 with ~ 1.5" stickout and are using the 1.25" adapter alone; or you have a TH400 with ~ 2.5" stickout and are using the 2.25" adapter/spacer combination, you will have to trim a little off the output shaft.

We'll cover how to calculate exactly how much and one other possible solution a little later.

This is the hardware kit for the input shaft swap and TH400 mounting.

Clockwise from top left we have:

• Six socket-head cap screws for attaching the adapter to the TH400.
• Six studs, lockwashers, and nuts for attaching the Atlas to the adapter.
• The new snap-ring we used when installing the new input shaft into the planetary housing.

Obviously, with the ~2.5" output shaft and only the 1.25" adapter installed there is significant stickout to deal with.

You can't just cut the transmission output shaft off flush with the adapter at this point, as that would leave you with barely any spline engagement.

Click on this pic to enlarge it, and you can clearly see the little bit of the shaft that needs to be trimmed off, even with the adapter and spacer used in combination - unless you were willing to gamble on the gasket thickness taking up this extra, which I wouldn't recommend.

There are no gaskets in place in this picture, but this is an area where it's definitely better to be safe than sorry.

A rotary cutoff wheel can be used to carefully trim the extra length.

BUT - there's a catch!

Because of the nonsymmetrical 6-bolt round bolt pattern of the Atlas, when bolting to a TH400 (with or without the spacer), the "flattest" you can clock the case is about 15° (for passenger drop case).

That's not too bad, but it does translate to a loss of about 1-1.5" of clearance under the transfer case.

In applications where every inch counts, as in a rock buggy, this can be significant.

There is a solution, and it even alleviates the requirement to trim the output shaft - but it comes at a cost!

Part no. 51-8603, pictured at left, is a clocking ring that is 3/4" thick and has five (5) different bolt patterns drilled in it.

This part can be bolted to the front face of the Atlas before the Atlas is installed and because of the multiple bolt patterns provides a number of different clocking options that wouldn't otherwise be available.

Clearly, if we use 51-8603 we will no longer have any issues with output shaft stickout.

We will also have the ability to clock the case at 4°, saving an inch and a half of of ground clearance.

But the cost for this is a slight loss of spline engagement (and a 3/4" longer drivetrain!)

With all 3 pieces on the back of the TH400, total spline engagement between the TH400 output shaft and Atlas-4SP input shaft will be 1-5/8”, limited by the depth the output shaft can reach inside the input shaft splines (because we have spaced the input shaft away from the output shaft), as shown in the pic at left.

If we delete part no. 51-8603, we have to trim 1/8" off the shaft to prevent it bottoming in the input shaft. This would leave us with total spline engagement length of 2-1/8".

So we have to choose between:

15° clocking, trimming the shaft, 2-1/8" spline engagement
or
4° clocking, no trimming, 1-5/8" spline engagement, 3/4" longer drivetrain.

Install the six mounting studs into clocking ring 51-8603 using thread locking compound; in the correct pattern to achieve the clocking angle you desire.

Make sure you get each one in the correct position relative to its neighbours (i.e. that there are the same number of holes between each pair).

The studs must be installed carefully and tightened securely without over tightening.

Getting them properly seated and tightened is a bit of a trick. The procedure I used is as follows:

First I threaded on two nuts back-to-back...

Then with a wrench on the inner nut to keep the outer nut from turning I used a socket on the outer nut to turn the stud into its hole.

It's a bit tricky as you have to hold the inner nut fast against the outer nut while rotating the whole assembly together, but you will soon get the hang of it.

Atlas-4SP with clocking ring 51-8603 installed, ready to be installed on TH400.

The first thing I did was remove clocking ring 51-8603 from the Atlas' planetary housing.

I then test-fit clocking ring 51-8603 mounted directly to adapter (AS-6401),and found that there were a couple of spots on the transmission casting that would need clearancing to accommodate the studs in clocking ring 51-8603.

One was on the lower right...

... which a little grinding fixed in no time.

And the other located underneath the transmission.

The spline length of the original, unmolested shaft is 2-1/4".

I marked the trim line with a green sharpie, and found that, using clocking ring 51-8603 mounted directly to adapter (AS-6401) I would have to trim 7/16" off the shaft, leaving me with 1-13/16" of splines to engage the Atlas-4SP input.

(the green line is actually a little deeper on the shaft than necessary here in this pic due to the difficulties in getting the marker inside the transmission to make the mark.)

Good enough, I though, so out came the cut-off wheel!

Okay, so nobody would accuse me of being the human lathe, and I trimmed perhaps a sixteenth of an inch more than I had to, but as I said earlier, better safe than sorry; and it's actually not critical to have a perfect cut.

That said, it would help you enormously in making a clean cut if you were able to turn the shaft while cutting by driving some sort of belt on the input end of the transmission.

After making the cut, chamfer the end a bit and then dress the splines with a small triangular hand file to ensure the transfer case input will engage the splines smoothly and easily (especially important when you are lying under the rig by yourself bench-pressing the transfer case into place - ask me how I know this!)

With the shaft cut and no turning back, it's time to explore the clocking options that come with using clocking ring 51-8603.

I started by making an indexing mark on the adapter and the clocking ring, inline with the centreline of the transmission. These marks therefore indicate the top centre of the components.

I then marked a similar line at the top centre position on the Atlas' planetary housing for reference.

Now, when test-fitting the clocking ring 51-8603 in different position on the Atlas, I know that if all the red lines line up, the case would be perfectly flat (clocked at 0°). And from there, I can quickly get an idea of the other positions (clocking angles) possible.

These indexing marks can now be used to give an indication of the different clocking angles possible with the different positions in which clocking ring 51-8603 can be bolted to the Atlas.

This is the "flattest" position possible. Since the line on the clocking ring is actually slightly to the left of the line on the case you can tell that this clocking ring position would result in the transfer case actually being clocked "up" slightly (about 1° as it turns out).

The next flattest position is shown here. The clocking ring line being to the right of the line on the case indicates that this is the more tradition arrangement where the transfer case would be clocked down (the front output lower than the input) - in this instance about 4°.

Admittedly, my old weld-spattered angle finder has teeny-tiny markings and I have tired old eyes, so you should probably read all of my angle measurements to plus or minus 1°.

In essence, this is the transfer case clocked flat.

This is the Atlas installed in the 4° down clocking position.

This is the Atlas installed in the 4° down clocking position.

The TH400, Adapter (AS-6401), clocking ring 51-8603, and Atlas-4SP combo is 40-1/2" long from bellhousing to rear output flange.

• TH400:                              24-3/4"
• Adapter (AS-6401):         1-1/4"
• Clocking ring 51-8603:    3/4"
• Atlas-4SP:                         13-3/4"
• = TOTAL:                           40-1/2"