Thanks again. More pieces falling into place.

To clarify something, once the pinion is engaged and latched, does working the pony clutch lever in either direction allow the drive to be released at all between pony and main? Or is it not a clutch in that sense? (I guess I wondering why Cat engineered it this way...why could they not have just had a simple clutch between the pony and main without the need for a pinion to be moved in and out of position. Of course it would still have needed some sort of locking feature so that you didn't have to manually hold the clutch engaged while starting. Maybe such a design would have needed a much bigger clutch....dunno).

Another question: if we manually release the latches, asssuming that the spring does in fact pop the pinion out, will that allow us to remove the whole pinion assembly, including the bits we can't currently see except through that little access hole? Or is the reason that it seems stuck more to do with the "O" ring that you mentioned? If in fact this is all rusted/swollen up, how do we get it out, and is the O ring replaceable? (What's it made of?)

Thirdly, how is this concealed part of the pinion lubricated? Is there a flinger that sloshes transmission oil on it from down below? It seems to be too "high" to be bathed in oil, and unlike the front part of the clutch which has it own oil reservoir, this part of it is on the other (in)side of the wall.

Pull starting the tractor is unfortunately not an option. It isn't moving freely enough on the tracks yet to be towed...we need the power of the main engine to overcome the frictions of the remaining dried up mud etc. So it looks like we just have to get the latches to let go.

Ah..I posted my last message before I saw your followup with the additional diagram. Yes that one is most helpful...now I can see the whole thing it makes a lot more sense. I had sort of guessed how the earlier diagram was showing the other end of the bits that we could see underneath the clutch inspection cover, but now that just confirms it concretely. Thanks for going to the trouble of posting that for me.

If you are still able to answer the remaining questions from my previous message I will be in your debt.

The clutch functions to engage and disengage the pony engine from the main. The pinion remains engaged (if set) and the clutch disconnect is at the pony drive gear.
The pinion is required to get the gear ratio between pony and main engine flywheel and must be disconnected when the flywheel starts driving the pinion with gear ratio reversed.
The assembly doesn't care if the pinion is engaged or not as far as removal but there is a clearance advantage to removing the assembly on later units. If the housing does not budge it is most likely due to the "O" ring seal (and rust)
Tapping on the housing, prying on the housing (but go easy on the bolt flange) rotating the housing.......and a lot of cussing....as well as penetrating oil are normally what is required. I've seen some elaborate pullers used but normally shouldn't have to go there. "O" ring seal is available from Cat.

The pinion and latch mechanism do not receive special lubrication. They survive on lubrication at assembly and what ever oil (and dirt) that is flung around in the main clutch housing.

The outer flange/o-ring is the troublesome bit. There shouldn't be anything serious inside of that. Lubrication and judicious prying on it should work it out. The latching of the pinion should only make the whole assembly shorter and is required on the later tractors to allow the assembly to clear the governor once the housing is removed

the second lever on all the older engines works a small clutch to first brake the windmilling pinion to allow it to be meshed with the flywheel and latched in with the rear lever, and then to engage the meshed and latched pinion to the starting motor gear train and make the main engine turn.

the latching and unlatching of the pinion gear is necessary to keep the pinion shaft/pony motor from being oversped by the main engine and turning all the time the main engine is running. There is a tall gear ratio (~10:1) there which allows the pony motor to have great advantage turning the main, but once the main cranks, it will quickly overspeed the pony motor and pinion shaft.

The pinion latching mechanism isn't lubricated after installation except by leaking engine oil slung from the rear main bearing as it leaks down the front face of the flywheel. the pinion clutch housing does have an oil bath and uses whatever you use in the diesel engine or ATF if you like it to be easier to brake it for meshing with the flywheel.

Before you turn the main engine over much, drain a small sample from the engine sump to check for water and also change/check the injection pump housing oil. It has it's own resevoir with a fill/check plug on the side and a drain on the bottom. It is also a good idea to bleed the air from the cooling system of the pony motor before you crank it. There should be a vent petcock on top. If you don't remove the trapped air, coolant doesn't circulate well or at all through the pony motor.

good luck

OldMagnet: One thing I can't quite see from these diagrams is how the pinion lever actually pushes the pinion into position...I can't quite visualise how this links up. To clarify, it seems that the pinion itself must move towards the rear of the tractor to engage with the flywheel. So does the pinion lever push it rearwards somehow, or does it pull it from behind?

ccjersey: thanks for the additional insights. When you say "latching of the pinion", do you mean latching into the engaged position? This intuitively seems to me to make the assembly longer. (What distance does the pinion travel to engage? Just a bit more than the width of the pinion, or a lot more? I must still not quite fully comprehend how this works.)

The latching is by moving the pinion an inch or so forward into engagement, not toward the rear of the tractor like a "normal" starter bendix.

The part that moves the pinion isn't shown on the diagrams, but is in the picture as I remember. It is a finger on a rod directly linked to the rear "pinion" lever and once the pinion is latched in, this lever is free to flop both forward and backward.

Yes the pinion actually travels to engage the flywheel. You can see the mechanism in the picture of the engaging latches. When you pull up on the lever the arm just beyond the latch mechanism moves forward setting the pinion (from behind the flywheel) and automatically engages the latches.

There is an echo on here...CCJersey...it's all yours

Ah you guys are great! Thanks for the photos and extra diagrams.

OK, now I have a much better idea of what's what, I think. Just one last lot of followups and I'll leave you alone for a moment.

In that last photo that you posted O.M., I can see (I think) the pinion engaged on the flywheel. Regardless, when it disengages, does the whole end of this .. including the latches .. move rearward by an inch? Or is there a slender rod with a stopper on the end that moves out to meet the pinion actuator lever? I can't quite determine this from the drawings already posted, but this may be important if I am to understand exactly where the latches are under the access hole.

Also, I presume that the latch spring can only be adjusted by pulling the whole assembly out, as we are attempting? (There being no way to remove the split pins etc through a half inch access hole!)

Is it possible that with the passage of time that some rust and gunge has glued up the latch hinges, in a way that regardless of the spring setting, they won't release by themselves when the main engine starts spinning? (And make it difficult to manually release)?

Lastly, the 7U diagram of the pinion control shows a return spring, but on the R4D I'm not sure there is such a thing? Unless the linkage between the lever and the actuator is not direct. The pinion lever is quite floppy when the pinion is engaged.