Quarter Scale Merlin V-12

A centrifugal supercharger takes in air along the central axis of its rotating impeller, and the fins use centrifugal force to rapidly accelerate this air radially outward to their outer tips. As the heavy air is slung away from the center of the impeller, a partial vacuum is left behind that draws even more air into the inlet. This high velocity air slams into the wall of the volute and piles up against the air that's already packed into the housing. This packing converts the kinetic energy of the high velocity air into static energy in the form of increased pressure, which is the goal for a supercharger. Compared with a Roots-type blower a centrifugal supercharger does a lot with its single moving part, and it doesn't heat the air nearly as much in the process. The downside, though, is that the kinetic energy imparted to the air varies with the square of the tip velocity. This means that boost falls off very rapidly with decreasing engine speed. (Glass-half-full people might say it rises very quickly with increasing speed.) The full-size Merlin used a variable pitch prop and operated over a modest rpm range which helped mitigate this shortcoming.
With the air-packed volute under pressure, a potential leak is present in the space underneath the impeller. This leak is created by the gap between the bottom edge of the impeller and the floor of the housing which is adjacent to the volute. The solution is a thin stationary ring bolted down to the floor of the housing that closely surrounds the impeller. The air flung off the impeller will then continue across the top surface of this ring before hitting the wall of the volute. This ring is referred to as a 'diffuser section ring' on the supercharger drawing. There was no detail provided on its design other than the assembly drawing showing its cross section. I had previously drilled clearance holes in the front-half housing for its eight 1-72 mounting screws. These holes were located and drilled through the centers of thin cast spokes on the front of the housing, and so their exact locations were not on a perfect bolt circle but were affected by the casting geometry. The first thing I did was measure their actual locations and compare them with my original notes.
Since the measured TIR of the mounted impeller came out to under a thousandth, I planned the gap between the impeller and ring to be .005" in order to add plenty of margin for hole uncertainties. I chucked up a 4" diameter aluminum round in the lathe and turned a portion of its end for the o.d. and i.d. of the ring while machining the workpiece extra long for a couple rings - just in case. The chucked workpiece was moved to the mill so I could drill and tap the array of holes. A witness mark was included to keep track of the orientation of the imperfect bolt circle pattern. After returning the chucked workpiece back to the lathe I parted off a ring using a right-side ground parting tool to minimize the burr on its inside i.d.. The first ring fit as hoped and was installed on the front-half housing using low–strength thread locker on the mounting screws.
Before finally assembling the supercharger, a few miscellaneous items had to be completed including the shaft keys as well as a nice streamlined spacer and bolt for the top of the impeller. For the time being I decided to not use any sealer between the housing halves since my previous flashlight test before adding the impeller showed no light leakage.
I plan to eventually do some systematic bench testing on the supercharger including an attempt to measure any boost it might possibly make, but this will require some special hardware to properly drive it. Rather than create an elaborate dedicated drive for a one-off test, I'd like to re-purpose at least a portion of what I have to machine for use in the electric starter. I'll let my brain work on this in the background while I continue on with the engine.
I couldn't resist, though, spinning it up by blasting compressed air into the intake. I have no idea how fast I got the impeller spinning, but it screamed like a turbine winding up and took nearly thirty seconds to slow to a stop when the air was removed. While holding the assembly in my hand I could feel no resonances or significant vibration throughout the entire rpm range that I ran it, and so for now I'll assume current impeller balance is 'good enough'. I could feel a lot of air leaving the outlet just after removing the air source, and so I know I at least have a decent diffuser.
After gaining some confidence that it probably wasn't going to blow up in my hand, I made several sustained stress runs lasting up to a minute or so. Neither the sound nor its feel in my hand seemed to change after accumulating some ten minutes or so of run time.
I found some 0W-20 full synthetic oil in a local auto parts store and dropped a dozen drops down the oil intake. The sound of the bearings immediately changed - not necessarily better, just different - and there was an obvious increase in drag. With the same compressed air spin up, the impeller came to rest twice as fast as it did without the oil. I don't know if any of this means oil is good or oil is bad in the ceramic bearings, but the extra drag has to mean more friction and therefore more heat. I left the oil inside the bearing cartridge, but right now I don't have plans to add anymore.
So far, I'd have say this supercharger has probably been the most fun portion of this build. - Terry

Hello Terry,
You mentioned the long run down time of the turbo impellor. This relates to me of the history of the Merlin engine where trouble was encountered with the drive shaft of the turbo breaking when the revs were reduced quickly. Have you considered this problem.
Grapegro

Terry, your work continues to be awesome!! The fits and finishes you are achieving
are amazing. I think it's safe to say than I'm not the only one anticipating the first
running of this fine machine.

Pete

grapegro said:

Hello Terry,
You mentioned the long run down time of the turbo impellor. This relates to me of the history of the Merlin engine where trouble was encountered with the drive shaft of the turbo breaking when the revs were reduced quickly. Have you considered this problem.
Grapegro

Click to expand...

I hope so ... see post 332.

Hello Terry!

I don't post much because there is only so many times you can say "holy crap!! but I want you to know that i'm still watching and enjoying the build very much. Keep up the the awesome work and Holy crap!!

I found some 0W-20 full synthetic oil in a local auto parts store and dropped a dozen drops down the oil intake. The sound of the bearings immediately changed - not necessarily better, just different - and there was an obvious increase in drag.

Click to expand...

When a perfectly spherical steel ball rests on a perfectly flat steel surface - its mass is supported by the "dent" it makes in the steel = area x compressive stress. For applied loads the dent gets deeper.

The stress at the middle is basically the compressive stress that the material can stand and is therefore always damaging.

Add into that the surface imperfections and geometrical errors and it gets worse.

For a ball in a groove or a roller in a ring the contact is a line rather than a point but the same rules apply. (the stress diagrams look like squashed jelly beans with the highest stress at the centre diminishing to zero at the edges)

From this it follows that there is no such thing a a ball bearing that can run dry - it is always at the expense of diminished ratings for speed, load & life expectancy.

The relatively high shear and dynamic compressive loads that can be borne by the lubricant provide a much larger support area at lower pressure - this is the principal role of the lubricant.

The role of lubrication in rolling element bearings is extremely complex but if you are interested you should find plenty of information on line.

While ceramic bearings are more tollerant of high heat, the rules of dry compressive stress support aren't going to change.

Just something to bear in mind.

Regards,

I agree the bearings shouldn't be run dry. I was working for SKF when they first started making them available for normal distribution and our training had us targeting food processing industries due to food safe lubrication issues. Basically we were told full ceramics could be successfully lubricated by things like orange juice, milk and even straight water if necessary. We were told they needed some lubrication though.

Interestingly, ceramics also come as full or part hybrids as well. A full hybrid has normal steel raceways but all the balls are ceramic. A part hybrid is basically a normal steel bearing with one of the balls being ceramic. The full hybrid was primarily for insulation properties from memory, but the part hybrid used the hardness of the one ceramic ball to 'repair' micro damage to the raceway surface as it occurred from particle ingress. It just crushed the dirt particles then flattened out the damage caused to the raceway. Greatly extended bearing life in demanding environments.

Lastly, I once met a guy who built his own gas turbines and used modified deep groove skateboard bearings (608's) for his turbines. He replaced all the balls with ceramics, with no cage, and ran them at ridiculous RPM (hundreds of thousands I think). He replaced raceways often but reused his ceramic balls many times with no apparent wear.

I just had an epiphany ..thank you very much!

The Quarter Scale's coolant pump is a dual output centrifugal unit with an impeller and volute that are very similar to those in the supercharger. Three castings were supplied for the pump including the main body, an inlet cover, and an impeller. All three took much more time to machine than expected due to their difficult fixturing and precision requirements. I was pretty happy when the time came to finally test the pump.
The pump body has a number of features that must be concentric with the impeller cavity, and a long spigot was cast into its top end to help with the machining. Similar to the supercharger's impeller, though, the body wobbled badly with the spigot chucked in the lathe's headstock. The inside jaws of my lathe chucks were too large to grip its interior so the spigot could be cleaned up. Instead, I inserted a tight-fitting Delrin plug into the cavity and supported it in a live tailstock chuck while the headstock chuck gripped the very end of the spigot. After truing up the spigot, I turned the neck of the pump body to its final diameter and added a groove for an o-ring. The pump body was supported by its neck for the remaining lathe operations, and the spigot ended up being used only to bring the neck to its final length when I parted off the body.
After the impeller cavity was bored to its final dimensions, a long narrow internal grooving tool had to be ground to open up the entry to the volute. The neck of the pump body was then drilled and reamed for a bronze bearing for the pump shaft. During these two operations I ran into the same soft metal issue encountered with the supercharger's impeller. The drilled hole had a bad finish, and even with a flood of coolant, the gummy metal fused to the flutes of the reamer. The final hole came out poorly finished, nearly ten thousandths oversize, and skewed to the axis of the body. Apparently, the metal in the thick spigot-like features on some of these castings didn't harden as well as the metal in the thin wall structures. Fortunately, this hasn't been an issue where it really matters because it could have been a real show stopper. I'm really glad now that I took the trouble to minimally anneal along only narrow bend lines when I earlier straightened the castings that were warped. Drilling and tapping all those mounting holes in gummy metal like I've seen in these spigots would have been a nightmare. In this case, the bronze bearing nicely buried the trashy metal out of sight and out of mind. Since the bearing was off axis from the pump body, I waited to finally bore it for the shaft until after the bearing was pressed and Loctited in place. The shaft bore, the pocket for a shaft seal, and the impeller cavity were all finish machined in the same lathe set-up.
The pump shaft was turned from 303 stainless. Its bottom end was tapped for a left-hand 10-32 thread for the impeller, and the unthreaded body of the shaft was polished for a nice bearing surface as well as the sealing surface for a standard rubber shaft seal. Since the shaft was also used as a mandrel for later machining the impeller, I waited to machine its top for the Oldham coupler until after the impeller was completed.
The coolant pump drawing called for a polished sleeve to be pressed onto the bottom of the shaft after inserting it through the bearing. This sleeve was to provide the sealing surface for a TC8x14x4 shaft seal that was part of a 2014 redesign. This sleeve, though, would have captured the shaft in the bearing between it and the Oldham tennon. I didn't like the fact that the shaft would no longer have been removable once the sleeve was installed. So, I modified the design for a TC6x15x4 seal that enabled me to use the polished surface of the shaft for the sealing surface and allow the shaft to be removed. Since the stock sleeve doubled as a thrust spacer for the impeller, I machined an integral spacer onto the bottom surface of the impeller as a replacement. I also bored a recess in the bottom surface for a steel spacer to provide a shoulder stop for the threaded impeller. The spacer was ground for an impeller thrust clearance of .003".
The spigot on the impeller casting was also considerably off axis to the impeller and had to be turned true so it could support the part while its center was being drilled and tapped. I first sawed off a badly warped portion at its very end which helped a little. I was worried that I was going to run into the same gummy metal problem during threading, but the portion I cut off appeared to be reasonably hard. A lot of material had to be removed to true up the spigot, and there probably wouldn't have been enough left for a repair bushing, but fortunately the threads came out looking great. The spigot was parted off, and the pump shaft was used as a mandrel for the remaining impeller machining operations. Since the shaft/mandrel was reverse threaded, rear mounted inverted lathe cutting tools had to be used for some operations, and the lathe was run in reverse with inverted front mounted tools for others. During an 'ah-ha' moment I discovered a power feed reversing lever on the Enco lathe that I've owned for some twenty years. Inlet fins were added to the impeller design in 2014 to improve the engine's cooling system. The fins were tediously profiled for a close fit to the pump body and cover just as was done for the supercharger.
The inlet cover casting required a number of machining operations with difficult fixturing requirements. Its small diameter spigot cleaned up with no issues, and its interior was machined to closely follow the contour of the impeller fins with the same clearance used in the supercharger housing. I added an o-ring groove to the cover so I wouldn't have to deal with a gasket or sealer on its narrow body mounting flange. The interior of the large inlet port was also prepared for an o-ring sealing surface. A dial indicator showed its rough cast i.d. was nearly perfectly circular. Instead of risking a boring operation I lapped it in my hand using 230 grit valve grinding compound on a wood lap chucked up in the lathe. The lap was easily turned to successively larger diameters as the lapping proceeded.
The final inlet cover operations included match drilling the two mounting flange pairs for some fourteen 0-80 mounting screws as well as the machining of stainless hose barbs for the inlet and outlet ports. Although the outlet barbs were threaded into the castings, there were too few threads in the casting for adequate seals; and so the barbs were permanently installed with JB Weld. Since John Ramm recently discovered that the i.d. of the coolant hoses had a significant impact on his engine's cooling system, I made the i.d.'s of the hose barbs as large as possible. The threads of the stainless steel pump shaft were coated with a Cu-Ni anti-seize compound before it was threaded into the aluminum impeller in order to reduce corrosion between the two dissimilar metals in the coolant.
Testing the pump by spinning its shaft with a battery powered drill through an Oldham coupler showed it to be quite capable. With the outlet hoses supported about twelve inches above a temporary one quart reservoir tank feeding the pump, it took about 20 seconds to empty the tank with the drill spinning at 1300 rpm. This corresponded to a 45 gal/hr flow rate at an engine speed of only 650 rpm. The pump also performed in reverse but with significantly less flow rate. - Terry

Fantastic.

Do you have anything assembled? How about a picture of the hundreds of parts that are done?

I agree with Steve. We've seen all these nifty bits and pieces being machined. How about a family photo?

Don

Terry, your wood lap has me curious & intrigued. I had a similar thought about making a simple/quick/disposable MDF lap finishing tool to touch up cylinder liners (OD). On your ID lapping tool, do you somehow progressively expand the lap tool boss with a center screw or something, or do you mean you turned a new surface a couple thou larger each time?

Instead of risking a boring operation I lapped it in my hand using 230 grit valve grinding compound on a wood lap chucked up in the lathe. The lap was easily turned to successively larger diameters as the lapping proceeded.

Based on the pics, I bet he just cuts off the used end and advances a bit more wood in the chuck before turning to the next size.

Peter,
As kvom said, I just cut old lap off, pulled a bit more dowell through the chuck and turned the new diameter for the next lap. I left the cross slide at the previous diameter and just added a thou or two for each new lap. I used pine. I'm not sure about using MDF since the part gets pretty hot, and I'm not sure how the MDF will hold up. I couldn't measure any circularity imperfection with a dial indicator when I was finished. I've used similar conical laps to remove machining marks from valve seats. - Terry

I've been testing the fits of all the parts as I make them, but I've just never taken a photo of any major assemblies. I'll assemble what I have and take some photos this weekend.

Mayhugh1:

It's been over a week! I'm Jonesing for my Merlin fix. Central Texas still there or did it get blown away?

Don

Don
I'm waiting my self to see it all bolted together.
Had to get in the loft and look at my merlin castings just for a fix.

Robbie

ddmckee54 said:

Mayhugh1:

It's been over a week! I'm Jonesing for my Merlin fix. Central Texas still there or did it get blown away?

Don

Click to expand...

This old saying is apropos I think: "The difficult we do immediately; the impossible takes a little longer."

I got to see Terry's supercharger first hand this past week and, like all his work, 'tis a thing of beauty. Flawless fit and finish, and it spins like a dream!

Chuck

The full-size Merlin had a dry sump oiling system that was fairly typical of the aero engines of its day. There was a pair of scavenger pumps located inside the rear end of the oil pan, also known as the lower crankcase, and a pressure pump located directly below them outside the oil pan. The second scavenger pump, which was not included in the Quarter Scale's design, drew oil through a suction tube from the front of the lower crankcase during extended dives. The Quarter Scale's pressure and scavenger pumps operate on common shafts that run between them through the oil pan. The bores for the shafts don't extend outside the engine, and so seepage around the shafts doesn't create external oil leaks. The pump's driven gear meshes with an idler that's driven by the coolant driveshaft inside the wheel case and provides a bit of mesh adjustment. Pressure reliefs will later be mounted to the external starboard side of the engine's crankcase, and copper oil lines with custom scaled fittings will eventually run over the outside of the engine.
After several days into this portion of the build I thought I'd be writing that I spent much more time deciphering the drawings for the pump bodies than I spent actually machining them. The pumps have complex shapes with several sides that require careful machining and cross-drilled holes running in and between them. Unfortunately, they're described by multiply-sectioned flat drawings intended for someone with more print reading experience than I have. All the needed information is in the drawings, but I had a really difficult time wrapping my head around the way it was presented. One or two trimetric views would have saved me days of headscratching. In the end though, while machining the pump bodies, I managed to generate enough scrap so the machining time came out greater after all.
Construction began by facing a mounting surface for the scavenger pump inside the rear of the oil pan as well as a second surface immediately under it on the outside of the pan for the pressure pump. These two surfaces must be absolutely parallel to each other to prevent the close-fitting shafts in the pump bodies from binding when they are assembled as a stacked pair straddling the pan. Their finishes also require care since they'll become the sealing surfaces that isolate the pumps from each other. The shape of the oil pan doesn't make the inside surface machining an easy task. The pumps must have gone through a redesign since the mounting bosses on the oil pan casting were designed since they don't match the envelopes of the pumps. This added to my confusion while trying to decipher the drawings.
Not part of the stock design, two pairs of phosphor bronze inserts were pressed and Loctited into the aluminum workpiece from which the pump bodies were milled. These inserts became the bearings for the pump shafts as well as thrust surfaces for the pump gears, but they created problems for some of the cross-drilled holes in the pressure pump body. A couple of these holes crossed the boundaries between the aluminum and bronze, and the drill invariably wondered into the soft aluminum. I had to plunge an end mill into the affected holes in order to straighten and prepare them for oil fittings. Compression fittings machined from standard 3/16" unions were installed in the scavenger's output port as well as the pressure pump's input and output ports. A similarly machined 1/8" compression fitting was installed in the pressure pump's gauge output. Scaled custom flare fittings will likely be used for the distribution tubing later.
A short length of 12L14 was machined into pump gear stock. The individual pump gears were parted off from it and their widths surface ground for .001" depth clearances inside the pump body pockets. A .002" clearance was machined into the pocket around the o.d. of each gear.
For the six pumps I've previously machined for my last four engines I generally start out with these optimistically close fits but usually have to clear the interferences left by my Tormach's backlash while machining the pump gear pockets. It's been my experience that the 60-70 psi that these pumps typically generate is much more pressure than is needed or even wanted in a model engine. Even though the scavenger pump has twice the capacity of the pressure pump, John Ramm mentioned that he had to throttle his pressure pump down to some 15 psi in order to better control the oil in his Merlin's crankcase. In both my radials I had to severely restrict the oil flow into the pressure pumps to avoid similar issues. I plan to later implement a relief valve in the pressure pump's output line with a return back to the oil tank similar to what John eventually did.
Even though I could probably be much more aggressive when fitting the pump gears, I prefer bluing and tediously polishing away only the high spots that create the interference. The stacked pump pair in the Merlin design could have created a whole host of additional problems with its pair of facing gear cavities, and so I was relieved to discover that I only had to open up the drive shaft bore in the pressure pump body for a .002" clearance in order to obtain a freely turning dual pump assembly.
The drive gears in both pumps were each broached with a pair of slots spaced 180 degrees apart to accept slip fit drive pins in the driven shaft. This provided some precision broaching practice for the operations coming up later on the rear crankshaft, but I was left with some concern about their impact on the gears' strength. These constant displacement pumps are capable of generating terrifically high stresses on the shafts and pump gears. I couldn't come up with a better way, though, to secure them to the shaft so the pump pair could still be assembled.
The driven gear for the pump assembly is secured to the pump driveshaft with a setscrew tightened against a machined flat, and a hold-down plate fits into a groove around the gear's neck to keep both shafts in place. The hole pattern drilled into driven gear is required for access to the plate's fasteners.
I machined a temporary spacer plate to simulate the oil pan surfaces between the two pumps so I could test the completed assembly outside the engine before drilling the holes through the oil pan. With the assembly immersed in a container of oil, both pumps generated vigorous streams of oil when spun with an 800 rpm drill. For a quick stress test I installed a gage on the pressure pump and blocked off its output line. The gauge pegged at 200 psi before blowing the oil line out of the gauge and causing me to lose my grip on the pump. The pint container of oil ended up in my lap, but I considered the test a 'pass'.
I promised to post some assembly photos of the engine's progress to date, but the large white Delrin fixture that I've been using as a temporary rear drive shaft inside the wheel case prevented it from being bolted to the crankcase. Since the wheel case and everything aft of it is where I've spent my time during the past several months, I wanted to include it in the photos. The next step will be to machine the actual dreaded rear driveshaft so I can finally assemble the wheel case and take some meaningful assembly photos. - Terry

In the full-sized Merlin, the crankshaft and wheel case gearing were connected through a thin splined shaft. This spring shaft used rotational flex to reduce the pounding absorbed by the gears from the engine's power pulses. It also helped to dampen potentially dangerous rotational resonances in the crankshaft itself. However, this shaft was not strong enough to handle the engine's starting torque requirements, and so a heavier outer shaft was designed to be engaged when the twist in the spring shaft reached a predetermined amount.
The documentation points out that a need for this additional complexity in the Quarter Scale has never been proven, but it's continued to be part documentation over several design revisions. A final design was included as part of the 2014 revisions. Both John and Gunnar included earlier revisions of this two-part shaft in their engines.
Reproducing a scaled version of the Merlin's rear main shaft in a home shop is difficult because of the requirement to machine three pairs of splines indexed across three separate parts without access to custom spline broaches similar to those used by the Rolls factory. If the splines are not precisely machined and accurately placed using the only tool likely to be available - a keyway broach - the rotational load will not be evenly distributed among them. In the best case, the final result could be inferior to a conventional solid shaft; but in the worst case, metal particles could be forever fretted into the oil system as the edges of mismatched parts vibrate against one another.
The rear main shaft documentation that I have feels like a confusing collection of drawings, some of which were obsoleted by later design revisions. A short adapter, fixed into the rear of the crankshaft, though, is common to all of them. In addition to a recess to accept the main shaft's centering bushing, this adapter has two sets of internal splines: an innermost set for the spring shaft and an outermost set for the main shaft. Splines on the front of the spring shaft engage the innermost splines in this adapter, and splines on the rear of the spring shaft engage a set of internal splines at the rear of the main shaft. When the spring shaft twists under load, it turns the outer main shaft with it via the rear splines until, at about five degrees of rotation, yet another set of splines on the front of the main shaft engages the outermost set of splines on the adapter. At this point, both shafts are locked to the crankshaft.
Three options were suggested in the notes for creating the troublesome internal splines that mate with the spring shaft. The option with the highest torque capacity, and the one that best resembles the shaft in the full-scale Merlin is a high-count multi-groove spline. This option requires the creation of a special broach or the use of an EDM machine, and so I ruled it out early on.
The second option was a scaled approximation to an SAE 6-groove spline cut using a keyway broach. This option produces a nice looking result that I very much liked, and I tried for several days to implement a version of it in SolidWorks. I used partial drawings that were provided in the documentation as starting points since I didn't feel a complete workable design had been included. I ran into several issues especially because I didn't want to re-machine some interrelated wheel case parts that had already been completed according to the original drawings. The shallow splines in the main shaft would have required me to butcher up an expensive broach, and a set of matching external splines with the proper angles would have required a custom ground tapered end mill. I also couldn't come up with a way to retain the spring shaft inside the main shaft without significantly compromising the strength of the main shaft since it was too late to increase its diameter. On top of everything else, I still had no real way to guarantee a precise and uniform placement of the splines by cutting the grooves individually with a keyway broach. At the end of the day, option two would have looked really great from afar, but I would have been left with a lot of uncertainty about its integrity. I think John was able to make this option work for him, though.
The third option, and the one I reluctantly adopted, was a simple square drive for the spring shaft. A complete design for this option was provided in the documentation, and it appeared to be the supported choice in the 2014 re-design. Theoretically, it offers less torque capacity than either of the first two options, but an actual home shop implementation could very well end up stronger. It clearly provides the desired loose coupling between the crankshaft and wheel case gearing, and fitting and indexing square shafts and square broached holes is much easier than dealing with keyway broached splines. My only real objection to using it in the Merlin is that It reminds me too much of the flexible drives in my gasoline-powered yard tools. Fortunately, it won't be visible after assembly, and my tired old brain will likely bury all images of it along side several other compromises I've had to make along the way.
Three parts make up the rear main shaft assembly, and I started by machining the crankshaft adapter from Stressproof steel. The end that fits into the crankshaft was turned for a close slip fit, and the pilot bore for the square broach was drilled and reamed in the same setup. The workpiece was then turned around in the chuck so the recess for the main shaft's centering bushing could be bored. I intentionally used the runout in the chuck to offset the bushing recess in hopes of canceling some of the TIR that I was previously left with in the end of the crankshaft. I was surprised when this worked because usually when I attempt something like this it backfires and makes things worse.
The broaching of the hole for the square drive was started in the lathe using the tailstock as a press in order to help get the hole started straight, but I completed the broaching in an arbor press to avoid abusing my lathe's headstock bearings. I then moved the workpiece to the mill. It was very important that the adapter's mounting holes and outer set of inner splines be precisely indexed to the square drive hole. To do this I inserted a tight-fitting square tool blank into the hole and then selected a gage pin to use as a shim for aligning the workpiece to the mill's vise. When the milling was completed, I moved the adapter back to the lathe in order to turn a taper on its rear nose. This taper was necessary for clearance to the pump shaft's bevel gear. I added a radially drilled hole in the bushing recess so the pressurized oil feed in the rear bearing cap would also lubricate the bushing. This hole will also be used later to index the bushing to its minimum TIR position when it is permanently installed.
The spring shaft was also machined from Stressproof. After turning a blank on the lathe with generous filets at the diameter transitions, I used the mill's fourth axis to mill the square ends and to insure they were aligned to each other. A corner radius end mill was to machine the flats to avoid any sharp corners at the intersections between the flats and the cylindrical shaft.
The last shaft component to be machined will be the main shaft, itself. This part will be a much more complex, and it has several features that must be accurately indexed to one another. Its TIR will also affect the fit of three major gears that are already a part of the wheel case. - Terry