Quarter Scale Merlin V-12

[mayhugh1]

The quarter scale Merlin's prop shaft is a scaled version of a design that's still in use today on full-size aircraft. It's externally splined, and its end is threaded for a castellated prop nut. In order to fit a propeller to this shaft I designed a splined hub that will become an integral part of a four blade prop. Its front also includes a threaded section for a spinner.
I've cut key slots before, but never splines - neither internal nor external. Cutting external splines looks fairly easy and is pretty similar to cutting gear teeth. Initially, the internal splines seemed more intimidating, and so I decided to start with the hub.
One of the options on the Merlin's prop shaft drawing is a variation of an SAE 16 straight tooth spline. I designed the hub for sixteen 3/4" long splines spaced 22-1/2 degrees apart. The splines in this hub can be cut using an ordinary 3/32" keyway broach. Since I had a .875" diameter broach bushing, this became the i.d. of the hub as well as the minor diameter of the shaft. Actually, I reduced the shaft diameter .010" to obtain a tooth end clearance of .005". I pulled this number out of the air with no experience or reference material to justify it. Broaching this i.d. creates sixteen teeth having a 22-1/2 degree included angle and a height of a .050". The prop splines must be cut using a form tool that matches the shape of these teeth.
My initial thought was to use a rotary to index the hub under a broach held in the locked spindle of my mill. I did a few experiments and realized the force required to drive the broach through a one inch piece of 12L14 was more abuse than I was willing to apply to the bearings in either my spindle or rotary. Instead, I manually indexed the hub, and I used a manual press to drive the broach. I CNC-engraved the top of the hub's blank with the 22-1/2 degree alignment marks and I scribed a corresponding reference mark on the broach bushing. Accurately aligning the broach for each cut was easy, and the whole splining operation took less than half an hour. After finishing the splines, the remaining machining on the hub was completed.
Instead of spending fifteen minutes grinding a single tooth form tool for the shaft splines, I spent most of an afternoon modifying a 3/16" HSS Woodruff cutter. The sides of the teeth were ground as close as possible to the theoretical 11-1/4 degree side angles, and the width of the flat end came out to .078". Normally, I would have turned a four tooth cutter from scratch as I did for the crankshaft gear for my two radial builds. In this case, however, I had a lot more material to remove, and I wasn't sure how well drill rod, even heat treated, would hold up against the Stressproof shaft especially if I had to make more than one shaft. The most difficult part of the Woodruff cutter modification was relieving the sides of the teeth. During testing I found this relief helped to reduce the height of some very tough burrs that tended to rise up on either side of the spline. These burrs interfere with the test fitting of the hub to the shaft while it's still fixture'd for splining.
I first tested the cutter by splining an aluminum test shaft. The splines looked great, and the shaft slipped into the hub with a close sliding fit. The clearances of the splines on the rear of the hub where the broach had entered looked nearly perfect, indicating that the spline cutter was accurate. However, the clearances between the shaft's teeth and the hub's slots on the front of the hub where the broach had exited were much greater. I checked the broach bushing and the perpendicularity of the press, and all was as it should be. I purchased a new duMONT broach since the one I had been using was part of an inexpensive import set. I made a new blank and reduced some clearances, but the results were essentially the same. Even with 3-4 teeth cutting at a time, the bottom of the broach drifted to the outside of the blank during every cut causing the splines at the bottom end of the hub to be deeper than those at the top end of the hub. Turning the set-up around on the table of the press had no effect. I examined several parts on other projects that I had key-broached, but none of them showed the drift I was seeing on these hubs. The only solution seems to be a rigid set-up to keep the bottom of the broach tight against its bushing, but since it wasn't obvious how to do that, I decided to save it for the next splining project.
Although not desirable, the tapered splines really don't create a significant problem in this particular application. Both hubs fit the test shaft snugly since the hub teeth fit into the shaft splines with their proper clearances along the entire length of the hub. It's only the shaft teeth that end up with excessive clearances to the hub slots. The two-piece propeller hub that I plan to make will sandwich the prop and should insure its perpendicularity to the prop shaft independent of the splines. I was concerned about possible run-out issues with the hub that would cause the spinner to wobble. But, with sixteen mating possibilities, it wasn't difficult to find one with near zero run-out.
I turned an 1144 prop shaft blank and adjusted the parameters of the splining program for the new material. When the part was completed it was immediately obvious that the splines weren't uniform. I had evidently over-tightened tailstock of the the fourth axis (again), and this caused the rotary to loose steps. My second attempt was more successful. Its fit inside either hub was smooth and very snug with no backlash.
The rest of the machining on the shaft was completed including a pair of shoulders for two ball bearings and a 7/8-20 thread for the prop nut. The bearing shoulders were turned concentric with the splines in a 4-jaw set-up. The center of the shaft was also drilled out in order to lighten it, and a mating position for the hub was found that produced a TIR of only .001". Punch marks were added to the hub and shaft for use during assembly. Finally, a bearing retainer was machined for the front of the gear case to limit the forward thrust of the shaft.
Manually rotating the shaft of the completed gear reduction assembly smoothly spun the crankshaft at nearly twice speed with no binding or rough-felt areas. Now there's even a flywheel effect due to the large prop gear. The next step, while I'm working at the front if the engine, will be to finish up the rest of the prop components. - Terry

 

[jimjam66]

That is just awesome machining Terry! Can't wait for more progress.

 

[Mattsta]

An absolute work of art

 

[petertha]

> sixteen teeth having a 22-1/2 degree included angle and a height of a .050".... Instead, I manually indexed the hub, and I used a manual press to drive the broach.

I don't know much about presses or broaches. Just out of interest, what kind of force/capacity is required on an operation like this? is it a once-through type deal, or a progressive sequence of cutters or shims to get to 0.050" depth? Very nice!

 

[mayhugh1]

Peter,
I was using an Enco one ton arbor press, and it took nearly every thing I had to push the broach through the hub. The broach was a progressive type with a number of teeth and about five inches long. It didn't require any shims and it was designed to cut a .050" deep keyway in a single pass. - Terry

 

[ddmckee54]

So you think maybe pushing/exceeding the limits of the press is what caused the broach to drift in the bore? I realize that you've got a work-around for the parts you've already made, but maybe for future reference?

Don

 

[ICEpeter]

Considering the direction the teeth of the broach are facing when broaching, it seems likely that the teeth and broach pull in the direction where the cutting force bites into the material. That would in my opinion cause the depth of the cut to be deeper at the bottom versus the top. A sort of reverse guide for the broach below the bottom of the piece that keeps the broach in line and counteracts the broaches teeth attempting to dig in deeper than wanted may do the job but is difficult to implement. Another way may be to make the bushing longer and part the excess off after broaching.

Peter J.

 

[ICEpeter]

The effect of the broach wandering to the outside maybe reduced if a multi step broaching process is used by using maybe two or three stepped broach bushings and taking lighter cuts and not one full depth cut with the broach.

Peter J.

 

[mayhugh1]

All good comments. The specs for the Dumont broach are a maximum length of cut of 1-1/8" and a required pressure of 780 pounds. I suspect the one ton press being max'd out was probably related to the broach drifting and taking a bigger byte than it was supposed to. It's probably also true that the press isn't really capable of its own rating.
The broach was rated to take the cut in one pass, but creating some additional bushings for a couple lighter passes might have helped. I thought about trying that at the time but cutting deep 3/32" square slots in steel bushings looked like another set of problems to deal with.
If you study the teeth on the broach you'll probably agree with PeterJ that the broach is just waiting for an excuse to drift out and take too big of a byte. Probably even an imprecise sharpening job would make the problem even worse very quickly.
After thinking about it for a few days, I'm beginning to think the problem was in my technique. I had only a quarter inch clearance in my press to start the broach; and I'm pretty sure that due to a lack if experience I wasn't paying enough attention to getting the broach started truly vertical with its spine totally against the bushing. This is a bit trickier than you might think since at the start of the operation the broach isn't yet sticking out the bottom of the workpiece, and at the top there are one or two teeth already down inside the bushing. So you need to pay attention to the angle of the broach with respect to the top of the workpiece which I was probably not doing. - Terry

 

[bigrigbri]

Bloody marvelous

 

[mayhugh1]

After finishing the prop shaft it seemed reasonable to complete the rest of the parts needed to mount a propeller even though I'm a very long way from needing them. The Merlin drawings don't include a prop, but it wasn't difficult to adapt a large scale model propeller to its prop shaft. During my last radial build I was given a 26"x10" four blade prop that I never used. It'll be a negligible load for the Merlin even at its highest rpm, but it happens to be a semi-scale prop for a P-51 Mustang, and so the Merlin will be a fitting home for it.
The center of the propeller was drilled out with a 1-1/8" Forstner bit to accept the splined hub I machined earlier. A second hub without splines was turned and bored for a close slip fit over the prop shaft. The propeller will be sandwiched between these two hubs, and the assembly held together with six 6-32 SHCS. The ends of the hubs bottom against each other inside the prop to avoid compressing the prop when the screws are tightened. Because of the clearances involved, the bolts need to be tightened while the assembly is aligned and on the prop shaft. I made this step difficult when I designed the assembly with the bolt heads facing to the rear of the prop. Fortunately, the splined aluminum test shaft I made earlier turned out to be a perfect tool for performing the assembly off the engine.
I machined a castellated prop nut to keep the prop assembly on the shaft. Its 7/8"-20 thread was cut for a very close, but smooth fit to the shaft. The prop assembly is captured between this nut and the inner race of the large gear case bearing. The rear side of this race is backed up by a machined shoulder on the prop shaft. An actual full-size prop assembly would have also included a pair of centering cones.
The front hub was previously threaded for a spinner to complement the P-51 prop. My decision to attach the spinner in this way wasn't well thought out, and it greatly complicated the spinner design after I decided to extend it between the prop blades. This required a two piece spinner with a rather complex rear half.
The spinner halves were designed around a pair of short 4" diameter 6061 drops that I had in my scrap collection. After truing them up, the machining on the rear end of the front half was completed while it was still easy to fixture. This included a 1-1/2"-18 thread carefully matched to the one on the front hub to minimize any run-out of the spinner. This hub was also used as a mandrel for turning the spinner's profile. A small radial through-hole was also drilled and reamed for a tommy bar. The profile was drawn in SolidWorks and turned on my 9x20 lathe using g-code generated by Sprutcam. I really wanted to make the front spinner slightly longer, but I continued with the material I had on hand.
The rear half was machined from the second drop. It turned out to be a fairly complex part for all that it does. Several boring operations were completed on the lathe before moving it to the mill to machine the prop blade cut-outs and internal contours. The deep and narrow slotted mounting holes were the most difficult features to machine, and I ended up writing scratch code to plunge mill them. Plunge milling is a feature that would be worth the price of an upgrade to my CAM software, if it were ever offered. During assembly, the rear half of the spinner with its four loosely inserted mounting bolts is slipped onto the prop shaft before adding the prop assembly. After tightening the prop nut, a shoulder on the front half of the spinner slides into a recess in the rear half as it's threaded onto the front hub. Finally, the four rear spinner bolts are screwed into threaded holes in the front half of the spinner. Removing the propeller will be a bit like opening a Chinese puzzle box. - Terry

 

[stevehuckss396]

Fan-freakin-tastic

 

[Swifty]

Your attention to detail is amazing, even your clamping down of the prop to bore out the centre is well thought out and executed.

Paul.

 

[Scott_M]

Holy Crap Terry !!

That is one impressive spinner. It looks really cool with the 4 blade prop.

That is definitely not one of those snap on plastic spinners from Tower Hobbies .

Scott

 

[Gerhardvienna]

Hi Mayhugh1
Who the hell will build the MK24 around it and fly that monster:hDe:
You are a real role-model to me! Thanks for showing, waiting for much more!!!!!!
Regards
Gerhard

 

[ddmckee54]

OK, I think I'm following how the spinner is mounted. The front half of the spinner is threaded onto the front of the prop hub, then the back half of the spinner is bolted to the front half, easey-peasey.

My question is how are you going to start this beast? If you use an electric starter pressed against the spinner isn't that going to loosen the spinner? I realize that it can't loosen too much, but won't that allow the slots in the back half of the spinner to rattle against the prop blades? What am I missing?

Don

 

[wirralcnc]

Scale coffman starter???

 

[mayhugh1]

Don,
You're right. I've not provided a way to start this engine using a starter on the front. This engine has a starter shaft that engages a gear on the rear of the crankshaft. It has a one way clutch and can be driven by a drill, but I'll likely adapt an electric motor to it. - Terry

 

[ddmckee54]

Terry:

I knew there had to be something that I wasn't seeing, thanks for showing me the light.

Don

 

[mayhugh1]

I returned to working on the heads and, in particular, the camshaft assemblies. I should start with the valves, but I have some more thinking to do about how I'm going to handle the seats. There will be nearly 300 parts in each top end, and so I'm going to be working on them for a while.
The Merlin is an overhead cam engine, and each camshaft in the full-scale version is connected to the crankshaft through nearly a dozen gears. Mercifully, the designers of the quarter scale version reduced the cam drive system to a much simpler chain drive, and they converted the model from four valves per cylinder to two. The quarter scale's top-ends are really impressive and probably why I've seen photos of the model displayed without its valve covers.
The camshafts are almost ten inches long. Each has seven 9/32" diameter bearings and will be turned from a single piece of steel. I thought the crankshaft would be the most difficult part of this engine, but making the camshafts will be another growth experience.
The first challenge, though, is to machine the bearing blocks that support the camshaft and rocker arms in each head. There isn't enough height clearance in the heads to properly line bore the bearings, and so great care is needed to individually machine these deceptively simple parts so they can be assembled in proper alignment.
The first step was to machine the mounting surfaces for the bearing blocks in each head. An alignment key slot was milled at each bearing location in addition to a pair of drilled and tapped holes for mounting studs. The key slots are especially critical, and they form the foundation on which the assemblies are built. Matching keys are milled into the lower halves of the bearing blocks, and these fit tightly into the alignment slots.
Each bearing block consists of an upper and a lower section that are dowelled together with two short lengths of #9 gauge hypodermic tubing. The mounting studs pass through these dowels. In order to make the parts as identical as possible the first stud hole in each bearing block pair was reamed for its dowel with a snug fitting gage pin inserted in the opposite side hole pair. The second hole pair was reamed in the same set-up with the first dowel set in place. Even though a floating reamer was used, testing showed the holes had to be reamed .001" over the diameter of the dowels in order to absorb machining errors and facilitate assembly/disassembly of the block pairs.
In addition to alignment keys, the drawing notes recommend machining the mounting studs with essentially zero clearance to the dowel i.d.'s. Following this recommendation locates each bearing block to its position in the head with three zero-clearance anchors. In addition, since the studs penetrate the coolant jackets, and so the drawings warn that the studs must be Loctited in the heads using a fixture to insure the studs are truly vertical when cured. It seemed to me that using the studs in addition to the keys to locate the blocks in the heads would create more problems than they solve. I could visualize ending up with fitments that were too snug to even be assembled, and most likely the bearings would have to be reamed over-size anyway.
Instead, I decided to mount the bearing blocks using SHCS's whose o.d.'s measured .005" under the dowel i.d.'s. I reamed the blocks a thousandth over the diameter of my camshaft test bar, but as expected the fit of the bar in the seven block assembly was too tight. After reaming the bores another thousandth over, the test bar turned freely. Both head assemblies were checked using the same +.002" bearing bore, and they were assembled/disassembled a few times to verify the alignment repeated. The test bar proved the block alignment was sufficiently precise, and its diameter set the target for the cam bearing diameters.
Although it wasn't necessary to select blocks for a particular location, I plan to engrave the blocks with their current locations so the same tested assembly can be duplicated.
The next step is to add the rocker arm assemblies. - Terry