Quarter Scale Merlin V-12

[mayhugh1]

The Rolls Royce V-12 Merlin, was one of the best known, if not most influential, WWII aero engines. It was deployed in the British Spitfire and later replaced the Allison in the American P-51 Mustang. I recently purchased a set of quarter scale castings from a small San Diego start-up that originally planned to build and sell completed quarter scale Merlins nearly a decade ago.
http://www.quarterscalemerlin.com
The parts I received were investment cast and can be best described as large pieces of (expensive) jewelry. They share most of the realistic features and intricate detail with the equivalent parts on the full-size engine. Photos of the castings are available here:
http://www.quarterscalemerlin.com/castings/
I've no experience in working with castings, and was a little taken back by the notes accompanying them. The notes warned, in several places, that being long, complex, and thin-walled, they will likely require straightening and, in some cases, heat treating.
The set I purchased includes castings for a functional supercharger, but it's not clear whether its scaled development was ever fully completed and just how much of it became a part of the prototype that was produced. The original designers opted for a glow plug engine, and so the magneto development may not have been completed. Finally, the notes mentioned fuel distribution issues with the Merlin's scaled-down intake manifold. The developers eventually designed an alternate configuration with multiple carburetors in order to get a running prototype, but the drawings didn't include information on its design. Over-heating issues were also mentioned, and a prop didn't show up in the published video of their running engine. Working these issues will add some interesting challenge to the project, but I'd rather additional development work wasn't going to involve very expensive and perhaps irreplaceable castings.
I've been able to find online evidence of three other builders who have tackled this project using these particular castings. One posted his crankshaft build on 'the other' forum but he never returned after creating his own piece of art.
My plan is to spend the first few weeks evaluating the castings I have so I can better understand the issues involved with getting them ready to machine. My first goal will be to see if I can get the major crankcase components straightened and fitted together with minimal machining. -Terry

 

[///]

I've been dreaming about these castings for a very long time but I just know it is well beyond my capabilities!
Really looking forward to following your build.

 

[Herbiev]

WOW!! That's some project. Will be following along for sure

 

[mayhugh1]

I'm going to try to justify some of the craziness that I'm about to apply to these very expensive parts with some background theory. Precipitation hardening is a common way of strengthening 356 aluminum which is a popular casting alloy. To promote this process, certain impurities such as magnesium and silicon must also be present in the aluminum melt. As the castings cool these impurities form simple compounds which gradually, over time, come out of solution (precipitate) and they end up distributed throughout the casting. These precipitates harden the casting by preventing its plastic deformation (bending or stretching) when it is put under stress. These precipitates strengthen the casting, but they also make it brittle. This hardening process is kicked off just after the castings solidify, and it continues for hours to days or even weeks later depending upon the casting's storage temperature. Because of this process's dependency upon time, it is also commonly referred to as age hardening.
A casting that has been age-hardened has little tolerance to bending, twisting, or stretching. If a 356 casting warped during its solidification, and if it requires straightening before it can be finish machined, then it must be annealed. This can be done by heating the casting to about 700F and then allowing it to air-cool. The common shop technique of using an acetylene torch to create a soot coating for use as an annealing temperature indicator also works for 356. If, after straightening, the part is left in its annealed state, significant strength will be lost. For 356 the tensile strength loss can be as great as 10,000 psi. Unfortunately, a 700F annealing is not high enough to kick off another age hardening cycle.
The casting can be re-hardened, though, back to its maximum strength by heating it to 1000F for a dozen or so hours and then quickly quenching it. What makes this difficult to do in a home shop is the fact that aluminum melts at 1035F, and so careful temperature control is required. There is also a chance that the casting will deform under its own weight if it isn't properly supported. In addition, air-cooling needs to be minimized which means the quench tank needs be located within seconds of the furnace.
Unlike the more familiar hardening process associated with steel, precipitation hardening does not occur immediately after the quench. The metal may remain soft enough to be straightened for up to a full day after the quench.
Production castings should normally be straightened by the foundry before age hardening has progressed to any significant extent. An even better solution, of course, is to design the part so warpage is a minor concern, and the casting can be corrected by finish machining. The Merlin castings were not straightened by the foundry, and their thin-wall and complex cross sections make them very susceptible to warping that may not be correctable solely by machining. Straightening, after the castings have been allowed to harden, was therefore left to the end-user.
Since this is a totally new experience for me, I felt it would be best to practice on some scrap cast parts I picked up long ago from my favorite scrapyard. My practice pieces, which are louvered vents, were sand cast from 356 and allowed to age-hardened for many years.
I decided to immediately answer a question that was in the back of my mind, and that was just how much could I permanently deform one of these castings without annealing. After breaking two practice parts, I realized the answer was 'pretty much nothing at all.' The rest of my practice was done using only annealed parts.
I eventually developed a process, after cracking a few annealed practice parts, for controlling the pressures I used to bend the castings. I learned to clamp the parts down firmly and to use positive calibrated stops to quantitatively limit the distance that an edge was being pushed. Rather than using a press I typically used my own strength and body weight in combination with fulcrums, levers, and clamps so I could maintain a hands-on feel for what I was doing. I found it was important to proceed in small deformation steps of .005" at a time and continually return to the surface plate to check my progress. I also decided it was best to not aim for perfection but to stop at the point where measurements showed I could machine the remaining defects away without negatively impacting the part's appearance. Since I had decided to not even attempt age-hardening in my shop, I tried to minimize the areas that I annealed. Before attempting any straightening, I located the major axis of the warpage using a surface plate, and I tried to anneal only a narrow region along that axis. I then applied my straightening efforts across this axis. After a full day of experimenting I had gained enough confidence to start on the Merlin parts.
I first selected the three crankcase-related castings. I was able to machine flat the bottom surface of the main casting with respect to the crankshaft bores in order to obtain a reference surface. The front gear case turned out to be the major problem area on this part. It was out of perpendicular by almost .050" over its 5" height. I annealed a line across the gear case just above the top deck of the crankcase. After clamping the crankcase with its reference surface down to my drill press table, I clamped a long piece of wood to the gear case to which I applied the straightening force. With a pencil mark on the wood as a moving pointer I carefully monitored the distance the part was being pushed. After a half dozen tries which included returning to the surface plate to check my progress after each push, I had finally bent the gear case to within .015" of perfection. At this point I was able to machine its cover mounting surface flat and perfectly perpendicular to the reference surface in order to meet the drawing dimensions with no noticeable impact on appearance. I then machined the rear of the crankcase to its finished dimension. The decks for the cylinder heads will be done later, since they can be cleaned up with just finish machining.
The gear case cover was relatively simple to correct because its major warp was also about a single axis. The documentation warned that this rather rigid part might have to be widened, and the drawings included the design of a complex 'stretcher' to attempt this. I'm thankful that my particular cover, which is a fairly rigid part, didn't require this really scary correction.
The oil pan was considerably more complex and problematic. Being rather flimsy by design, it was warped across two separate axes. In addition, it's width had to be spread to match the crankcase. When checked on the surface plate, one corner of this part was initially almost 1/8" higher the other three. This part required almost a full day to correct, and I ended up annealing practically the whole casting. Fortunately, the oil pan is not a structural part, and the loss in strength that it likely suffered is not important. In the set-up for its final flange machining, the pan had to be packed with plastic modeling clay in order to dampen the chatter created by the mounting flange machining. I used high relief aluminum-cutting Korloy carbide inserts for all the machining operations and was able to obtain excellent surface finishes on both the annealed and un-annealed areas of all three castings. - Terry

 

[kvom]

Quite an interesting process. If I understand properly you only needed to apply the torch to anneal along the axis of warping, while the rest of the casting was not heated.

 

[Scott_M]

Nice work Terry, those really are some good looking castings.

Scott

 

[michael-au]

This is a big project, I know someone that has completed the same engine, I have seen it run
It is an amazing project

Here is a picture of his engine

View attachment ImageUploadedByModel Engines1424169717.048976.jpg

 

[Ethan D]

This is going to be amazing! the detail is crazy! Can't wait to see updates

 

[gus]

Mind boggling and only a true-blue blood HMEM Forum Member will attempt to build this engine. 1,000,0001 parts to machine. This engine will take at least 2 years to complete and run.

 

[Ripcrow]

Following along. Can't wait to hear it run. The most famous engine in the world. To me anyway

 

[mayhugh1]

Commercial parts recommended for purchase by the documentation include 25 ball bearings and 24 gears. That's a lot of shafts. I decided, for now, to purchase only the bearings associated with the crankcase group since I'm going to need them shortly, and the rest are commonly available. Gear cutters, though, are a different matter. My own experience is that actual in-stock gear cutters are becoming hard to find, and they now seem to be manufactured only in Poland and China. I've never had a problem with the Polish cutters, but out of several Chinese cutters I've purchased, two were so poorly made that they were unusable. I've previously purchased the Polish cutters from McMaster-Carr, but back-order times have sometimes stretched out as long as six months. As luck would have it, I have on hand only four of the seven cutters that I'll eventually need. I placed an order for the remainder, but only one was in current stock. The engine's documentation spec'd the gears only as part numbers from a specific U.K. gear manufacturer, and so I used their online catalog to extract the cutter requirements. Purchasing the gears from them outright would have cost nearly $700 compared with less than half that amount for all the necessary cutters. If the back-ordered cutters don't arrive before I need them I'll make my own as I did for the CAM drive gears on my radial engines. Most of the gears in this engine are cut from steel; and so I'd rather start out with quality commercial cutters, if possible.
The next step in this build is to mate the crankcase to the gear case cover and oil pan which, in Merlin parlance, is the lower crankcase. I decided to do this with 2-56 SHCS's even though the plans suggest an elegant but much more complex 'to scale' option. There's very little excess screw boss material on either part; and so cover needs to be precisely located over the gear case, and the spindle needs to be carefully centered over each boss before drilling. The bosses in the cover have cast-in dimples to help locate a drill, but not all of them are actually centered; and so I decided to re-spot them all with a 2-flute carbide v-drill. My eyes continued to trick me into positioning the spotting drill over the off-center dimples instead of the true centers of the bosses. So, I put a small length of close-fitting brass tubing over the spot drill in order to hide them. With this combination I was able to easily position the spindle over the center of each boss so I could re-spot its center (after removing the tube, of course).
In order to locate and temporarily secure the cover on the gear case I turned close-fitting Delrin alignment plugs for the as-cast openings in the prop and drive shaft bores in both the gear case and gear case cover. These openings were amazingly round - and indicated to within a few thousandths. I drilled four .070" holes through the pair which is the tap drill size for 2-56. After removing the cover I reamed the cover holes to .087" to clear the screws and then I tapped the four holes in the gear case. After bolting the cover back into place on the gear case using these four holes, I drilled and tapped the remainder of the holes similarly. Match drilling the holes in this way eliminated transfer errors and allowed minimum clearances around the cover holes. A simple shop-made alignment tool was used to hand-tap the 2-56 holes. The 45 holes securing the oil pan to the crankcase were drilled and tapped similarly.
The crankcase was then mounted vertically on the mill table against an angle plate with the gear case facing upward in order to bore the openings for the drive shaft and the prop shaft bearings. (Drive shaft is Merlin-speak for the front end of the crankshaft.) The center of the drive shaft opening was indicated and used as the reference datum for the next four boring operations. It will also be used later when the crankcase is line-bored.
With the cover bolted in place, the opening for the driveshaft bearing retainer was bored. The spindle was then moved into position to bore the front prop shaft bearing opening. The distance between the prop shaft and drive shaft is critical for proper mesh of the gear pair that will eventually connect them. I had to rely on the print for this distance as I don't yet have the cutters needed to make these particular gears.
The cover was then removed and, without moving the spindle, the blind recess for the rear prop shaft bearing was bored. Leaving the spindle in place for both boring operations insured alignment of both prop shaft bearings. A shop-made bearing puller was needed to check the fit of the rear prop shaft bearing in its blind recess. Finally, the spindle was moved back to the driveshaft reference point where the lower gear case opening for the driveshaft retainer was bored.
Being overly cautious about cracking one of these castings I opted for an interference fit of .0002"-.0003" instead of the .001" fit I would have normally used for these bearings in aluminum. The result is a snug fit using my upper body weight as a press. All four bearings will eventually be held in position with retainers. The circular hole patterns for the retainers' mounting screws were finally drilled in the cover. The prop shaft retainer uses 2-56 but the drive shaft retainer requires 1-72.
There are a lot of tiny tapped holes in this engine. I've barely scratched the surface of this build, and I've already drilled/tapped ninety-two 2-56 /1-72 holes. And, there's a lot more 1-72's and 0-80's ahead. - Terry

 

[wirralcnc]

Terry great progress. I will be following with great interest,and following your techniques for when I machine mine. I actually purchased the full set of gears from hpc gears. Great quality and good delivery time.

 

[mayhugh1]

Wirralcnc,
One thing I haven't mentioned is that you'll need to spend several hours with dental picks carefully going through the castings to remove all remaining traces of investment stuck in the numerous hidden interior corners. Digging around in there will give you even more appreciation for the effort someone put into their design. On mine, the water passages in both cylinder blocks were totally blocked. I discovered this while trying to understand how the liners were actually cooled. You don't want this stuff circulating in the oil or coolant loops. - Terry

 

[petertha]

Wow Terry, you don't mess around. Impressive start to a complex project.

Newbie question: I've noticed variants of your tapping method for small, finicky size threads - freehand locating a tapping block jig over the pilot hole & some sort of knurl wheel handle affixed to tap. I've got the same job in front of me.
- is the tapping block hole a close sliding fit to tap OD or also partially threaded to assist engagement in the part?
- do you use the block just to the point where tap is nicely established, then remove bock? Or better to keep block positioned the whole operation?
- any words of wisdom on threading blind holes in aluminum (tap type, cutting fluid, swarf ejection etc)

 

[digiex-chris]

I've got a thing for the Merlin engines after working on a full size Lancaster bomber at the Nanton air museum as a kid. Recently got to hear them run. Can't wait to hear this one run!

 

[digiex-chris]

Wow Terry, you don't mess around. Impressive start to a complex project.

Newbie question: I've noticed variants of your tapping method for small, finicky size threads - freehand locating a tapping block jig over the pilot hole & some sort of knurl wheel handle affixed to tap. I've got the same job in front of me.
- is the tapping block hole a close sliding fit to tap OD or also partially threaded to assist engagement in the part?
- do you use the block just to the point where tap is nicely established, then remove bock? Or better to keep block positioned the whole operation?
- any words of wisdom on threading blind holes in aluminum (tap type, cutting fluid, swarf ejection etc)

I use a tapping block like that all the time. A close fit to the tap thread major diameter seems to be fine. If the hole I need to thread is so deep and the material so hard that I'm concerned, I'm locating the tap with a spring center on the mill. Especially in aluminium, I almost always spend the extra cost on spiral flute taps. Also interested with the huge experts here suggest.

 

[mayhugh1]

Wow Terry, you don't mess around. Impressive start to a complex project.

Newbie question: I've noticed variants of your tapping method for small, finicky size threads - freehand locating a tapping block jig over the pilot hole & some sort of knurl wheel handle affixed to tap. I've got the same job in front of me.
- is the tapping block hole a close sliding fit to tap OD or also partially threaded to assist engagement in the part?
- do you use the block just to the point where tap is nicely established, then remove bock? Or better to keep block positioned the whole operation?
- any words of wisdom on threading blind holes in aluminum (tap type, cutting fluid, swarf ejection etc)

Peter,
The block isn't threaded but is a close fit to the body of the tap. I used it to go full depth with WD-40 for a lubricant. The depths of many of these holes were about the same as the total length of the threaded portion of the tap. I typically tapped them in two steps of about half depth each. I cleared the chips and re-lubed the tap for each cycle, and I use compressed air to clear the chips from the hole. For these holes I just used a hand plug tap. - Terry

 

[XD351]

It takes a truly special machinist to take on building that big daddy and i'm really looking forward to watching the build progress !
It inspired me enough to hunt around on you tube to find anything i could on a Merlin engines and surprisingly enough there is some fabulous footage of the full size engine being made .
It amazes me that men worked in factories wearing ties ,vests ,suites and the ladies wearing floral dresses !
Some of the machinery just blew my mind and much of it was designed specifically for this engine.
I never knew Packard made them also or a version of them anyway .

Ian.

 

[gus]

I use a tapping block like that all the time. A close fit to the tap thread major diameter seems to be fine. If the hole I need to thread is so deep and the material so hard that I'm concerned, I'm locating the tap with a spring center on the mill. Especially in aluminium, I almost always spend the extra cost on spiral flute taps. Also interested with the huge experts here suggest.

Hi Chris,

Good and wise decision to invest in quality spiral taps that gives good thread and chips are brought up and not pushed down into tapped hole and causing obstruction. Gus invested in Japanese Taps and Dies. Why save pennies and risk breaking a tap and lose the job piece at the very last tapped hole?? ( I have yet to see any M.I.C. Spiral Taps in the market)

 

[Cogsy]

( I have yet to see any M.I.C. Spiral Taps in the market)

Just had a look on ebay Australia and there are plenty of Chinese spiral taps in metric at least - looks like they've caught on.

Terry - Another excellent build for me to follow. I love the Merlin and I once got to fly 'hands on stick' in a very rare twin seat trainer P51D. Cost $2000 for a 20 minute ride. Over 4G in the loops, 500+ KPH flyover the family with me at the controls and that beautiful noise. Love it.