Quarter Scale Merlin V-12

[DICKEYBIRD]

I can't help but say this project for one man to accomplish is no less a feat than NASA's moon landings in the 60's!

 

[ddmckee54]

Terry:

If they don't like your oil lines, let them build their own 1/4 scale Merlin. Then THEY can try making scale oil lines and you can complain that the resulting oil leaks aren't scale sized.

Don

 

[tms6401]

Absolutely incredible build!! There are no other words that haven't been used many times already. You must get an incredible feeling of satisfaction every time you look at or touch this engine.

Tom

 

[camm-1]

What can I say more than amazing beautiful and exelent skills of everything Terry!

 

[Blogwitch]

Terry,

As they say, it is just as easy to make two as it is to make one.

So if I send you a couple of hundred bucks, will you knock one off for me?

John

 

[mayhugh1]

Thanks all for your kind comments ...

In order to wrap up the coolant system components that are actually mounted on the engine, a means must be provided to return the coolant from the outlets on the fronts of the heads to an external off-board radiator. In the Merlin's aero applications the full-size engines used a header tank for this.

For me, the header tank was one of the most difficult to understand (and ugliest) components on the Merlin. The Quarter Scale documentation contained little information about configuring a coolant system for the model and nothing at all about this tank. Dynamotive's Youtube video shows a scratch-built tank installed on its prototype, but this was a late attempt to solve the engine's overheating problems.

The full-scale tank, hidden beneath a plane's cowling, took on a number of different shapes depending upon the model of the engine (there were 57 variants) and the airframe in which it was installed. On an engine stand it would have been situated well below the prop wash and appear to provide little cooling benefit.

A revelation concerning its real purpose occurred when I came across a re-builder's online photo showing its internals. The tank wasn't at all what I had expected and was filled mostly with air. Its actual purpose was to isolate a pair of large diameter (expansion) coolant return lines from the engine's heat so they could de-aerate the coolant before it was returned to the radiator.

Since I plan to display and (hopefully) run the Quarter Scale on a test stand, I chose to fabricate a functional header that was more appropriate for that set-up. The engines I've seen displayed in simulated Spitfire or P-51 mounts are very impressive, but they greatly limit access to areas of the engine that will likely require some fiddling to get it running for the first time. These mounts would certainly demand a header tank, though, that was more reminiscent of the one in the first photo.

I started fabrication of my header by forming a length of 3/4" diameter stainless tubing into a 200 degree five inch diameter bend. The 3/4" die set that I own for my tube bender happened to be very close to the required diameter, but the last 20 degrees of the bend had to be muscled in using a vise and a large clamp. In order to prevent the tubing from deforming, I filled the starting workpiece with Cerrobend.

I've learned through experience that a lot of Cerrobend headaches can be avoided if the time and effort are taken to use it properly. In order to avoid overheating the metal, it was melted in a beaker sitting in a pan of boiling water on our kitchen stove. While the metal was being heated, one end of the tube was tightly corked and filled with ordinary cooking oil. The oil is really necessary to keep bits of Cerrobend from later sticking to the interior of the tube, and one of the reasons for not overheating it with a torch is to prevent scorching the protective oil. The oil was poured out of the tube just before filling it with the molten metal. The other end of the tube was then quickly corked so the workpiece could be plunged into a sink filled with ice water. The fast chill added some ductility to the Cerrobend. After the bends were completed the tube, with its two open ends pointing up, was re-heated in a large pan of boiling water. The Cerrobend poured out cleanly with no 'cling-ons' and was reclaimed in a scrapped muffin mold.

Machining began by notching the tube for a pair of inlet fittings that will eventually connect the header to the engine's outlet fittings through a pair of short pieces of flexible hose. These fittings were lathe-turned from 303 stainless and then tack-welded to the tubes. This particular stainless alloy isn't really weldable, and so the fittings were soldered to the tubes. I used a 96% Sn, 4% Ag solder alloy available from TM Technologies (kit #ABS-0065 includes the flux)
https://www.tinmantech.com/html/soldering.php
I previously used this particular product, recommended to me by Petertha, to fabricate the fuel inlet tubes for my 18 cylinder radial.

I wanted the tube and its fittings to appear as a single sculpted part, and so I buttered the solder onto the assembly as though I was doing auto body lead-work. The next several hours were spent with files and emory paper metal-finishing the result. It's less frustrating to perform the initial shaping with a file that doesn't tend to load up with the soft solder. I have two round chainsaw files that work nicely, and one of them was used for most of the work.

A pair of outlet fittings was next machined for the bottom ends of the header. Each of these two piece fittings included a hose barb that was permanently threaded into a lathe-turned 45 degree elbow which, in turn, was joined to the header tube with 620 Loctite.

The flexible hose couplers would not, by themselves, provide adequate support for the header tank against the engine's vibration. And so, two band clamps were formed from .010" stainless shim stock and added to the lower ends of the header. These bands secure the header to bolts already present on the front of the prop drive cover.

The color of the Ag/Sn solder wasn't a perfect match to the stainless steel and was something of an annoyance for an assembly that was going to be in full view at the top of the engine. While trying to decide whether or not to paint the header, I noticed a slight shadow under one of the front fillets that I thought had been previously polished out. This made me suspicious of my soldering that, up to that point, I had been so happy with. So, I decided to pressure test the header. The disappointing result was a very slight leak at the edge of the fillet. This indicated that the solder had evidently not wetted the seam which was located a good eighth inch under beneath the fillet, and so there was likely contamination as well.

Merely re-heating the joint would probably not have been a reliable fix because there was no way to clean and re-flux the affected area. So, I unsoldered the whole assembly, ground away the weld tacks, and separated the parts so I could start over. After removing all traces of the soft solder, I brazed the pieces together using a gap-filling (35% Ag, 26% Cu, 21% Zn, 18% Cd) brazing alloy available from McMaster Carr. Large thick fillets can be obtained with this particular alloy, and they flow out more controllably doing away with the need for soft solder 'buttering.' The final metal finishing was made more difficult by the harder filler, but the assembly didn't leak when It was completed.

The now yellowish fillets left no doubt about whether to paint the header, and so I used the remainder of the matte black Gun-Kote purchased earlier for the valve covers. The full-scale header tanks were typically painted white in order to reduce the absorption of the engine heat surrounding them. In my case, the coolant will likely be hotter than the immediate surrounding area, and so the black paint seemed more appropriate. - Terry

 

[ddmckee54]

Judging from the amount of work that you put into your "simple" coolant header I'm guessing that your last project for this engine will be an in-flight variable pitch prop with forged prop blades, maybe even de-icing boots. Just funnin' with ya Terry.

I have learned so much just from watching you build your engines and the testing you do to build them that it isn't even funny anymore. You just continue to amaze.

If you're going to build one of these beauties for John, how many orders do you need before we can get a group discount?

Seriously, how are you ever going to top this one with your next project?

Don

 

[DICKEYBIRD]

Seriously, how are you ever going to top this one with your next project?

I think if I were Terry, I'd probably retire from being retired for a while before starting something else!

 

[Barnbikes]

Judging from the amount of work that you put into your "simple" coolant header I'm guessing that your last project for this engine will be an in-flight variable pitch prop with forged prop blades, maybe even de-icing boots. Just funnin' with ya Terry.

Don

I thought he was going to build a 1/4 scale plane to put the engine in.

 

[DiegoVV]

Everytime I see posts from Terry or from Keith5700 I can´t help to envy their knowledge and craftmanship level. Every single piece they make deserve to be exposed in a museum. Go ahed with these builds!

 

[wirralcnc]

1/6th scale pratt whitney r1830. Combine the skills of machining castings and radial engines.

 

[kvom]

Learn something every time. I have a Cerrobend ingot but have never used it. The trick about oiling the inside of the tube before pouring in the metal is useful info.

Waiting anxiously for the first run video.

 

[cfellows]

Really incredible work, Terry. The fit and finish on that pipe-work is unbelievable.

Chuck

 

[mayhugh1]

The full-size Merlin's carburetion evolved continuously during the engine's development which seemed to span most of its life. Dual venturi updraft units manufactured by SU, Stromberg, and Bendix were used over time. In 1943, a form of throttle body injection replaced the Merlin's updraft carburettor, and metered fuel began being pumped directly into the supercharger.

Similar to the full-size Merlin, the Quarter Scale's supercharger inlet casting was designed to accommodate a (scaled) Stromberg carburetor that had been planned for the engine but never actually implemented. I've included a photo of an early carburettor mounted on the full-size Merlin's inlet casting for comparison with the Quarter Scale's casting. Since the Quarter Scale
Stromberg never materialized, the metering area on the front of the casting was available to builders for their own carburetors.

The Quarter Scale documentation included some discussion about selecting a carburetor for a scaled engine. The only concrete recommendation that came out of it, though, was a Honda GX120 unit manufactured for an industrial 4 hp single cylinder engine. This carb is still available, but its .456" diameter venturi feels a bit oversize. It wasn't at all clear whether one of these units had actually ever been tried on Dynamotive's prototype. The notes do mention that it had run reasonably well, although without a supercharger, using a commercial RC carb with a .350" diameter Venturi.

I had good luck with the Perry carburetor that I used on my 18 cylinder radial:
http://www.homemodelenginemachinist.com/showthread.php?t=21601&page=36
The model 1401 with its .312" venturi came up quickly and was not overly difficult to tune even on gasoline. The engine idles and transitions well, and the settings have remained stable over time. So, I decided to try a similar unit but with a .340" venturi on the Quarter Scale.

My plan was to adapt the Perry carb, running on gasoline with an alternate idle disk, to the metering area on the front of the supercharger inlet casting in such a way to retain continual access to all the carb's adjustments. It was also important that the carb be easily removable for experimenting with other size units since the Quarter Scale's huge induction system and supercharger are still big unknowns for me. Finally, the end result had to look like it belonged on the the rear of a Merlin rather than the front of an RC plane.

I discovered that a 1/2" solderable copper pipe elbow perfectly fit the standard .550" diameter o-ring'd neck of the Perry carb. This elbow became the smooth transition that I needed between the input of the supercharger and the output of the carb. The large venturi area in the starboard-side inlet passage was reduced and sealed to the elbow so the supercharger can draw only through the carb. A tapered aluminum reducer was machined to provide this seal as well as the transition between the output of the copper elbow and the input of the supercharger. The assembly was made permanent by backfilling the rear of the reducer and elbow with JB Weld.

The carburetor will draw its fuel from a bowl whose level will be regulated by a recirculating loop driven by an electric pump. A constant level fuel bowl will allow a lot of freedom later when selecting a location for the engine's fuel tank. It will also provide consistent fuel delivery to a large engine that will quickly consume a lot of fuel. I've used a similar scheme on the last four multi-cylinder engines I've built. A float-regulated bowl could work as well, but tiny floats can be difficult to get working reliably.

The fuel bowl was to be tucked inside the casting's port-side inlet passage adjacent to the carb. The throttle shaft had to be lengthened with a two inch extension in order to clear the bowl and to extend the throttle lever beyond the inlet casting for access from the rear of the engine.

The carb bowl was machined from a single block of brass, and then four inlet/outlet tubes were soldered to it. The most critical tube is the return outlet whose height establishes the level of fuel in the bowl. The fuel's surface tension interacts with the edge of the return tube to stabilize the level somewhat above the end of the tube, and this can become an annoyance in a limited capacity bowl. I've attempted to solve this numerous times in the past with the design of the return tube, but I usually end up shortening the tube anyway. Arbitrarily reducing the fuel level can create problems in a small bowl since not only is the bowl's capacity reduced, but the turbulences generated near the inlet and outlet tube can aerate the fuel before it reaches the carb. The carb's inlet should be located at the bottom of the bowl and as far as possible away from these tubes.

This time I tried rolling over the end of the return tube to form a thin-edge bell-mouth, but I still ended up having to lower the height of the tube more than I had hoped. The bowl's inlet tube is designed to supply its fuel to the rear of the bowl and away from the carb inlet and the bowl return. The carb inlet at the bottom of the bowl allows the bowl be easily emptied for storage.

The goal was to stabilize the fuel level at a quarter inch below the carb's spray bar, and this was verified by setting up a temporary alcohol recirculating loop. The fuel pump, commonly available in RC hobby stores, is the same one used on my other engines. Its components will later be re-packaged in an aluminum machined housing. This pump is a high volume 6V-12V unit designed to quickly fill the tank in a large RC plane. For my application I inserted a .022" diameter orifice in series with the input line to the carb to limit the flow rate and reduce the turbulence generated inside the bowl. For fine tuning, a 50 ohm rheostat will also be placed in series with its 6V power source. In operation, the pump is easily tuned by listening for a consistent drone before the engine is started. This particular pump is designed and recommended by the manufacturer only for methanol, but as far as I can tell it contains no components chemically incompatible with gasoline. I've run four gasoline engines using these re-packaged pumps and have had no issues ... so far.

At this point, the carb and its fuel bowl are sitting, unsecured, in the metering area on the front of the supercharger inlet casting. The final step will to be to machine a cover or covers to secure the assembly and shield its components from the prop wash. - Terry

 

[Ghosty]

Terry,
Got to love copper plumbing fittings. Have used them for a lot of things over the years. Still love the work, and will keep looking.

Cheers
Andrew

 

[dsage]

Great work as usual Terry. It looks like the fuel comes up very close to the top of the reservoir. I'm wondering if, once you put the top on the reservoir, capillary action will attract the fuel and form a meniscus with the lid and the side wall and interfere with the drain. Perhaps if the drain were more in the middle of the reservoir or the reservoir were much deeper than the normal fuel level this could be avoided. Tough to tell from pictures. Not sure what the lid looks like either maybe it adds to the height of the reservoir interior?

 

[mayhugh1]

Great work as usual Terry. It looks like the fuel comes up very close to the top of the reservoir. I'm wondering if, once you put the top on the reservoir, capillary action will attract the fuel and form a meniscus with the lid and the side wall and interfere with the drain. Perhaps if the drain were more in the middle of the reservoir or the reservoir were much deeper than the normal fuel level this could be avoided. Tough to tell from pictures. Not sure what the lid looks like either maybe it adds to the height of the reservoir interior?

Dave,
Once the gasket is added, it lifts the lid the thickness of the gasket. I should have mentioned that I ran the loop again with the lid assembled, and it seems to run OK. Because of space limitations I just didn't have much choice on where to place the tubes on this one. I probably should have butchered the casting and made the bowl bigger and more square, but I wasn't sure how well I'd be able to hide the crime scene. - Terry

 

[petertha]

Very nice Terry. Once you have a handle on running characteristics, I'm sure you can replicate the important dimensional guts of the Perry for a 'period look' of the outer body if you so choose.

Entertainment side note - I just stumbled on a Netflix 6-part series called Plane Resurrection. Part 1 is a P-51 Mustang, part-2 is a Hurricane. Amongst the rebuilding, some teaser shots of the engine & ancillaries. Gives some appreciation of engineering mechanics of that era. And the correspondingly huge work effort to those who restored them back to flying conditions.

 

[mayhugh1]

In order to wrap up work on the carburetor, I needed only to secure it to the engine. In homage to the Merlin's original designers, I turned a simple sheet metal bracket into a half dozen complex machined parts. I drew a line, though, at using the same number of fasteners they would have used to hold everything together.

The fuel bowl was secured to the metering area of the inlet casting with a pair of interlocking machined brackets that also form a bearing (of sorts) for the throttle shaft extension. Thanks to some poor planning, the fuel hose between the bowl and the carb ended up with a tight bend that could have collapsed over time, and so a spring was slipped inside the hose.

The carb cover was more involved because of the complex shape of the carburetor body. I wanted the carb's inlet to end up vertical and pointing downward, but I also wanted to be able to choke the carb by holding a finger over its intake. Another copper elbow fitting, o-ringed to the carb's inlet neck, solved both requirements nicely. The fitting was soft soldered to a machined brass top plate which, in turn, was bolted to a pair of endplates attached to the inlet casting. The covers were bead blasted and painted with flat gray Gun Kote which, after its oven cure, was impervious to gasoline. Other than the throttle shaft extension, the only other modification made to the carb was the machining of a new and longer-than-stock idle stop screw. Finally, a gasket was cut from rubberized automotive gasket material to seal the inlet casting to the bottom of the supercharger.

An important loose end that I tied up while waiting for the Gun Kote to arrive was the testing of the oil system including all the plumbing and the pressure relief valves. There was no practical way to include the engine's oil pump in this particular test, but it had already been thoroughly exercised in a previous test.

For a pressurized oil source I cobbled a 50 ml syringe into the input of the pressure relief valve, and I attached gages to the high and low pressure lines so the operating points of the relief valves could be set. The first test identified a fitting whose 1-72 mounting screws needed a bit more tightening. Before any relief valve adjustments were made, the high and low pressure lines each hard-pegged the 15 psi gages I was using. After setting the high pressure line to 15 psi and the low pressure line to 6 psi, I found obvious evidence of oil reaching all the crankshaft bearings as well as those in both camshafts. The output of the wheel case oil feed line was buried deep inside the engine by this time, but since the timing chain seemed to be wet with oil I assumed it was working. The oil feed to the prop gear was also hidden, but there seemed to be oil returning from it to the bottom of the crankcase. I was happy to see the pressure readings responding to the adjustments as expected and to see oil flowing in the low pressure tank return line. Up until this point I wasn't sure that I fully understood how the tandem relief valves were going to work. - Terry

 

[dalem9]

Truly amazing . Great work .