Quarter Scale Merlin V-12

[gbritnell]

Terry, BREATHTAKING!!!!!!!

 

[rdefrei1]

This is a "dream engine" what a beauty.

Congrats.

 

[dsage]

WOW.

That's a beautiful sight. So much so I had to go back several times to look at your assembly pictures and marvel at the fact that it looks so perfect. Here's hoping that it fires right up and you never have to open it up again. Especially after the procedure you went through to put it all together.

:thumbup:

Amazing work.

Sage

 

[mayhugh1]

Thanks, all ... -Terry

 

[kvom]

Here's hoping that it fires right up and you never have to open it up again.

The dream would be to cast all the exterior cases in transparent plastic to be able to see what's inside.

 

[kuhncw]

Looking at your assembly photos, it is easy to think I'm looking at a full scale engine build. Beautiful work, Terry!

Chuck

 

[Hopper]

Meticulous work, mate. I stand in awe, at both your skills and your patience!

 

[neubert1975]

Meticulous work, mate. I stand in awe, at both your skills and your patience!

what the gentleman just said :thumbup:

 

[mayhugh1]

With the wheel case and heads assembled to the crankcase, the timing chain and its cover could be finally assembled and installed. The Merlin's overhead cam drive system was greatly simplified for the Quarter Scale version, but it was designed more than a dozen years ago around a 3/16" steel roller chain that has since become obsolete. A year ago, I purchased an engineering sample of this chain that had been gathering dust in a Nordex salesperson's office. (Nordex.com was the recommended supplier for this chain in the Quarter Scale documentation.) I was told that this particular sample as well as a pair of connecting links were the last of their kind in their inventory since this chain had long ago been superseded by .1475" pitch products. The Quarter Scale's documentation mentions the .1475" chain as a possible alternative, but I wasn't sure that its fit had actually been verified in the cover design supplied in the documentation. So, I machined the timing sprockets in my engine around my piece of obsolete chain.

Just after machining the sprockets and chain cover components a year ago, I tested the fit of my chain in a partial assembly of the aft portion of the engine. At that time, I was mainly concerned with verifying the clearances of the chain inside its cover, and so I took a few liberties with some of the difficult accesses required for the trial assembly. I wasn't able to re-use them during final assembly, and so the chain and cover installation turned into a bit of an ordeal.

I spent the better parts of several days working out a sequence of steps for their installation. I kept notes along the way so I wouldn't have to re-derive the process should the engine ever need to be disassembled. I've included those steps in the next paragraphs even though I realize they will be of interest to only a couple readers who currently own the Merlin castings:

The retaining screws on the hubs of the port and starboard cam drive sprockets are first loosened so the sprockets can spin freely of their camshafts. The two idler sprockets are installed in the wheel case, but the central tensioner must initially be left out to get access to the crankshaft sprocket. The rear half of the main cover is installed on the top of the wheel case and temporarily retained with a single center mounting screw. The port-side head cover with its two cover tubes is then installed on the port head, and the lower ends of the tubes are inserted into the rear half of the main cover. The tubes are a snug fit in the head cover and, with care, will remain in place during during assembly.

A piece of stiff chord (I used waxed cable lacing chord) is then fed down through the left-most port-side cover tube and down over top the idler sprocket. Some deft manipulation with a long probe and tweezers will be required to get the chord routed around the crankshaft sprocket and up and out of the wheel case. Because of very limited clearance around the crankshaft sprocket, any knot tying the lacing chord to the chain will likely be too large to pass around the sprocket. So, a length of thin high strength thread (I used .010" diameter upholstery thread) is tied between the lacing chord and the chain. While rotating the crankshaft, the chain can be pulled down through the leftmost port cover tube, over the idler, around the crankshaft sprocket, and then up and slightly out of the wheel case. Again, some assistance from a long skinny probe will likely be required to get the chain properly started on the crankshaft sprocket. Once on the sprocket, though, the chain is pulled up about 1/4" above the wheel case chain cover mounting flange. The starboard-side head cover along with its two cover tubes can then be installed on the starboard head. The rest of the screws may then be inserted in the rear half of the main cover. Again, while rotating the crankshaft, the chain is pulled by the thread up and over the idler sprocket, through the right-most starboard-side cover tube, and then over and around the starboard cam drive sprocket.

The central tensioner sprocket can then be installed in the top of the wheel case. This has to be very carefully done to avoid dropping its parts down into the wheel case which will then have to disassembled to retrieve them. The chain may then be pulled completely around the starboard cam sprocket, down through the inside starboard cover tube, and around the bottom of the tensioner sprocket while rotating the crankshaft. The chain is then finally pulled up through the inside tube of the port-side cover and onto the port-side cam sprocket where it will meet up with the rear end of the chain.

The chain must now be shortened to its final length which is 116 links excluding the connector link. With the tensioner sprocket set to minimum tension and after a final check for tautness, the roller pin to be removed is marked with a Sharpie. With the engine well covered with a protective cloth, a Dremel tool can be used to grind away a peened end of the roller pin so it can be pressed out and the excess chain removed. The connector link is then installed while the two ends of the chain are held in alignment on the cam drive sprocket with a pair of adjacent sprocket teeth. The pins on the connector link are .030" longer than the pins on the chain. I was concerned about them rubbing against the inside of the chain cover, and so I ground them down to about .015". I don't know if there is a preferred orientation for the connector link's keeper that's dependent upon the chain's direction of travel, but I oriented mine as shown in one of the photos. Testing showed the chain and connecting link moved smoothly through the cover with no sign of rubbing - a result that continues to astound me. As I've said before, my hat's off to the designer of the chain cover. It's not only functional, but for a component that wasn't part of the engine's original design, it wound up asca nicely integrated sub-assembly. Finally, the front half of the main cover can be installed along with the remainder of the mounting screws.

With the timing chain installed, the valves could be timed to the crankshaft. First, the lash on all the rockers was adjusted to .004". The engine's cam card shows the intake valve beginning to open 10 degrees before its piston's TDC. The firing order is 1A-6B-4A-3B-2A-5B-6A-1B-3A-4B-5A-2B using the Merlin convention that 'A' is the starboard bank, and 'B' is the port bank. The crankshaft was rotated so that piston #1 (front-most piston) in the starboard bank was at TDC and the degree wheel (http://www.homemodelenginemachinist.com/showthread.php?t=24153&page=40) on the prop shaft was zero'd. The crank was then rotated back 30 degrees or so to remove backlash and then rotated forward to 10 degrees BTDC. Using a wrench to turn the milled hex on the front of the starboard cam, the cam was rotated until a D.I. indicated the #1 intake valve had opened .005". The retaining bolts on the starboard side cam hub were then tightened to lock in the starboard side valve timing. The crankshaft was then rotated 60 degrees until piston #6 (rear-most piston) in the port bank was at TDC, and then the process was repeated.

With the chain cover installed, the air/fuel connecting tube between the supercharger and the intake manifold could be machined and installed. This nearly one inch diameter tube runs through the center of the timing chain cover with very little clearance. It delivers the pressurized air/fuel mixture to the intake manifold from the supercharger. This tube wasn't machined earlier because its design depended upon measurements involving the finally assembled manifold, chain cover, and supercharger.

The front and rear ends of the tube were grooved, respectively, for -022 and -023 Viton o-rings. Its o.d. was turned for a chain cover clearance of .010". The height of the mounting flange of the connecting tube's elbow was finally machined in order to set the tube parallel to the axis of the engine. A thin linen paper gasket was cut to seal the elbow to the top of the supercharger. - Terry


Addendum ...
I've been doing more searching for the 3/16" pitch roller chain, and I may have found some in stock at:
http://www.powertransmissiondepot.com/artifact/3113516/ .
I ran across this website after reading a request for help on a pocket bike forum. It seems that this particular chain was used by a Japanese manufacturer in one of its pocket bikes, and an owner was trying to find a source for it so he could repair his bike. Another reader pointed him to this particular supplier who claims to currently have it In stock.

 

[Cogsy]

Looking fantastic Terry! If the chain rotates clockwise in the picture showing the connector clip then the clip is on backwards (at least that's what I was taught with motorcycle chains). It may not be important to you with your fully covered chain but with an exposed chain it's possible for a piece of debris to contact the clip in motion, putting force on the clip in the direction to remove it.

 

[mayhugh1]

Cogsy,
Thanks for the information. I forgot to draw a direction arrow on the photo, but the sprockets do turn counter-clockwise. - Terry

 

[mayhugh1]

The supercharger has been completed for some time but can only now be finally bolted onto the rear of the engine. It'll be held there with a couple dozen 1-72 SHCS's that will have to be inserted through the front side of the rear wheel case flange amid numerous obstacles. Most of the screws will have to be tightened a fractional turn at a time using a ground down hex key, and so this will be another assembly step that I prefer to do only once.

My original plan was to perform some bench tests on the supercharger before it was installed on the engine. There's an SAE standard that deals with such testing, but it's a heavyweight and well outside the scope of this little project. My primary goal was to stress the supercharger with several minutes of full-speed full-load running before installing it on the engine. However, with a few shade tree measurements I should also be able to gain some insight into its effect on the engine's performance.

To spin up the supercharger for testing I used the Nichibo 775-8511FDAS dc brush motor (http://www.homemodelenginemachinist.com/showthread.php?t=24153&page=43) that I purchased earlier. I found an inexpensive speed controller on eBay that was adequate for some rpm control. I machined a 32 pitch 48 tooth brass driving gear for the motor's shaft as well as a mounting plate to secure the motor to the supercharger's housing. The number of teeth on the drive gear was selected to spin the supercharger at twice the rpm of the motor in order to obtain a maximum 36k rpm. This rpm will match the speed of the supercharger running on the engine at a maximum crankshaft rpm of 3600 rpm.

For my first measurement, I sealed a 42 gallon (5.6 cubic ft) plastic trash bag to the end of the supercharger's outlet tube. After a 36k rpm spin-up, the bag was fully inflated in just under 10 seconds. Although this may seem more like a parlor trick than a meaningful measurement, it allowed me to estimate the supercharger's maximum flow rate at (5.6 cubic ft)/(10 sec) x (60 sec/min) = 33.6 cubic ft/min.

The engine's airflow requirement can be computed using CFM = engine displacement x rpm x volumetric efficiency / (1728x2). The 1728 converts the cubic inch displacement to cubic ft, and the 2 corrects the four stroke engine's rpm to account for its two revolutions per cycle. My engine's displacement is 21 cubic inches. For a maximum crankshaft rpm of 3600 and an estimated .8 VE, the engine's maximum air flow will be around 17.5 cubic ft/min. Since supercharger's flow rate is about twice the airflow requirement of the engine, the supercharger should at least not act an obstruction in the engine's induction system. The question, though, is whether the supercharger can increase the engine's volumetric efficiency by increasing the pressure in the intake manifold above atmospheric.

For the next measurement, I blocked off the supercharger's outlet tube with a rubber stopper. The stopper was through-pierced with a barb to which was connected an automotive-type boost pressure gage. At 36k rpm, the gage showed the supercharger building one psi of boost.

One psi above atmospheric isn't much, and as several with experience in this area have previously commented, 36k rpm isn't sufficient to build significant boost from a supercharger of this size. While playing with the test setup I felt like I could have easily achieved another pound of boost with another 10k rpm or so. However, I'm not willing to extend the engine's maximum rpm to 4600 rpm.

In any event, I managed to accumulate some ten minutes of running time on the unit which improved my confidence in its ability to hold up to the stresses that it will endure in the running engine. I expect any real effect on the engine's performance beyond improving the air/fuel distribution in its huge intake manifold will likely be marginal, but for talking purposes I can honestly say the supercharger does generate measurable boost. The supercharger was then finally installed on the engine. - Terry

 

[aonemarine]

Frick man, hurry up! Quit you job, work 23.5 hours a day or what ever...I need to hear this baby run!

 

[Ken I]

Flow and pressure curves for vane type charges show a fall off in pressure at stall - running it with a closed port might be doing it a disservice - it may do better dynamically.
All said its not going to be much but as you have observed, it won't be an impediment.
Can't wait for this to roar into life.

Regards,
Ken

 

[dsage]

A new series just came to Netflix (here in Canada anyway) called "Plane Resurrection". The first episode is on the Mustang (the British Merlin Engine version). There is a guy in England that took on restoring a Mustang and the Merlin Engine in it. Now he finds himself rebuilding a couple of engines per year and having done 30 or so, so far. Lots of interesting clips of work on his engine, old clips of original engines and comments about them. A wonderful episode.
Terry: Don't get in any dog fights with yours. Apparently since the cooling system is on the bottom, they were vulnerable to being damaged by enemy fire.

Sage

 

[Blogwitch]

I have watched those Sage.

The one that caught my eye was the battle of Britain Hurricane which had a very early Merlin fitted, cylinder block and heads, all one casting.

John

 

[dsage]

Yes indeed. I had only started watching the series while I posted about it before and was surprised that the second episode (the one you mention) was about another Merlin equipped aircraft. A very interesting and well made series (so far).

Thanks

Sage

 

[mayhugh1]

The only paint I have planned for the Quarter Scale is a matte black finish on its valve covers. For this I used an oven-cured product called Gun Kote which is a durable ultra-thin finish available from Brownells as well as others. After being cured for an hour at 325F, the finish is impervious to engine fluids, and there are several Youtube videos attesting to its durability when applied over a properly prepared surface. I've used this paint on custom motorcycle parts that have been in the weather for more than a dozen years and they still show few signs of wear. The only issue is its current cost. A 6 oz. rattle can is around $32 which is 300% more than I paid several years ago when I last bought what I think is the same product under the label 'Brownells' Baking Lacquer.' A less expensive paint could certainly have been used, but I was concerned about a thicker coat washing out the shallow 'Rolls Royce' badges cast into the sides of the covers.

I made a pair of gaskets for the valve covers as well as a gasket for the lower crankcase. For these I used 1/64" rubberized automotive sheet gasket rather than the shrink-prone brown fiber material used earlier to make the intake gaskets. Of course, these gaskets also required clearance holes for a truckload of 1-72 mounting screws. In order to get them all in the right locations I first set the valve cover down on top of the gasket. Using a length of pointed drill rod with a close fit to the covers' bolt holes, I transferred the locations of the holes onto the gasket as light dimples. A simple hole punch was made up from a short length of brass tubing by sharpening one of its ends with a small 45 degree countersink. After removing the valve cover from the gasket, the punch could be precisely centered over each dimple with the help of a second piece of pointed drill rod slipped inside the tube punch. Once in position, the rod was withdrawn and the punch given a couple light taps with a small hammer. With the gasket backed-up with a sheet of thin cardboard between it and an anvil, the tiny holes came out perfectly punched. The punch had to be resharpened every dozen holes or so, but two of them lasted through the hundred-plus holes that I ended up punching in the three gaskets.

The magnetos (actually, the distributors) can't yet be installed because once they are in place they will complicate the installation of some of the engine's oil lines. These lines will emanate from a pressure relief valve housing that's yet to be installed on the starboard side of the crankcase. The next step, therefore, was to machine this assembly so the engine's oiling system can be first completed.

The valve housing, which actually contains a couple functions, is the last component to be machined from the Quarter Scale's documentation. I spent several days studying its drawing and trying to decipher its functionality, but I wasn't able to make my peace with it until after I had actually completed its machining. Then, I had to make some modifications to it to correct for my misunderstandings of its operation.

Part of my initial confusion was caused by the mounting hole pattern that I had already machined into the crankcase. I had faced, drilled, and tapped a number of mounting hole bosses that were originally cast into my particular crankcase; but evidently it had been designed for an earlier version of the valve housing. The housing in my drawing requires only two mounting holes, and its backside was intended to be machined to fit the complex contour of the crankcase. I began by modifying the design of the valve housing to match my already machined crankcase. A much bigger problem, though, was that I didn't appreciate what the designer of the valve housing was trying to accomplish, nor did I completely understand the interaction between the valves inside the housing.

The housing contains two valves which, in the documentation, are labeled as high and low pressure relief valves. Pressure regulation is a good idea in an lubrication system using a constant displacement oil pump because of the extremely high pressures that can be generated (80-100 psi in my radials, for example). And, it's almost mandatory in a dry sump system in order to prevent the pressure pump from getting too far ahead of the scavenger pump. It's easy to provide a tank of froth-free oil for the pressure pump to draw from, but the scavenger pump is left to fend for itself in the engine's sump. In a model engine it spends a lot of its time sucking air and re-priming until the engine's sump is filled with waste oil that has drained back from the engine's upper end. The only way the scavenger pump is then able to catch up is if its pumping capacity is higher than that of the pressure pump. Fortunately, the Quarter Scale's pump designs try to provide for this.

Relief valves work by blocking off oil flow using a spring-loaded plunger sitting in a seat. When the oil pressure on the plunger exceeds the spring pressure, the plunger is lifted, and a portion of the oil is allowed to escape through a 'return to tank' line. This action creates an adjustable pressurized source of oil that can be used to lubricate a portion of the engine; and the excess is simply returned to the system's oil tank.

The Quarter Scale's valve housing, however, contains two pressure relief valves in series. A high pressure relief valve supplies oil to the crankshaft, but its 'return to tank' line is actualły the input to the second valve. This second valve is a low pressure relief valve that supplies oil to the rest of the engine. In the stock design, its 'return to tank' line is actually a return to the engine's sump through a passage in the rear of the valve housing. This creates a potential problem because it means that the entire output of the pressure pump, including the 'tank' returns of both relief valves, ends up inside the engine where it can overwhelm the scavenger pump if the pressures and flows are not precisely balanced.

John Ramm mentioned that the sump in his engine routinely over-filled causing excess oil to be spit out the engine's exhausts. He improved the situation by adding a third pressure relief valve with a true 'return to tank' just ahead of the stock high pressure relief valve. Feedback concerning this issue may never have made it back to Dynamotive for an ECN because John may be the only builder who only recently got one of these engines to run long enough to see it. I've attempted to solve the issue, however, by modifying the design of the stock valve housing so the low pressure 'return to tank' line realły does return its excess oil to the tank rather than dumping it into the engine's sump.

The Quarter Scale's high pressure relief valve also happens to function as an anti-drain-back valve by preventing oil from draining back out of the engine when it's shut down. Before I understood the interaction between the two relief valves, I thought this was its primary purpose; and I mis-labeled it as such on the housing that I machined. I added a gage port to the housing so the low pressure relief valve would not have to be blindly adjusted. When I machined the pressure pump long ago, I included a gage port on its body which will be used to set the high pressure relief valve.

The next step will be a tedious one that includes making up and routing the various external oil lines. I've not been looking forward to this because I can already tell it's going to be one of those tasks that I'm never quite satisfied with. I've bought a lot of miniature copper tubing to play with, but I'm sure most of it will end up as scrap. I've gotten a head start on the easiest one - the high pressure line feeding the crankshaft. It required only three tries to get a satisfactory result, but I still have to machine and solder a mounting flange on its far end. -Terry

 

[Buchanan]

Absolutely fantastic! May I ask, with the greatest respect , Will you try to make the oil control block look like a casting. It would blend in with the the lost wax castings of the engine so much better .

 

[mayhugh1]

Hi Buchanan,
I agree with you that it would look better if it were blended into the crankcase to look like part of the casting. I added it as a bolt-on, though, because that was the the way it was handled on the full-size engines. Maybe it should be painted, though. - Terry