(CAD) 28 Cyl P&W R-4360

[Rayanth]

Well, I promised (threatened?) that I would start a WIP for my probably long-running design process for a scale Pratt & Whitney R-4360.

For anyone curious, the R-4360 is considered the largest piston engine to ever go into production for airplanes. It was used on a variety of aircraft at and after the end of World War II, and while thousands were manufactured, most were scrapped when the military no longer needed them due to the introduction of the jet age. Few remain, and fewer still are still flying. Most are in museums as static displays, many were even cut away to show the interior workings.

In the past couple of years I gained a sort of passionate love for this 28 cylinder radial engine that my girlfriend is probably a little jealous of. I have done a lot of research on the engine, and am always doing more. I feel the best way to show this beast a proper respect is to try and develop a working scale model.

Countless design decisions will be forthcoming, and I welcome input from all corners of the forum on anything that sounds or looks questionable. While I intend to build this behemoth of a model, I currently have no shop, no machines, and very rusty skills. If something looks like it will be impossible to machine, please don't hesitate to speak up.

The full-scale R-4360 sports some utterly astounding statistics. 4,363 cubic inches of displacement, 28 cylinders arranged in 4 rows of 7 cylinders, with 4 magnetos driving them. It 'sipped' upwards of 100 gallons of 108 octane fuel an hour (try that with today's prices!) all metered from a single pressure carburetor. Each cylinder sported a 5.75 inch bore with a 6 inch stroke, giving a 6.7:1 compression ratio and fired two spark plugs to ensure the cavernous compression chamber ignited evenly and fully. Intake ports were mounted on the top of the chamber, with exhaust out the side (analysis done since have suggested it would have been far more efficient to switch these two, and would have solved some of the intricate heat problems they had). This entire machine was air-cooled, requiring an astonishing amount of air to flow through the cooling fins, and indeed a lot of designs were tried to get the optimal cooling. The 3000-3500 horses were caged in a package that measured 55 inches in diameter and 96.5 inches long, weighing 3,870 pounds dry.

In military applications, the Wasp Major as it was known, was used in a variety of aircraft, from experimental to production. A list of aircraft that used this engine can be found on the following wikipedia article: http://en.wikipedia.org/wiki/Pratt_%26_Whitney_R-4360_Wasp_Major

Perhaps most notably was the F2G "Super" Corsair (Not the much more popular F4U Corsair, but very similar). Postwar, Several F2G's were converted into racers that dominated the air races until a tragic accident shut down the Unlimited Class until the 1960's. Since then, one Super Corsair named the Dreadnought took away numerous finishes, until it crashed in the mid 1980's. I believe Dreadnought is nearing rebuild, but that information may be inaccurate.

My goals with this design are to create a 1/6th scale version of the Wasp Major, though many design concessions will have to be made to ensure that it will be a running model. I will attempt to document the process of development, and as I am using a CAD program it should be easy for me to go back and make adjustments as required (provided I follow proper CAD procedures, something I often fail to do)

Again, if anyone sees anything questionable, has input, or any questions in general, please don't hesitate to jump in and help me. I apologize for this being a CAD work (for now!) and not a real machine, but some time down the road I will start another WIP when I have sufficient tools and skills to give the real thing a try.

Cheers,
- Ryan

 

[dmac]

They are an impressive engine.

 

[dsquire]

Ryan

Congratulations on the start of what is a very ambitious project. Myself and many others will be watching your progress. You have got the hardest part of the project done already, you have it started. :bow: :bow:

Cheers

Don

 

[stevehuckss396]

I got my chair pulled up ready for this one.

 

[steamer]

I got my chair pulled up ready for this one.

Me Too! Blaze a way Rayanth!...Looking forward to the drawing!

Dave

 

[mu38&Bg#]

Very cool. I always wonder what this sort of project uses for reference. Do you have some drawings you're using, access to an original, or is it going to be a semi scale sort of thing.

 

[b.lindsey]

The drawings alone are an ambitious project...I'm definitely along for the ride!! Should be a great journey too.

Bill

 

[Tin Falcon]

They are some engines I remember many years ago hearing a ww II bomber flying over the house at night. leaving after an air show. I will never forget th sound of the four radial multi stack engines what a drone.
Probably wright cyclones but what a sound.
Tin

 

[Rayanth]

The majority of my references come from a book by Graham White, with whom I am also in email contact on occasion, titled 'R-4360 Pratt & Whitney's Major Miracle.' This book is published by Specialty Press, and is no longer in active print. The hard cover is very hard to find, and as a result is quite expensive when you do find it, but I was fortunate to stumble on a softcover copy on Barnes and Noble's website for $75.

I also have a copy of the service manual for the engine, and a list of patent numbers related to it, which I use a free patent viewing website to download in PDF form and view.

Lastly, I do volunteer work at the restoration center for the Seattle Museum of Flight, and in one of our hangars we have an R-4360 sitting on a handling dolly in the back corner. It seems we'll never use it, or ever refurbish it, and I'm not permitted to take it apart, but I am more than welcome to take measurements of the outside and any photographs I want. I'm going to ask if I can take off one of the exhaust cooling manifolds so I can measure some of the stuff under it, I might be able to convince them to let me do that.

Starting the project wasn't very hard for me... in fact, I could say I haven't actually started yet. The problem I'm facing is WHERE to start. This is an astonishingly complex project, and internal combustion engines in general have everything relying on everything else. I know certain dimensions, from which using standard I.C. formula I can extract other dimensions, but ultimately it boils down to finding a spot and starting there.

I've been turning this issue over in my head for the better part of a week. At one point I decided to do a mockup of the crankcase in my CAD program (Autodesk Inventor 2011), just to make sure I could get the polygons drawn to proper angles. It was far easier than I anticipated - one of the best features of this program is that you can plug in an equation as a dimension, and it will use the result.

Note, the holes 'drilled' in the crankcase are obviously not the full size for the piston. they were only put there to locate the cylinders easier.
Also note that I have the entire engine offset by a rotational factor. As seen, row 'C' is facing directly up, when in fact row 'A' (the aftmost) should be. I was in a hurry.

But a crankcase's dimensions are determined by the size of the components inside of it. So I can't just start at the crankcase, if I'm going to maintain a 1:6 scale. I have the known dimensions of my bore and stroke, 5.75 bore and 6 inch stroke, scaled, are 0.9583~ and 1.0 inches, respectively. From that, I can infer many things about the crankshaft, and ultimately, a radial engine is built around the crankshaft. So a crankshaft is where I will start... problem is, not a great deal of info is readily available about the crankshaft as far as dimensions. From the patent and Graham White's wonderful book, I know how far the offsets are between the master rod journals on each row, but the rest it seems will be up to me.

 

[Rayanth]

Crankshaft, Revision 0

After reading all of the fun language that is a patent, two sayings came to mind. One, I've heard a few times in the various places I've been reading about machining engines. The other is a fairly popular one. Nike's "Just Do It", and "Start with a chunk of metal, and machine off anything that doesn't look like a [insert part, in this case crankshaft]"

A Crankshaft is effectively a long cylinder, with additional cylinder offset from its center line by 1/2 the length of the stroke. Where these offsets are, the main cylinder is cut away, to permit the rod for that crank throw to pass through the crankshaft's center line (If it couldn't, your engine wouldn't run very long. Half a turn at most.)

I have no information to go by for how thick the rods are. But what seems like a good guess to me, is about half of the diameter of the piston. My pistons are 0.9583 inches bore, so I'll use 0.479... naw, we'll just make it a half inch thick. I can always drop it to 3/8's later if I need to. If my rods are a half inch thick then by necessity my master rod journals need to be a half inch long, perhaps slightly longer. I'll add 1/32th of an inch to either side as a sort of buffer. Now, the patent drawing for the sideview of the crankshaft (patent is no. 2,426,879, I'm referring to Figure 21) shows lots of fine details, but being a patent, doesn't have dimensions. One notable thing is that the centerline bearing journals are bigger than the master rod journals, thus the 'long cylinder in the middle' is bigger than the master rod pin diameter. I can infer from a scale, assuming the patent drawing is to scale, that the diameter at the bearing is about 35% larger than that of the master rod journal.

... I talk a lot.

Anyway, while measuring that tidbit I found that at the scale I happened to print this image, the width of the crankshaft master rod journal pins is actually a half inch. this is a bizarre coincidence, but I'll take what I'm given! I can use this to get some other basic measurements with ease.

Anyone who saw my original post in the Questions thread regarding split gears will understand that I'm going to likely run into issues if I make this a single piece crankshaft. For now, I will design it as single piece, because it will give me a better visual idea of how I might be able to split it for a multi piece.

A basic layup of all those 'cylinders':

Obviously something's missing - there's nothing nearly strong enough actually holding the m/r journals in place but a thin piece of metal. I did warn that it was a basic layup. Of note in the second image is that each row is offset by 180 degrees, plus the offset between cylinders, of 12+6/7 degrees. this works out to 192.85 and change degrees, which allows each bank to fire on the opposite side of the bank before it, thus evening out vibration. This will make more sense when we get to firing order much later. (A preview of the firing order is visible at http://moozorzica.com/engines/4360order.jpg)

After some basic touchup, we have something that looks a bit more like a basic crankshaft :

This seems workable for now, so I'll leave it at this. Being as this is the first part, it's subject to great changes... hopefully not too badly, though.

I tried to keep the design minimal. I could have done the interconnects in that crazy half-round style they like to do, but this method will be way easier to machine, and will also lend itself to breaking down into multipart easier.

Based on this, I have : master rod thickness: 0.5", master rod journal diameter 0.435", master bearing diameters 0.5", 1/32" buffer on either end of master rod journals, all appropriate angular offsets between rows, etc. The overall crankshaft ends up being 7.200" long in its present design.

Next up: master rods. These are two-piece designs, that clamp together around the master rod bearing (which are also split bearings but I am leaning towards a multipart crankshaft, in such a way that i can slip the m/r journal bearings in one piece, way cheaper that route) and will require attach points for the other 6 slave rods, in each row. The Wasp Major has two of these actually passing through the bolts that hold the m/r halves together, which I'm sure I can manage, but will require custom made bolts.

This is enough for tonight. Thoughts on whether this would be machinable as one piece? a lot of pieces? Think I'd just end up making a bunch of scrap pieces in the process? I'd love to hear from you.

Cheers,
- Ryan

 

[ozzie46]

This is one very ambitious project! :bow:

Can't help with your questions as you are way outside my comfort zone, but know that I'll be watching with great interest.


Ron

 

[gbritnell]

Ryan,
I have been reading this thread from the beginning and would like to make a comment or two. First, although I applaud your desire to build this thing you have no idea what is involved. Reading posts by other fellows who have built single row radials will give you an idea of the work involved in producing such and engine. Now multiply that by 28 and even for machinists with years of experience it would become a daunting task.
Having designed several engines in my time I can tell you that there is much more to it than coming up with numbers. There's so many problems to be solved when miniaturizing that these issues alone can seem sometimes unsolvable.
The second part of my review is that of the actual machining. As you have found out by your initial postings about the gear mountings and how to split or not split them, for someone as you say with 'rusty skills' this should not be taken lightly.
I have somewhere upwards of 2500 hours in producing my V-8 engine and I would think that Steve has well over 1000 in just one of his so to build something like this my biggest concern would be the dedication to see it through.
You asked for any and all suggestions when you first posted and my advice to you would be, if you really enjoy aircraft radial engines why not design your own 5 or 7 cylinder engine? To design and build even one of these would be a great accomplishment.
Just my two cents worth.
gbritnell

 

[metalmad]

u better bet ill be watching :bow:
pete

 

[Admiral_dk]

I will certainly like to see you complete it, but can't help thinking that you ought to do it as a three step approach.

First a single cylinder IC engine with the same design and measurements of the parts as on the multi, as a "fast" way to confirm that you got all your calculations an the design right - it sure beats changing one cylinder head as opposed to 28 if you don't catch an error or the like before trying to run it.

Next step, make the 7 cylinder single row Wasp, again same design parameters. When this is a success, increase the number of rows as much as you desire.

 

[steamer]

I would second Georges comments. It's hard to grasp skill level on a forum., but this engine will take a lot of it and time to boot!

....but then again, so does building a lot of projects.....I guess it depends on who's up for the challange and how stuborn you are. ;D

As Aeromotor's post states, Lee's engines are very well designed and documented. If I were you, I would buy the 14 cylinder two row engine DRAWINGS and have a good read. That would, in my book, constitute an investment in your education....and a very worth while one at that. You don't live long enough to make all your own mistakes....learn from others......Dave

 

[Rayanth]

I appreciate the comments, folks. I really thought I'd made something clear, but I will try again.

I don't intend to jump into this as my first machining project. I don't expect to even have the CAD work done before I get my shop set up. I know this is ambitious, but I can' keep myself motivated unless I have something special to strive for.

Building this machine is certainly going to be nearly as daunting as designing it at scale. I understand it's not 'just numbers' but a great deal of it IS the numbers. My CAD program is incredibly powerful, and will give me the rest - i can run all of the simulations on it to determine if my numbers work... and redesigning a part in Inventor isn't much more than simply changing the numbers - it does the rest. For example, critical dimensions that might change, have been given a custom label. All other dimensions on ANY part that relies on that dimension uses the label in the formula. So, down the road if I find that my crankshaft's main diameter is too big, All I have to do is change the initial value for it, and everything else that relies on it, the openings in the crankcase, the bearings, the gears, will all change for me - and those that are 'purchased parts' will kick back an error if the new dimension isn't available.

My goal is to design this, while I gain my machining skills. I will be changing a lot as my skills evolve and I learn what can and can't be done. I will certainly build a few other engines before I tackle this one, and yes, 28 cylinders is a great deal of complexity. I am already working on designing a test rig for each cylinder. after each cylinder and piston is built, it'll be tested on its own as a one cylinder engine, before adding it to the 28 cylinder behemoth.

I'm even considering adapting plans for smaller radials to use the cylinders I come up with for the 4360 - this would let me build a few at a time, and if nothing in the design changes, use them later.

Re: Lee Hodgson - In a sense, he gave me the idea! I visited him personally at his shop in Cincinnati and saw his machines first hand. His father had begun work on designing the 4360, but ran into certain complexities and never finished the project. Those complexities were mostly in the very strange angles and offsets that are very hard to obtain - but with CNC, they should be quite doable. I respect Lee and all of his beautiful engines, and do fully expect to be building either the 14 or 18 double row from him before I start making chips for my 28.

I'm sorry to say it, but I won't be dissuaded! I have (hopefully, I'm only 31) a lot of years left to tweak and fine tune details if I must, but this machine will see life some day. I would hope that you all have experienced a project for which you had such dedication, and help me out when I stumble, on this one.

- Ryan

 

[stevehuckss396]

I'm sorry to say it, but I won't be dissuaded!

You have about 50 years before the body starts to fail to the point where model building is no longer possible. You better get going!!!

I am looking forward to seeing progress and if you never build a single part, this will still be a great thread.

 

[Rayanth]

... It seems I woke up on the wrong side of the bed. (I was still IN bed when I wrote my last reply). I apologize if I was 'snarky' in my last reply. My life has been riddled with people telling me I can't do something, and nobody has ever told me I can. It seems everywhere I go, I meet with resistance. But I digress.

I am fully dedicated to making this thing run. It's been in my head for well over a year, and led me to many other radial engine designs that I have also toyed with in CAD while learning my CAD program as well as more about how the engines work. I've built the Edwards Radial 5 3 times in my CAD program now, making minor changes here and there so things worked better, and run all the simulations to ensure my changes didn't affect the running of the engine, etc. I have also done extensive reading on I.C. Engine design, and am fully familiar with the physics behind them.

As I did say earlier, I know better than to make this my first machining project. I agree, it's beyond my scope for now. I do intend to make the Edwards 5 (I haven't seen anything that makes me say "I can't figure out how to machine that") as one of my first projects, perhaps with a single-banger before that. Lee Hodgson's engines were suggested as a step up, and I've been considering those as well. I was leaning towards the 18 cylinder, but it seems the 14 cylinder will give me more information on the strange angles involved.

The biggest problem with the R4360 is that it's 4 rows. No other production radial engine is 4 rows. Others were developed, but never saw production. The added two rows add a ton of problems that can't be solved by gaining experience with one or two radials - even in a two row, the bearings and cams are on either end of the crankshaft... nothing has to worry about splitting bearings, splitting gears, and splitting the master rods (V8's might have to split the master rods, I'm not familiar with those - if anyone has a recommendation for a starter V8 I'd be glad to hear it). THAT is why I'm tackling the design phase now - the earlier I can identify the problems, the more time I have to figure out how to get around them.

Again, I apologize if I sounded snarky. This engine has become a dream of mine, and I was feeling rather defensive. I will see it through to the end, though it may take years. I will build my skills along the way, and probably make a million modifications to the design as I learn what will and won't work. I know it will be rough, but that's part of why it appeals to me - if anything, I am more determined to see something through, when I've been told "you can't" or "it's too hard" - I love to prove the naysayers wrong ;D

I've heard Bruce Satra may have made one but can't find anything on it, and only one other static, non-running model was made that I can find. This is entirely new waters, and I will be glad of any help you fine folks have.

 

[el gringo]

Off topic in a way but I have to mention the feeling transmitted through my body whilst sitting amongst four Major Wasps running @ 60 inches just before the A/C released the brakes for the takeoff roll. As I think of it now I feel it. I was also fortunate enough to run them up for ground maintenance as a crew member (maintenance scanner, waist blister) RB 50f

Hope this dosn't screw up the thread too much!

ray M

 

[Rayanth]

Ray-
not derailing my thread at all =) I love to hear the stories. I've heard it said that at a META power check in a maintenance hanger, if you were standing in the wrong spot you would be struggling just to breathe...one day I hope to experience the real thing, but until then, this is what I have =)

- Ryan