Edwards Radial 5 build thread --- PHOTOS!

JRaut

H-M Supporter - Diamond Member
H-M Lifetime Diamond Member
Joined
Mar 9, 2018
Messages
465
Episode 1 || Introduction to My Build Thread


Background

After a brief search through the H-M.com archives, I was a bit surprised that nobody had documented their journey building an Edwards Radial 5-cylinder engine. These engines are reasonably complex and time consuming to construct, but as (1) detailed drawings are freely available and (2) it's a 'billet engine' requiring no castings, it tends to be a project commonly tackled by hobbyists like us (here's one nice example). There are several build threads for the Edwards in other forums, but apparently not here. UNTIL NOW!

The engine is a 5-cylinder radial designed and developed by Forest Edwards in the late 1980s. Edwards built about 30 of them for use in large-ish scale model airplanes. Robert Sigler drew up a set of plans (39 sheets total) for the engine in the early 2000s, which will form the basis of my build.

Overview of Engine
The Edwards is a methanol/nitromethane-fueled, glow-plug ignition, 5-cylinder radial engine with an overall diameter of about 10" and total displacement of 3.4 cubic inches (56cc). Running optimally, it reportedly spins a 22" propeller something like 6,000 rpm*. It packs a serious punch! The Edwards relies on a dry sump lubrication system to pump castor oil to the moving bits. The valve train in comprised of an intake and exhaust valve in each head controlled by pushrods activated by a 'cam ring' running concentric with, and geared to, the crankshaft.

Intent of this Build Thread
I hope that this build thread will serve several purposes:
  • First and foremost, I hope that the mere fact of starting it will keep me motivated throughout the process, and help ensure I end up finishing it.
  • Second, everyone on H-M likes a good build thread chock-full of photos, both for the entertainment value and for helping others undertaking a similar project; I intend to include lots and lots of photos along the way.
  • Finally, one thing that I wish other build threads had is an accounting of how much darn time everything takes. Is it just me, or do even seemingly simple components take a very long time to make, particularly if there's any special setup requirements? I'm reasonably new to the hobby, so it could be that I'll get faster as I go, but it's hard to believe I'll get much faster. Anyway, I intend to keep rough track of how much time each component takes, and a running total of how long the whole build is taking.

Side notes:
(1) I started this build about a month ago, so several components are already done. Accordingly, my first few posts after this one will be in pretty rapid succession, as I have time to write them up. Things will slow down precipitously after that.
(2) I have a full-time job, a 2.5-year-old son, a 1-month-old daughter, and a honey-do list a mile long. Shop time is hard to come by. Bear with me if there are stretches of inactivity...

*Edit: I originally said that this motor can produce 6HP based on something I recall seeing online. I can no longer find reference to that stat anywhere. I don't think I made it up, but perhaps I did. Let me know if you see the peak power referenced anywhere online by a reputable source.

PK2a.jpg


Drawing snip 01.png

Drawing snip 02.png
 
Last edited:
Wow, this engine looks really cool! I cannot wait to see how it turns out!
 
Popcorn ready!!! Really looking forward to this.
 
Episode 2 || Autodesk Fusion Model

The first thing I like to do when starting a project (particularly complex projects, but usually the same holds for simple projects as well) is to create a computer model of what I intend to make. And my modeling software of choice is Autodesk Fusion 360. Because let’s face it: Fusion 360 is awesome!

I won’t dwell on exactly why Fusion is so damn awesome, but here are a few top-of-mind reasons:
  1. It’s incredibly user friendly and lots of people use it. That combination means there are tons of tutorials online for folks just learning the ropes.
  2. It’s FREE to home gamers (not going to debate whether it will always be free).
  3. The free version available to us is FULL FEATURED, meaning that we get all the bells and whistles that someone paying full retail price for it gets.
  4. Not only can you create a model with Fusion, but you can also use it as a springboard for 3D printing and CAM routines (neither of which I’ve utilized to date).
  5. You can import parts directly from McMaster. Insanely useful for bearings and fasteners and the like.

I modeled just about every part, save for most of the fasteners and the propeller blade. And not just that, I used the joints feature within Fusion to make everything mechanically linked much the same way it will be in real life. Meaning: if I click/hold/drag the crankshaft, all pistons and other moving bits move to the right place at the right time*.

While the available CAD drawings for this engine are superb, having this 3D model available to me as I’m making parts is invaluable. I can quickly query any dimension/distance I so please, can evaluate clearances and look for interferences almost instantly. And if I modify a part while machining it (whether on purpose of accident…) I can hop into the model and see how it effects things.

Anyway, not sure what else to talk about here, so I’ll just toss in a bunch of screenshots of the Fusion model.


TIME ON FUSION MODEL: 21 hours
CUMULATIVE TIME: 21 hours



*The joints feature in Fusion allows a user to directly couple relative movements (rotations, displacements) between components, and allows for things like gear ratios to be incorporated. What joints do not allow, however, are any sorts of nonlinear motion relationships. So, for example, cams pushing on pushrods cannot be incorporated with the joints feature because the camming action is a nonlinear function. Accordingly, the valve train in my model is not ever set into motion. Bummer.

There is an alternative within Fusion that makes this (sort of) possible, and that’s by using Contact. My luck using Contact on complex models has been limited, mostly (I think) due to Fusion taking too long to calculate everything when contact sets go into contact. But you should give it a try for yourself if you’re interested, as it may be a computer limitation rather than a Fusion limitation.

While I’ve never used it, I understand that Autodesk Inventor is like Fusion on steroids, and has much more capabilities for joints, including cams and whatnot. But it ain't free, so I ain't using it.

Fusion_01.PNG
Fusion_02.PNG
Fusion_03.PNG
Fusion_04.PNG
Fusion_05.PNG
Fusion_06.PNG
Fusion_07.PNG
Fusion_08.PNG
Fusion_09.PNG
 
Episode 3 || Crankcase (Attempt 1)

I figured I’d start out with the Crankcase of the Radial 5, since it forms the backbone of the whole motor. It’s quite a complex part to make, requiring several different setups on both the lathe and the mill with a dividing head. And about 60 holes, many of which are tapped #2, #4, or #6.
Fusion_01.PNG

Fusion_02.PNG Fusion_03.PNG


I started out by chucking up a chunk of 5”-diameter stock in my 4-jaw chuck. A 4”-diameter piece would have worked, but I had some 5” on hand. So it took quite a bit of roughing to get it down to size. Note that there’s a fair amount overhang; that’s because I’m trying to get the crankcase, cam housing, and back plate all out of this one piece and barely have enough length. Would have wasted too much chopping it into three separate chunks.
IMG_6959.JPG IMG_6960.JPG

You can see in the fusion model that there is an ‘undercut’ deep within the inside of the housing. It proved rather difficult to remove this material, and I had to grind up a trepanning tool to do so. I’ve never ground a trepanning tool, and this one had to have quite some overhang to get deep down in there. It ain’t pretty, but it worked. Sorry for the photo quality, was hard to get a good photo.
IMG_6973.JPG IMG_6970.JPG

Once I finished roughing everything out, I gave it some time to cool down, then finish turned everything to final size. The central bore accepts a deep groove ball bearing so had to be pretty precise. Lots of chips were made over on the lathe!
IMG_6978.JPG

The next step was to get the finish-turned crankcase mounted up to my dividing head. That wasn’t the easiest process. In an ideal world, I would have parted off the crankcase and bolted it straight to a H/V rotary table / dividing head mounted to my mill. I don’t have one of those. And because of the geometry of the crankcase, I didn’t really see a good way to grab onto it over at the dividing head, so I opted to leave it in the 4-jaw chuck in an attempt to maintain concentricity. My lathe spindle doesn’t match my dividing head spindle, so I had to make an adapter. All that added up to a tremendous amount of stick-out. Not even close to ideal…
IMG_6987.JPG IMG_6990.JPG

In an attempt to control runout and deflection while milling the flats, I set up a clamping system and indicated in each position. It was a huge hassle, but I got close enough (within a few thou) in every setup.
IMG_6994.JPG IMG_6997.JPG

Once I had the 10 flats milled to dimension, I drilled, bored, tapped what felt like a million holes.
IMG_6998.JPG IMG_7003.JPG

BUT!!!
I ended up breaking off a #4-40 tap in the second-to-last hole that required tapping. Huge bummer.
More on that in the next post. I reckoned I’d be able to get it out, so I kept on with things.
IMG_7037.JPG

After mounting back on the lathe and parting off, I had just a few more drills / counterbores and it was finished (except for removing the broken tap).
IMG_7005.JPG IMG_7023.JPG

There were a few issues with the part (I accidentally oversized one hole; broken tap issue, some non-concentric features, etc.), but those were generally on non-critical features. Overall I was pretty proud of what I had accomplished. It was a very complex part and I had knocked it out with pretty darn good success.

Next episode will be on (trying unsuccessfully) to remove the broken tap.


TIME ON CRANKCASE: 22 hours
CUMULATIVE TIME: 48 hours
 
Last edited:
This is a seriously impressive project.

Finally, one thing that I wish other build threads had is an accounting of how much darn time everything takes. Is it just me, or do even seemingly simple components take a very long time to make, particularly if there's any special setup requirements? I'm reasonably new to the hobby, so it could be that I'll get faster as I go, but it's hard to believe I'll get much faster. Anyway, I intend to keep rough track of how much time each component takes, and a running total of how long the whole build is taking.

It is not just you. I think we all watch sped up YouTube videos of people machining stuff or CNCs cranking out tons of parts, but doing it yourself is a labor intensive process. I've been building a steam engine which is much simpler than this project and recording the time for each part. What I have found is I spend 60-75% of my time doing fixuring, clamping, and setup. Some parts require 5 or more setups and have taken me 10, 15, or sometimes 20 hours. I also find that I am pretty likely to make at least one piece of scrap no matter how hard I try to be perfect. Like you said, much of this is non-critical, but I am making it for a friend so I am being extra tough on myself over mistakes.

Please keep up the excellent content. We love pictures. I'll be following closely! Good luck on the tap
 
Back
Top