E.P. Allis 3000hp Quadruple Expansion Corliss Engine, 1893 in 1/12 scale

[bananarchy]

Introduction

Here begins what will be a very long and daftly ambitious undertaking, but something that, when reduced to its individual elements, at least appears to be feasible. Some years ago, I happened across mention of this engine in Wonders of Machinery Hall, an account of one of the main exhibits of the 1893 World's Columbian Exhibition in Chicago. There was all manner of fascinating engines and manufacturing equipment, but one in particular stood out. This was purported to be the largest mill-style engine ever built, at least at the time, and was used to drive dynamos and generate electricity for the fair. The cylinders measured 26", 40", 60" and 70" with a 72" stroke, and the engine ran at 60rpm.

Even more interesting than the quadruple-expansion setup and the massive scale of the engine, though, was the valve mechanism. I managed to find an original copy of an arrangement drawing from a period publication, and noticed that in addition to the regular eccentric which drives the wrist plate as usual, there is a second, smaller one which feeds into the governor linkage somehow. Strange! More detail about that in a subsequent post, as it needs a bit of a treatise all on its own.

Over the course of the last couple years, I got involved with an organization that has iron foundry capabilities, plus some machine tools, and I also managed to acquire and set up a machine shop of my own (something I've been fantasizing about since I was a teenager). I had started modeling the engine in solidworks, partly for giggles and partly to try and reverse engineer the valve mechanism. A few months back, I realized I actually had all the resources and capabilities necessary to take a stab at really making this thing.

[bananarchy]

The Miniature

So if we're going to build it, the next question is, how big? The answer, effectively, is as big as possible. I just like big engines to begin with, but additional scale also allows things to run more slowly and more smoothly, and be able to do real work. Also, one of the most interesting aspects of this engine is the trip gear and all the associated fiddly little linkages. I would like those parts to be scaled to the rest of the engine as closely to real practice as possible, and making everything bigger makes those parts more practical. The big parts can't be too big, though.

The determining factor ends up being the flywheel - this will be made as an assembly of castings, just like the real one, and it must be turned as a finished assembly. I have access to an Axelson lathe with a maximum swing of 30 inches, and the flywheel on the original was 30 feet in diameter, so bam. 1 inch to the foot. Perfect. At that scale, the trip gear parts are small, but totally doable, and a good bit bigger than plenty of things I've worked with professionally (I design robotic surgical instruments).

The engine will drive two dynamos with two concentric belts, which will allow load to be put on the engine to demonstrate the operation of the governor and trip gear under varying loads, and allow the compounding to properly work (or as much as it can without the vacuum from a condenser). Including the dynamos, the model will be about ten feet long, and the cylinders are 2.125", 3.3", 5.0" and 5.8" with a 6" stroke. The flywheel will be 30" in diameter and 6" wide on a 1.75" diameter crankshaft, and will weigh a bit under 200 pounds when finished.

The large bits are pretty much sorted, design-wise. There's still a fair bit of detailed design work to be done on the trip gear, and the governor will be a whole project in itself. The flywheel, though, is straightforward to design, and is a nice self-contained project, so that's where I'm going to start.

[peatoluser]

Looking forward to following this build.
I notice the drawing is from 'The Engineer'. By chance I have a pdf from the rival periodical 'Engineering' that covers the same engine and you're right, from the description, it is a complicated valve arrangement!   
Good luck with the build

[crueby]

Looks like an excellent engine to model!  Will be watching along...   

[Sanjay F]

Wow - what a feat of engineering the original is and so too will be the model. 30" flywheel, that's going to be something to behold!

Very best of luck with it and I'll be following along intently! 

[crueby]

Yes! It hadn't really sunk in how BIG this will be, it will definitely be able to do some real work!

[CI]

A twin tandem; very interesting project !
Typically the second eccentric controlled cutoff, and the first eccentric was for admission; at least that is my guess.
Corliss valve gear can get extremely complicated, such as compounding the action of two eccentrics to give a variable cutoff.

I like larger scale model builds, for the same reason as this poster, because the small parts and fasteners are very difficult to make when the scale gets small.

If you intend to cast the parts, keep in mind draft angles and machining allowances, and a shrinkage factor.

While this will be a very large engine, I don't think it is too large, but you definitely will not be carrying it out and putting it in the car.

Great subject for a build.
Good luck.


.

[Jasonb]

It's a big project for sure.

You mention your design background but have you done much machining in the past?. Also do you plan a large steam boiler to run this from as air won't work well with the compounding and would need an equally large compressor.

[Kim]

Looks like a very interesting project.  And quite big!  I assume you won't be moving it around much? 

Will be following along.
Kim 

[bananarchy]

Transport wise, I'm figuring the engine will be built on top of a table-height steel frame with some beefy casters on the bottom to facilitate moving it around, but transport is definitely going to be a trailer job. Boiler wise, the organization I'm with has multiple boilers of varying sizes that we take to shows and events, so thankfully that's not a worry. This is my first foray into designing castings, but I've spent a good while in machine shops. Some of the work will definitely be stretching my skillset, but hey, that's part of the fun!

[bananarchy]

First Part: The Flywheel Hub

The flywheel is to be constructed in a fashion as close to the original as possible (which is pretty much the easiest way to do it anyway). It is an assembly of the hub, 12 spokes and 12 rim segments bolted together. The original hub was split down the middle, but I've elected to make it a single part for the sake of simplicity and strength. This could have been done as an iron casting, but given that basically every surface would need to be machined in order to make things run true, and with the extra effort of making a pattern, I decided to machine it from steel. This was also helped by a fortuitous find at the local steel yard. The flywheel hub finishes at 7" in diameter and 3.5" wide, and I found a hunk of steel round bar in a corner that was 7.5" in diameter and 3.8" or so long. Perfect! This lump is 47 pounds to start, and the finished part will be about 15 according to the cad model.

My machining workflow on this ended up being pretty complex, and probably moreso than really necessary. This is largely due to the fact that the big lathe I have access to (30x120 Axelson) is a bit iffy in the precision department (or so I thought when I started) and I wanted to do all the precision finish machining on my Hardinge, which does not have nearly enough chooch to do the rough machining on a part this size. Since I didn't trust the big lathe to cut the bore as straight as I wanted, the first step was boring the 1.75" center hole on the mill. A long and slow process, since the length of the boring bar means a very low RPM to yield a good surface finish, but I was happy with the result.

Once I had my finished bore diameter (a few tenths under the nominal), I went ahead and turned a mandrel that will be used for the majority of the machining. The initial roughing and grooving is done with the part in the chuck in the Axelson to maximize rigidity, but everything else will have the part held on the mandrel shaft. This made use of the custom steady rest I made for the Hardinge, which was the first real project in the new shop earlier this year. I managed a nice snug fit on the diameter, about 0.0005" clearance.

That done, the part is moved to the big lathe (which makes even pretty decent sized parts look tiny). The mandrel was slid into the part for the purposes of indicating, and jacking bolts were placed between the backside of the part and the chuck to make getting the bore axis running true easier. I'm leaving a healthy finishing allowance during the roughing, so this doesn't have to be dead nuts.

Next: Time to make some real chips!

[bananarchy]

Flywheel Hub: Heavy Turning

This setup was chosen to provide maximum rigidity for as much of the material removal as possible. The rough machining of the outer faces is pretty straightforward - I've got some nice carbide 1" shank lathe tools which made quick work of skimming down the OD and the face, then facing down the outside features, leaving 0.050" on every face for finishing in the Hardinge (this turned out to be much more than necessary, but better safe). I was running the lathe somewhere in the 200s RPM for this part.

The real kicker on this part, though, is the central groove. It is 1" wide and nearly 2" deep radially, and it needs to have a flat bottom. This wasn't remotely feasible with the carbide I have on hand, so I tracked down a king-size candy bar of high speed steel - 3/4" wide and I think 1" high. It even already had a pretty close grind to what I needed, I just did a bit of cleanup and added some relief here and there. This looks right at home in the E size Aloris tool holder - those things are no joke!

RPM was reduced to 59 (speeds available on this lathe run from a max of 555rpm to a minimum of either 6 or 9rpm, can't remember) for the high speed steel, and the grooving tool was just plunged right in there, hand fed. This is easily the heftiest cut I've ever done, but the lathe never blinked and it worked wonderfully. Quiet, smooth, good surface finish, not a hint of chatter anywhere. This is where having a big, heavy, rigid as hell machine makes a big difference. You know you're doing something serious when the swarf makes a solid *thonk* when it hits the chip pan.

[crueby]

That done, the part is moved to the big lathe (which makes even pretty decent sized parts look tiny).

That lathe makes my LATHE look tiny!  Great start!

[bananarchy]

Flywheel Hub - Finish Machining

Once the big groove was roughed out, the first keyway was broached in the hub to allow it to be driven on the mandrel (plus a 7/16-20 set screw hole), and it was mounted in the big lathe to finish roughing the backside. After that, it goes into the Hardinge for finish machining. This ended up being a bit of a puzzle to figure out combinations of tools/holders/compound positions etc to be able to machine all the necessary surfaces, especially in the groove. I had 3d printed the finished part previously just for giggles, but this actually ended up being really helpful - I just put the 3d printed part in the toolmaker's lathe before I even started the part in order to make sure it was feasible. I was a bit concerned with the tool stickout, but I ended up finishing the groove width out to 1.001" with a taper of a couple tenths from top to bottom. Good enough!

One turning feature remains - the radius between the inner and outer ring of fasteners on the hub. This was also too beefy of a cut for the hardinge, so back into the big lathe with a hand ground tool (using a 3D printed radius gauge) to finish out the turning. The surface finish wasn't quiiite what I was looking for initially, but a bit of sanding and scotchbrite took care of it.

Next up: MANY HOLES

[bananarchy]

Flywheel Hub - Bolt Holes

The last big step on the hub is the addition of the bolt holes - each spoke is affixed to the hub with three 1/4" studs, so that's 72 holes in total, in two rings. These are all drilled from one side to ensure alignment, and reamed 0.251 to keep everything nice and snug. The initial setup is slightly ticklish, as the bolt holes actually require specific clocking relative to the keyway. The idea is that the keyway is lined up with the plane of the joint between two spokes, and a small semicircular channel is milled in each spoke to allow an allen wrench to sneak in and get to each of the two set screws (these two set screws will be the only OTS components in the flywheel). A piece of keystock was placed in the keyway, and I indicated a parallel clamped to the keystock to clock the keystock with the Y axis of the mill to provide that angular reference (with the rotary table set to 15 degrees, so that the bolt holes and spoke centerlines land on 0, 30, 60, etc).

Each hole was center drilled, drilled, reamed, and countersunk. The bottom side holes were also centerdrilled before drilling to ensure accuracy, which required an extended centerdrill. This was several hours of rotary table cranking, but everything turned out great. To deburr the holes on the inside of the slot, I got a Noga reversible countersink tool that worked like an absolute champ - this would have been a nightmare otherwise, especially with the inner ring of holes. And that's a (virtually) done part! The second keyway gets broached and gets its setscrew, then this one is finished. Oh look, that's a fastener! More on those soon.