Conversion of 4 Cycle Utility Engine to Steam

[vtsteam]

I've begun work on a project to convert a Tecumseh H50 5 horsepower horizontal shaft motor to steam. It's been a great learning process for me. The goal of the project was to take one of 3 defunct engines I received in a trade and do the conversion. I didn't want to do just a simple "see it rotate" conversion like many of those shown on Youtube, but wanted to have proper valve events for both inlet and exhaust.

To explain that, many of the conversions I've seen are inlet open for 360 degrees, and exhaust closure is the only thing that provides drive during part of the cycle. The usual procedure for the simpler conversions is to add weld metal to the back side of the exhaust cam lobe and file it to a profile to match the existing lobe.

The camshaft on utility 4 cycle engines rotates at half the rate that their crankshaft turns, and this is accomplished (on my Tecumseh engine) by a set of gears of 30 and 60 teeth with a pitch of  5 TPI. Because single acting steam engines are essentially 2 cycle engines, at least by analogy, the existing cams on a 4 cycle IC engine are rotating at half the rate needed, and in addition, the valve events themselves are quite a bit too long for an efficient running steam engine. In other words the cam lobe lift(s) are too wide.

With suggestions by an experienced steam engineer, my intended valve timing for the steam engine eventually resolved to an inlet opening at 10 deg BTDC and closing at 90 ATDC (100 deg dwell) and exhaust opening at 15 deg BBDC and closing at 15 deg BTDC (180 deg dwell).

Before moving forward with my present plan, I'm going to first retrace here all of the steps (mostly missteps) I've taken so far, including the many mistakes and changes of direction as I gradually discovered problems with my various ideas.

I'm sure this is all old hat to those more experienced in steam and IC engine design, but it may be entertaining for that reason! I'm definitely beyond Plan B -- maybe I'm now at Plan F. But it has been wonderfully educational, particularly with regard to cam design, for me at least, and I enjoy learning this way. It's the old take apart the alarm clock, and try to put it back together before the folks get home kind of approach to education!

Below:
1.) Candidate engines -- all Tecumseh 1 seized, 2 with thrown rods.
2.) The star pupil -- a seized 50H type
3.) Crankshaft and camshaft gears
4.) Camshaft and lifters
5.) The existing IC timing marked off

[vtsteam]

My first plan was to cast a new cylinder head, with a boss for a piston intake valve that would be activated by a bash type pin welded to the exhaust valve. This would still require adding a second lobe to the exhaust valve cam location -- opposite the present one. At the time I didn't realize that the exhaust valve cam would be far too wide to work well for this. The old inlet valve was to remain closed, achieved by removing the inlet valve tappet.

I drew out a wooden pattern for the head to be cast in aluminum, tracing around the old head gasket with a magic marker -- the thick marker outline was a simple method for adding a shrink allowance for casting the actual head.

Approximate head bolt locations were marked from the gasket, although these would change due to shrinkage. But they were close enough to give me locations for adding bosses to the casting at the bolt locations. I also drew out the cylinder and valve spaces.

[vtsteam]

I carved out a passage in the pattern for the new single inlet/exhaust valve, and added bosses to the pattern for head bolts and a base for the new piston valve and its cylinder  -- which would be added as a separate assembly.

[vtsteam]

The valve boss was a little tricky to locate and align because the exhaust valve is inclined 5 degrees toward the engine cylinder. I also added a boss for a relief valve over the cylinder center.

Finished pattern below, contrasted with the old IC finned head.

[vtsteam]

At about this time I began to realize that a combination exhaust and inlet valve actuation had some limitations. It seemed that the inlet event would have to be symmetrical with the center of the exhaust event. They didn't have to overlap, but a single actuator simply means that those events must be tied in regard to their center points in a cycle. Or at least that's what I thought. This stuff can get tricky to imagine.

I also started to think that a 3/4" thick aluminum casting would absorb a lot of steam heat, at least initially, even if lagged. So I began to think about adopting a suggestion I'd had to fabricate a steel plate head with a bash type ball valve for inlet, actuated by a pin added to the top of the valve, and retaining the old poppet exhaust valve.


This would require adding two new lobes to the cam -- one each for inlet and exhaust. It would also require filling in the carburetor manifold with something (lead was suggested) to reduce the head space left by actuating the inlet valve, but using it only as a pin to actuate then newer ball valve.

Because my sadly rusted used milling machine purchase was in a somewhat exposed storage shed at the time -- surrounded by layers of odds and ends and unwired, and because temps have been in the sub zero (F) here in Vermont lately, I wasn't able to use it at the time. (Happily, I've since got it wired, cleaned, and operational, as well as closing in the shed it is in)

Instead I decided to make the plate head as two laminated pieces, brazed together. This would allow me to make a pocket for the exhaust and inlet valves by drilling the bottom piece -- the other would act as a cover plate.

[vtsteam]

Instead of filling in the intake manifold space with lead, I decided to use pot metal (die-cast) -- zinc based alloy from some old auto door handles. It has a little higher melting point than lead, though well below that of the aluminum engine block. It is considerably harder than lead, and it was quite a bit more difficult to file back after the pour. I'd have have poured short, if I'd been a bit smarter about it.

The intake carb seat was simply blocked with wood, and I used a candle sooted piece of 5/16" drill rod (silver steel) to act as a core for the intake valve. In fact I had decided to use this same material as the bash valve lifter, instead of adding a pin to the top of the steam inlet valve.

The valve guide in the engine kept the drill rod at the proper 5 degree inclination toward the cylinder center.

[vtsteam]

I originally intended to follow suggestions to weld a standard cast pipe tee to the head as a valve body and use a stainless ball and spring for the valve. A hole in the head itself would work as the valve seat after tapping lightly on the ball to form the seat. But later decided to use a threaded connection to the head -- which would allow some adjustment of the valve assembly height, among other advantages. It would also have been tricky to weld and center the fitting at the 5 degree angle required. Not impossible, of course.

Instead, I turned a separate valve seat out of brass for a press fit (later to be soldered) into a red brass pipe nipple. I used a pipe cap on the end, and silver brazed an inlet connection nipple threaded end into a close fitting hole in  the valve body. A spacer at the top of the valve and the pipe cap would  allow some adjustment of spring pressure. The valve was therefore a separate assembly that could be threaded into the cylinder head.

[vtsteam]

A pattern was made of the head with bolt hole locations by placing a piece of paper over the engine cylinder and tapping it very lightly with a brass hammer -- similar to the old trick for cutting out a paper gasket. This marked the important location details on the paper -- one advantage of a sligthly dirty block -- an automatic grease pencil!

I also punched holes at the head bolt locations with a round pencil -- the tapered point was larger than the hole so it gave a nice locking paper rivet to keep the rest of the paper from moving while I did the rest. A piece of drill rod was also used to locate the center of the 5 degree inclined valve passage.

I now had a pattern for marking out the drill locations for the head bolts and valve.

I marked and cut out the second piece of steel sheet of the head. Most of the steel I used was scrap material.

[vtsteam]

I decided to use braze instead of welding for much of this project -- I do own a welder, but need to get more proficient with brazing. It does use a lot of gas however, so I decided this time to try pre-heating the parts. I have a wood stove running this time of year, so why not put the pieces in the coals for a preheat?

I first pinned the two plates together with a couple of 1/8" steel dowel pins. Then popped them into the stove for a bake. The coals away from the vent kept the metal from oxidizing, and when it looked somewhat red, I pulled it out with tongs into a bucket of wood ashes, and trotted down to my brazing site (outdoors, cold) and brazed the part. The ashes definitely retained the heat, and braze was flowing shortly after lighting the torch.

[vtsteam]

After brazing the parts together, I used my paper pattern to locate the valve hole, and set the drill press up to drill at 5 degrees. The drill rod and new head lined up perfectly when in position. The head bolt holes matched up.

[vtsteam]

After testing the valve location, I enlarged the valve hole and tapped it to 1/2" pipe thread to fit the valve body.

I also chipped out steam passages in the head using a chisel and continued shaping with an electric die grinder. I later polished the grinding marks out and assembled the valve on the head.

Getting close now, I thought. All I need to do is braze a couple lobes on the cam shaft opposite the present ones, and I'm done.


Not.

[vtsteam]

So with valve and head pretty much completed it was time to simply add some braze metal to the camshaft, as secondary lobes, file them off to profile and I was good to go.

Just had to figure what that profile was. I marked the cam gear with the existing valve events after turning the crankshaft through 2 revolutions, expecting the timing to be reasonably close to steam timings. I figured it must be in the ballpark because so many people on YouTube and elsewhere have simply duplicated the existing exhaust lobe at 180 degrees, hooked the engine to their shop air compressor and spun up their new steam engine.

But after looking at the timing marks on my gear, I realized things were very different. 4 cycle timing is MUCH broader (naturally) than 2 cycle timing. And those cam lobes were a lot wider than what I needed. To figure the actual dwell for the crankshaft you have to double the dwell you get from the camshaft.

On my engine this meant that the existing cam would open the inlet for about 200 degrees, instead of the desired 100 degree dwell, and the exhaust would be open for about 300 degrees instead of the desired 180. Not only that but the two events would overlap by about 60 degrees. No good.

The photo below shows the measured valve events on the stock cam.

[vtsteam]

This seemed to mean I'd have to modify both the old cam lobes and the new ones. Well modify how much? I spent a lot f time sketching out cam lobes, and found a cam coordinate program on the web.

I wanted at least one eighth inch of lift on the inlet cam (for the ball valve) and I wanted about a quarter inch of lift on the exhaust cam for the poppet valve -- roughly the same as it had been on the IC engine. When I tried to sketch this out, or enter it into the program, the inlet cam wouldn't work out unless I made its "base circle" much larger than the original cam. I could not get .125" lift for 100 degrees of dwell on anything smaller than about a 2" diameter base circle. The original cam lobe base circle was 1".

I learned a lot through this exercise. The main reason the base circle had to be so large was that the camshaft was running at half the speed of the crankshaft. And the "mushroom" tappets will bridge any hollow in a cam profile. They will always maintain a tangent to whatever two points they rest on -- in a proper cam they will only rest on one point at any time, but it is possible to design a new cam profile that isn't correct -- in  that case, the tappet will "correct" it by ignoring the dip and bridging it.

A small dwell angle like the 100 degrees we need on the inlet of a steam engine (a short cutoff for steam economy) means you need a very big cam to get even a small amount of lift. Particularly on a half speed camshaft.

Anyway, I eventually drew out both an inlet and an exhaust cam that seemed like they might just work, assuming I didn't try to round off much of that profile. Pretty much straight sides to the top of the bump, a sharp angle there and a straight side back down. Not an  ideal profile as far as wear goes, or for valve train acceleration. But the best I could do within the constraints of the size of the valve chamber, which I measured carefully. The inlet cam would just fit in. The exhaust cam was somewhat smaller.

So I drew these out as full round cam blanks. I would have to turn off the old cam blanks with the lathe (also rescued from storage) and then braze on the new cam "donuts".

I figured I'd make the exhaust cam bore a little bigger than the inlet cam's so it would slip on the camshaft easier. The camshaft was a solid piece of cast iron, including the cam gear.

Below are drawings of the cam blanks I decided to use. The method of construction was to to turn the blanks to base circle diameter, and then braze on pieces of .125" dia. and .25 dia drill rod. These would set the max lift, and I would fillet them with braze and file to profile. When the file touched both the rod, and the base circle blank, that would be the profile.

(Both show only one lobe for clarity, though each has two. The inlet cam doesn't illustrate the .125 drill rod pin, the exhaust cam does show it and the intended braze fillets)

[vtsteam]

The coordinate program I found on the web gave me a bunch of coordinates that looked good as a table, but were totally wrong when I plotted it out. My pin and braze method was the one I chose to make the cams.

I turned down the camshaft lobes, made the cam blanks and bored them, and then brazed the pins in position after carefully marking out. Then I spent an afternoon carefully filing the braze down until I had my new cams.

Below are the cam blanks from scrap 1/2" plate -- interesting patterns on the face from exposure.

And the filed cams ready for mounting on the camshaft. The big one is the inlet cam.

[vtsteam]

A chance reference on an internet account of another conversion (the usual weld blob camshaft conversion job) mentioned having to narrow the width of the new cam lobes, because other wise they would interfere with the crankshaft counterweights when they swung up to the cam area.

When I read that I ran out to the work shed even though it was a chilly night, assembled the cams onto the shaft and installed them into the cam space in the engine. Then I rotated the crank.

 

Uh...........no go.

Big time no go.

Not even "trim a little here, slide this over a bit" go.

I mean NO GO. The space for the camshaft in the crankcase was obviously sized to the crankshaft throw. So unless the cams were original diameter, they would interfere.

[vtsteam]

So that brings us up to the present

I have a new idea.

 

Uh, external crankshaft mounted eccentric driving a valve in a new head casting.

Well maybe not so new.

Been around for at least a century or two.

[vtsteam]

Actually, external valve gear had been my original plan before I started searching around the web for conversions other people had made. The internal camshaft modifications seemed to be the most common, and promised a quick and simple conversion. It certainly would have been nice to package everything inside the engine by modifying a camshaft. And these modifications are certainly do-able, elegant, and probably fun. But they don't result in an efficient steam engine, or one with much power output.

Well I should modify that statement -- if there is clearance for sufficient sized cam lobes on a half speed camshaft, then it would be possible to get good timing figures. But on the Tecumseh, at least, it isn't possible.

My main interest are the Westinghouse style twin engines, like the Stuart Turner Sun and Sirius, and other model boat engines built in the early history of model steam boat racing. I own two antique model Westinghouse style engines, and once started construction on a new one.

Westinghouse twins almost always have external valve gear (though there are exceptions) and something similar to this for valve actuation was what I had originally imagined for the Tecumseh single conversion. That conversion was intended to give me practice and performance figures for a subsequent scratch built Westinghouse style twin.

Well, now it's back to that original idea, external valve gear and a head casting to receive the valve.

My plan now is to try a shaft type rocking valve coupled to an eccentric on the crankshaft. This seems to me to give a very simple linkage, allowing me to use pressure to keep the valve faces steam tight, as well as compensate for wear, while reducing wear over a full rotary valve.

A piston valve could also be used, but these must be very well fitted, and it seems to me that any wear, no matter how well fitted originally, would lead to steam leaks.

Traditionally, a D-valve was used on top of the head, with a linkage composed of either bevel gears and a vertical crank shaft, or a long bell crank and a ball end somewhere in the linkage to take up side motion.

I think my choice is the simplest I can think of for linkage, and offers some of the advantages of all of the other valve types.

So, on with the show......

[vtsteam]

I had a little time today to work on the head pattern. Here it is being tried on top of the cylinder.

[DavidF]

Looking good,  not 100% sure what your up to yet but judging by the bosses in the side of the head it should be interesting.

[vtsteam]

Hi David. The bosses on the side of the head will form the valve housing. The bosses will be bored through the head to fit the valve.

The valve itself will be a shaft which rocks, driven by the eccentric on the crankshaft directly below those bosses. A pushrod will connect the eccentric and a horn on the valve shaft.

The shaft will have milled areas to pass steam to and from the cylinder head beneath.

The long oval boss on top of the head will be drilled and tapped for a steam inlet pipe and an exhaust pipe.

Two ports directly under these, below the valve body, will be the steam in and exhaust ports.

It's all very simple.

[DavidF]

sounds like you have your project figured out more than I have mine.  Im stuck looking at some scribble on paper atm.  Ill sleep on it and hopefully dream up some solutions...

[vtsteam]

I'm sure you'll work it out David.

I painted the head pattern this morning. Getting close to the point where I can cast some metal.

I pulled out my old core boxes today, and will likely bake a 1/2" dia. rod core. My wife is planning on baking some chocolate chip cookies, so I'll put it in the oven with them. I use molasses water, so the cores smell good while baking -- makes it easy!

[Meldonmech]

Hi
      Pattern is looking good.

                                               Hope you get a good cast      Cheers David

[vtsteam]

Thanks, Meldonmech! 



Here's the baked sand core. Cookies were good too!   

[black85vette]

Head pattern is looking good.   Interested to see how it looks.