Heisler Engine Design Part III
Cylinders & Heads
Nelson Riedel, Nelson@NelsonsLocomotive.com
10/12/2004, last updated
05/17/2006

Caution:  The design shown in the following pages has not been constructed yet so it likely contains errors as well as design faults.  I will come back and update these pages as errors and faults are discovered

The layout of the cylinders and valves is more complicated than just scaling the full size locomotive.  For example, direct scaling of wall thickness is not practical unless very close tolerances are employed.   Another complication is the steam passages between the valve and the steam input/exhaust and the valve and between the valve and the cylinder.  On the full size cylinders these passages are made as part of the casting process by using sand cores that are shaken out after the parts are cast.  These type of cores are not normally used in the investment casting process that will be employed to make the model cylinders so the design process must include a technique to machine these passages.

 I lack the knowledge and experience necessary to make the rule of thumb judgments that are probably second nature to both mechanical engineers and machinists (my background is electrical engineering).  To compensate, I've used Ken Schroeder's Shay design as a model and tried to make all components at least as robust.   Also, emphasis was placed on making all components as tolerant as possible of my kind of machining (sloppy).  

Another area of design concern is engine efficiency.  I've found that my little shay has much more power than needed when running on a flat track or tracks that model a mainline where the grades are on the order of 1.5% or less.  However, the geared locomotives typically were used on branch and logging lines where 3% grades were the norm and 6% and greater grades were not uncommon.  I usually run my shay at the Mill Creek Central track that has 3% grades.  On that track it's a constant effort to maximize power to climb the hills and then try to slow the thing down when descending ---- which makes it really fun to run.  With this in mind, I made the steam passages and ports as large as possible to improve efficiency in that area.   The operating steam pressure of the model will be 100 psi vs. 200 psi for the full size prototype.  The correct 1.6" scale pressure would be 27 psi so the model will use much greater than scale pressure.  This greater pressure will permit a somewhat smaller that scale steam and exhaust passage cross section.  The valve design was done using a reprint of design techniques described in Locomotive Design Part II - Valve Motion.  This pamphlet (No 28) published in1910 was originally published in Railway Machinery.  The pamphlet was purchased from Nation Builder Books,  www.nbbooks.com.  

The design process involved numerous iterations where everything was sketched out and then modified to achieve the objectives outlined above and sketched again.  Hopefully, this process will minimize revisions after the engine is built. 

The table below summarizes parameters arrived at during this iterative process.  The decision on the bore and stroke were based on reasonable power and achievable steam as describe in the Heisler Engine I page.  The crosshead guide ID decision was also described in the Engine I page.   The valve-cylinder spacing and the valve stroke were determined before the crankcase,  crankshaft and eccentrics were laid out, also discussed in the Engine I page  The rationale for some of the other compromises are described in this page.   Note that some values in the MRSR 91 column are preceded by the "~" symbol.  These values are estimates based on other measurements or estimates made from  photographs.  For example, the MRSR 91 cylinder heads were hidden by the head covers. The head OD value of ~ 21" was based on the overall cylinder OD, the head cover OD and the head diameter of a slightly small Heisler engine.                   

  Heisler Model Engine Parameters

Parameter MRSR #91 1.6" Scale Actually Used
Cylinder Bore 18.25" 2.45" 2.125"
Cylinder Stroke 16" 2.14" 2.0"
Crosshead Guide ID ?   1.75"
Left Valve-Cylinder Spacing 17" 2.28" 2.3125"
Right Valve-Cylinder Spacing 24" 3.216" 3.25"
Valve Stroke 4.5" 0.60" 0.625"
       
Cylinder Height 23" 3.08" 3.25"
Cylinder OD 21" 2.81" 2.91"
Cylinder Liner OD NA   2.375"
Head OD ~21"   2.875"
       
Valve Head OD ~12" 1.6" 2"
Valve Piston OD ~7" 0.94" 1.125"
Valve Liner OD ~8" 1.07" 1.5"
       
Steam Input Port ID ~4.5" 0.60" 0.5"
Steam Exhaust Port  ID ~6" 0.80" 0.625"
       
       

Squeezing it all in:  Note from the table above I chose to use a slightly oversize cylinder OD ( 2.91" vs. 2.81" scale).  The unusual 2.91" dimension is a 3% shrinkage from the 3" OD rods used for the pattern.  The cylinder length was stretched a little over 5% (3.25" vs. 3.08" scale).  This slightly longer cylinder together with a slightly shorter stroke gives enough room to make the piston with a flat top and bottom instead of having a recess in the piston bottom and a bulge in the piston top.   The lack of a bulge in the top allows the upper head to be flat and eliminates a bulge in the head cover.   Before fixing the cylinder length the overall engine was sketched out and the relationship to the frame and boiler examined as shown in the next drawing.      

The cab width of  9.5' scales to 15.28" which was increased to15.5" inches.   The inside dimension between the upper and lower frame bars of 40" scales to 5.36" which was increased to 5.437" for the model.    The outside edge of the cab should  line up with the center of the cylinder head cover and the outer edge of the walkway should line up with the upper corner of the cylinder castings.  The drawing above shows this to be the case on the model once the cylinder head and head cover are added, so the esthetics are preserved.   Another concern was that there was room between the cylinder casting and the boiler for the exhaust pipe.  The drawing shows that there is plenty of room so we should be OK.   

Steam Passages: The valve performs the function of connecting the steam input and exhaust ports to the two ends of the cylinder at the correct time --- or more accurately the correct position of the piston.  Before designing these passages it's useful to get an general idea of how it was done on the prototype.  The following are photos of a smaller Heisler engine that show the general structure of the cylinder, valve and passages.

This is the right side cylinder casting.  The larger cylinder on the right is the main cylinder and the smaller cylinder is for the piston valve.
The slit along the upper edge is the upper cylinder port.  It connects to the upper middle of the valve cylinder.    There is an identical port on the bottom of the cylinder.
This is the valve cylinder.  The upper slots connect to the exhaust as do a similar set of slots on the bottom of the valve cylinder.   The next set of slots (the narrow ones) connect to the top of the main cylinder.   The big slots in the middle connect to the steam input port. The next set of narrow slots connect to the bottom of the main cylinder.  
This is a sketch of the valve cross section with valve centered and both cylinder ports covered.  If the valve moves up or down slightly one of the cylinder ports will be opened to the associated  exhaust chamber and on to the exhaust.  This means that except when the valve is exactly centered which occurs twice per cycle, one end of the cylinder is open to exhaust --- or each end  is open to exhaust essentially half the time.

The small lengths of the valve identified as lap insure that the valve must move a small distance off center before opening one of the cylinder ports to steam admission.  For example, if the valve moves up a distance greater than the lap distance, the upper opening leading to the upper cylinder port will be uncovered slightly and steam will travel from the admission port, around the narrow middle part of the piston and out the partially uncovered port to the cylinder.  

Left Cylinder Casting:  The next drawing shows the left cylinder casting which contains both the main cylinder and the valve cylinder.  Both cylinders are oversize to accommodate bushings (sleeves). 

The cylinder casting is basically a chunk of cast steel with two holes for the cylinders and a hole for the admission pipe out the side. The cylinder heads are attached with 6-32 screws while the valve heads, steam admission pipe and exhaust manifold are all attached with 4-40 screws.  The depth of the treaded holes for the cylinder heads and valve heads were not shown to reduce drawing clutter.  A depth of 0.375" is reasonable for all these holes.   The holes for the exhaust manifold attachment screws were also omitted.  The locations of these holes are shown on the manifold drawing.

The investment casting process should produce a casting very close to the size indicated on the drawing.  The main cylinder and valve cylinder holes are undersize and will need to be bored to achieve the exact size shown with a smooth surface.  The hole for the valve cylinder is considerably undersize because I increased the valve diameter after I made the pattern and then realized that I had room for a larger valve.   This extra material will give some room to make necessary adjustments to insure the two cylinders have the correct offset.  Skimming cuts will also be required to provide a smooth sealing surface on the surface that mates with the exhaust manifold and on the admission port.  The ends should also have a skimming cut to provide a smooth surface for the heads. These cuts should be made after the bushings are installed.    

The steam passages add a bit of character to the cylinder casting.  The steam admission is via the 0.5" hole in the side of the casting.  The exhaust passage is via the four  0.375" rectangular holes, two near each end of the valve cylinder.  There is a recess in the 0.470" thick exhaust manifold that routes the exhaust to the exhaust port on the manifold.  

The passages between the valve cylinder and the main cylinder will be machined.  The first step (after the cylinders are bored to the correct size) will be to machine the 0.3125" high horizontal passages.  A 1.25" diameter woodruff key slot cutter will be used inside the valve cylinder to make the 1.75" OD groove in the valve cylinder and then the cutter will be moved as far as it will go toward the main cylinder to open the slot toward the main cylinder.  The horizontal slots will be completed by using a 2.25" diameter slitting saw inside the main cylinder extended as far as it will go toward the valve cylinder.   The final step will be to use a 1.125" woodruff key slot cutter in the main cylinder to cut the partial moon-shaped vertical recesses in the main cylinder sides from the horizontal passages to within 0.125" of the cylinder ends.   These passages do not require close tolerances; the openings in the bushings are the critical dimensions.   The most critical part of machining the steam passages is to avoid cutting through the side of the casting.             

Right Cylinder Casting:    The  distance between the valve and main cylinder centerlines is 2.3125" on the left cylinder.  This dimension is increased to 3.25" on the right cylinder.  The woodruff cutter and slitting saw used to make the horizontal steam passages on the left cylinder don't cut deep enough to make the passages on the right cylinder.  The two holes with ends  threaded 3/8" NPT shown on the adjacent drawing are used to joint the cuts made by the cutter and the saw. The ends of these holes are plugged and covered by the exhaust manifold.  Note that the width of the recess for the exhaust manifold has been increased to cover the plugs.

Additional holes are required in both cylinder castings for the cocks and valve observation holes used to adjust the valves. The exact location of these holes will be determined after the castings have been examined.

 

Cylinder Bushing:  The drawing on the right shows the cylinder bushing (sleeve).  The bushings for the two cylinders are identical and will be machined from gray cast iron rod (McMaster-Carr).   The bushing and cylinder can be machined to a press fit if desired.  Heating the casting in an oven to ~ 450 degrees F and cooling the bushing in a food freezer can ease the pressing of the bushing into the cylinder.  I'll probably make the bushing a very tight slide fit and use Loctite 620 to retain the bushing.

 The slots on the ends line up with the vertical steam passage slots machined in the inside of the cylinder casting.  The slots in the bushing will be cut with  a woodruff key slot cutter after the bushing is secured in the cylinder casting. 

The heads also serve to retain the bushing so there will little or no force on the bushing.    However, it is import to make sure that the outer busing is sealed to the cylinder around the steam passages to prevent steam flowing between the upper and lower passages.         

 

Upper Head: The upper head shown in the drawing on the right is rather plain and may be machined from steel or brass.. The 12 attachment holes match with tapped holes in the cylinder casting.  The tapped center hole is for a stud to attach the head cover.  This stud will be sealed with Loctite 620.   The 0.625" diameter recess provides clearance for the locking nut on the top of the piston.  The 2.123" diameter shoulder aligns the head with the bushing ID.   The 1.75" diameter shoulder is meant to reduce the free space at the top of the cylinder --- space that would otherwise need to be filled with  steam every stroke.  

The two upper heads are identical. 

Lower Cylinder Head & Packing Gland:  The lower head shown above will be fabricated from steel or brass similar to the upper head.  The 2.123" diameter shoulder aligns the head with the inside of the main cylinder bushing and the 1.748" diameter shoulder that fits inside the cylinder reduces free space inside the cylinder. .  The 0.125" thick flange is sandwiched between the cylinder casting and the flange on the upper end of the crosshead guide.  The 1.748 diameter shoulder on the lower side of the head aligns the head with the inside of the crosshead guide.  The head bolts pass through the flange on the crosshead guide and the head flange into the cylinder casting.  The packing gland is similar to the gland Ken Schroeder used on his shay and will be machined from bronze rod. The packing will be a pair of  5/16" ID, 9/16" OD Viton O-Rings.  The hex head cap screws that hold the packing gland will be lined up horizontally.  Hopefully it will be possible to access these screws a long hex driver up through the crankcase. 

Cylinder Patterns: This shows the aluminum patterns for both cylinders.  The patterns will be coated with epoxy to make a mold. That epoxy mold will then be used to make wax images of the patterns or wax patterns. 

The pink stuff is Bondo body filler used to fill voids and make fillets.   

This shows one of the patterns with the cores in place.  The cores will be in the position shown when the epoxy mold is made.   The mold will part in the same plane as the core rods.  When the mold is separated, the cores and pattern are removed.  The cores are then removed from the pattern and put back in the mod.  The mold is then filled with hot wax.  After some cooling, the wax with cores are removed from the mold.  The cores are then pulled out of the wax pattern leaving cavities.   There is also a core that goes in the steam admission port that I forgot to insert before taking the photo.  

The valves and exhaust manifolds are described in the next part.

Heisler Project
NLW Home