Valve Gear Design for Steam Engines - My approach

Home Model Engine Machinist Forum

Help Support Home Model Engine Machinist Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

GreenTwin

Well-Known Member
Staff member
HMEM Supporting Member
Global Moderator
Joined
Jul 2, 2021
Messages
4,563
Reaction score
1,802
Location
MidSouth, USA
I became interested in steam engines at a young age, as I assume as many folks did, by operating a Wilesco steam plant and engine, which was a single-action oscillating engine.

I went so far as to build a boiler and single-acting steam engine for a 12 grade science project.
Most folks who saw it operate had no idea what it was, or what it was used for.

Fast forward many years, and I finally had a bit of time to study steam engines in general, and I started with the Audel series, which covered many steam engine and boiler types in detail.

I recall naively selecting one of the complex patent valve gear designs, with the intent of building a model engine using that style.
I did not get anywhere with that design.

I then read that most patent-style valve gear designs never came into widespread use, I think due to their complexity, and the difficulty of maintaining the precison of all the joints/bearings required to make them work accurately.

I decided to go back to the fundamentals, and start with the simple D-valve (which is really not that simple in my opinion).
I designed a D-valve with basically little or no cutoff, ie; the just full travel on the valve.
A simple D-valve design does work very well on a model engine, but I really wanted to get closer to a real valve configuration.

I began studying the Stanley 20 hp steam auto engine, since drawings are available from the Stanley museum.
The Stanley engine used a Stephenson's link, which was a very popular valve gear design, and a design that allowed the engine to be reversed.

I consider the Stephenson's link to be one of the less/least complex of the reversing/adjustable designs, but even then I discovered there are many variations just with the Stephenson.
You can have open rods, crossed rods, links curved inwards or outwards, suspension points, marine links, non-marine links, etc, etc.

I could feel myself going down a rabbit hole again, and so I studied the Stanley-style Stephenson's link only, which seems like a fairly typical Stephenson's configuration.

The trick with the Stephenson gear is to get equal motion of the valve in any of the gear postions, and I have read white papers about how to achieve this, but still don't comprehend it entirely.

One thing I did learn is that for locomotives with Stephenson link valve gear, the gear could be positioned to give a very late cutoff when starting the train, such that the locomotive produced maximum power and maxium torque as it began its motion.

Lucky for steam engines, they can produce 100% torque at zero rpm, but as the Stanley auto manual warned, don't open the throttle fully with no engine speed, else you will bend the crankshaft.

Once the locomotive/train was moving, the Stephenson's link could be adjusted to give an earlier (and variable) cutoff.
One desirable feature of the open-rod Stephenson's link is that you can increase the advance with it, which is similar to advancing the spark timing on a gasoline engine as the engine runs faster.

As the locomotive traveled at high speed, a very early cutoff could potentially be used along with an advanced admission, thus saving much steam/fuel/coal.

An early cutoff turns off the steam to the cylinder while the piston is early in its travel down the cylinder, and thus the expansive power of the steam could be used to finish moving the piston the remainder of its stroke.

.
 
Last edited:
Going back to the simple D-valve, even that has a lot of hidden and subtle design considerations.

I learned the main valve events were admission, cutoff, release, and compression.

Manipulating the various valve surfaces/geometry, and the eccentric that drives the valve, can affect one or more of the above valve events.

There is an exact way to design a D-valve and the associated valve face and eccentric(s), but first you have to decide exactly what it is you are trying to achieve, and there does not appear to be any one correct D-valve design, but rather various designs that are more applicable to certain applications, like a stationary engine which may be run in either direction often, or a marine engine which may be operated mostly running forward.

I select a non-exact generic approach to D-valve design, which is basically a 20 hp Stanley valve and eccentric design.

I also somewhat understand piston valves, which I consider just a D-valve wrapped into a cylinder.
With piston valves, you can have inside or outside admission.

The passages get a bit complex with piston valves, and the valve and sleeve can also be complex, and so I have stuck with the D-valve design.

One thing I have included in my D-valve design is a balanced D-valve, which is a method to relieve the pressure from the back of the valve, to prevent excessive wear between the valve face and the port face.

If you calculate the pressure times area on a typical D-valve, even at 100 psi you can have a great deal of force, pressing the valve against the port face.

Below is a typical unbalanced Stanley D-valve, (and my version of a balanced D-valve design in the next post, which is pretty much straight out of an Audel's book).

Stanley-Valve-01.jpg
Stanley-Valve-02.jpg
Stanley-Valve-03.jpg
Stanley-Valve-04.jpg
Stanley-Valve-05.jpg
Stanley-Valve-Seat-01.jpg







Stanley-Valve-01.jpg
Stanley-Valve-02.jpg
Stanley-Valve-03.jpg
Stanley-Valve-04.jpg
Stanley-Valve-05.jpg
Stanley-Valve-Seat-01.jpg
 
Last edited:
Here is the balanced version of the D-valve.
Not that there is a hole in the top of the valve dome, which is necessary to keep the pressure off of the top of the valve.

There is typically a ring (like a ring on an automotive piston) in the slot in the round part of the valve to provide a seal with the sliding piston-like part above the valve.

A balanced D-valve requires a precision-machined interior valve cover surface, which is not really difficult to achieve.

The percentage of force reduction is directly proportional to how much of the area of the valve you cover with the relieve dome.
In the case below, probably 2/3 of the force will be relieved, while leaving some force to hold the valve securely up against the port face.


Balanced-Valve-02.jpg
Balanced-Valve-03.jpg
Balanced-Valve-04.jpg
Balanced-Valve-Assembly-01.jpg
Balanced-Valve-Assembly-02.jpg
Balanced-Valve-Assembly-03.jpg
Balanced-Valve-Assembly-04.jpg
Balanced-Valve-Top-01.jpg
Balanced-Valve-Top-02.jpg
 
Last edited:
And an assemble of the engine that I will use the balanced valve on (bottle engine), showing the beginning of the Stephenson's link design.

The Stephenson's link is an eloquent and relatively simple solution to valve gear control, and it was widely used and liked for many years on locomotives and other engines, due to its simplicity, and due to its excellent function.


And so this is the limits of valve and valve gear complexity that I am willing to engage in, for model building constrution.



Valve-Gear-05-10-15.jpg
 
After studying the dual-valve arrangement on the Monitor ironclad steam engine, and studying Rich Carlstedt's drawings, I still don't feel I have a complete understanding of how all that vavle gear operates in every scenario, such as foward, reverse, mid-gear, slightly advanced gear, etc.

So is is most impressive to see folks building Monitor engines, and making them work accurately such as with Rich's accurate valve gear.

This is a realm that I avoid (very complex engines and valve gear).

I follow the KISS method of steam engine design, which as most know is "Keep It Simple Stupid", and simple fits my brainpower very nicely.

In the world of engine design and technical things, if you could compare people's minds to engine types, I would consider Rich's brain to be like a super turbocharged Merlin V-12, or better.
My brain would be like a Briggs and Straton lawnmower engine, if that good.

But I enjoy this hobby as much as anyone, and I build engines because I like to design and build engines.
I was building engines long before the internet, and if the internet all stops working one day, you will still find me in the backyard casting my own engines.

.
 
Last edited:
The only other semi-complex valve arrangement I have studied in detail is for the Soule Speedy Twin steam engine, which was mainly used for sawmill carriages, due its very high torque, and very fast reversing under load.

The Speedy Twin steam engine was used by sawmill operators even after electric motors began being used to power sawmill carriages, since the Speedy Twin could outlast and outperform a typical electrical motor configuration.

The money that could be made by operating a sawmill was directly proportional to how many times you could reverse a sawmill carriage in a day's time, and the Speedy Twin was the king of that contest.

The Speedy Twin steam engine was produced up into the 1950's, and the foundry and manufacturing plant that produced this engine have been turned into a museum, which is in Merridian Mississippi.

The Soule Live Steam Festival is the first weekend of every November.
I highly recommend it.

This year it is on November 4-5, 2022.

https://www.industrialheritagemuseum.com/exhibitions-events/
 
The Speedy Twin valve arrangement appears to be unique, as far as anything I have ever seen.

I studied the Speedy Twin steam engine for several years before I finally figured out how all the passages, valves and eccentrics work.

The Speedy Twin uses semi-balanced multi-chamber D-valves, and the pressure is reversed on these D-valves without them lifting off of their seat (when reversing the engine), which took me a year to figure this one feature out.

The Speedy Twin is a two-cylinder engine, fully and very quickly reversible, but has two eccentrics only, and no Stephenson's or similar links.

I was lucky to have access to disassembled Speedy Twin engines, and the knowledge of the staff at the Soule Museum, as well as knowledge from one of the original owner's/operators of the Soule foundry/factory (Bob Soule).

Below is an almost completed 3D frame for the Speedy Twin.

My 3D model is full size, which I think is the way all old engines should be modeled.
I can scale and 3D print this engine to any size.


3D-PRINT-01.jpg
rImg_3154.jpg
rImg_3161.jpg
rImg_3164.jpg
rImg_3166.jpg
 
I have never found anyone outside the Soule Museum who understands how the Speedy Twin valves and valve gear work.

Here is a video of an actual Speedy Twin running.

Understanding how this valvegear arrangement works is the limits of my excursion beyond the balanced D-valve and Stephson's link.





 
Here are photos of the Speedy Twin semi-balanced muilti-D valves; one for each cylinder.

There is separate smaller single-chamber D-valve on top the engine for reversing (shown in one of the pictures below).

The top of the Speedy Twin has a number of complex passages in it, an casting those while holding the multiple cores accurately in place was a real trick.

rImg_5186.jpg
rImg_5193.jpg
rImg_5195.jpg
rImg_5199.jpg
rImg_5204.jpg


The mechanism shown in the photo above is the actual setup (driven off of a line shaft) that was used by the Soule factory to lap the reversing valves.
 
Last edited:
Interestingly, it appears that all of the original Speedy Twin patterns still exist at the Soule Museum, as well as some of the core boxes, and so technically, the Speedy Twin could be molded and cast again, assuming one had the knowledge of how to do such a thing (we do).

rImg_2155.jpg
rImg_2156.jpg
 
Glad to share it, and glad someone is interested.

I studied valve diagrams, and there are polar diagrams, and typical X-Y graphs for valve design.

The polar and X-Y diagrams contain the same information, but I find the X-Y graph to be far easier to understand.

I discovered that the X-Y plots were used in the 1,800's, in a book by Dalby.
Dalby created a machine that graphed out the valve travel mechanically on paper, shown below.

Dalby calls the X-Y plots "displacement curves", since they physically trace out how far the vlave is displaced first in one direction, and then in the opposite direction.


DALBY-Diagram-01.jpg
DALBY-Diagram-02.jpg
 
So I decided to create my own displacement plot for the Stanley 20 hp steam engine, and this is what I came up with, using an Excel spreadsheet.

What you are seeing on this plot is the movement of the piston, drawn against a background which represents the steam/exhaust ports in the port face, and also traces of the movement of the inside and outside edges of the D-valve.

Using this diagram, you can determine admission, cutoff, release, and compression for a steam engine (the four critical points).

r2Displacement-Diagram-Stanley.jpg



Basically you are tracing the path of the valve over the top of this port face.
The openings in the port face which lead to the passages below are shown in white above.
The blue areas are the raised metal portions of the valve face, upon which the valve rides.

Bottle-Assembly-04.jpg



Cylinder-Pattern-Right-01.jpg


Final-Cylinder-08.jpg
 
This is what is happening inside of the cylinder.

Note that there are passages from the port face to both ends of the cylinder (I omitted one passage in the second screencapture, but should have shown that).

The engine is exhausted via the purple exhaust passage, and out the exhaust pipe.

Often the steam from the boiler entered the side or top of the steam chest.
The steam chest contains the D-valve.


Cylinder-Cores-01.jpg
Cylinder-Cores-02.jpg
Cylinder-Pattern-Right-01.jpg
Final-Cylinder-13.jpg
 
In another book, I think called "Verbal Notes for Marine Engineers", or something similar to that, there is an excellent illustration of a steam cylinder, and the critical four events related to both valve travel, the port locations in the valve face, and piston travel.

Also shown superimposed inside the cylinder is the "card", which is a tracing that is created by a mechanical device called an "indicator".
The indicator was basically a mechanical pressure-reading device, with a round drum wrapped with paper, and a pencil that traced out the pressure inside the cylinder as the piston traveled down its stroke.

This is the best steam engine illustration I think I have ever seen.

rVerbal-Notes-Image-01a.jpg
 
The Wilesco steam engines I have seen are single-acting, ie: the steam acts on the top of the piston only, pushing it to the end of the stroke, and then the momentum of the flywheel mass returns the piston to the top of the stroke, pushing out the exhaust steam as it travels back to top-dead-center.

Almost all of the old steam engine designs (with a few exceptions) were double-acting, ie: the steam from the boiler pushes the piston down the stroke of the cylinder, and then the valves swap the steam and exhaust ports, and the steam pushes the piston back to the top of the stroke.

The double-action effectively doubles the power output of a steam engine.

If you compare a double-acting single-cylinder steam engine to a single-cylinder gasoline engine, the steam engine produces two power strokes per one revolution of the flywheel.

The gasoline engine produces one power stroke for every other revolution of the flywheel.

So I often see folks exclaim how a very small steam engine can produce such a large amount of power and torque, and the answer is that a single-cylinder double acting steam engine is roughly equivalent to a four cylinder four-stroke gasoline engine.

A two cylinder double-acting steam engine is equivalent to an eight cylinder gasoline engine (roughly speaking).

And unlike the gasoline engine that does not begin to produce torque until the rpm's are raised to some level, the steam engine can produce 100% torque without the engine even rotating.
This very high torque at no rpm is how steam locomotives were able to pull very heavy trains with only two cylinders.

.
 
Hi GT,
First I must compliment you on writing this text book of explanation. Brilliant!

However, being an odd-ball engineer, I look at things sometimes from the obscure angle - and often go chasing ghosts for my trouble:
But...
I have a notion on Single acting engines that exhaust to atmosphere?
Could it be that the steam - that has lost much of its heat and has partly condensed during the power stroke - is expelled when the exhaust valve opens, but this expansion causes a further drop of temperature - and condensing - so that the pressure in the cylinder (at some indeterminate point in the stroke) drops below that of the atmosphere. Thus the piston appreciates the "partial vacuum" and atmospheric pressure difference to thus "Power" the engine towards the end of the stroke? This is based on understanding of higher speed engines such as 2-stroke petrol engines where this "partial vacuum" that comes from the shock (pressure) wave accelerating down the exhaust pipe naturally causes a low pressure in the cylinder for exhaust purging/assisting inlet feed.
The "gas" events happen at sonic speeds... so steam ejected from the cylinder will expand far more rapidly than the piston can "push" it out... - I think?

Just an hypothesis, but it does strike me that regular "mechanical" engineers (Like what I try to be..) think not of the "gas dynamics" that exist, just of the changing volumes and consequential pressures thereof. (Been there, made that mistake! - along with a dozen other Mechanical engineers!). Boyles law doesn't work with any gas around the liquifaction point. - e.g. condensing Steam. And steam in passages is limited by the speed of sound - which varies as the pressure drops in the passages - and is partly "choked" by condensate - but is partly helped by condensation being a dramatic reduction of volume of the fluid.
The maths of the gas dynamics (of the condensing steam) are beyond my experience and knowledge so I cannot model them to see what is happening. (Suggestions please?).
I have tried modelling exhaust gases in silencers, to some (poor?) degree of success, but did manage to make an air silencer >10dB better than the previous design... (Hmmm. 136dB to 123dB... Maybe not so poor?) and have roughly modelled an Atkinson cycle engine (simply by an excel spreadsheet). But these were gases alone, not gas/liquid mixtures.
I worked with a Doctor of Maths who wrote (Mainframe) computer models (c: 1983~86: Pre-PC!) of an air motor and the intake and exhaust cycles from a gas movement perspective. He modelled the real thing to within 5% of the real acceleration of piston that we measured. (Which was not predicted by the 20 years of work by previous mechanical engineers!). We accurately designed and predicted motion when we tuned the damper (orifice size with stroke). And produced a design of a pneumatic shuttle valve that was much less sensitive to manufacturing limits and fits. - So "any part combination" worked correctly, instead of laboriously selecting parts that worked at the right speed. This all taught me that "gases" do not work by "volume" alone. And Steam works by non-linear steam tables to make things more complex!
Perhaps I should try and figure out the SA or exhaust of a DA engine?
Time it dig-out the Wilesco model...?

Any thoughts?
K2
 
My background is electrical (power distribution), but the sadistic teachers did make us take thermodynamics, I think to watch us twist and turn as we tried to understand it.

All I recall is PV=mRT

I also know from using my pressure washer that if you have enough pressure and velocity, water droplets act very much like a solid abrasive material.

I have read that the bane of steam engine design was condensation of the steam, and that some of the more efficient designs kept the live steam and the exhaust steam well separated, such as in the Uniflow design.

As I understand it, the efficiency of a steam engine can be determined by the printout from the indicator (according to Charles Porter).
I am not sure exactly how this is done; perhaps measuring the area under the curve? and I am not sure how this shows efficiency.

Here are prints of a couple of typical indicators.
There is a piston inside the indicator cylinder, which drives the arm/pencil-pen assembly.
Paper is wrapped around the drum, and the drum rotates as the pencil/pen draws its graph.

The small indicator piston is piped to one end of the cylinder.
A "card", which is what they called the printout, was taken for each end of the cylinder, one card at a time.

This is basically a mechanical version of the electrical oscilloscope, and this device is critial in diagnosing how a steam engine functions.

An indicator can diagnose all sorts of steam engine problems, such as leaking valves, broken piston rings, etc.

Reynolds-04.jpg
Scranton-18.jpg
 
Steam was often admitted to the cylinder slighly before the piston reached top-dead-center, I think to give the steam time to flow and build up full pressure in the cylinder.

As a D-valve begins to open, the initial opening is a thin line, and thus the flow is restricted until the opening becomes larger, and I think this is another reason for using some advance on the admission point.

Between the admission and cutoff points, the D-valve has to move to its end of travel, reverse direction, and then at the cutoff point close the admission of steam.

The earlier the steam is cut off during the stroke, the more energy that can be saved, with the understanding that if you cut off the admission of steam too early in the stroke, the engine may not produce sufficient power or torque for a given load.

Here are some good sections of a typical steam engine cylinder that uses a D-valve.

Audels-Cylinder-Section-01.jpg
Audels-Cylinder-Section-02.jpg
Scranton-20.jpg

.
 
Last edited:
The most fundamental arrangement of valves on a steam engine, and the easiest to understand, is a 4-valve Corliss arrangement, shown below.

The function of a Corliss valve arrangement is as follows:

1. Open the top left steam valve, force the piston down the cylinder.
2. Close the top left steam valve, and open the lower left exhaust valve.
3. Open the top right steam valve, force the piston back in the opposite direction.
4. Open the lower right exhaust valve.

The Corliss engines were very efficient, and much of that efficiency was due to the separation of steam and exhaust passages, and the very quick and often totally independent operation of the steam and exhaust valves.

Shillitto-26.jpg
 

Latest posts

Back
Top