Shay Steam Brake Valve
Nelson Riedel Nelson@NelsonsLocomotive.com
Initial: 10/26/2003 Last
The day after I published this page Mike Green from up in Ontario emailed and
pointed out that I had the steam and air brake controls on
Cass 5 backwards. I've corrected the following to
reflect Mike's input.
| Cass 5 Brake Valves: This photo shows the two brake valves
on Cass No 5. The valve on the right controls the locomotive
steam brakes and the valve on the left controls the train air
Note that the air brake valve is located under the throttle handle and
beside the reversing lever.
The November/December 2003 issue of Live Steam has a neat
brake valve design for a three-truck Climax. The design by Bob Reedy
is really two valves, a locomotive steam brake valve and a smaller train air
brake valve. The locomotive brake uses O ring seals whereas
the train brake uses a Teflon plug. I decided to use a physical appearance
similar to Reedy's locomotive brake but try a Teflon sleeve similar to the design
used for the cylinder
cocks. Recall that I had the steam and air valves reversed so
I modeled the air brake valve thinking it was the steam brake.
|Test Model: The first thing was to make the
model shown on the right. The plug has two holes at a
90 degree rotation of the plug that go half way through. The
body has three holes at 90 degree rotations. The plug has
three positions ---- in the first position the plug
holes line up with the
steam supply and brake cylinder holes ---
this is the brakes applied position. Rotating the plug 45
degree counterclockwise puts the plug holes between the body holes
---- brake hold position. Rotating a further 45
degrees lines the plug holes up with the brake cylinder and
exhaust holes --- the brake release position.
The model worked and made a pretty good seal.
Next, a model was made using a 3/4" sleeve OD and a 7/16" plug
OD. The valve worked pretty well, leaked slightly at room
temperature with 125 psi air applied and leaked only a little in boiling water. The handle was a
little hard to turn. Tried another sleeve with a smaller ID to put more
pressure on the sleeve and body ---- no difference in force or small
leak. Tried yet again with similar results.
Many years of engineering experience have taught that there is
a time to leave the lab and do a bit if thinking about the engineering principles
---- and that time had arrived for the brake valve. Some
thought yielded the following:
Teflon is pliant: Teflon is a very tough
material, hard to tear and slippery which are properties that make it
good for sealing. Another property is that it is pliant
which is also necessary for a good seal. However, the pliability
means that as pressure is applied on the sleeve between the body and
plug, the Teflon tends to ooze out the ends of the body. That is
why no improved sealing was achieved by forcing the plug into a smaller
ID sleeve --- the extra Teflon oozed out the ends of the
body. The maximum pressure is related to the sleeve
thickness ----- the thinner the sleeve, the greater the maximum
pressure. The cylinder cocks sealed pretty well with a 1/32"
sleeve thickness. Several different brake valve sleeve thicknesses
were tried. A thickness of 1/64" worked but was
touchy; sometimes the sleeve tore. A thickness of
1/16" didn't seal very well at 125 psi.
Conclusion --- 1/32" is a good workable choice. (A
different approach would be to seal the ends of the body so that the
Teflon couldn't ooze out. That was not used because there was no
obvious way to accommodate sleeve swelling under higher
Pressure maintains the seal: The seal is
maintained by pressure, not force. That's sort of obvious --- the
steam pressure is tending to compress the Teflon and the resilience of
the Teflon is pushing back. The area of the seal is less
important except that a larger area will reduce leaks due to small imperfections.
Plug Rotating Force is Proportional to Plug Surface
Area: The same sealing pressure is required to maintain a seal
independent of the plug size. The friction force between the
plug and the sleeve is proportional to the pressure on the surface
multiplied by the surface area. The plug surface area
and hence the friction force is proportional to the plug
radius (and diameter).
Torque: The plug is rotated by a lever so the
force of interest is a rotating force or torque. The torque
required for a given force on the sleeve-plug interface is proportion to
the plug radius (also diameter). Hence when 3 & 4 are
combined, the torque to rotate the plug is proportional to the square of
the plug diameter.
Pressure Between Sleeve and Body: The
total force (pressure X surface area) between the plug and sleeve is the
same as between the sleeve and the body. On one model with a
3/4" OD sleeve and 3/8" there was a leak between the sleeve
and body. After trying several sleeves a little thought
revealed that the sleeve OD area was nearly twice the ID area hence the
sealing pressure on the OD was roughly half the sealing pressure on the
ID. No wonder the the sleeve-body interface was the most likely to leak.
After thinking about the above points, , the following dimensions were
1/4" plug diameter (smaller is better but limited
by hole diameter)
5/16" hole in body (this is sleeve OD which
gives a 1/32" sleeve thickness)
1/2" long plug working surface (this size looks
5/64" holes in plug and sleeve
holes in body and plug at 120 degrees instead of the 90
degrees of first model
1" body OD.
|Plug: The plug was made from 1/4" diameter 303
stainless steel. The stem is 3/16" diameter and the end is
0.133" partially threaded 6-32. The holes are
5/64". The body was initially made with a 1/4" hole
and the plug put in the body and the holes in the plug partially
drilled using two of the holes in the body as a guide. After
the holes were drilled to the plug center, the surface around
the holes was smoothed with a fine file and then the plug was
polished with 400 Emory cloth.
|The three holes in the body that access the plug were drilled
3/32" and threaded 1/8" MTP and plugged with pipe
plugs. The three ports are in the bottom and drilled up to meet
the three horizontal holes. The input and exhaust ports
are 3/16" MTP and the exhaust port is 1/8" MTP. The holes for attaching the top are 2-56.
The body ID was bored to ~ 0.305" and an old 5/16" reamer
driven in the hole ~ 1/8" to score the hole to keep the sleeve
from rotating. The sleeve was turned from 3/8"
Teflon rod to a diameter of ~ 0.320" leaving a head on one
end. The sleeve was then driven into the body with
A 3/16" hole was drilled in the sleeve and then 5/64 access
holes drilled through the 3/32" holes in the body.
The body was then bored to 0.010" less then the plug
OD. (The pug was 0.249" so the sleeve ID was 0.239")
A handle was attached to the plug and the plug forced into the
sleeve (by hand) while the body was rotating very
slowly. After the initial insertion, the plug was
removed and the ends of the sleeve trimmed flush using a utility
knife. The plug was then inserted again (easy the second
|This is the body with sleeve. Note the 1/8" plug in the
side hole. The input
port below is fitted with a 3/16" MTP nipple and a 1/4" to
3/16" bushing. This was used to test the seal.
|The top was also turned from 1" brass and is about 1/2"
high. The 6 holes were drilled in the top first and then
the top was used as a guide to drill the mating holes in the
body. The upper part of the hole for the plug shaft is
3/16" The lower part is drilled 1/64" larger to
The top was later notched to allow rotation of the handle while
providing stops to limit rotation.
|The (air brake) valve on Cass 5 is mounted at about the height of the top of
the rear engine cylinder which scales to about 4" above the cab
floor. That height was tried and looked awkward so the
valve was lowered to about 2" off the floor as shown on the
A brass rod screwed into the input port is the support bracket.
The tee was silver
soldered on the rod and the top part of the rod was drilled and outside
threaded 3/16" MTP. As a result, the top part of the rod
is actually a pipe while the bottom part is a fake pipe serving
as the mounting bracket. The lower end of the rod is tapped
4-40 and held by a flat head screw on the under side of the
floor. The copper pipe pointing into the camera is
the input pipe that will connect to the steam turret.
The vertical copper pipe to the right of the input pipe is
from the cylinder port. The pipe goes through the floor and
connects to a pipe running across under the frame I beams to the cylinder.
This view is from the left side of the back of the valve.
The valve handle is in the BRAKES ON position.
The smaller copper pipe on the rear side is from the exhaust
port. The pipe goes through the floor and has an open
The valve and cylinder were temporarily connected with
3/16" plastic hose and the system tested using compressed
air. Everything worked as expected.
OK, now that I've learned I had the two brake valves mixed
up what a I going to do? I have a valve that can be used to control
the steam brakes but is in the air brake position. Of course, it can control an injector for train vacuum
brakes. I think I'll make another valve ---- at least
another valve mounting arrangement in the more correct position for the
steam brake. But there is more to it than just the valve location.
The following is part of the note that Mike Green sent describing the steam
Just opened up your site today and noticed that
you have added the Brake Valve installment. Having ridden on #5 in the
past I think that you have the two brake valves reversed in your captions
for the first pictures. The valve with the long stem directly in front
of the Engineer is the loco steam brake. The larger valve unit under
the throttle lever is the Westinghouse valve for the train air brake
control. I didn't follow the three pipes descending from the steam
valve but one applies the brakes, while another feeds the back side of the
cylinders to keep the brakes off, the third one is the exhaust. The
Engineer was constantly moving the valve probably because of a disc leak at
that time and the application side was showing some pressure. The
horizontal pipe would be the steam supply leading up to the stop valve on
the main manifold. The reason I was told for the extended valve
spindle was to keep the heat of the hot valve from getting into the handle
and allow a machinist to be able to re-pack the top gland without
dismantling the whole thing, with the gloves of today, the hot handle wouldn't be
too much of a concern.
I haven't seen any models that use steam release (backoff) of
the brakes. My brakes hang up a bit on release and was thinking of
using heavier return springs. However, the thought of using steam release
is really intriguing........
Mike also sent a copy of Lima Instruction Sheet No 10
that describes the operation of the Lima Steam Brake. Turns out that I
also had a copy. Both our copies came from a booklet of Lima
Instruction Sheets published by Kyle Neighbors of Cass WV. The
following is a scan of the two page Instruction Sheet No 10.
Update (12/12/03): I fabricated a steam brake valve
along the lines of the valve describe above shortly after Mike pointed out
my errors on the first try. I delayed updating
this page until the valve was installed and operating the cylinder. The
boiler arrived in the mean time so I was off working on getting it installed
and plumbing the water and steam. The locomotive is about ready for
the first run at the track and assuming that it actually gets moving, brakes
might be a nice feature to help stop it. So, hooked up the
|Brake Pipe Routing: The pipes run from the valve on the top
of the shelf down the shelf front, through notches in the platform
top and then bend and go under the frame just behind the
firebox to the brake cylinder on the left side.
The pipes are identified on the photo on right.
The photo below of the left side shows the cylinder and the other
end of the pipes. The pipe to the cylinder center is
3/16" The other pipe is 3/16" to the tee at
the big end where it converts to two 1/8" pipes.
There is a drain cock at the end of the small cylinder. The
union in the upper pipe appears to be oversize and in fact is
a 1/4" union. I had run out of 3/16" unions so
the 1/4" union is a temporary
|This photo shows the cab interior with the installed brake valve.
The valve works great! The return springs on the brakes were
removed. (The return spring on the park brake lever was left
in place.) It takes about two seconds for the brake pistons to
move after the brakes are applied. The speed is controlled by
the pipe and valve size. This speed is probably good; we weren't
designing a steam catapult. On the test stand it's
possible to control the braking force by moving the valve on and
off. It'll be interesting to see how the brakes work on an
engine and string of cars with a little momentum.
Thanks to Mike Green for steering me to this neat design.
Update 1/30/2004: The valve
worked OK on compressed air but sometimes would not pass steam, especially
under lower pressure such as less than 50 psi. I modified the
the valve by drilling the passages larger. This fixed the blockage
problem but the clearances were reduced with the larger passages and the
valve developed a small leak. One solution would be to make a
new valve designed with larger passages. A valve with 3/32"
passages would be satisfactory and this size passage could probably be
achieved reliably with a 3/8" valve plug. Another
solution would have been to ignore the leak. Instead, I tried a different design using O-Rings that worked pretty well. That
valve is described in webpage titled Another
Shay Brake Valve.