Wayne Roderick, 3rd Division, PNR, NMRA (life)
1/05/99, rev 04/20/08

The Teton Short Line has been assisting a major customer in the design and construction of a rotary car dumper. We looked at the prototypes (12 inches to the foot scale) and the only one we know of in HO scale by Walthers and decided they were all to big for our fleet of MDC ore cars. The TSL does operate in the '50s & '60s but we're not above borrowing a bit of "future" technology if its fun to do so. This project is well underway, with much of the construction complete, but we're posting this on the chance that you might find it useful or maybe have experience that we can use.

If you have no idea of what a rotary railcar dumper is then click for a prototype picture used without permission ;-)

THE BIG PICTURE: Ross mine ships nominal 15 car trains of ore to Coots Mithril and Leverite Mill at West Grand for processing. Up to now, these cars were loaded in an unprototype way with spoonful of ore at the mine and dumped with an unprototype rotary device commonly known as the "wrist", a device connected to the human hand, into a nearby coffee can at the mill. You can certainly see room for improvement here! :-)

Filling the cars at the mine is a challenging job of switching loads around empties, but at the mill, We've had a lot of labor problems and even now with the new ROTARY DUMPER coming on line, nobody wants the job of shoving the cars into it- "repititious and boring" they say. It looks like we'll have to have a TRAIN POSITIONER too so that our feather bedding (Thursday night operators) crew can run a train in and sip coffee and tell lies while it unloads. You can see the challenge faced by our engineering department.

West Grand is capable of receiving up to 15 new loads, but the previous train of empties must be pulled out when the new loads arrive.

West Grand


The design of our dumper and postioner involves the obvious modeling challenges but the the real work is in four mechanical systems. When that's done, we'll add the modeling gingerbread.

Overhead view of the rotary dump Click on the thumbnail for a VGA (640x480) Overhead view of the rotary dump facilites at West Grande on the Teton Short Line"

Rotary dumper under construction Click on the thumbnail for a VGA (640x480) close up view of the rotary dumper under construction at West Grande on the Teton Short Line"

1. The BARREL must rotate up to 180 degrees to insure complete dumping of ore. Most dry product will dump at 135 degrees, but consulting my mining engineer friend, he says they rotate it 180 to insure dumping of wet product. Barrels seem to fall into three general classes. One type simple tips the car over the edge of the track, spilling the load into a bin alongside the where loaders can handle it. This take a lot of horsepower and very heavy above ground construction. A second type apparently rotates the car around an axis somewhere near the center of gravity dumping the ore into an underground hopper where a conveyor system hauls it away. This doesn't take much horsepower, and the mechanical stresses are less, but it takes a lot of underground construction. A third type rotates the car around an axis in line with the coupler and provides for quick unloading of a train equipped with rotary couplers without separating the cars. The offcenter load would likely take more horsepower, and heavier construction. These observations are those of an electrical engineer that has seen a few photos and manufactures literature, but has zero experience wih the prototype, so if I'm too far off base- sorry about that :-)

Our model will be of the third type, not because of the rotary couplers, but because the car sets higher in the barrel, giving us more undertrack space for mechanical stuff.

2. The CAR CLAMPING SYSTEM holds the car in place as it's rotated. In all cases, I found that four or more clamps are used. They may be separate, or paired with a span across the car. One system, Svedala, uses a hinged "breakaway" clamp that will swing out of the way if a car or loco were to hit it while it was lowered. For simplicity of construction, we have chosen to use the paired system that spans the car with the mechanics contained completely within the barrel.

3. The TRAIN POSITIONER. The prototype usually uses cables, hydraulics, or rack and pinion devices to move the cars into and out of the dumper rather than a loco. Sometimes they'll use a rubber tired tractor to shove with. Visiting with an "old hand" recently, he described the fine art of shoving cars into the dumper with a loco while dealing with the problems of slack in each coupling. We are using a closed circuit wire rope system enhanced with gravity.

4. The EMPTIES. If you've got a long shallow grade away from the dumper, it's just no concern, just let'em roll away, but we don't. The mill sets right at the edge of a deep gorge (its handy for waste- don't tell the EPA), so we'll let the empty car roll down a grade and through a trailing spring switch to a bumper. Another cable system then moves it out of the way to the left, assembling it into a train of empties.

If all continues to go well, we expect this to be a major animation attraction on the TSL. Just run a cut of cars into it and watch it happen.

(03/22/99) The TSL has turned the dumper over to our customer, Coots Mithril and Leverite products. It is a huge success, handling the unique ore- a granular royal purple and red product with rock ranging up to 5 inchs. This unique ore has not been seen since the Dwarves mined it at Moria in ancient times.


If we've still got your interest then here are some of the details and engineering sketchs. As you read through the following, please remember that I'm from the electric farm where they use solder- not the mechanical one where they use bolts and nuts. For the destitute TSL, I tend to design and build using the materials at hand, so this is NOT a do-it-yourself article- just take some ideas and run with them and if you've got some suggestions then e-mail me.

Dumper sketch- Side View

The BARREL is 4" wide and 4-1/2" long with the ends and floor made from 1/16" sheet brass salvaged from an old door kick plate. The sketch is to scale. The end discs were cut, trimmed and drilled together to insure alignment when they were soldered together with four 5/32" steel rods between them. The floor has a pair of code 100 rails soldered to it and is bolted to the ends at an elevation that puts the coupler center at the axis of the barrel. The NMRA standard S-1 specifies the HO coupler 25/64" above the railhead. Don't try to conserve on weight- the heavier barrel rolls smoother and is less affected by the car and load. Gingerbread will be added to the barrel for modeling purposes when the mechanics are working well.

A BOX FRAME of 1/8" plexiglass with two 1/4" cross members provides a support that can be bench assembled and later dropped into a hole in the layout. It will be sceniced to represent a concrete wall where it projects above the terrain. A catch pan underneath will collect the dumped product.

Dumper sketch- End View

(02/20/99 update) We initially used a "bull" gear on the barrel that was driven by a single pinion below, making the barrel free to lift out. It seemed to work reasonably until we slowed down the drive motor for realistic appearance and then it tended to lift the barrel and jump a tooth once in a while. The bull gear teeth salvaged from a K-Mart toy were triangular shaped causing the lifting action- Scrap that idea!

Phase two- After considering alternate drive schemes such as VCR belts, multiple pinions etal, we found a 2mm pitch timing belt, and a 14 tooth sprocket in a dark corner of the roundhouse. To turn the 4" barrel 180 degrees requires the belt to travel about 6.3" or about 160 mm. That means the sprocket must turn 160/(14*2)=about 6 times.

The BARREL BEARINGS. Looking around for support bearings or rollers, we went to the TSL wheel shop and came upon a source of plastic HO wheels and axles (typical Athearn) that were left over from wheel upgrades. The wheel flanges do their thing to keep the barrel in place. Drill the 1/4" plexiglass for the appropriate size brass tube bearing and reassemble the wheels and axles in it. Use ACC to hold the bearing in the plastic. We also used these wheels and axles to help guide the timing belt as you see in the drawing.

The BARREL DRIVING MOTOR is a hacked model airplane servo. These very popular devices usually rotate over a range of about 180 degrees in direct proportion to an input pulse that ranges from 1 to 2 msec, repeated about 50 times/second. They need 5 volts and have torque values ranging from 10 to over 200 oz/in. We chose one with 53 oz/in. It's quite easy to open up the device and disable the feedback potentiometer that senses its position and replace it with a pair of fixed resistors. The servo then rotates continuously, seeking but never finding it's commanded position, the speed and direction being controlled on one wire simply by varying the pulse width from our control circuitry. The typical maximum speed is about one rev/second (60RPM). That would rotate our barrel 180 degrees in 6 seconds, too fast, so we command it to run slower.

The BARREL ROTATION LIMITS are controlled with a pair of Photo-Transistors (PT), a tiny incandescent bulb and a semi-circular plastic vane that blocks the light from both PTs when the barrel is traveling between "Home" and "Dump" positions. The PTs are mechanically placed so that one is illuminated and the other is dark at "Home". The opposite is true at "Dump". If both are illuminated, the barrel has traveled beyond its normal 180 degree range and is in a fault condition-the motor is stopped, and an alarm actuated. Work area lighting is provided at both ends of the barrel with pole mounted lamps, but the real purpose of the lamps is to illuminate PTs set under the center-line of the tracks just outside the barrel ends to insure that a car is not partially in the barrel when the rotation occurs. These PT's are also useful to shutdown the works when we run out of cars.

The CAR CLAMPS are assembled from square brass tubing and 1/16" steel music wire as shown (red color) in the sketchs. The bell cranks are cut from the 1/16" brass sheet and are attached to the vertical steel rods with "solder quick links" from the model airplane world. Brass tube bearings and music wire shafts under the floor deck couple the clamp pairs. Guides made from 1/4" square brass tube and 1/16" i.d. round tube are soldered to the 5/32" steel rods for upper support. The resulting mechanism is coupled to an unmodified "standard size" model airplane servo that is positioned with a 5 volt variable pulse width, as described previously. The controller moves the clamps in a slow and smooth manner.

Two of our powered mechanisms are now working.

The POSITIONER. Most prototypes are built on flat ground, but we can use gravity to greatly simplify our system. A ramp slopes up about 4% and has a cabled car moving system embedded in it. The ramp is cut from a piece of knot free wood. The cable system is constructed in the center of the track using pulleys turned from 1/8" plastic with music wire shaft and and brass tube bearing. For the MDC ore cars, the cable system must engage an axle and move forward about five inchs. This will push a new car into the dumper where two Kadee delayed action (the wide one) uncoupling magnets, laid end to end, will keep all couplers in the delayed mode. They also tend to center the car because of the magnetic attraction to the steel axles.

Train Positioner

The new car will shove the old one out where it encounters a down slop of 4% and away it goes. Now the cable system reverses direction allowing gravity to back up the cut of loaded cars on the grade. After backing clear of the barrel, about an inch, leaving a car isolated in the dumper, a simple brake system (see sketch) stops the pusher car and the entire attached train, while the cable system returns to its starting position. If too many cars are NOT on the 4% grade, then the train might not roll back. Free rolling wheels are important, but we haven't found it necessary to change out the orignal plastic ones, and we surely don't want to lose the steel axles that aid in centering the car in the barrel.

April 2008. In order to finally win the battle of dirty track that is associated with plastic wheels, we converted the fifty MDC ore cars to Intermountains brass wheelsets. Centering became a problem but we overcame it by slightly notching the rails to fit the wheel contour. A simple fix that works even better than the magnetic centering

The POSITIONER BRAKE. Operated by gravity, the tines extend just high enough to block an axle, contacting it just behind the inside of the wheel. It's important to avoid interferance with the coupler.

Brake or Retainer

The POSITIONER DRIVE. We need five inchs of linear motion. The sketch shows the basic mechanics of a 4 RPM (salvaged) motor that cranks a lever through 60 degrees. The lever in turn is linked to a nylon slider on a piece of 1/8" steel welding rod. Crudely assembled on a piece of 24 guage galvanized steel metal using music wire for shafts and levers and brass tubing for bearings. A micro switch (not shown) is operated by a cam on the motor to identify the stopping location for the control circuitry. The motor current is monitored to detect a jam up and shutdown if needed.

Positioner Drive

The POSITIONER GUIDE between the tracks. One side is ground off of a square 3/32 tube (K&S) making a U-piece. Two of these are attached to the ties between the rails, positioned like a guard rail, and form a pair of guide channels for a piece of 1/2" x 3/4" brass sheet "boat". The boat is driven by the cable system (1/2A model airplane lead-in wire) that passes over the top of it and is attached with a spot of solder. (sketch)

The POSITIONER AXLE ENGAGER. Operated partially by gravity, this mechanism is assembled on the "boat" that slides between the rails. When moving up the slope, it engages and pushes the first axle that it encounters.
(03/03/99) If There are several cars in the string, gravity alone will cause them to roll back, but if only one car remains, the Kadee uncoupling magnet tends to hang onto it, so the pusher has been modified to drag the axle back a bit until the "brake" captures the car. The tiny coil spring on the pusher boat then yields allowing the pusher wires to drop free of the axles and be dragged back..

Boat-axle pusher

Positioner boat Click on the thumbnail for a VGA (640x480) close up photo of the train positioner mechanism between the rails

Three of our powered mechanisms are now working.

The EMPTY CAR MOVER. revised April 2008. The empty car that has been shoved out of the dumper rolls downhill (to the right in the track plan), through a spring switch where it encounters another car positioner mechanism. This device will push the car to the left about 12-1/2" where it encounters a string of accumulating empties. The empty string is pushed far enough to clear the turnout. The pusher device then returns to it's starting place. We could not make the spring switch (turnout) reliable with the light weight empty cars, so it is mechanically thrown with linkage to the cable system underneath. Again, a cable system is used, but the pusher is a freight truck riding on the rails. We didn't use the boat idea because of complications with the turnout. The cable lies between the rails and must pass thru the turnout so it requires another frog per the sketch. Part of the cable that wraps around the motor capstan is 30 pound fishline, The remainder, between the pusher truck and limit switchs is the same same steel cable used in the car positioner, to maintain accuracy.

Empty- track & cable
The cable has to travel about 12-1/2" so it was impractical to use a cyclic actuator such as we did for the car positioner. This final mechanism has been through several iterations to get it reliable. Like the barrel drive, we uses a "hacked" servo to drive it. The pusher travels the distance in about 11 seconds.

Pusher truck

The pusher truck has a specially designed shock absorbing Kadee coupler that works much like the prototype systems used to reduce cargo damage. The Kadee #5 draft gear box slides about 1/4" in a pair of brass U channels against a coil spring. Some of our cars roll easier and faster than others. The pusher might derail if hit too hard so it parks under a small frame (not shown) to prevent that from happening. As the pusher leaves home, the switch is quickly set NORMAL with a linkage underneath.


The empty car pusher is viewed from underneath in this photo. First note the pulleys on each end. On the left end you see the hacked model airplane servo. The servos internal stops and feedback potentionmeter have been disabled so that is now a simple gearhead motor. It's an economical device that we can be easily control, in speed and direction with the firmware. The motor capstan is turned from a scrap of nylon or teflon, not sure which. The cable is two pieces. From the pusher cart above, 30 lb nylon fish line comes down around the left pulley and makes four wraps around the motor capstan and then connects via a tension spring to the slider. The four turns allow the line to slip on the capstan without damage if something goes wrong up top. The other piece of cable is metal control line used in model airplanes chosen for dimensional stability. i.e. it won't stretch and lose calibration. From the pusher cart it comes down around the right pulley and attaches to the slider. The slider is two small brass plates soldered to a short piece of brass tubing. It slides on the 3/32" rod to impact the micro switches that limit the travel. Look closely at the left microswitch and you'll see another thin rod that extends to the right with a figure ess bend. This works the track switch when the slider reaches its maximum left position, and the push cart is at rest awaiting an empty car to roll into it. The firmware permits the motor to run slightly longer than it takes for the push cart and slider to move from one end to the other in case of a jam-up. Normally a limit switch stops the action.

(03/09/99) The fourth and final powered mechanism is working.

CONTROLLING the whole mess without manual intervention is a challenge because we have to consider such faults as a derailed car, failure to uncouple, a car part way in or out of the barrel et al. The area will be flood lighted for day and night operations, so photo sensors work nicely to verify car position(s). A jam up in the positioner mechanism will be detected with motor current monitoring. It is necessary for an overseer to be able to PAUSE the operation if he sees a problem so a RUN/PAUSE switch is provided. The only significant long range problem has been keeping the Kadee couplers on the 50 ore cars in near perfect condition. The current program takes a little over two minutes per cycle, so a 12 car train takes about a half hour to unload.

We're using a Parallax Basic Stamp2 (BS2) controller This nice little device has 16 I/O terminals that can input our photo sensor and limit switch info and output the various motor and servo commands. You program it with a simple BASIC like language. If you're into this sort of thing and would like to look at the BS2 program, drop me an e-mail- I'll send the most current version.

As of now (03/09/99), the car clamping and barrel rotation is working perfectly and has many hours of excercise. Even took it to the Hostlers big show in Odgen UT last week. Ran it continuously all day on a ni-cad battery. Work continues on the rest to increase reliability. Gee- this is more like robotics than model railroading (VBG) Stay tuned- we'll keep you informed and if you got any good ideas, please share them.

(03/22/99) Today, we commissioned service with the new dumper. Works great.

THE ORE was difficult to find. Many materials were explored to simulate the Mithril ore. Most of the popular commercial scenic materials are either light weight foam or mineral (rock). I felt that the rock products might cause excessive erosion and the dust could clog the clamping mechanism. The ultimate answer was crushed and screened walnut shells used to polish brass. I found them in a gun shop. They are heavier than water and take RIT dyes very nicely. When I found a color that my mining and geologist friends could not relate to any know product, I was satisfied that the mythical Mithril ore from ancient times of middle earth (Tolkien) had been re-discovered VBG .

MITHRIL is a very unusual material used in ancient times for the very best swords, armor and especially impervious chain mail. Legend tells us that it was mined by the dwarves at mines of Khazad-dum (later called Moria) in the Misty Mountains of Middle Earth and traded with elves and men. Unfortunately according to the legend, the dwarves mined too deep and released an undescribed terror that destroyed the under-earth clans. Rediscovery of mithril after ages of platonic action would be a matter of rare chance, but it looks like our customer, Coots Mithril and Leverite Ores and their subsidiary Ross's Rolling Rock Mine may have lucked out.

Return to Teton Short Line Home Page