I haven't seen any discussion on this but there's an excellent 14 page article in Classic Trains Steam Glory 3 (the one with a NYC Niagara on the cover) about Loco Valve Pilot.

They don't go into too much detail about how the cam at the heart of the Valve Pilot unit was designed. Can anyone elaborate?

The author says that only one Valve Pilot-equipped engine (B&LW 643) has operated during the preservation era but two may in the future (B&M 3713 and SLSF 1352).

I don't suppose anyone has explored the idea of moving a unit to an operable engine, provided someone could figure out a way to machine a new cam. This isn't so much to improve the performance of an engine or its crew but to help preserve this amazing bit of technology which was a sort of mechanical computer. Hate to see it fade away.

Another idea would be to simulate a Valve Pilot digitally using a rugged tablet computer with speed, throttle, and valve gear setting inputs. It could be self-learning. Whenever the engineer feels that they've got the engine hooked up in an ideal way, just reach over and touch the screen and eventually a mass of data will be accumulated that could later on offer an advisory about where the valve gear should be in a similar situation. (Of course a digital version could include boiler and back pressure and other inputs.)

Anyway, just thinking outside the box.

Steve

Hi Steve,
Glad you enjoyed the article! Excellent question on cam design. The first step of the design process was ascertaining the position of the tumbling shaft (or lift shaft, if you're an N&W man) for each different cutoff percentage. Basically, the tumbling shaft is the piece that connects the reverse lever to the valve gear proper. Valve Pilot Corp. would send a paper graph to the railroad. They would mount the graph paper on the tumbling shaft and put the engine on valve setting rollers, which allow the main driver, main rod and valve gear to operate while the engine stays stationary. They would then put the reverse lever "down in the corner" (full forward gear), mark the paper, and determine the cutoff percentage by looking through the steam chest inspection ports and measuring the crosshead travel from top dead center. They would then hook the reverse lever up a notch, mark the paper and repeat the process until they got to center (neutral, essentially). Due to the geometry of outside valve gears, the cutoff changes very little when you are close to full gear, and it changes dramatically close to center. They would then mail the graph to VPC.

Once VPC had the graph, they had one of two options. If extensive dynamometer test data for that class of locomotive was available, it was relatively straightforward to lay out a cam that would move the cutoff hand to the proper speed on the instrument dial (say 75% cutoff at 17 mph, etc). If the locomotive was new and no testing had taken place, then alchemy and a bit of inaccuracy entered the picture. First, VPC engineers would use driver diameter and stroke to figure piston speed for a certain locomotive speed in mph (if driver diameter stays constant, and stroke is increased, the piston will be moving faster at the same locomotive speed and vice versa). They would then go to published test data, such as the books published by the Altoona test plant and other railroad test data to find a locomotive with similar bore, valve size and front end arrangement. They would determine what specific cutoff (say 30%) at a certain piston speed (say 1400 ft/min) produced maximum horsepower on the test locomotive (say 56mph), then determine how fast the new locomotive would be traveling when the piston was moving at 1400 ft/min, and that was the proper speed to be running a 30% cutoff for maximum horsepower. Due to differences in design of steam passages and exhaust nozzles, among other things this would be in the ballpark, but not perfect.

You are correct that you could run a Valve Pilot on any locomotive just by changing the cam. You could work out the new cam profile in the same manner as the VPC engineers did. If you wanted ultimate accuracy, you would have to run dynamometer tests with the locomotive in question. Conceivably, you could do this by coupling to a modern diesel that has computer readout of dynamic braking tractive effort, and pulling against the dynamic brakes.

Using a computer to log when the engineer thinks he has the engine "hooked up just right" would not replicate Valve Pilot readings, as each engineer has his own idea of what "just right" is. One source I read dramatically illustrated this point. The guy was part of the team at the Purdue test plant helping WFM Goss test the newfangled superheater back in the early 1900s. He claimed the superheater could show a 20% gain in efficiency, no gain in efficiency, or a 20% loss of efficiency depending on which crew was running the engine. That is why the Valve Pilot was so important. It allowed every engineer to get perfect "test plant" results out of a locomotive every time. Thanks again for your kind words and your interest. -Chris Zahrt

Thank you! Very interesting.

Steve

One of the bests posts and replies I've seen in a long time.

Good morning!

I found references on Google Books to the following while trying to find more information on the Valve Pilot:

Locomotive Control: The Valve Pilot and how it Increases Efficiency in the Operation of Locomotives
by William J. Gaugh
New York University College of Engineering - 4 pages

http://books.google.com/books/about/Locomotive_Control.html?id=olNqHQAACAAJ

Valve Pilot Tape Talks
by William Bell Wait
Valve Pilot Corporation, 1943

http://books.google.com/books?id=ITMPGQAACAAJ&dq=%22valve+pilot%22&hl=en

You may be able to find the books through an inter-library loan. E-books are not available.

You can find other references in Google e-books at:

http://www.google.com/search?q=%22valve+pilot%22+steam+locomotive&tbm=bks&tbo=1&gs_l=heirloom-serp.12...351428.355863.0.357949.13.13.0.0.0.0.349.993.12j3-1.13.0...0.0...1c.1.3KHNd8Ya9cU&oq=%22valve+pilot%22+steam+locomotive

Have a joyous day!

PR