The Kerbin Locomotive Works

[Ultimate Steve]

Hello! After watching James May make the longest ever toy train journey for a TV special, it has re-awoken my fascination with steam locomotion and trains in general. I got out my old electric train sets and messed with them for a bit, I learned some amount about the real life Flying Scotsman, and I've attempted to create vaguely locomotive like vehicles in KSP. Originally I wanted to do it with actual pistons and stuff, but the only requirement now is that it must use engine thrust to move mechanically (that engine thrust may not actually be used to move forwards).

This is probably going to be a one off. Maybe I'll make a challenge based off of it.

My early experiments took the whole "train" thing very literally. The train was to be powered by pistons actuated by toggling twitch engines on and off with an action group. This was incredibly finnicky but I was able to get it working reasonably reliably, although I had to hit "1" repeatedly at the right time. The first issue was that fairings, while they are great at actually rolling, will explode if you go above 7m/s.

My next attempts used these Mk 3 to 3.75m adapter parts, as they are the only large circular parts with an impact tolerance above running speed. However, the ride proved incredibly bumpy as the hitbox doesn't have enough sides to make it a good enough cylinder for this purpose.

I tried covering the wheels with flags for a more circular wheel, with the knowledge that they made good takeoff gear, but in that application, they were constantly on the ground, not repeatedly impacting the ground. The flags blew up once speed was made.

I attempted to automate the piston changing with the KAL controller, as there's no way I was going to be able to stand pressing "1" every few meters for an entire long journey. Once you get it at the right speed (easier said than done), it can keep going that speed more or less, but asking the KAL controller to change speeds, even slowly, led to it desynchronizing and slowing down instead. This inability to change speeds was the downfall of the KAL method. Even if you could slowly change speeds, going up a hill would change the speed more rapidly than it could likely cope with it.

The solution would ideally be some mechanical interrupt connected to the wheels that blocked the thrust of the engines at the right time. You wouldn't even need a KAL. But I really did not want to deal with moving parts more than I already had, so, in sadness I abandoned the piston based approach and instead attempted a sort of paddlewheel based approach. While this is a departure from conventional steam locomotives, steam turbine locomotives were also a thing, briefly, so this is technically still reasonably realistic!

The early attempts were simply the previous vehicle with one heat shield paddle per wheel.

This worked pretty well and I was soon able to surpass my speed record. This is however when I found out that flags didn't like going that fast.

I also discovered that the vehicle would just refuse to exist going faster than about 20m/s. This turned out to be because of the common axle on the rear wheels, which wasn't needed any more because I had abandoned pistons (for the time being, at least).

The next prototype upped the number of heat shield paddles to 8, 4 per side, and this, unsurprisingly, worked significantly better (and was significantly more fuel efficient).

This version tested structural panel paddles and smaller engines, evidently without much success, as I quickly changed back to Thud engines.

At this point I should probably say that I was having fun giving the engines increasingly British names. The lineup so far was:

L01 Hufflepuff

L02 Raleigh Flow

L03 Prince Edward

L04 Reginald Langley

L05 Rupert of Birmingham

L06 Bristol Faerie

L07 Newcastle Challenger

L08 Newcastle Challenger II

L09 Spirit of Southampton

The L10 Brisbane Challenger (I did not know Brisbane was an Australian city until I looked it up just now haha) was a pretty substantial redesign. While it kept the paddles from the previous locomotive(s?), the wheels were changed to structural panels and grip pads, as they had pretty good impact tolerances. The drive system was two Thud engines, this time moderately angled down so they wouldn't be so low as to be in danger of hitting the ground. They were fueled by just over an entire orange tank of fuel, in anticipation of a long range journey, as fuel economy was significantly better than previous piston based craft.

The Brisbane Challenger quickly surpassed the current speed record and was easily on track to break the distance record!

However, she quickly ran into a new problem: The vehicle was going so fast that the paddles didn't have time to cool down before hitting the engine again! Thus, the throttle had to be significantly limited.

Like the Flying Scotsman becoming the first locomotive to reach 100mph (about 45m/s), I wanted similar glory, so I pushed my luck. After the loss of several paddles, the engine just barely didn't reach that speed (or maybe just barely did, this is the only screenshot I have, it could have gone slightly faster). It attempted to stop before the end of the runway, but in my efforts to make the vehicle realistic, the only brakes were the main wheel bearings with the sliders set low.

Thus, the vehicle did not stop in time!

The Brisbane Challenger II was slightly changed, removing the grip pads on the wheels, and slightly increasing the paddle and wheel diameter for more torque, which resulted in improved performance. It soared past the end of the runway at nearly 50m/s, and promptly crashed into the water. 

The Brisbane Challenger III placed heat shields over the paddles for better protection in an attempt to reach a higher top speed. During speed trials, it reached a speed of at least 56.8m/s, but at the cost of destroying nearly the entire paddle system. Jeb had to bail out with his parachute when the locomotive couldn't stop before hitting the water.

After another iteration, the Brisbane Challenger IV tested out some stuff, the design was upgraded into the L14 William Tell Overture (which is also not a British song like I thought it was, it is an Italian song for a French opera about a Swiss folk hero).

This was not a speedster, it was a long distance vehicle. The initial goal was to reach the end of the runway, but now the goal was to go as far as possible.

The craft reached the edge of the space center and encountered a slope for the very first time, which it tackled fairly easily.

Throttle was limited to a little bit over 2/3 for most of the journey in accordance with the results observed by the Brisbane Challengers. 40-45m/s was about the safe maximum speed on level ground. Going higher in speed or throttle risked paddle damage.

The William Tell Overture made it a whole 21 kilometers before running out of fuel!

I decided to sacrifice vibes a little bit and switch to jet engines for their vastly superior efficiency, though rockets kinda seem significantly more steampunk to me. Maybe it is the noise. Maybe the inefficiency is the appeal. Ideally this whole thing would have been rocket pistons, but I wanted to reach the mountains!

Initially I powered the vehicle with twin Wheesleys, but the exhaust temperature was too high and it kept melting the paddles. At this time I also started looking at adding more train cars. In this case, just a fuel tank. Partially to test towing capacity, partially to make it as far as possible.

However, I had a really hard time keeping it straight and powerful enough to go up hills, especially given that I had to limit the throttle on the Wheesleys. Weirdly enough the Thuds gave me a lot of advance warning before the paddles exploded, but not the Wheesleys. After a lot of research and development trying to get the tender to stay straight and the engines powerful enough (L15 Willaim Shakespeare, L16 William Penrose), I found the solution: Cluster a ton of Junos together. I think I ended up with 8 per wheel. The temperature is not an issue (you can go full throttle), it looks neater than the Wheesleys). To keep the trailer straight, the two rearmost wheels had their friction controls turned up and were canted outwards by 5 degrees to provide some ground friction/drag. To maintain steering control, the friction control on the front pair of wheels was increased to 0.7 and the steering limiter was disabled. Previously all wheels had zero friction control as that helped solve other problems. The main drive wheels are very slippery side to side but very draggy forwards-backwards, at least compared to normal landing gear.

The L17 William Tell Overture II set out on a journey to the mountains, modifications in hand.

The locomotive made good time, and, at least for the flatter bits, a combination of MechJeb Smart A.S.S. and MechJeb Rover autopilot was able to keep it within 5-10 degrees of where it was supposed to be pointed, at least on the flatter areas. However, a new problem appeared. With a load this large, and going fast over terrain borders, the wheels liked to skip (either due to being momentarily airborne, or just slippery in general). The wheels were also kind of floppy, getting crushed like tires. While this was good for grip, it was not good for longevity, as wheel segments would also have some vertical velocity when slamming into the ground, increasing impact velocity.

Ideally the pieces should hit the ground with zero impact velocity each time under normal circumstances, up to any speed, but as we saw with the flags earlier, KSP still considers them to be impacting at least as fast as the vessel is moving for weird KSP reasons.

After it became evident that wheel lifetime would be the limiting factor for this mission and not fuel, power and speed was reduced somewhat to extend the life of the wheels (which were already significantly damaged 13 minutes in). At this point only about a quarter of the orange tank had been used up. The rear tank had enough fuel inside to refill the orange tank twice (it was half empty to reduce mass). If treated gently this thing could go very, very far.

The train made it to the base of the mountains without much trouble. I believe I slightly reduced throttle after it became apparent the wheel issues weren't one offs.

The new goal for the Kerbin Locomotive Works was to make it through the mountain range. Not around, through! This may be not a reasonable goal, and would probably require significant scouting, because I struggle to make it up that mountain range in a normal rover.

These dreams were dashed fairly quickly as a 20 degree hill (ish, the locomotive is at a slight angle usually that changes with how fast the wheel is moving because it is stretchy) quickly killed the locomotive's momentum and brought it to a standstill. As it started to roll backwards, fuel was transferred from the tender to the engine, and the tender was cut loose.

The L17 William Tell Overture II once again began to accelerate up the hill, an impressive feat for real life locomotives, but probably less than what is required to make it over this mountain range.

However, with already missing wheel segments, sharp transitions between terrain tiles, and the high power setting, the wheel slip problem only got worse.

The vehicle also slipped to the side and began going back down, the wheels unable to correct.

The vehicle came to a stop under a kilometer from where the tender ended up, maybe even lower than when the tender was released, the wheels thoroughly shredded. I nevertheless adjusted the steering wheel settings and turned back towards the mountains.

But the trip proved too much for the engine, and the wheels soon gave out, causing the locomotive to flip over and become destroyed.

Note the now non existant wheel tread. After a large enough portion fails, the rest will soon follow as the wheel will accelerate a lot without resistance (being in the air because there's no wheel tread in that area).

Nevertheless, I loaded a quicksave and tried again.

This time, an even steeper slope was chosen, and the locomotive appeared to handle it well, until it crested the top. Its rear wheels became momentarily airborne, and the left wheel was completely shredded.

The vehicle tumbled down the hill a little, but came to a stop in remarkably good condition.

There was another engine, the L18 Spirit of Luton, which was a test to produce more durable wheels. Unfortunately the modifications (a second layer of plates and autostrut) didn't appear to help much, only managing to limit the vehicle's speed to 35m/s, and a breakthrough in wheel durability would have to happen for me to consider longer journeys. Nevertheless, the concept is likely capable of at least reaching the north pole. The current locomotive is probably good to 500 kilometers. The pole is only 950 (probably more accounting for routing to avoid mountains).

I briefly considered making train track, but the part count would be prohibitively high for even 100 meters of track unless I designed it to run in something really weird looking like an open cargo bay (and even then, the cargo bay would probably explode before the wheels do). Kerbal Konstructs may be viable but I would probably have to go through the hassle of installing a 3D modelling program again and actually making the track pieces. There is at least one other person doing Kerbal Konstructs trains right now but from what I know, it will not release for a long time.

Part of me wonders if this technology actually has any practical use cases. Assuming the heading control and wheel durability problems can  be solved, I think I would rather drive one of these than a solar-electric rover, as they tend to eat power like crazy when encountering even moderately steep slopes. They also tend to want to flip over a lot. Except for not being able to accurately stick to a heading, the L17 was a joy to drive. I never felt in danger of flipping (except when tumbling down a hill), and I didn't have to stop to charge, or pay much attention to how I was driving, as the efficiency curve is significantly flatter for this design compared to rovers. I have not yet attempted to physics warp with this design, though, as I fear it would go poorly. Granted, that is a problem that stock wheels also have, you have to tune them to behave during time warp. Fuel cell electric is probably better than this.

I'm not 100% sure, I'd have to do some more tests, but it might not be any better (or it might be significantly worse) than just using 16 Junos for their thrust directly. But it sure looks a lot cooler!

And then, of course, their efficiency goes down significantly if you use them on another planet.

 

That concludes my research into mechanical engine driven craft for the time being! I don't intend to do much more research at the current time, but if you have any ideas for making the wheels more durable, or know of any Kerbal Konstructs train track stuff, let me know! If the problems can be fixed I may attempt a run to the north pole.