Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[omega_dwarf]

Has anyone ever attempted a fully-automatic minecart machine gun?

Shotguns usually work with a minecart slamming into an obstacle and spraying its contents at a target. That, however, requires a lot of manual cleanup. What if, instead, a boatload of heavy-material minecarts were created, stored, and unleashed in rapid succession at Very High Velocities^TM along a narrow corridor or - even better - stairway? An automatic collection system could be set up by laying tracks along said corridor and having a collection chute at the end.

Corollary: Can minecarts be loaded into other minecarts for rapid manual transportreloading, or would we have to make a powered minecart elevator system as well?

The prize for creating such a weapon is set at 1,000,000 Urists. Bonus 500,000 Urists if you can get the carts to travel a parabolic arc from a firing hole into the kill zone. Bonus 500,000 Urists for completion without a horizontal array of impulse ramps. Bonus 750,000 Urists for an infinite magazine (replacing carts at exactly the rate they're being used, no gaps. Only using one cart is cheating )


Mostly I like this because it's less glitch/feature abuse than shotguns (to my understanding.) Cleanup from a shotgun could in theory be automated using sloped surfaces and liquids to concentrate the spent ammo into a quantum stockpile for rapid loading into a minecart, which would become the next bullet. (It's a shame you can't dump items into a minecart from above with floor hatches/retracting bridges. But you could still do cool things like using magma to push the items, reducing goblinite to useful iron/steel/bronze/silver/copper bars, using gold bars for ammo, and sorting the items at the quantum stockpile into two separate minecart routes, one to the forges and the other back to the shotgun.)


Oh, and also, has anyone ever completed the water shotgun described here?

That same page also claims you only really need a three-z ramp to reach lethal velocities. Anyone know the truth of this?

[Delioth]

I mean, shotguns aren't really a glitch.. they just make sense. Minecart full of spears travelling at high velocity + Sudden stop for the minecart = bunches of spears with enough momentum to fly a decent distance with kill potential because sharpness. Same concept as people w/o seatbelts flying out windshields with high-speed collisions.

I believe minecarts will 'stack' up a z-level or two (so if I put one minecart and drop 7 on top of it, there'll be 1 minecart on the track and 7 on top.) I think they share velocity until the top ones or the bottom ones hit a wall, but I've not done research on this.

Also if they follow an arc, your kill zone will be very small. Minecarts are relatively accurate (although flying spears are as well, but that's because 1000 spears over a 5x5 area gives you a giant sample size.)

[omega_dwarf]

Eh, maybe I misunderstand shotguns. I always thought the minecart hit a wall and its contents warped through the wall and shot out the other side, while the minecart remained behind. If there's some other way to stop it that suddenly, then yes, it makes sense.

So with the stacking, physically in the game, there would be 8 z-levels occupied by minecarts supported by each other, in your example? Or 2 z-levels? Or just 1?

So minecarts launched in a straight path can diverge from that path? I meant for the minecart to land on tracks and then hit the target, not to land on top of the target. But that was just for style points anyway

[Delioth]

Well, some shotguns use a wall to hit, and projectiles go through an empty space above where the minecart was, but I use a fortification at the end of my barrel. Makes the whole thing nice and 1-z so I can watch the flechettes go from cart into hippie caravan all at once. Cart stays safe behind fortification, and by the time the thing's shot there aren't any threats (I had 2 of them set up with a crossfire approach. ~2000 spears went through the trade depot, one of them shot into the 1-wide fort entrance, the other shot into a carefully constructed wall ~30 tiles back to catch all the spears. Made clean-up much quicker (didn't have to go 1/2 way across the map to get all dem spears) and was still effective.

I believe (like I said, I never actually did this, just read somewhere) that it stacked them up to 2-z total- 1z of the first minecart, infinitely many on 2z.

Yes, they can diverge if given sideways momentum. However if they're launched without any, they will follow a singular path. You can actually use this and the fact that different weight carts will travel different distances to make a sorting system- ie wooden carts with wood in them will fly this far and then turn off into this stockpile, while carts with steel axes will only go this far and turn to this stockpile &c. Or you can use it to make the entrance to your fort a minecart leap at full speed across a 7-wide magma moat populated with magma sharks. with lasers on their heads. (lasers may break time period immersion. I take no responsibility for immersion breakage.)

[omega_dwarf]

That sorting method is like mass spectrometry. Is the sideways momentum killed off by riding on straight track for a while, or does it glitchily still exist after that?

I've never done much with dwarves riding minecarts...perhaps this fort will be the time for it.

Regarding shotguns: ah, okay. So it's not 2000 quantum-tunneling spears, it's either 2000 perfectly-aligned spears that slide through the arrow-slits in fortifications, or 2000 spears which all jump over the wall and out (the second one is possible with a sloped forward interior of the minecart, I guess.) What pop is that fort, out of curiosity?

Interesting tidbit about the minecarts. I could definitely use that to my advantage. Will those minecarts stay stacked while a dwarf guides the bottom vehicle? And how would the game treat 1000 minecarts suddenly occupying the same minecart track (if I bumped the bottom cart into a wall to dislodge the top ones onto a track 1-z above)? Are they allowed to just clip harmlessly, or will they explode along the tracks in both directions, Minecraft-esque?

[gchristopher]

The wiki does link to QuantumMenace's Wave Cannon. In particular, that design was first to demonstrate the 1-z barrel firing through a fortification, fully-automatic repeat fire, and removing water from the firing range using overhead pumps pulling up through grates. My squirt gun vs. clown rush is slightly more "compact" (though still huge) and demonstrates the effectiveness against otherwise difficult opponents. Also, "squirt gun!" There are a lot of other good water cannon designs.

Several parabolic designs have been made, but I've never seen one that could reliably aim outside a relatively small pre-set target area. I think the problem of a broadly aim-able parabolic minecart shotgun is unsolved.

Automatic recovery of ammunition from a non-liquid shotgun will have some problems from bridges and hatches being targeted by building destroyers. Moving items via water flow over long distances or through fortifications had problems in the past. Is the bug where items moved by water flow become invisible resolved in 40.24? I haven't tried.

One really interesting, hard-to-implement idea is to use syndrome-tainted water or coated projectiles. I remember seeing a lot of science on the topic, but don't know how successful anyone was.

[gchristopher]

Interesting tidbit about the minecarts. I could definitely use that to my advantage. Will those minecarts stay stacked while a dwarf guides the bottom vehicle? And how would the game treat 1000 minecarts suddenly occupying the same minecart track (if I bumped the bottom cart into a wall to dislodge the top ones onto a track 1-z above)? Are they allowed to just clip harmlessly, or will they explode along the tracks in both directions, Minecraft-esque?

As of 34.11, if there are 3 or more carts above the bottom one, exactly one will fall from the stack of "ghost" minecarts. (They're actually projectiles in flight, held up by the bottom cart, according to dfhack.)

If there are 2 carts, they both fall, become real in the same tick and interact push each other away.

That was 34.11, and pretty quirky. I'd be unsurprised if it changed.

[Delioth]

I believe the sideways momentum is killed pretty quick (10 tiles or so seems like what I remember, but that could be wildly inaccurate.) Sadly, I cannot give a population of that fort. I think it died and I'm pretty sure it was .34 anyways. Either way it was culled when I cleaned out my hd some time ago. Seems like the fort may have been 80-120ish though. That's usually about the time I implement a militia (since I get annoyed at having so many idlers). Cranked out spears like a boss though.

[Larix]

Using weighty minecarts themselves as weapons is feasible and in fact makes the easiest minecart defence mechanism - if you use one of the "minecart grinder" concepts. The core premise of those is to have a fairly short path on which a fast cart cycles or reciprocates; any "shot" models work very poorly, because they still require "reloading" after the deed, while a grinder can stay active and effective forever - you might even need to install a dedicated _stopping_ mechanism to interrupt it for maintenance. Grinders with multiple carts on the same track also tend to work poorly, because once a goblin gets inside, the cart colliding with the goblin will slow down, and even a short slowdown will desynchronise the carts, leading to jostling and pileups.

"Aiming" ballistic shots through sideways velocity appears to be rather difficult - in my experiments, minecarts given lateral speed only rarely recognised launch ramps. They typically failed to launch entirely, banging into the adjacent wall instead.

Cart weight doesn't affect movement, btw. - the movement parameters of a cart depend only on friction, acceleration and current speed. A platinum cart will fly/roll exactly as far as a featherwood cart. Weight makes a difference when "momentum" changes, which afaik only happens through collisions (cart vs. cart collisions can give momentum, collisions with creatures take away momentum; changes in momentum have a greater effect on ligher carts). Ramps, rollers and dwarven pushes directly grant speed and don't take weight into account.

Incidentally, i'm not aware of _any_ notable changes to the minecart rules since 0.34.11. If you've observed some quirky minecart behaviour in that version, chances are very good that it's still the same now.

[Sanctume]

A water filled minecart, used in a shotgun fashion is fatal.
So just put that in a circuit to auto fill and auto shoot and you have a machine gun with infinite ammo.
I think it's referred to as the wave cannon.

[Delioth]

Ah, I believe that I had a detail wrong in the sorting mechanism- with your details, I remember that it wasn't differing fly distances, but rather it triggered pressure plates to lower drawbridges until it was the weight it needed to be, wherein it ran into a bridge-backed curve to divert to the proper pile.

Another attempt at aiming that I've heard of but never actually seen in action is to actually launch the minecart from the cannon... and then hitting it mid-air with another minecart. The collision will send projectiles flying (along with both minecarts) along the roughly combined trajectory. I believe this one could be customized with different cart weights (momentum changes differ, so higher weight carts in relation to your launch cart would divert further from the straight-and-narrow, while featherwood carts should divert relatively little.). Requires significant timing and science to make the two carts actually collide midair. Even moreso if you want there to be different aims with wood, iron, and platinum minecarts... and both directions too. Lots of work.

[omega_dwarf]

That's all interesting...it would be really cool to get that mid-air collision thing working (wouldn't it be as simple as dropping two minecarts onto the same length of ramp, and placing the ramps perpendicularly at two diagonal corners of a square?) I didn't really need the gun to be aimable, though. I'd actually rather not go through the trouble. I'm fine with a small, pre-made kill zone in the approach to my fortress. It would just be style points if the carts could be launched from an ominous hole in the ground. Somehow futuristic if that could work. Like rapid-fire minecart artillery, or something.

As for the water guns...that feels very glitchy (the way the water is treated as a single boulder), and the grinders lose the cool factor and also feel a bit glitchy. That's why I'm interested in making a machine gun with legit rollers, downhills for acceleration, a large magazine of carts, etc. The reloading is only a problem if it's slow. My idea is to keep the carts moving at a rapid pace. That's why having the gun shoot down a sloped rampway is possibly the best setup - even if the carts stall or even reverse, they'll eventually head down to the collection chute.

As for building destroyer issues, the sensitive parts can be blocked off by bridges until the building destroyers are dead from shotgun. But ideally, my project wouldn't require any cleanup stage, and should be less vulnerable to the buggers.

[Larix]

Relevant reading:
http://www.bay12forums.com/smf/index.php?topic=126669.0
Ballistic shooting of minecarts, including using them as shot containers and some thoughts on aiming.

Less talking, more building

PS: the likely biggest headache will be carts getting desynchronised and jostling each other, leading to newton's-cradle stop-and-go patterns or full pileups, often requiring a lot of work to disentangle/reload the thing after use.

Sending carts back to a stack magazine puts a lower bound on the distance between carts, as well - carts take time to settle into a stack, and that time gets longer the more carts are in the stack - 12steps +2 for every cart already present i think. I managed to stick 17 carts into a one-track magazine-fed minecart magma lifter (powerless, thus exploity as hell), but had to use a 52-step interval between carts to avoid pileups: http://mkv25.net/dfma/movie-2733-multi-cartmagmadelivery

[omega_dwarf]

Yeah, that's why the slope is important; it should naturally resolve any pileups that happen.

If that's an issue, then it sounds like I probably won't be using stacked carts. That could cause difficulty in keeping the weapon firing continuously, anyway, especially if the stack is dwarf-hauled.

Unfortunately, time for building is in short supply until the weekend, so it has to be just plans for now.

Can one consider an impulse ramp array (say 50 tiles long) to be the equivalent of accelerating a cart down a 50-z ramp at 45°? That would help with determining what the size of this mechanism would be like.

[gchristopher]

Relevant reading:
http://www.bay12forums.com/smf/index.php?topic=126669.0
Ballistic shooting of minecarts, including using them as shot containers and some thoughts on aiming.

Thanks! I wasn't able to find that thread.

Sending carts back to a stack magazine puts a lower bound on the distance between carts, as well - carts take time to settle into a stack, and that time gets longer the more carts are in the stack - 12steps +2 for every cart already present i think. I managed to stick 17 carts into a one-track magazine-fed minecart magma lifter (powerless, thus exploity as hell), but had to use a 52-step interval between carts to avoid pileups: http://mkv25.net/dfma/movie-2733-multi-cartmagmadelivery

Really? I need to go test that now. I was pretty sure that stacking and de-stacking weren't the rate limiting factor, at least for water cannons? My memory is unclear. For a cart to drop down the end of a barrel takes 11-12 ticks? But I don't recall the stack being a time issue. Are you talking about the time for a cart to fall from the stack tile into the newly-vacated bottom tile, or for a cart to fall down the barrel and merge with the stack tile?

The pass-through track stop where you dump the magma works very smoothly in that video. I never got dispensing to run that cleanly.