Print

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

[QuantumMenace]

In my first watergun thread, I mentioned working on an updated design. After numerous delays, including being driven to make the whole thing out of clear glass, it's finished:

Spoiler (click to show/hide)

The 3 layers of the Wave Cannon setup from top to bottom.


The chamber stays pathable. They walk in but they can't even get close to the other end.

It has a total of 50 40 acceleration ramp tiles (including the regular ramp where it actually goes down a level) and most of the resulting water globs reach the end of the 56-tile long chamber. More lanes can be added to make it as wide as you want.

Here's a side view schematic of the cart track:

# = fortification  W = stationary cart  H = hatch
L = exploit curve track ramp  [] = wall

  LLLLL
/# _H[]\LLLLLLLLLLLLLLLL  W#
       \_________________[]

Where the ramp to the left of the fortification has a roller that rolls right at highest speed, and the bottom level has rollers that roll left on all tiles including the ramp. The last 4 tiles of this, including the ramp, have "low" speed, the rest are highest.

What makes it work without needing the track to curve around is the part on the left. After the hatch allows the cart up the ramp, it jumps over the gap, flies through the fortification and lands on the roller ramp, which pushes it up that ramp instead of back through the fortification. From there, it accelerates along the curve ramps, strikes the stationary cart to release the water and falls back in the water.

There's somewhat of an odd hazard with this setup, after flying through that fortification on the left, it appears to check collision with anything below the ramp tile, meaning creatures directly below the ramp may get crushed. Before you try crushing invaders with phantom minecarts, be warned that sometimes this gets the cart stuck in the fortification where it's hard to get it out.

This setup is designed for high volume and range (not sure how many ramp tiles it takes to reach the speed cap), and carts stay more or less synchronized in waves. The type of water filling mechanism probably wouldn't work well with poisoned water. Maybe we can find a better way to fill a large number of carts with contaminants.

Either block access to the chamber or enable the alert before firing, because they're morons and will walk right into a death machine to clean up a spot on the floor.

Save is now up here.

Other things to note about the map:
- The shooting range. I tried making a 4-way repeater so a goblin always remains targetable in an attempt to stop armed civilians from fleeing, but it causes massive lag and doesn't seem to help.
- One of the caged forgotten beasts. "Amas", has a gas that causes blindness and enough drowsiness to make things sleep that otherwise would not. Maybe try to get vampire migrants and see if gassed prisoners make valid vampire food?
- The semi-molten rock layer is extremely thick.

I've heard that TinyPirate made a version of the watergun, is it on the forum or only in his book?

[My Name is Immaterial]

Yes.
PTW.

[gchristopher]

I like how it avoids the potential jams from carts being pushed into turning tracks.

How high a rate of fire have you achieved? It seems like the the long acceleration ramp, you'd need a lot of carts per barrel to keep it firing quickly. Though, with the bottom ramp mostly at High speed, it's, 80-ish ticks for the return trip?

Any insights on drainage? I haven't made one completely building-destroyer-proof, yet. I discovered that an artifact grate can be destroyed in the explosion of a clown made of fire. Maybe I need to look for an artifact furniture based on a magma-safe material?

[QuantumMenace]

I think it fires about every 15-18 ticks, with a slightly larger gap after the series of 13. You can add more but I think they start becoming desynchronized instead of staying in waves.

I drain by sucking the water out through bars in the ceiling. See the rows of pumps?

[gchristopher]

I think it fires about every 15-18 ticks, with a slightly larger gap after the series of 13. You can add more but I think they start becoming desynchronized instead of staying in waves.

Starting from a full stop, the rate limiter is carts climbing up the ramp out of the water? There's probably some added impetus from rollers farther back in the queue being pushed by rollers. I think that's how magazines with turning tracks become jammed.

Another theoretical fire rate limiter might be the direction reversal ramp. I'd expect that to be fairly fast, maybe 4-5 ticks to pass through the fortification and get up one block and out of the way of the next cart?

I drain by sucking the water out through bars in the ceiling. See the rows of pumps?

Oh, DUH! I love it! Embarrassed that I didn't get that.

I like that traders get to enter through it. If only that had an impact on price negotiations.

Having the firing range floor be bridges has worked well for me. Then I can drop all of the goblinite down to a lower floor for leisurely collection, with minimum downtime of the cannon. Non-magma-safe bridges or mechanisms also had a problem with fire creatures exploding. I'm not certain what the conditions are for surviving the explosion, or whether a high melting temperature or nether-cap material is better.

How does it affect large mounts and other large creatures at the higher muzzle velocity? Is there more knockback effect? The longest acceleration ramp I've tested was 6, and larger creatures we able to make forward progress even when hit repeatedly. Riders were still knocked backwards and off their mounts.

[pop123321pop]

How do you do the minecart speed exploit? I've seen it in the minecart shotgun video but have no idea.

[HmH]

Spoiler (click to show/hide)

Am I seeing things, or is that a tsunami on the far side of the room?

[QuantumMenace]

The "low" speed ramp along with water friction regulates the rate of releasing the carts. If you try to make carts release faster, they will pile up at the "barrel" end where the water is shot out and cause a jam. Trying to haul carts away from a pileup on the exploit ramp is dangerous.

Maybe an alternate way to regulate the speed is by having the carts pile up like this:

____WWWWW|
                   W____

Where all the tracks have "fastest" rollers on them. The carts would be delayed by the fall after the cart on the bottom moves away, which might precisely avoid jamming the barrel. Needs testing.

The ramp speed exploit is this: You build track ramps like this, where L is a curved track ramp with the two directions being the direction that you want acceleration in, and the direction against a wall "[]"

[][][][][][][][][][][][]
=LLLLLLLLLLLLLLLL=

When a cart derails from the left and ends up on the track ramp, it acts as if it's partway "up" the ramp and rolls "down" the direction that is not against a wall, gaining speed. This can be repeated many times.

I haven't seen its effects on mounts since I tend to cap the fort's population. I tried releasing a few FB's in the middle of the room and they didn't last very long, much less get far.

The "tsunami" is a standing wave caused by piled-up water released from the globs. The mist is constant and might be caused by the globs entering the deep water. It also tends to push goblins away from the wall so they can be smacked back again for extra skidding damage.

The save is up, tell me if it works! There should be some ambushers on the map as of the save, so feel free to invite them in.

[gchristopher]

Trying to haul carts away from a pileup on the exploit ramp is dangerous.

Extremely! Definitely forbid all carts that are part of the system and/or prevent dwarves from entering. I've mangled an awful lot of dwarves in cannons not even as deadly as QuantumMenace's.

I know a few ways to clear a pileup:
- If they're forbidden, then you can remove the carts one at a time, starting from the rearmost position.
- Beforehand, build a retracting bridge over the exploit ramps. Extending it will not atom-smash the carts, but will cause them to stop pushing. Lots of people warn that it might move the carts around, but I have yet to witness this in a way not explainable by simple derailment. (A bridge counts as a NSEW track, I think?)
- Exploit a bug in 34.11 in the way deconstructing buildings is made safe. I've observed that items near a deconstructed building are teleported to the same location as the building block. It is possible to teleport all of the carts out of the cannon into a square this way. Just deconstruct something close by and all the carts will be moved to that location. If that square has a track roller or ramp, they will immediately squash the dwarf who just deconstructed the building. It's pretty annoying if you want to make small changes to a minecart design without first moving all the carts a safe distance away. I don't know how consistently this bug happens, but it seems 100% in my current game.

Maybe an alternate way to regulate the speed is by having the carts pile up like this:

____WWWWW|
                   W____

Where all the tracks have "fastest" rollers on them. The carts would be delayed by the fall after the cart on the bottom moves away, which might precisely avoid jamming the barrel. Needs testing.

I think that will jam. I've seen carts over open space being pushed into a wall by a high-speed roller be "held" there and fail to fall, even with completely empty space below. Two carts is enough to cause the jam, and cutting power to the roller will allow the floating cart to fall.

I haven't seen its effects on mounts since I tend to cap the fort's population. I tried releasing a few FB's in the middle of the room and they didn't last very long, much less get far.

Sieges are even more fun than invaders. I've let a clown rush in, and they all died pretty quickly too. Sadly, I haven't gotten any really nasty procedural creatures that are blobs of metal or other highly resilient compositions.

[Urist Da Vinci]

...
The save is up, tell me if it works! There should be some ambushers on the map as of the save, so feel free to invite them in.

This is a device of awesome power. Using DFHack, I have verified that the water projectiles are being launched at the known maximum possible velocity of a minecart. The velocity decreases across the room as the water projectile skips across the ground. This made me curious what would happen if I placed a target creature at point blank range, with a solid surface behind it:

Spoiler: Hidden for awesome! (click to show/hide)

This will probably kill anything except a giant sponge

[gchristopher]

This is a device of awesome power. Using DFHack, I have verified that the water projectiles are being launched at the known maximum possible velocity of a minecart.

Any idea after how many acceleration ramps the minecart reaches maximum cart velocity?

[Di]

Awesome as always

[Awessum Possum]

This is awesome! However, you lost major geek points for not calling it a Wave Motion gun.

[Itnetlolor]

I gotta watch this in action. Got a recording of it in action made? I don't mind ramping the rate to 1000+ to view it full-speed, I just want to watch the carnage.

[gchristopher]

This is a device of awesome power. Using DFHack, I have verified that the water projectiles are being launched at the known maximum possible velocity of a minecart.

dfhack crashes if I try printall on a projectile item reference, so I can't measure the projectile speed. But I was able to measure the cart speed over the length of the barrel. Acceleration is linear over time and the cart reaches a top speed of 245,590 just before impact. This is slightly slower than the reported max of 270,000. The cart takes 42 ticks to travel the length of the barrel, accelerating 4,890 per tick.

So, it would need another 5-ish ticks for the cart to reach max speed. The cart is going faster than 2 tiles/tick at this point. The actual number of tiles moved per tick seems a little glitchy, possibly because it's running inside a cellular automata at faster than the spatial resolution. That makes it hard to estimate the additional barrel length required to reach theoretical max speed. At least 10 more, possibly as many as 15.

Here's the raw data for those observations:

Spoiler (click to show/hide)

Tick	Position	speed_y	Notes
0	Ramp	-13690
1	1	49990	50,000 is the reported fixed velocity of a high-speed roller
2	1	54880
3	2	59770
4	2	64660
5	3	69550
6	3	74440
7	4	79330
8	4	84220
9	5	89110
10	6	94000	Non-exploit down ramp
11	6	98890	Non-exploit down ramp
12	7	103780
13	8	108670
14	8	113560
15	9	118450
16	10	123340
17	11	128230
18	12	133120
19	13	138010
20	14	142900
21	15	147790
22	16	152680
23	17	157570
24	18	162460
25	20	167350
26	20	172240
27	22	177130
28	22	182020
29	24	186910
30	24	191800
31	26	196690
32	26	201580
33	28	206470
34	29	211360
35	31	216250
36	31	221140
37	33	226030
38	34	230920
39	36	235810
40	38	240700
41	39	245590
42	41	245580	Still contains water
43	41	245580	Water has been ejected
44	41	245580
45	41	245580
46	41	245580
47	41	245580
48	41	245580
49	41	245580	Falls below

I think the fire rate is close to maxed out. Carts sit at the end of the barrel waiting for the cart underneath to move. It takes the high-speed roller under the end of the barrel 14 ticks to clear the tile.

Awesome job, QuantumMenace!