Print

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

[Snaake]

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.

Too tired to dig up the source of the info now, but iirc carts attain maximum speed far before 50 tiles of impulse ramps or 50-z of downramps. It's possible to temporarily (friction will of course start slowing it down immediately) get a cart above said max speed by bumping a heavier cart at max speed into a lighter cart, or at least was in 34.x. There was some science on that and I remember speculating if it would be possible to use artifact gem (or other esoteric materials) minecarts to bridge the gap in the weight sequence of materials, to get a full sequence of successively-lighter bumpers, all the way from platinum down to feather tree wood. Any excess weight over 2 times the weight of the lighter cart didn't count for the momentum transfer iirc, or something like that, and the weight difference between the lightest metal and the heaviest wood was/is over 2 times, hence the need for artifact materials in between.

[omega_dwarf]

Interesting. And funky. But using a bumper isn't a bad idea; even without going over max speed, I could use heavy carts to bump lighter carts to get the lighter carts to max speed. That being said, that's more useful with shotguns, since we need a lot of momentum for the carts to actually mow down invaders (not as much if they're just crashing into a wall and spraying their contents.) But yeah, that could be used to make a much more compact shotgun (with, presumably, a higher rate of fire.)

I've seen sources say that maximum velocity occurs after about 50 impulse ramps. Was your shorter distance only possible with the bumping? Or, maybe, exceeding roller speed at that shorter distance?

I'll have to try it out, of course, but I have very few possible targets.

[Delioth]

I seem to recall a need for something like 47 impulse ramps for max speed (From when there was a megathread for all Minecart science since they had just became a thing.)

[Hetairos]

I once made a small scale water gun I have linked in my sig... no idea if there is anything helpful there. I remember there is a minecart speed calculator spreadsheet on the forums somewhere. Minecarts can definitely be put inside other minecarts, even a few at a time. Only manually though, like anything that isn't a liquid.

My English and facts might be a bit off, I have a bit of an infection going on.

[Snaake]

Interesting. And funky. But using a bumper isn't a bad idea; even without going over max speed, I could use heavy carts to bump lighter carts to get the lighter carts to max speed. That being said, that's more useful with shotguns, since we need a lot of momentum for the carts to actually mow down invaders (not as much if they're just crashing into a wall and spraying their contents.) But yeah, that could be used to make a much more compact shotgun (with, presumably, a higher rate of fire.)

I've seen sources say that maximum velocity occurs after about 50 impulse ramps. Was your shorter distance only possible with the bumping? Or, maybe, exceeding roller speed at that shorter distance?

I'll have to try it out, of course, but I have very few possible targets.

I could be wrong on the number of impulse ramps needed, maybe even on vertical ramps. The latter are less cheaty anyway, just a bit more fiddly to build because you have to jump around z-levels.

Iirc you actually can't use the bumping method to power shotguns, since the bump knocks out any contents. It's just a minecart accelerator. So a railgun, basically. An automatic railgun could probably be built, of course.

[omega_dwarf]

Interesting. And funky. But using a bumper isn't a bad idea; even without going over max speed, I could use heavy carts to bump lighter carts to get the lighter carts to max speed. That being said, that's more useful with shotguns, since we need a lot of momentum for the carts to actually mow down invaders (not as much if they're just crashing into a wall and spraying their contents.) But yeah, that could be used to make a much more compact shotgun (with, presumably, a higher rate of fire.)

I've seen sources say that maximum velocity occurs after about 50 impulse ramps. Was your shorter distance only possible with the bumping? Or, maybe, exceeding roller speed at that shorter distance?

I'll have to try it out, of course, but I have very few possible targets.

I could be wrong on the number of impulse ramps needed, maybe even on vertical ramps. The latter are less cheaty anyway, just a bit more fiddly to build because you have to jump around z-levels.

Iirc you actually can't use the bumping method to power shotguns, since the bump knocks out any contents. It's just a minecart accelerator. So a railgun, basically. An automatic railgun could probably be built, of course.

Oh, okay. But the question still has to be begged of whether the faster-but-lighter cart is going fast enough to do as much damage as the slower-but-heavier cart (in DF physics; irl, especially with air resistance and inelastic collisions, the heavier cart solo would probably be better.)

And about the compact water gun, Hetairos, I'll check it out!

[Larix]

I fiddled with my minimal watergun design and reduced it to the smallest/fastest-repeating single-cart guns i could come up with.

Powerless:

Code: [Select]

.WF        WF       
 ╔▲W       ╔═W       WWW
 ▼▲W       ▼╚W       ~║W
W▲▲W      W╗╚W       WWW
 ╚╝        ╚╝

z+0     z+0,track     z-1, track

W = Wall
F = Fortification, north of shot location
~ = constant water source

This setup just uses ramp oddities to load water and accelerate the cart. The NS ramp in the channel is enough for reliable loading - the cart enters at ~25k speed, gets slowed to loading speed (<10k) by maximum-depth water, then slowly crawls out and onto the adjacent impulse ramp. Since the loading ramp has no downward connection, the cart faces no ramp deceleration. All ramps on the upper level accelerate the cart normally. The three angled ramps give enough speed to enable a shot (~55k speed are needed to shotgun the cargo). The straight EW ramp re-accelerates the cart after the shot and re-cycles it onto the loading path.

Tested and proven, needs exactly one cart and delivers one shot per 35 steps. If there's no ceiling above the fortification, the water is expelled on the z-level above. If there's closed ceiling above, the water is pushed through the fortification itself.

Powered:

Code: [Select]

.WF       WF       WF
 RR       ╗═       ew
~▼▲W     ~║╚W     ~▼▲W
 RR       ╔╚       nw
 W        W

visible     track   rollers
Fairly tricky, because i optimised it for the fastest single-cart cadence i could conceive. It uses roller-on-corner features to shave off a few steps. Loading water happens just like in the powerless version, on a non-accelerating ramp in a water-filled trench, which is passed from north to south.

The cart then rolls onto the north-pushing roller on the SE track tile. That one turns the cart back north and "around the corner" immediately, at highest speed (50k). I.e. it leaves to the east on the very next turn.

On that tile, the west-pointing roller once again tries to turn it around instantly and, again, pushes it around the corner, this time to the north. That roller, too, works at maximum speed.

The cart now rolls onto an impulse ramp. That's required to get it to shot-worthy speed, and it must get three steps worth of acceleration there. One step will get neutralised by the checkpoint effect upon leaving the ramp and >55k speed are required for a shot. The chain of two opposing highest-speed rollers moving the cart onto the ramp means that it enters at a sub-coordinate where it has enough space to get the full three turns of acceleration.

The cart rolls off the ramp and onto a medium-speed west-pointing roller. On the next step, it will crash into the fortification and releases its load, in a northwesterly direction (~58k northward speed, 30k westward speed). It technically stops at the fortification but is instantly re-accelerated to the roller's speed, which now moves it off to the next tile west.

On that tile, there's yet another opposing-mode roller, pushing east to move the cart south via corner redirection. That roller works at "high" speed. That's the highest speed compatible with loading water onto a cart in a one-tile trench loader.

All this effort achieves a single-cart cadence of one shot per thirteen steps. And this is what it looks like in action:
http://mkv25.net/dfma/movie-2743-mini-watergun

That one's rotating in the opposite direction due to space constraints. The minimal powerless watergun can be seen a bit to the east and north, currently not active.

EDIT: In the powered design, time consumption breaks down as:
one step on each corner/roller tile - 3
two steps on the shot tile -> 5
three steps accelerating on the ramp -> 8
five steps loading in the trench -> 13

Loading is the biggest time sink, at five steps per cycle. This should also be the theoretical cadence limit for a cannon with multiple carts per barrel.

The powerless cannon has of necessity a longer cadence, since it can only accelerate through ramps, at 5k/step. And since in every cycle of a watercannon, the cart must be accelerated from <10k (slow enough to take on water) to >55k, that translates to 10+ steps of acceleration. The cart also needs to be re-accelerated (via ramps) after delivering the shot to send it to the loader, and the loader itself consumes time, too. I actually managed to accelerate the cadence of the powerless gun to 31 steps by _increasing_ the footprint slightly:

Code: [Select]

. WF        WF
 ╔▲▲W      ╔╝═W
~▼+▲W     ~║+╚W
W▲+▲W     W╗+╚W
 ╚▲╝       ╚╔╚
  W         W
visible   track
After the shot, the cart not just rolls off a single ramp, it goes through another full impulse ramp, accelerating it to about the maximum speed legal for loading water. Loading at this speed takes only four steps (instead of over ten when coming off a single ramp). Due to the larger footprint, the circuit contains four acceleration ramps for the actual shot and not the minimum three. All corners are passed in single steps thanks to the checkpoint effect. The faster loading more than compensates for the longer time spent on ramps, so this cannon has a slightly faster cadence than the smallest design presented above. I re-tooled the powerless cannon for this design and tried it out (that's how i know of the improved cadence ).

[Max™]

Objects moved by liquids do not become invisible.

Well, objects moved by magma which are magma-safe do not become invisible, I assume water works the same way.

[Snaake]

Objects moved by liquids do not become invisible.

Well, objects moved by magma which are magma-safe do not become invisible, I assume water works the same way.

This is not the same as saying the bug was fixed. It was never intentional behaviour, your response mostly gives the impression that you've never run across the bug. I've seen it a couple of times, and would like to clarify that it didn't happen every time, most of the time, or even necessarily very often. But it did use to happen on occasion. The "invisible" bit refers to the fact that said objects could still be found and their tile zoomed at using the stocks menu. IIRC it was possible to use dfhack to manually move the object, at which point it would become visible again.

[Hetairos]

And I believe dwarves were able to pick them up once the tile was accessible again.

[gchristopher]

Argh, this is hard to keep up with when I haven't been actively playing with minecarts. Okay...

EDIT: In the powered design, time consumption breaks down as:
one step on each corner/roller tile - 3
two steps on the shot tile -> 5
three steps accelerating on the ramp -> 8
five steps loading in the trench -> 13

Loading is the biggest time sink, at five steps per cycle. This should also be the theoretical cadence limit for a cannon with multiple carts per barrel.

That's neat. The fastest possible firing rate in a design where the cart has to come to a stop and fall onto a lower tile is one shot per 11 or 12 ticks, IIRC. So that's a 2x speedup over most designs, which is great. It'd be very cool to see a proof of that concept. For clown rush and other extreme uses, that might be a big reliability improvement.

~55k speed are needed to shotgun the cargo

It's been a while since I tested, but I thought momentum counted for how effective shotgun damage is. That's about 20% maximum cart speed. I know this post was focused on size/firing rate, but would it be worthwhile/interesting to revisit that question while requiring higher velocity shots?

[Max™]

Objects moved by liquids do not become invisible.

Well, objects moved by magma which are magma-safe do not become invisible, I assume water works the same way.

This is not the same as saying the bug was fixed. It was never intentional behaviour, your response mostly gives the impression that you've never run across the bug. I've seen it a couple of times, and would like to clarify that it didn't happen every time, most of the time, or even necessarily very often. But it did use to happen on occasion. The "invisible" bit refers to the fact that said objects could still be found and their tile zoomed at using the stocks menu. IIRC it was possible to use dfhack to manually move the object, at which point it would become visible again.

I meant like this, most of this stuff was from removing the hillside to the right, but the magma flow pushed it to the left.

[Larix]

[motorised water-sprayer, cadence with a single cart 13 steps, theoretically capable of cadences down to five or six steps per shot]

The fastest possible firing rate in a design where the cart has to come to a stop and fall onto a lower tile is one shot per 11 or 12 ticks, IIRC. So that's a 2x speedup over most designs, which is great. It'd be very cool to see a proof of that concept.

Yep. Letting a cart fall down always struck me as horrendously inefficient - come on, a whole eleven steps wasted!!

My designs also work without a stationary cart in the gun's barrel for releasing the shot: when you don't recycle the cart by dropping, you don't work with a flying cart, so no passing through the fortification at collision. Sideways roller means it pulls pretty strongly to the side, though. Might be possible to circumvent by putting the shot point two tiles past the last ramp while working with a high-speed (over 1 tile/step) cart.

[used minimum shotgun-able speed of just over 55k]

It's been a while since I tested, but I thought momentum counted for how effective shotgun damage is. That's about 20% maximum cart speed. I know this post was focused on size/firing rate, but would it be worthwhile/interesting to revisit that question while requiring higher velocity shots?

The motorised model can fit two carts, and they're capable of about doubling the cadence if you have a very massive water supply. The single-cart 13-step cadence already threw the occasional blank since it hit the limit of what my brook could supply. Having both carts load in the same cycle was a rare occurence, but the occasional double shot within six steps was delivered.

Range and effect depend on momentum, that's right. For a full military installation, i'd suggest a dozen or more ramps for a decent speed. The motorised gun can handle multiple carts in a single barrel, at high cadence. I tried out a non-optimised version sometime, which used five carts and averaged one shot every 11 steps. It had water supply issues already - good runs gave cadences of nine steps.

What i like about my design is that i find it extremely uncomplicated. I've occasionally built such a thing as a pure ornament, set up to irrigate some random plot of land i don't care about. With a river or brook, you can set one up in a matter of game weeks, in the first year. Making it useful takes a _bit_ more work...

[gchristopher]

My designs also work without a stationary cart in the gun's barrel for releasing the shot: when you don't recycle the cart by dropping, you don't work with a flying cart, so no passing through the fortification at collision. Sideways roller means it pulls pretty strongly to the side, though. Might be possible to circumvent by putting the shot point two tiles past the last ramp while working with a high-speed (over 1 tile/step) cart.

Yep, that's a nice feature. Also, with a cart going full (270,000) speed, the 50,000 speed from the roller only creates a 10 degree deflection, which isn't too huge. For some firing ranges, that could be a bonus, letting you converge fire at an entrance area and effectively doubling the firepower of the design. Procedural material clowns are actually a bit weaker from the tendency of body parts to be fatally blown off, but ones with normal body structures are tougher and that increase might make a difference.

I haven't gotten any REALLY fun clowns made of metal yet, so 100% effectiveness is still an open question.

if you have a very massive water supply

I keep forgetting to mention in this thread, for these designs, ALWAYS use a pressurized water supply. Relying on water flow will inevitably move carts around and mess up timing. That can be as simple as having a water reservoir one tile higher than the loading area, or (better) having powered pumps pull from the bottom of a reservoir to provide pressure. Typically this requires an aquifer or very large reservoir.

What i like about my design is that i find it extremely uncomplicated. I've occasionally built such a thing as a pure ornament, set up to irrigate some random plot of land i don't care about. With a river or brook, you can set one up in a matter of game weeks, in the first year. Making it useful takes a _bit_ more work...

hahaha, "Take that, PLANTS!" Love it.

Also "take that, FPS! I'm flowing some water!"

[Max™]

Larix: I like to build ornamental minecart waterguns when I'm not working on a turing complete dorfputer capable of running df.