Print

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

[Larix]

A quick test says perpendicular rollers indeed appear to work like that: just a simple 40z downward ramp ending in a S-N track, with a E->W roller at one point. The cart went off the rails after passing the roller and moved with a NNW-ly heading. A bit surprising, seeing as how rollers just *set* speeds for non-relativistic minecarts instead of adding speed. An array of (different-speed) rollers on a flat track right before the launch ramp should allow shooting at a multitude of angles. As long as the sideways component doesn't slam the cart into nearby walls or makes it miss the ramp, of course.

I hope the poster who speculated about Emu Radar and fire control is taking notes...

Edit: and yes, adding more rollers results in a stronger deviation, the principle is sound. It takes four maximum-speed rollers to deflect a 40z-accelerated cart by ~15°, though. And if you want to work with varying firing arcs from a single cannon, you'll have to put the rollers apart so that you can activate/deactivate them separately. This means you either adjust the arc of a high-speed minecart cannon by a whopping 5 to 10 degrees, or you need to work with a much lower 'muzzle velocity'.

[Urist Da Vinci]

Hint: you can even build the perpendicular rollers ON the ramp, as long as they are built next to a mechanical power source on the same z-level.

[Sergarr]

Launching a dwarf using the single (standard) shotgun seems to break reality, as the dwarf will teleport while in midair between several possible trajectories, as well as continue to take damage from the cart that he left behind. This also means that the dwarf will land from the fall multiple times. I haven't seen anyone survive it. It doesn't happen with the puppies, so I assume it is a bug with riding the cart.

Quantum Dwarfs?

[gchristopher]

I was thinking the same thing with perpendicular rollers at the end of the barrel. Good to know that they only affect velocity in the roller direction.

Though, I don't see how multiple rollers would achieve much more deflection than just one, since the roller just imparts a fixed velocity.

[Larix]

Oh right, Good catch.
Yes, sideways rollers impart a fixed sideways velocity, you can't stack them. The 'control' over the firing angle is restricted to choice of roller speed.

A weird effect of carts moving on a non-orthogonal vector is that they appear to ignore tracks: a cart moving at a SE to NW heading and meeting a straight E<->W track will just cross it and keep its heading, even when only propelled by 'low' speed rollers, which means a speed far below derail velocity. Seems you need to completely stop the cart before you can again propel it along a track.

[anzelm]

Instead of aiming a single cannon sideways, why not make an impulse ramp cyclotron with 4 exits triggered with pressure plates placed at the targeted area? Put a few such gun emplacements in the field and watch the invaders get minced.

[Suoli]

Instead of aiming a single cannon sideways, why not make an impulse ramp cyclotron with 4 exits triggered with pressure plates placed at the targeted area? Put a few such gun emplacements in the field and watch the invaders get minced.

Cannons with a conical firing arc would have a better coverage than cannons with a 1-tile wide straight firing arc. Think of how many cannons it would take to completely cover an NxN size area. For straight firing cannons, you'd need at least N, assuming you completely line one end of the area with them. For hypothetical cannons with a 90 degree firing arc, the number would be 4, with each one pointing at one of the cardinal directions from the center of the area.

[gchristopher]

A weird effect of carts moving on a non-orthogonal vector is that they appear to ignore tracks: a cart moving at a SE to NW heading and meeting a straight E<->W track will just cross it and keep its heading, even when only propelled by 'low' speed rollers, which means a speed far below derail velocity. Seems you need to completely stop the cart before you can again propel it along a track.

Yeah, there's a lot of confusing derailment behavior. Maybe once a cart is derailed by entering a tile without an associated entrance track, it's treated as skidding? I wonder if there's something in the vehicle entry in lua that stores an "on-track" state? I know there's velocity, offset and a "stopped" counter.

Instead of aiming a single cannon sideways, why not make an impulse ramp cyclotron with 4 exits triggered with pressure plates placed at the targeted area? Put a few such gun emplacements in the field and watch the invaders get minced.

I think the biggest reason many people aren't playing with cyclotrons is rate of fire. If you have to wait for a cart (or set of carts) to speed up to firing velocity, then the cannon barrel isn't firing for that many ticks. That'd be okay for carts loaded with bins full of weapons/trap components, but for a squirt gun, it doesn't fire fast enough.

[gchristopher]

I just thought of a derailment behavior that might somehow be useful for aiming cannons. A curved track on an up ramp has different physics than a curved track on flat. The up-ramp curved track will still turn a high-speed cart as it goes up one level, but the cart will retain a small amount of its previous velocity component. This can break cannons, (since the cart ends up going in a non-orthogonal direction) so you have to avoid curved tracks on ramps for cannon barrels, but maybe it'd help get a little more velocity mix?

Here's a bit of imaginary cannon barrel that goes up one level:

Code: [Select]

z: 0   z: +1
 ↓
░║░    ░░░░░░
░╚░    ░▼╚╚╚╚
░░░    ░░....
       ░░░░░░All the curved tracks are ramps. If a high-speed cart enters from above, at the arrow, it will appear to turn the corner correctly and keep accelerating along the upper row of impulse ramps. However, it will retain 5,000 south velocity after going up the level and eventually fly off to the south, even striking the south wall and ejecting its contents. Maybe being continuously derailed lets it keep that small south velocity?

Maybe there are some tricks to be discovered with curving track up-ramps that can build diagonal velocity?

[Suoli]

Here's a bit of imaginary cannon barrel that goes up one level:

Code: [Select]

z: 0   z: +1
 ↓
░║░    ░░░░░░
░╚░    ░▼╚╚╚╚
░░░    ░░....
       ░░░░░░All the curved tracks are ramps. If a high-speed cart enters from above, at the arrow, it will appear to turn the corner correctly and keep accelerating along the upper row of impulse ramps. However, it will retain 5,000 south velocity after going up the level and eventually fly off to the south, even striking the south wall and ejecting its contents. Maybe being continuously derailed lets it keep that small south velocity?

YES. I wondered about something like this but decided it couldn't be even remotely possible. Stacking "drift" with multiple such ramps could be the key to a 360 degree firing arc from only 4 barrels.

[gchristopher]

YES. I wondered about something like this but decided it couldn't be even remotely possible. Stacking "drift" with multiple such ramps could be the key to a 360 degree firing arc from only 4 barrels.

Well, don't get too excited yet? All I managed to do with it is break otherwise good cannons. But there might be something useful there.

[Urist Da Vinci]

If a minecart hits another minecart that is travelling in a perpendicular direction, what happens to the component velocities? Is the behaviour different if the carts collide while in midair? Call it the T-bone aiming system if it works.

[gchristopher]

If a minecart hits another minecart that is travelling in a perpendicular direction, what happens to the component velocities? Is the behaviour different if the carts collide while in midair? Call it the T-bone aiming system if it works.

I don't care if it works or explodes all over itself. That idea is beautiful and I love it.

[anzelm]

Cannons with a conical firing arc would have a better coverage than cannons with a 1-tile wide straight firing arc. Think of how many cannons it would take to completely cover an NxN size area. For straight firing cannons, you'd need at least N, assuming you completely line one end of the area with them. For hypothetical cannons with a 90 degree firing arc, the number would be 4, with each one pointing at one of the cardinal directions from the center of the area.

A standard single-barrel cannon already has significant spread (about 10 degrees), so the coverage should be much larger:

Code: [Select]

+ - four direction turret
|- - fort walls
. - firing coverage

    ...    ...
    ...    ...
    ...    ...
   .....  .....  ..
...................
.....+......+......
.......... ........
   ....|--|....  ..
    ...|  |...
   ....|--|....  ..
...................
.....+......+......
...................
   ...    .....  ..
    ...    ...
    ...    ...
    ...    ...
   .....  .....
   .....  .....
This assumes a range of 7 tiles (as in the drawing, not ingame) and an exaggerated spread of ~15 degrees. Not all of the map gets coverage, but you can add more turrets.

As for the 90 deg cannon, four cannons in the centre have less coverage if you take into account their minimum range (what is the smallest muzzle velocity that still causes considerable damage?).

[gchristopher]

If a minecart hits another minecart that is travelling in a perpendicular direction, what happens to the component velocities? Is the behaviour different if the carts collide while in midair? Call it the T-bone aiming system if it works.

Okay, tried it, with two perpendicular 10-tile impulse ramp accelerators that converge on a NSEW flat track. Both carts start with 0 offset on the exact same tick.

Prior to colliding, both carts have the same speed, 112,470, in their respective directions.

On the tick where collision occurs, one cart enters the point of convergence of the tracks. The other comes to a halt one tile before that point. The cart that enters the point of convergence does retain the full velocity of both carts, in addition to an extra 5000 from the other cart, so it has slightly more speed in the direction of the non-continuing cart. That cart is now traveling at a 45 degree angle.

That's the good news. The bad news is what happens to the contents of the carts. Both eject their contents on the tick where collision occurs. Both ejected contents appear to have their velocities based just on the speed of each cart before the collision, so both are still traveling orthogonally.

The tick after collision, the stopped cart moves into the tile at the point of convergence and stays there. The other cart moves away at 45 degrees and all the ejected contents fly along orthogonal trajectories, not showing any effect of the collision.

Summary, you can get a cart going in a non-orthogonal direction with a collision, but this test didn't reveal a way to fire non-cart projectiles at an angle.