Print

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

[Fleeting Frames]

Trivial findings on bouncing minecarts, thanks to watch-minecarts.

A typical wood minecart (weight 24 urists) coming off a bouncer pit (falling in design from the side - interesting, it seems simple double ew has much greater impulse) hitting equal one filled with water on low friction track stop moves it 464 sub-tiles per bounce, for a repeat cycle of 108 hits/movement.

If the one being hit is suddenly platinum, it moves only 38 sub-tile positions per step, for repeat cycle of 1316 hits/movement.

If it's hit by, say adamantine minecart carrying iron nestbox, it'll only move 5 sub-tile positions, for 10k repeat cycle.

Replace iron nestbox with cobaltite one, and you have it moving only 1 sub-tile position per cycle, for 50k repeat.

(All assuming sideways return, of course.)

Of course, this last bit can probably be obtained cheaper by using two slightly different weight carts hitting from opposite directions.

Still, perhaps another use for platinum if you want to use platinum for something.

But I wonder....As you can nest carts infinitely and fruits are really light (plus you can use two opposite slightly-different weight pushes for further accuracy), a hypothesis: it is possible to build any arbitrary repeat cycle that can be expressed as ceiling(100000/n) by just selecting cart weights and track stop friction correctly.

But would it be possible to boil that down to a rule so simple you could easily calculate it in your head for setting up timers quickly?

EDIT: I have learned that minecart collisions only use whole urist weights by cart. A cart with 1 or 0 wood nestboxes inside gets same push, and cart with 2 acts as if it's weight was exactly 1 urist larger. If they also floor sub-tile positions, the above cobaltite nestbox in adamantine cart pushing on platnium-water cart checks out.

Disappointing, though.

[taptap]

Good thinking.

You may be interested in the "weight-difference-detector" / adjustable length repeater cycle I made with similar weight minecarts and Larix' more general scale. (Both in the dwarfputing experiments thread in my signature.) But I certainly did not consider using very light vs. very heavy collisions with multiple attemps. Please make something workable out of it.

[Fleeting Frames]

Making something workable out of it isn't hard. I have created a flawed worksheet to quickly calculate min and max weights for all the interesting cycle delays, impulses and possible frictions (really just 4/5 of track stops and a floor tile), but...

Lets take a really simple cycle:
buildings over tracks and ramps:
╔^
▲S˘▼

track
╔ ╗
||+==

Carts on Track stop and hatch, walls as necessary. Cart paths should be obvious.

When hatch is opened, the cart takes 26 steps to deliver the first push and then delivers one every 18 steps thereafter.
With this design, the input for equal weight is 29310 - just have to do floor(29310×weight of light cart/weight of heavy cart) for the speed of heavy cart. Then heavy cart moves as it's speed dictates, first losing friction and then taking the move if there's anything left over for each turn.

Technically an arithmetic sum, but for low cycle lengths(<33) you can use high friction track stop and higher cycles generally work best with medium track stops as far as minimizing weight required goes, which conveniently makes the arithmetic sum 1 deep, meaning just heavy cart speed - friction.

Ok, so, lets say I want to use it for waving a bridge. 200/18=11,11....So 12 pushes per cycle. 50000 length, so the cart must traverse between ceil(50000/12) and floor(50000/11) tiles each time - that is, between 4167 and 4545 tiles. All friction options below highest can work (if it was elastic, highest would work too for heavy cart pushing on light one), but a high friction track stop will work for just a single push with the lightest cart - required speed being 10000 greater, so for standard 24 urist wooden minecart the respective max and min weights are 49,6364...(floored to 49. You cannot use minecart collisions to measure the weight of single ultra-light thing) and 48,3465...(ceil'd to 49).


Tested this with loading the heavy cart with a bucket and second minecart. Cart moves 4355 sub-tiles per push, the heavy cart moves onto ramp 211 steps after hatch opens, and takes 11 steps to hit the plate....Excellent for a timer originally conceived for locking dwarves in rooms for long periods, but if you realized that it will vary due north-south sub-tile positions and fail to accumulate enough acceleration to complete cycle in less than 18 steps on 3rd bridge wave, give yourself a cookie


Of course, you can also take given convenient heavy cart you have and alter how often or how big you deliver an impulse, or alternatively use the heavy cart to propagate another signal. If I had used platinum cart for the above, I'd have gotten 321 sub-tiles per move with medium friction stop, for 156 cycle length - ok so that'd need a roller to deliver signals at reasonable bridge-wave pace, which would alter push impulse due the tiers not fitting, and...

It might be easier to mess with weight, though - more compact, and can just look up wiki's weight and density for most things, instead of fiddling to get a push with correct impulse and delay.

EDIT:29310, not 29300. Forgot to account for floor friction with younger cart pushing on older ^^;

[taptap]

Most untrivial things on trivial findings today.

Not sure I can follow all the numbers, never used that script and mainly playing 43.05 currently. When talking about light weight and being unable to measure, you mean sub-1 Urist differences, right? Well, I can live with that, a single metal bolt is heavier after all. (One application I dream of would be small footprint, minecart repeaters tuned by adding weight, i.e. I could look in the minecart and see: 4 copper bolts = 120 steps. Or we should just make a table of different measured minecart repeaters of definite length.)

You can use an element that overrides North-South position, e.g. a roller pushing west on a N-E-track-corner (in track corners the WE position overwrites the NS position when leaving the tile or vice versa, if I understand correctly).

[Fleeting Frames]

The script was mostly useful for double-checking my guesses/math and finding the things being floored. Don't really need it if you're sure of how things work, but it'd be a pain to double-check a cycle that takes 50076 steps (156*321) to go through it. Ask about anything in particular that you don't understand, I'm not the most clear person :V

Could also alter it to have the acceleration ramp push sideways or onto another ramp or what have you. My initial design was ^S˘▼ for the whole thing and I guess it shows

I consider it pretty trivial, though - For the locking in room application, can just lock and unlock the door through overseer interface or use a guided cart, and it can't produce optimal hatch open delay, repeating spikes with minimal items of what have you for typical useful repeaters. Neat for dwarfputing but not much else.

[taptap]

I don't think custom length delays/repeaters are trivial. I more often than not end up carving tracks, refitting multiple times and still not getting the precise delays I hoped for, but something that is close enough to work with. Yes, it is all deterministic and less lazy people could figure out the exact results for a specific track before the test, but even then fitting a track suitable for the specifications isn't trivial. This whole multi-collision setup is certainly an innovation.

[Fleeting Frames]

Eh, I just have limited use for timers/repeaters tbh, hence a novelty They're essentially oscillator circuits adapted to work with sub-tile scale.

As an aside, diagram of my first and probably favourite impulse setup:
╔ ╗
▼ ╝
˘

Toss cart in, when hatch is released provides 4860 impulse if dry or 4460 if filled with 2/7 water (other water levels provide too much friction, or dry out). First, low speeds are very useful for low weight ratios (though don't really work for low repeat delays), and second..



Push delay is easy to add, while the above cycle still has the 4460 input.

(The track stop on corner is just another way to stop cart / test wall-hitting input speed)

EDIT: The previous numbers were wrong for empty cart on empty cart as they were done with 49 urist minecart (forgot to empty it after testing ^^;. The proper numbers have now been added.

Also, it seems this setup is useful for providing 29300* pushes on longer cycles for those push numbers where high friction track stop is useful.
=?╗
╔║╗

All track ramps expect upper right corner. ? is either NS (299 repeat) or NE (301 repeat) [almost useful for a clock...].

Note that if you enter on hatch it will stop the cart on S-SW tile, thus requiring ~2 cycles to go through (something like 586? steps from open to first signal?), suggesting a NSW in upper left corner might be useful for roughly doubling the delay.

* Yeah, this means the dual EW pit is 29310. Measured after a single movement earlier, due move order, oops.


Now, for something completely different: For the first time, I see a native platinum vein in olivine. This one in particular stretches "inside".

So that's what Toady meant by "Stopped veins in large mineral clusters from being improperly restricted" in 43.03 half a year ago.

[Libash_Thunderhead]

All the dorfs who lost their feet were using silver crutches in my fort.
I noticed recently they could hit enemies with their crutch.

[rhavviepoodle]

I switched galena ore to be usable when my stonecrafter procced a strange mood. He proceeded to bring one galena and one copper bar to the craftshop he'd claimed, and the end result was... an iron crown? This has to be a bug, right? Mostly I'm double checking because I haven't seen this posted in the bug tracker yet and it's so absurd that I'm kind of confused right now.

[Bumber]

I switched galena ore to be usable when my stonecrafter procced a strange mood. He proceeded to bring one galena and one copper bar to the craftshop he'd claimed, and the end result was... an iron crown? This has to be a bug, right? Mostly I'm double checking because I haven't seen this posted in the bug tracker yet and it's so absurd that I'm kind of confused right now.

Sounds like: http://www.bay12games.com/dwarves/mantisbt/view.php?id=5625

[Fleeting Frames]

I love that bug. Gives crap materiel moods a chance to be elevated to military-worthy gear.

[Libash_Thunderhead]

Dwarven alchemy, of course!

[carnivorn]

Cave adaptation also affects the outpost liaison. Not sure how long it generally takes but I'm on my eleventh year in this fort and he's been doing nothing but loiter in my tavern during his stays. I just noticed him leaving a vomit trail down my stairs by chance.

[FantasticDorf]

Because immigrants are sourced from population, they have a chance to appear with BP parts missing if they engage in wars a lot.

This means that if you play with a alternative race, who are very active militarily, they will sometimes already have non-fatal injuries from a long career as ex-soldiers on the frontlines. Dwarves never fight wars proactively (in this version might change for artifact) so are beside from a small range of accidents, always BP anatomically whole and un-scarred.

I once had a goblin spearman (i think it was) with only 1 eye, i've had one or two also turn up without a arm (never seen both missing though its possible).

[mikekchar]

I've actually had a dwarf show up with a missing nose.  I checked legends mode and it was taken off by a roc when she was a child!  Somehow she escaped and eventually made her way to my fortress.