Print

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

[Tacomagic]

I'm working on a new magma collector design (well, new to me anyway, I'm sure somebody else has thought of this) and came up with a question pertaining to screw pumps:

How much do screw pumps pump per tick?

Here's the context:  I want to create a chamber of magma with a pressure plate at the top level.  I'd set the pressure plate to 7/7 magma so that it would fire off when the chamber is full.  This pressure plate would connect up to a disengaged gear assembly so that it would toggle to engaged when the plate fires.  This gear feeds a pump so that, while engaged, it pumps the chamber somewhere else.  100 ticks after the pressure plate turns off, the gear disengages.  49 ticks later, the screw pump stops.  How much will that pump have pumped in those 149 ticks?

I looked around for this information, but couldn't find anything.  I can do the science for myself, but I was hoping to save a little trouble if this is something somebody knows off-hand.  If this isn't something that anyone has worked out, then...



Cheers,
Taco.

[Girlinhat]

A pump will move the entire contents of the tile ahead and under it, instantly, I believe every other turn.  Which means ideally 7/7 but also less, because pumps work MUCH faster than water flows, so you have to get creative to get adequate flow.

[Tacomagic]

A pump will move the entire contents of the tile ahead and under it, instantly, I believe every other turn.  Which means ideally 7/7 but also less, because pumps work MUCH faster than water flows, so you have to get creative to get adequate flow.

Doesn't vertical flow happen instantaneously, or does it just seem to because falling happens so quickly?

My plan involved a tall 1x1 shaft filled up (well, 1x1 aside from the pump level which would necessarily be 2x1).  My hope was that when it pulled the magma, the magma above would fall almost instantaneously into the square until eventually the magma level reached the level of the pump.

[Tacomagic]

*Sigh*

DF 1, Science 0.

I decided to create a new world as a testing platform to see how much a pump will move when gravity assisted.  Picked a cold area with a stream because I wanted it to freeze for quick and easy damming.

It says frozen year-round.   .

Time to pick a different spot.

[Tacomagic]

OK, I successfully SCIENCED!  And the results were very surprising, in that they surpassed Grilinhat's predictions... by a lot.

Here's the setup:

Underneath a river I carved out a 20x30x2 reservoir for the river to fill.  Below that I dug a 1x1x10 shaft that dumped directly into a channeled square.   Next to that channel, I put a pump powered by a water-wheel, but inactivated by a disengaged gear.  This created a pressurized shaft of water that, concevably, should fill the lowest square instantaneously once it's emptied.  Something that did seem to be true during the test.

The outlet of that pump was aimed at a 3x3x10 collection chamber.

The disengaged gear was hooked up to a pressure plate set to activate on citizen dwarves. The plate was situated near the start of a long in a 1-tile-width hall sandwiched between 2 doors.  At the end of the hall was a dumping zone.

Once the reservoir (and shaft) were completely full, and the river had recovered its flow, I ordered a single thing for dumping.  Once the dwarf passed through the doors and over the plate, I forbid both doors to prevent re-triggering.  From there I waited until the the pump stopped.

The hypothesis was that the pump should fire every other tick for 149 ticks.  If that was true, the best-case-scenario amount of water pumped would be, at most, 75 squares at 7/7, which would fill the shift completely up to level 8, and then fill alittle of level 9.

However, in reality, the pump filled the entire 3x3x10 chamber, having pumped at least 90 squares of 7x7 water.  That's quite a bit more than even the best-case predicted.

So, now I need to go larger.  Luckily, I had the forethought to include a way to empty the chamber and have service access.  I'm going to drain the chamber and extend it out to 4x4x10, increasing the total capacity to 160 squares of 7/7 water.  The absolute best-case scenario would be that the pump empties 7/7 water every tick, providing 149 (or 150, depending on my build order) of water.  So, even in the best case, this new cistern should not fill to capacity during the test.

ONWARD TO MORE SCIENCE!

EDIT:  Unfortunately, I didn't have the forethought of using a gravity drain.  Time to wait for everything to drain and evaporate.  There's always that one square of 1/7 water that doesn't want to evaporate. Le sigh.

[Tacomagic]

So, the results of the second test are in!

The pump pumped a total of 160 x 7/7 water, and 1 x 4/7 water (no idea what that was about, probably a partial tick action from when the plate was activated).  I forgot that the dwarven movement speed entered into how long they would be on the plate (10 ticks, 1 to open the door, 9 to move off the tile).

Based on this test, I'm pretty confident that a pump tries to pump every tick and is only actually limited by the need to have at least 2/7 on the input tile (thus it is essentially inflow limited, outflow unlimited).  By using a gravity fed system, a tremendous amount of flow through a pump is possible.  Granted, for high-flow situations, it'd be far easier to just use a gate or bridge, since they don't require power and gravity feed is extremely fast anyway, but for very specific applications (such as my magma collector), this is a very important behavior to know about.

Now, back to my magma collector design.

The other interesting thing I noticed was that water has a pretty decent hang-time after it's pumped.  It could be possible to build a pump stack without any containment walls and still have it function perfectly.  Granted, things could get kinda fun when you turned it off.

Actually, I think I'll make that my next bit of SCIENCE! I'll see what happens if you build a pump-stack without any walls.

[billybobfred]

The other interesting thing I noticed was that water has a pretty decent hang-time after it's pumped.  It could be possible to build a pump stack without any containment walls and still have it function perfectly.  Granted, things could get kinda fun when you turned it off.

Actually, I think I'll make that my next bit of SCIENCE! I'll see what happens if you build a pump-stack without any walls.

I do believe that one has already been tested and confirmed to work.

For magma, too.

(Can't imagine how you'd turn the thing off safely, but that's details, right?)

[Tacomagic]

The other interesting thing I noticed was that water has a pretty decent hang-time after it's pumped.  It could be possible to build a pump stack without any containment walls and still have it function perfectly.  Granted, things could get kinda fun when you turned it off.

Actually, I think I'll make that my next bit of SCIENCE! I'll see what happens if you build a pump-stack without any walls.

I do believe that one has already been tested and confirmed to work.

For magma, too.

(Can't imagine how you'd turn the thing off safely, but that's details, right?)

Aww.  Okay, I won't re-invent the wheel for that one, then.  Later tonight I'll add this information to the wiki.

Actually, assuming you've got a way to keep the dwarves out from under the thing, you could probably just toss a pair of drains under either side of the pump stack.  Falling liquid would immediately drain itself when you turned it off.

[billybobfred]

Actually, assuming you've got a way to keep the dwarves out from under the thing, you could probably just toss a pair of drains under either side of the pump stack.  Falling liquid would immediately drain itself when you turned it off.

Now this sounds like a good time to run some ‼science‼. "What's the most effective draining design for an open-air pump stack?"

[GavJ]

(Can't imagine how you'd turn the thing off safely, but that's details, right?)

Just shut off the liquid source at the bottom a little while before you turn off power. It should suck up all the remaining water in the column and leave dry or 1/7s only.

[khearn]

Sucking water out of the air is the whole basis for the standard mist generator.

BTW, it's possible that your pumps were moving more water than normal because of the pressurized input. You would have probably moved just as much with a door in place of the pump. Since magma doesn't pressurize the same way that water does, you might get different results with magma. I'd do a test before getting too far along with a design that is intended for magma.

   Keith

[Tacomagic]

BTW, it's possible that your pumps were moving more water than normal because of the pressurized input. You would have probably moved just as much with a door in place of the pump. Since magma doesn't pressurize the same way that water does, you might get different results with magma. I'd do a test before getting too far along with a design that is intended for magma.

   Keith

I have considered that, but magma testing is far more time-consuming, so likely I'll just go right ahead with the live test in an established fort to see if the behavior is any different.  Mostly I wanted to make sure that if I had a stack of say 5 7/7 magma tiles, I could expect a single pump activated by a pressure plate to pump the whole thing, or if I would need to latch the system.  Initial testing with water indicates that I'd probably be okay without a latch provided that the magma fell fast enough.  I'm kind of assuming magma and water fall at the same speed, which may not be true, but it seems to at least be reasonable.  Again, water was mostly an initial test because it was fast and easy.  Magma testing is next.

I should do a comparison of the full setup, too once I know how this kind of pump setup works with magma.  There are a few issues I'm worried about using just gravity feeding with a door (mostly evaporation related) but those may only be theoretical issues.

Anyway, more science on the horizon.

EDIT:  *Sigh*  Long day and I'm too worn out for science.  Somebody else science and then let me steal credit.

[Miuramir]

Based on this test, I'm pretty confident that a pump tries to pump every tick and is only actually limited by the need to have at least 2/7 on the input tile (thus it is essentially inflow limited, outflow unlimited). 

That's what has been generally understood.  The interesting thing about pump stack construction is that order of construction is important; do it right, and a block of liquid can teleport up the whole stack in a single tick; do it wrong, and the block will move up only one per tick.  (The game evaluates pumps in reverse construction order IIRC.)

One trick that can be helpful for moving magma, which is slower to flow, is to have three intake pumps feeding the bottom of the stack pump

[Tacomagic]

Based on this test, I'm pretty confident that a pump tries to pump every tick and is only actually limited by the need to have at least 2/7 on the input tile (thus it is essentially inflow limited, outflow unlimited). 

That's what has been generally understood.  The interesting thing about pump stack construction is that order of construction is important; do it right, and a block of liquid can teleport up the whole stack in a single tick; do it wrong, and the block will move up only one per tick.  (The game evaluates pumps in reverse construction order IIRC.)

One trick that can be helpful for moving magma, which is slower to flow, is to have three intake pumps feeding the bottom of the stack pump

I've used that trick quite often, and it's on the wiki.  I just couldnt find the information of how quickly they pumped.  Granted the forum search is rather... yeah.  In any event, I've added this information to the wiki so that the next person who has this question can get their answer without having to reinvent the wheel... as I apparently did.

[jcochran]

That's what has been generally understood.  The interesting thing about pump stack construction is that order of construction is important; do it right, and a block of liquid can teleport up the whole stack in a single tick; do it wrong, and the block will move up only one per tick.  (The game evaluates pumps in reverse construction order IIRC.)

The issue with build order doesn't affect how much fluid over time that can be pumped. It affects the latency and storage capacity of a pump stack.

For instance, assume you have two pump stacks of 100 levels. One stack is built from the top down, the other is built from the bottom up. Also assume that the bottom pump of each stack has a somewhat plentiful supply of fluid. Say 3/7 every tick.

Finally, assume you have a floodgate at the output of the pump at the very top of the stack, and that you also can turn on or off power to the stack via a gear.

So you open the floodgate and turn on power to both stacks.
The top down stack immediately starts to supply 3/7 fluid every tick.
The bottom up stack will take 100 ticks before is starts to supply fluid. However, once it does start supplying fluid, it continues at the rate of 3/7 per tick (same rate as the top down stack).
Now close the floodgate at the top of each stack while leaving the power on. Both stacks will eventually fill up with 7/7 fluid at the output of each pump.

Now open the floodgates and see what happens.
The top down stack will supply 7/7 fluid every tick for 100 ticks, then slow down to 3/7 fluid per tick. Additionally, there will be no fluid in the stack itself since the stack was unable to pull fluid from the source until the stack itself was empty.

The bottom up stack will also supply 7/7 fluid every tick for 100 ticks and slow down to 3/7 fluid per tick. However, each level of the stack will have 3/7 fluid so when the output floodgate is turned off, the bottom up stack will fill up to 7/7 fluid on each level sooner than the top down stack with was effectively empty while supplying 3/7 fluid.

Frankly, both stacks have the same long term capacity for moving fluid and deciding on top down or bottom up is merely an exercise in perversity. What I do when building a pump stack is to build a scaffold gear or axial every 3 levels, then build a pump attached to those gears or axials. That lets be build quite a few pumps simultaneously. After those pumps are constructed, I'll then build all the other pumps for the stack in an other mass designate (macros are your friend). Finally, I'll remove the scaffolding. Total construction effort on my part is simply 4 designation sessions for the entire stack instead of build a pump, wait for the pump to be constructed, build next pump, rinse, lather, repeat.