Print

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

[zchris13]

Repeater.  Not mist.  Multi sequence repeater, to be exact.

[Nightwind]

ah, you've fixed the original design to do what it was ment to be in the start, got it

[AncientEnemy]

This actually started out as something completely different (a repeater that didn't function) but it turned out to make one of my favorite things: LOTS OF MIST!

Actually after lots of tinkering I finally managed to make a perfectly working and reliable repeater based on your design. :-)

The system design is the same as for magma-killing machine. However:

1. Each pump is powered through a separate disengageable gear (this is the important bit).
2. On the bottom of each of 4 pits there's a pressure plate, set to trigger on water 5-7.
3. Pressure plates are linked to gears powering the pumps in the following fashion (assuming the water rotates anti-clockwise):

  a. NE pressure plate is connected to W pump.
  b. NW pressure plate is connected to S pump.
  c. SW pressure plate is connected to E pump.
  d. SE pressure plate is connected to N pump.

System initialization:

1. Disengate power to the entire system.
2. Fill one of the pits.

After the pit is filled the pressure plate inside the pit triggers. As the result, the pump that will drain this pit is connected, but the next pump in the queue is not.

Then power up the entire system. The water is immediately shuffled to the next pit but not further because pressure plate resetting takes some time. By the time the first pressure plate resets (activating the second pump), the second plate has already triggered, thus deactivating the third pump. And the sequence goes on and on.

oh cool. I thought of a design like this but i didn't realize that gear assemblies were on reverse settings (ON = disengaged). how many steps is it between each ON and OFF (assuming you link one device to all four plates)?

[Lav]

oh cool. I thought of a design like this but i didn't realize that gear assemblies were on reverse settings (ON = disengaged). how many steps is it between each ON and OFF (assuming you link one device to all four plates)?

Well, you can look for yourself. I couldn't upload a movie to DFMA yesterday for whatever reason, but now it works for me:

http://mkv25.net/dfma/movie-1370-pump-basedautorepeater

Actually it's perfectly possible to use only two pressure plates (for example, NE plate controlling S pump and SW plate controlling N pump). Or three plates.

As for gear assemblies, that's what I didn't understand myself at first. I got one idea correct, that you must power each pump separately so it could be controlled on an individual basis but I thought that pressure plates only sends one signal, when triggering conditions are met, and no signals thereafter. I quickly found out I was wrong however, and that the pressure plates actually send two signals: ON signal when conditions are met and OFF signal when the plate resets after the conditions are no longer met (with a resetting delay).

Anyway, now it is the time for the Megaproject of my dream: Dwarven Binary Summator. I already have the schematics for AND, OR and NOT elements in my mind, but proper triggers were out of reach without a synchronizing sequence generator.

P. S. I also made an autorepeater based on the mini-waterfall schematics. Both pumps separated by floors (so power does not transfer) and individually controllable. The theory was that the water falling two tiles instead of one should generate some mist and so the mini-repeater will double as an erzats-waterfall. Well it does, but unfortunately it's a very poor waterfall. :-(

[AncientEnemy]

i'll have to build my own when i get home from work. i have no doubt that it functions as a cool sequential repeater. my initial goal was to create an ultra-fast repeater. i've been finding it difficult to make a design that breaks the 100-step-cycle-barrier. im too tired to math in my head right now but i don't think the sequential-pump thing we've got here is going to do that, althought it still aught to be slightly faster than my other ultra-simple repeater with the floodgates and bridge.

what i really want is to figure out a controlled way of getting two independent repeaters going, each offset from eachother by 50 steps.

i really have no idea what the applications would be, sincemost devices are limited to 100 steps anyway. except max-speed spike traps, or wildly flailing doors. i'm just obsessed. and im sure if i come up with the design, someone else will find a suitable use for it.

somehow this has become my reason to DF. I don't really care about megaprojects, but I get wayyy into micro-projects as it were, finding the simplest and fastest possible devices like these.

[Lav]

Any single repeater based on pressure plates will be limited by the speed of the plate resetting, that much I have already learned. :-(

Hooking up several repeaters might work though. Assuming we have several sequencers with interconnected plates/gears:

Seq A advances to step 1, sends signal to Seq B Pump 1.
Seq B advances to step 1, sends signal to Seq C Pump 1.
Seq C advances to step 1, sends signal to Seq D Pump 1...
Seq N advances to step 1, sends signal to Seq A Pump 2.
Seq A advances to step 2...

Of course, if you want to get maximum frequency signal from this system, you will have to hook every pressure plate in every sequencer to the target device. With four sequencers each containing 2 pumps this amounts to 8 pressure plates and 16 mechanisms.

althought it still aught to be slightly faster than my other ultra-simple repeater with the floodgates and bridge

I'm not sure it's faster at all. But this design does not need an infinite water source and can be placed anywhere, which is the main selling point. :-)

P. S. Bridge flinging does not work as I expected... damn. I so much hoped to make a Dwarven Rabid Transit System... But at least the Atom Smasher can be automated (though it's no big news).

The day when bridges will fling objects in the direction they actually should will be the happiest in my life.

[Vactor]

I think you may be able to use the time it takes water to fall as a way to offset a set of repeaters. using varying heights you can create different delays.

[chaos985]

question, If you hooked up two pressure plates to the same device, one set to off, and one set to on.  And disabled reset, could you set both of the plates on a simple repater system, and have it go fast, or would it just cycle on/off once, then be done?

[Itnetlolor]

Ooh, this just gave me an idea for a defense gauntlet, or prison escape gauntlet.

not only are the pressure plates linked up to the gears for the pumps, but each individual pressure plate is designated to a device in the gauntlet.

You know, 2 for floodgate alternating path mazes, and another 2 for spike strips. Of course, I'm thinking small, but the repeater has potential as a repetitive sequence trigger. Doors A (Started closed), then spikes A (Started Down), then doors B (Started Open), then spikes B (Started Up). Opening and closing in sequence, of course.

Make a defense grid a work of art.

[eerr]

is the reset speed taking into account the fact that masterwork/artifact mechanisms and linkages trigger faster?

[zchris13]

is the reset speed taking into account the fact that masterwork/artifact mechanisms and linkages trigger faster?

LIES. They don't do that.
Point me to your research.

[Akhier the Dragon hearted]

is the reset speed taking into account the fact that masterwork/artifact mechanisms and linkages trigger faster?

A mechanism's quality has no effect on how quickly it operates.

http://dwarf.lendemaindeveille.com/index.php/Mechanism

[ItchyBeard]

oh cool. I thought of a design like this but i didn't realize that gear assemblies were on reverse settings (ON = disengaged). how many steps is it between each ON and OFF (assuming you link one device to all four plates)?

Well, you can look for yourself. I couldn't upload a movie to DFMA yesterday for whatever reason, but now it works for me:

Gears assemblies are kinda screwy in that they don't behave the same way as the on/off lever semantics used everywhere else (gears toggle). You can verify this by just hooking two levers up to the same gear and then flipping them both (with a slight delay between flips). You'll end up with the gear in the original state. The same works for pressure plates set to creatures (use military dwarves to test). I believe it works the same for liquid plates though I haven't actually tested it properly.

I ended up using levers to turn my gears off then setting the pressure plates in my design to 'on' at 0-1 water. Each pressure plate is set to drive the next pump in sequence. When water lands on the square, it sends an 'off' signal to the disabled gear which re-enables it. When the water gets pumped out again, the plate sends the 'on' signal to the running gear disabling it again. It sounds like you get the exact same behaviour from using 'on' as 2-7. Hurts my brain. 

Anyway, I've also been fiddling around with a repeater based on the mist generator but with a different intent. I was aiming at creating arbitrary timing mechanisms (mainly for traps, but a seasonal clock reset when water freezes in winter would be very dwarfy).

While it's hard to speed up the repeaters, slowing them down is easy. I made a video of one of my experiments - this one showing 4 of the repeaters hooked up in sequence, each repeater operating at half the speed of the last one. You could get very precise timing by adding more repeaters in sequence (which would again halve the speed) or by increasing the size of the loops (for more fine-grained control).

Movie here: http://mkv25.net/dfma/movie-1385-clockmechanismtest2

Apologies in advance for movie length - first movie I've posted. The upper level (with my horrible test setup) is shown towards the end of the movie.

I experimented with having a single plate in the first repeater drive all of the pumps in the 2nd, but that resulted in the water spinning too quickly in the second (like the original mist generator). The design in the movie uses 2 plates in the first repeater, each driving half of the pumps in the 2nd (so that the water will only ever move a single place at a time in the 2nd repeater). The 2nd in turn has 2 plates which each drive half the pumps in the 3rd, etc.

[ItchyBeard]

Gears assemblies are kinda screwy in that they don't behave the same way as the on/off lever semantics used everywhere else (gears toggle).
...
The same works for pressure plates set to creatures (use military dwarves to test). I believe it works the same for liquid plates though I haven't actually tested it properly.

I just exhaustively tested this, just to confirm it for myself.

Specific to gear assemblies, the concept of 'on' and 'off' signals do not exist like they do for other mechanically operated things (e.g. floodgates, doors, spikes). Both 'on' and 'off' signals from a pressure plate or lever are treated as 'toggle' by a gear assembly.

I have tested this with:
* Dual levers - triggered individually
* Dual pressure plates (configured for creatures) - triggered individually
* Dual pressure plates (configured for 'on' being 0-1) - triggered individually, both 'on' and 'off' signals.
* Dual pressure plates (configured for 'on' being 2-7) - triggered individually, both 'on' and 'off' signals.

I every case, the state of the gear changed when the state of one of the pressure plates or levers changed, irrespective of what it changed to.

As an additional test, I set up a situation where both liquid pressure plates (configured for 0-1 being 'on') would trigger at the same time. If both plates trigger at the same time, the gear did not appear to change state (or more likely it went through 2 state transitions to arrive at its initial state). So it's safe - you're not going to lose a signal due to timing when operating gears.

As a final test, I checked the speed of the signal from the pressure plate to the mechanism - it appears to be instant (the gear changed from disengaged to engaged in the same frame as the water on the plate dropped to 1, as expected).

Tested on 40d.

[Lav]

I haven't had time to test if gears work on toggle signals yet. But if you are correct, then a high-frequency repeater might be possible. I'll have to think about this.