Print

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

[suokko]

I didn't notice any clock generator optimized for upright spikes. I decided to experiment my self with minecart clock generators. Design target is to generate alternating open-close signal every 41 ticks.

My first attempt with roller was somewhat bad direction but it takes only 5x5 spaces plus power source. I hid it in case someone wants to check it.

Spoiler (click to show/hide)

Here is my best design so far that has only 14 ticks tolerance.



Left is complete build for signal generation. Right is the track only view.

Requirements are power source (blue gear) to a low speed roller (blue N->S). There is 4 separate presure plates to generate 4 different output signals. There is also two low friction stops adjusting the timing for signals. Micecart insertion is simple push  towards north from pictured position (it will turn back to south from the roller)

There is 8 signals generated from 4 pressure plates but each pressure plate generates one signal that is ignored because spike already has another signal operating it.

To use the combined signal one has to link all pressure plates to all spikes requiring 8 mechanism per spike. That might be quite expensive requirement if you ever end up having stone shortages.

I suppose it might be possible to adjust the timing if moving roller and adjusting stops again. But that feels to me very hard operation because each signal has to be timed perfectly to two other signals requiring multiple adjustments to avoid break low tolerance signals. In short I feel it is very hard to improve this with only minor changes.

Time	Signal
0	Bottom pressure plate sends open signal
8	Left pressure plate sends close signal (ignored)
26	Top pressure plate sends open signal (ignored)
41	Right pressure plate sends close signal
82	Left pressure plate sends open signal
100	Bottom pressure plate sends close signal (ignored)
115	Right pressure plate sends open signal (ignored)
126	Top pressure plate sends close signal
174	Back to begin repeating the signal

First two signals haver very littler tolerance if there is minor variance to times because cart enters to round at different position. But if there is random variance that breaks one of signals for a round the broken spike state last for very short time because this design corrects the state automatically when signal is back to normal.

I made some extra reach based on helpful links by Larix. I managed to get my first 164 tick repeater to generate 4 signals with only two pressure plates. I suppose with some minor tweaking spaces requirement and stop requirements could be improved. But I decided to document the improved prototype version here in case someone else likes to design something better than I can come up tomorrow.



That was built on top of my test version of http://www.bay12forums.com/smf/index.php?topic=130767.msg4579632#msg4579632 that I used to verify that I understood the impulse ramp control correctly.

Track stops are medium frictions. The cart leaves second time the pressure plate 23 ticks after  entering it. Whole round trip is 164 ticks.

164 tick two pressure plate design

I took the knowledge gained from the prototype design  to come up with a smaller design that generates the 82 tick cycle signal using 164 tick track with two pressure plates. This is reducing mechanism requirements for actual use to half what 4 plate designs require.



Top right and bottom left halves are mirrir images of each other. The motion for the cart is generated from z-1 level in corners. There is two ramps (one under the wall in corner). Pressure plates are in crossing that will be left second time 23 ticks after first the plate was tricked. After loop there is 4 floor tiles slowing down the cart a bit so half section runs at 82 ticks. I first carved the complete track and the smoothed those 8 tiles.

Possible control mechanism is hatchet cover with lever like the Larix's design tells.

This design is quite easy to tweak making. If someone wants it shouldn't be hard to shrink the required floor space a bit. Another possible optimization is to drop cycle to 160 ticks and depend on build order to make pikes run a bit faster. That would require loop run to drop from 23 ticks to 20 ticks which should be about 2 corners less before entering the loop.

This design is heavily influenced and helped by Larix's earlier published work. Following his links gives information why and how this design works and what are limitations if wanting to improve this..

[Larix]

Well, there are these two (both in the same thread):
http://www.bay12forums.com/smf/index.php?topic=130767.msg4579123#msg4579123
and
http://www.bay12forums.com/smf/index.php?topic=130767.msg4579632#msg4579632
(the second has a ramp underneath solid ceiling next to each visible ramp; track on ramps is plain NS).

Both designs have the minimum repeat time of four signals in 164 steps and a full extend/retract cycle of the spikes in 82. They're both using ramp bugs to get perpetual motion without power. I've built and used mine in a fortress and it worked perfectly.

A powered design would have the benefit of being bugfix-proof

[suokko]

Thanks. That was what I was looking for. But I guess I could then continue my design towards 164 tick cycle trying to get stable speed from the ramp bug and with only two plates. It should be possible to make track cross a plate twice to generate close signal 123 ticks after open signal.

With my initial attempts with ramps I had fairly unstable signal making it impossible to aim for 41 ticks signal difference.