I made a clock on pixel packs

[Hokaeru]

Hello everyone,

I was wondering what can I use automation for in my bases, I thought of a clock. It is not very useful at the moment. It can go from 00 to 99. I built it in debug, and I will try to make it as compact as possible and to build it in survival. At the moment it is quite big.
It is based on shifted "advanced mode" signal counters and memory toggles.
After finishing it and searching on the forums if it had already been done, I have seen the signal counter stopwatch : https://forums.kleientertainment.com/forums/topic/119551-signal-counter-stopwatch/?tab=comments#comment-1348535  but I think this is somewhat cooler on pixel packs. 
What would you use it for ? I was thinking of a cycle count, and every multiple of 365 it would send a notification.

Have a nice day ! 

P.S. : This is my first GIF, and I was limited by size, so sorry if it's fast and short.

[darknotezero]

47 minutes ago, Hokaeru said:

What would you use it for ? I was thinking of a cycle count, and every multiple of 365 it would send a notification.

A while back I posted my version of a clock/timer that I was using to accurately track geyser active/dormancy periods using a very simple water clock and a lot of automation:

Your use of pixel blocks is much much more elegant looking and has the added bonus of +decor, so i might try to tinker around with that.

If you're not already familiar with the issues of long-term clock tracking using buffer and filter gates, make sure you know about it, and do a search for water clocks.

[JRup]

@Hokaeru: How are you generating the single seconds? Are you using the timer sensor or some other interesting means?

[Hokaeru]

@darknotezero thank you, it is indeed very interesting ! 
@JRup The first goal was for it to display properly, so at the moment it is using timer sensor 1s/1s. However you can trigger the ticking with any green signal you can think of. 
It could be used to count the number of critters dead for the colony by ticking each time an egg goes into the water 

[LiveActionPixel]

I'd love to see the signalling for this if you get chance to take any more screenies   

[Hokaeru]

@LiveActionPixel of course if I can be of any help
A timer is plugged to the 10 signal counters which are shifted, are on "advanced mode", and count to 10. They have a memory toggle that is reset by the "next" counter to reach 10. Then from the memory toggle it goes to a lot of ribbon writers, respectively to bit 1,2,3 -> 5 times. I deleted the useless ones (to shape numbers).
Then the one responsible for the "0", pluggs into 10 other counters that work the same but for the tens digit.

As I said, it was kind of a challenge for me to understand the whole thing. I think it is pretty "brute force" and messy. There is lots of room for compaction and improvement. 

[darknotezero]

aaaah interesting. 

true that it's a little brute force, but the way that automation works there aren't a lot of options other than brute force to accomplish what you were able to accomplish.

All about it. I may mess around with this design some and see what I can do with it, and then incorporate it into my own sorts of big timer designs.

Well done.

[JRup]

6 hours ago, Hokaeru said:

so at the moment it is using timer sensor 1s/1s

Hm, at some point I had wanted to build a timer with this sensor. I could recommend using 0.5s/0.5s. That would give the actual period of time for 1s. What 1s on/1s off would do is provide a period of 2 seconds total.

[Hokaeru]

darknotezero : I'm glad that you like it, I have very bad understanding of electronics so when I see people making a 16 bit ROM on ribbon, I am quite... puzzled. 
JRup : Yes for the timers it was not the important part since it would be to count cycles, not seconds, but noted, thanks !  

[JRup]

Here's a power consuming-heat generating water clock that has a T=2s... I'd made it just for fun in sandbox. Turning it on for the first time will give an inaccurate result because of the water in the piping. Valve is set to 5000g/s and total liquid used inside the pipes is 20kg (or just two blobs)

The transformer must be used to not throw the voltage sensor off...

Point is, any idea is good as long as it works.

[Saturnus]

1 hour ago, JRup said:

Point is, any idea is good as long as it works.

This is an old, very old, favourite of mine. It's a 100% accurate 1s off/1s on clock but can be expanded to other settings by using a longer loop. It uses two different blobs of liquid that just circulate via the bridge and an element sensor to sense when the triggering liquid passes.

[JRup]

4 minutes ago, Saturnus said:

This is an old, very old, favourite of mine. It's a 100% accurate 1s off/1s on clock but can be expanded to other settings by using a longer loop. It uses two different blobs of liquid that just circulate via the bridge and an element sensor to sense when the triggering liquid passes.

But but but, where's the over-complication?

Now seriously, wonderfully neat! Now we just have to find a way to divide T for the miraculous 1s without the timer sensor... (Over complication returns! - maybe) 

[Saturnus]

15 minutes ago, JRup said:

But but but, where's the over-complication?

Now seriously, wonderfully neat! Now we just have to find a way to divide T for the miraculous 1s without the timer sensor... (Over complication returns! - maybe) 

Just make an edge detector with a NOT and an XOR gate. This will produce a negative pulse every time the clock output changes state. (So it's on for 0.9s and off for 0.1s recurring).
 

[darknotezero]

53 minutes ago, Saturnus said:

This is an old, very old, favourite of mine. It's a 100% accurate 1s off/1s on clock but can be expanded to other settings by using a longer loop. It uses two different blobs of liquid that just circulate via the bridge and an element sensor to sense when the triggering liquid passes.

 

This is exactly how i have my water clock set up except it has length of pipe to cover one-tenth of a cycle. Which primarily came from me looking at a lot of other people's water clocks design and thinking, "this is waaay beyond me right now, i need to simplify this to its basic elements so i can understand it."

[JRup]

The signal distributor seems a bit overkill, but I guess you have your design reasons.

[darknotezero]

14 minutes ago, JRup said:

The signal distributor seems a bit overkill, but I guess you have your design reasons.

yeah it doesn't really need to be there. it's a half-realized concept that i haven't finished building in my actual colony and may never get around to because the automation required to do what i want to do would take up a large footprint unless i can condense the design, and it's less important now than it initially was when i started conceiving of it.

Spoiler

 

[JRup]

Clocks get crazy.

[darknotezero]

so i messed around with the design of this some and came up with this:

mostly the same concept with a few differences:

1. didn't use Advanced Mode which took out the need for the memory toggles.
2. For each pixel "row" configuration of on/off, i created 'sets' of ribbon writers for each configuration, and labeled them with a ribbon reader as either set 1, 2, or 3, and then for each number, i would set the ribbon writer to that set. This meant that I only needed to create, say, for the middle row, a 000 config for 2/3/4/5/6/9, a 0x0 config for 0/8, and a xx0 config for 1/7 once, and then just assign the proper set for each number.

It's still a bear of a set up. I mucked around a bunch trying to see if I could get the tens place to leverage the same set up instead of having to duplicate everything, but until there's a way to use a signal distributor or selector for 4-bit info rather than just 1-bit info, i don't think it's possible. I haven't completely given up on it yet though, so we'll see if i can figure it out once i come back to it.

 

As an alternative, i used @Hokaeru's nifty offset-counters-to-create-isolated-green-signals design to make a less flashy but still functional pixel clock that's more like a 'grid pixel' clock.

less easy to read, but much simpler and smaller footprint.

I also tried to get it so that it would be a 'color wipe' instead of a color replace where a pixel stays green until the whole thing resets. i know how i would do it but it would increase the footprint a great deal so it doesn't feel worth it, especially if i were to expand this to 3 or 4 digits

Next step is to integrate this stuff into my automated dormant/active geyser timer.

[kbn]

The disadvantage is that it takes about 0.5 seconds to display, but you may be able to simplify the display part a little by using bit shift.

The ten Ribbon Writer next to the Signal Selector are the memory with the input and output connected. Each contains font data from 0 to 9.
The binary counter circuit that combines AND and XOR at the top generates signals from 0000 to 1001 and operates the Signal Selector.
The many switches on the right are the devices used to create the font data from 0 to 9 stored in Ribbon Writer's memory. It is unnecessary after storing the font data in memory.

Save data:Abyss.sav

[Hokaeru]

Darknotezero and Kbn, very nice !  

[darknotezero]

11 hours ago, kbn said:

The disadvantage is that it takes about 0.5 seconds to display, but you may be able to simplify the display part a little by using bit shift.

 

This is exactly the sort of thing I was trying to sort out but I couldn't wrap my head around it.

Also I HAD NO IDEA YOU COULD SEND RIBBON DATA THROUGH THE SELECTOR. That's a game changer.

Love this design, well done.

[Saturnus]

7 hours ago, darknotezero said:

Also I HAD NO IDEA YOU COULD SEND RIBBON DATA THROUGH THE SELECTOR. That's a game changer.

You should probably check this thread out then.
 

 

[Hokaeru]

Yes it is very very clever, and at the moment, still quite beyond my grasp, I will have to tinker with this to get familiar with it.
A question, are you familiar with computing or did you trial and error in ONI ? 

[OxCD]

On 12/4/2020 at 5:53 PM, Saturnus said:

Just make an edge detector with a NOT and an XOR gate. This will produce a negative pulse every time the clock output changes state.

Thought about that before reading this from you. Just because learned this years ago from a topic about edge detector from you ^^ 

 

On 12/4/2020 at 5:53 PM, Saturnus said:

So it's on for 0.9s and off for 0.1s recurring

Just a quick question, is automation state change an half-tick ? I may be mistaken, but as far as I remember there's 5 ticks into 1 sec, so your sentence would mean that automation pulse is half-ticked. Did I get it correctly ?

[Saturnus]

1 hour ago, OxCD said:

Just a quick question, is automation state change an half-tick ? I may be mistaken, but as far as I remember there's 5 ticks into 1 sec, so your sentence would mean that automation pulse is half-ticked. Did I get it correctly ?

No. An automation tick is 0.1s. It's been consistently proven.