Rotational Logic Gates (Proof of Concept) | Lego Computer #2
ฝัง
- เผยแพร่เมื่อ 4 มิ.ย. 2024
- Part 1:
• A Computer made of Leg...
Hi,
this is a proof of the concept of rotational logic gates. Please don't take it too seriously, and please don't build them (if you want that your motor keeps working).
I hope you enjoy watching and get some ideas,
2in1
--------------------------------------------------------------
Links:
Rebrickable (part list and .io file):
rebrickable.com/mocs/MOC-105123/
rebrickable.com/mocs/MOC-105124/
rebrickable.com/mocs/MOC-105125/
rebrickable.com/mocs/MOC-105126/
Music:
Underwaterbeats - Neon Lights | • Underwaterbeats - Neon...
--------------------------------------------------------------
Timestamps:
0:00 Intro
0:06 Rules
0:30 OR
1:58 NOT
4:13 AND (really bad)
8:00 XOR
10:14 Thanks for watching!
--------------------------------------------------------------
#lego #technic #computer #logic #mechanism - วิทยาศาสตร์และเทคโนโลยี
Watch the first video:
th-cam.com/video/n_A5mzgXFLY/w-d-xo.html
Just a note for everyone who wanted to write a comment that I could make True clockwise and False counterclockwise: I know, it would be really easy to make a NOT-gate, but if the output has to rotate counterclockwise when being false, it's almost impossible to make AND or OR. If you have any ideas how to, make sure to share them!
Maybe utilise a simple differential mechanism with one input and gravity latches on both outputs, such that clockwise input results in one output rotating, and anticlockwise input results in the other output rotating. Then build on this. You will need multiple differentials to.build an and/or gate i think.
Nice idea, that's kind of a converter so that you can use these ones :)
If you make bitstate not continuous ccw/cw rotation, but limited transition it effectively will be equivalent to linear translation logic.
Linear translation can controll switches, so it possible to amplify signals.
Also for complex block you can use rotational analog calculations with DAC/ADC converter on input/output (ADC for few bits can be done like caterpillar with two pin on it for every bit. One pin code 0 other 1(it's basically sequential ROM). Two chained 4bit adder made that way probably more efficient in terms of part count than binary one. DAC is expensive one. It can costs one differetial for one bit of data minus one... So two 4 bit inputs needs 6 differencials at least. So summator probably can be done with 7)
Sorry for long post.
I actually put a lot of work into this a couple years back - I think this is the best way actually since it accommodates repeaters which are incredibly important if you want to move to sequential logic. Anyway it turns out that a multiplexer is a complete set for binary logic and they’re pretty easy to built with minimal friction. You do a similar thing to the absolute value gate you had for the xor - the selector has some minimal friction which turns an arm with two free gears on either end which connect an output to two independent inputs. There’s an issue with the gears detaching themselves and the solution is to have two arms on either side of the system. (This is a pretty poor explanation - I leave it to you to figure out, I also leave it to you as to how to use a mux for a complete logical set)
One more thing, working through this has made me realize that logic gates are not the most efficient computational system for legos. A friend of mine has been considering a Turing machine using chains which seems like a good idea with significantly more power per unit lego
4:13 for the *AND-gate,* you could make one motor start the rotations and the second motor closing a switch (with a clutch gear). The switch consists of a liftarm and two gears, with one of the gears constantly connected to motor 1 and the second gear only connected to the output if motor 2 turns clockwise, rotating the liftarm and pushing the second gear against another gear, passing on the rotation to the output.
(Use a rubber band to keep the switch open while motor 2 doesnt turn.)
The rotation speed of the second motor doesn't influence the rotation speed of the output yet, but if you want, you can simply make another one of the same contraption, but for the second contraption, motor 2 does the rotations and motor 1 closes the switch with a clutch gear.
If you connect both outputs in the end, both motors are responsible for rotation speed. But as long as one of them doesn't turn, the output doesn't turn, resulting in a very simple AND-gate.
Basically, that's exactly what I did. The part with the rubber band and the clutch gear (in my case differential) is really hard to design without much friction... It could be solved by adjusting the rubber band strenght and gearing next to the differential.
BTW the lengh of your comment broke youtube :D
The next step is to combine the rotational logic with the mechanical logic. Soon you will have a super computer.
soup computer*
(joke)
For the AND gatter u can use the demorganic rule if u put 2 NOT in front of a OR and 1 NOT behind u get a AND it also works in the opposite direction (2NOT in front of AND + behind --> OR)
yeah same thought
additional power
yo, do I have your permission to try and make a 2 or 3 bit simple binary adder? I would create my own mechanisms obviously but proof of concept is useful for knowing it's possible lol.
Feel free to do that! It won't be easy, though... I'm looking forward to your version :)
@@in1 it will be hard, but I’ll get there lol
@@in1 I'm finished. 9 marbles to add two 3 bit numbers, so anything from 0+0 to 7+7. works super well, had to make the mechanisms more optimised by combining gates in ways a computer wouldnt realistically do but ah well. thanks for the idea! video soon
@@puzzLEGO Amazing! When building the stuff for this video, I also considered marbles, but I was only able to build a flip-flop :) I'm looking forward to your video!
Great designs! currently building a LEGO computer myself and these are great references and ideas, to get inspiration from.
How far are you?
@@schwarzerritter5724 Well I build a simple adder using rotational logic gates and it actually worked. I just never got around to making the final video showing it off in its completed state.
@@Axiom12543 I am still in the early planning stage.
Do you know the Aerial Tram from the 8891 Idea Book? The most underrated official technic model ever. It has a mechanical switch I believe I can use. Although the book was written before the clutch gears with driving ring you use in your builds where a thing, so that might be the better option.
@@Axiom12543 I believe the fundamental problem with all mechanical logic gates I see on the internet is that the output of each logic gate provides the power for the input of the next logic gate, so mechanical resistances and tolerances keep adding and the question is just which makes the build unusable first.
Electronic logic gates by comparison are connected to a power sources. The power at the output comes from the power source and the inputs are just switches that decide if the power gets put through and I believe that is how mechanical logic gates need to be build.
@@schwarzerritter5724 definitely would be the best way to do it, although the amount of power functions needed is a turnoff. I ended up making and-gates with nor-logic and every not-gate would have a build in motor, so they provided the extra power needed with all the friction. I would recommend if you try to build it, that you construct and-gates this way. I fiddled around with countless designs and that was definitely the easiest way and worked the best.
Starting my first year of engineering and this is such a a great way of visualizing logic gates, I imagine it'll all come in handy later
A differential gear on its own may serve as the core of a working AND gate.
Although I'm thinking of a system where one direction could be considered 1 and the opposite direction 0.
Such a system would need a continuous source of rotation to power each gate, but this is analogous to many CMOS circuits.
The problem with that would be that if one gear is true, it would output something in-between true and false. I don't really know a fix for that, but there could be one.
The mechanisms are always so creative and interesting. Good job 👍 keep it up
That XOR gate is one of the coolest things I've ever seen made of Lego
Hey! So for the end gate - maybe because I don't have that LEGO piece - you could use the torque piece or what. So my idea is to have two axis'(inputs) with the clutch gear on them, and one big wheel in the middle as an output. If only one axis is rotating, then the torque wheel clicks in and doesn't allow the wheel to rotate, because the other torque wheel is functioning as a break. If both axis' are rotating in the same direction, then nothing is braking so they are free to go, and the big wheel gets rotated.
I'd wager you could make a simpler AND gate with the help of some one way clutches. I'll try it out in the afternoon and share it if it works.
Also a question. Let's consider logic gates that are able to take three inputs simultaneously. In your personal opinion, what would be the output of a three input XOR gate if all three inputs were true?
Keep in mind the existence of the odd parity function which outputs true when the number of true inputs is odd, and the non-equivalence function which outputs true whenever all inputs don't have the same value. With only two inputs these two functions along with the XOR function have the same behavior, but they are not the same
I stumbled over that topic a recently when I programmed a python script... What I found actually was what you found: the output is true if an odd number of inputs is true.
If I personally created the gate, I would propably say that there's no XOR gate with three inputs, similar to there's no NOT gate with two innputs. But that's the definition, I guess.
Another great video! I love your AND gate, it's really cool, nice job :)
using the differentials is a very creative solution
There are already some designs that allows you to build push and pull logic gates without usiung springs / elastic bands or additional power sources. I wonder if we can build flip-latch and flip-flop with these.
amusingly I actually needed something like this, though more a rotary adder than and gate
Then a simple differential does the job...
So I engender on Legos and your vids have helped me out so much I think you will get 1 more subscriber today keep up the good work!
So cool dude keep it going, I love logic and I really like your bulk head style xor gate, as a co.puter programmer I think this is really awesome, I dig your channel, thanks!
Glad you like it :)
Wow, really cool! I wish I had as much Lego skill as you!
Alternatieve xor gate: reverse turn of one input. And add an direction correction to the output. Typically the latter is called 'two-way to one-way' on TH-cam an Rebickable. This might also help inspiration for an AND gate alternative.
That's basically what I did :) The mechanism after the differential is a modified "two-way to one-way"-converter. I needed a 2:1 transmission, so I couldn't use my other designs.
Oh my word... He's done it with gears. 😯
Hello, for the XOR, use the 2 planetary sides as inputs and 'usual' input as output.
input ---|o|--input
out
Only problem : output will rotate half speed if only on motor is driving.
I could probably fix that, for example by taking the not-gate and by replacing the motor with an output. The problem would be, that I would still have to handle the problem that the output direction is different depending on the input.
To make the best logic gates, you should simply make some "Switches" (What transistors would be in a real computer) and assemble them differently for each gate.
I know quite a bit about computer science, so ill probably comment on next episode too. Hope you see this! Love your work, keep it up!
Thanks! Basically, a transistor is an AND-gate. Theoretically, you could build the most complex computer only using AND and NOT
@@in1just a NAND also suffices. If you could design a NAND with little slack that would be great
For the NOT gate, would a True input that mechanically disconnects the output from the continuous power source be a more efficient way to do it?
It definetely would be, it would have the same problems as the AND-gate then, though
For the And gate, you could use a servo. To power it, you could use a regular motor as a generator.
Great idea :D The servo works with a special frequency of binary signals provided by the infrared reciever, though, so probably you'd need a different one. Would still be better than my current solution :)
How long does it take to make this great videos? Like, finding the idea, building the idea then, filming?
Probably like 15 hrs, this time a bit more because it's much longer than the average video. Probably something like that:
7 hrs building
2 hrs filming
3 hrs editing
2 hrs building the models in 3d CAD (thumbnail etc.)
1 hour uploading
@@in1 Thats a lot of time! So glad you do it though, cause it makes GREAT videos!
for the AND, u could connect the 2 inputs to the output of a differential, then add a one direction mechanism then a not gate
The problem could be, that the output already turns at half speed when only one input is true
@@in1 oh yh. Thanks for replying
Germany’s Enigma machine used this exact principle to encode their messages
An and gate could use the idea that one input make sure the other one has a connection, by using makeshift to push?
I tried that, actally when there is a small friction on the output, it slips, so the gearbox design worked a littlt bit better (and has the same problem that there has to be a rubber band that pulls it back)
I’m a ‘starting TH-camr’ making Lego videos too, but I found it very difficult to find a good background music for my videos, so I would like to know the name of your background music and the editing software you use (I use IMovie, but don’t really like it).
Thanks for answering my question
And I really like your videos with very interesting mechanisms.
Music is in the description, for editig I use Hitfilm Express
Thanks for your answer
idk for the and gate you could hook up one rotation to the gear inside the differential and hook the other rotation to the outside then put a not gate on the end using the gears as power instead of an extra motor.
I've got an idea for an AND gate: what if you connected both inputs to a rod which is held back by friction if only one input is turning?
Nice idea, the problem is that it's almost impossible to find a rubber band with the right strenght to pull the arm back...
You could make an AND gate by having an OR gate with NOT gates at all inputs and outputs (I realize this would take a quite a few motors)
Yep, you could connect them that you only need one motor
@@in1 you could of course also connect the connected input with an OR gate connected to both inputs, that way you don't need any extra motors
And gate is also a Or gate with all inputs/output reversed
Yep, you would need 3 motors, though.
I think for an AND gate, I would try the same mechanism you used for the OR gate, but instead of both inputs each driving the output, one drives to a middle-gear that has an interface for the second input to mesh into... i'm not sure that's making sense. the second input will just move the lever arm over but won't actually drive anything, the arm will connect the gear driven from input A into an output gear. I can try to throw it together or sketch it up if that's not clear... :/ (I love your videos btw!)
Glad you like my videos! I get what you mean, it's actually something similar to my design: the biggest problem in both of our ideas is that the lever has to go back when the input doesn't turn anymore. I solved that using a rubber band, but obviously this isn't the perfect solution. It's hard to convert continuously rotating movent into mechanical movement without a lot of friction. If you have an idea about that, please show me :)
@@in1 I was thinking if nothing else, it could simplify the overall design without sliding linkages, motor stress, etc. I'll see what i can throw together to test out my idea...
How about make the gate react to a different rotation depending on the directions for example clockwise = true counterclockwise =false the reason being that motor have 3 states forward reverse and stop
Stop could be three state
Very nice!! As for the tricky AND gate, you could also just make it out of NOT and OR gates, after all A AND B = (NOT (NOT A) OR (NOT B))
Yep, you would need 3 motors for the not-gates then, but it would still be better than my current design
@@in1 Oh, you could have a single motor for all the NOTs, with maybe a "master shaft" that goes through the gates providing power to all the secondary "not" inputs. Much in the same way as in old factories there was a single master rotating shaft that provided power to all the different machines in the factory :)
@@SantiagoOntanonVillar Oh, yes, for some reason I didn't think of that...
I've been building lego mocs and sets for over 30 years and even I cannot seem to figure out how to encode rotational direction as true/false.
Would it be simpler to make a NAND gate instead? It stops spinning when both inputs are true. Dunno, just curious.
Do you have an idea for that? There has to be some kind of additional motor, because the output has to spin when no input is spinning
@@in1 Hmmm, I would not know how to explain it in lego terms :) But, it would be a differential adder, like the construction at this link: th-cam.com/video/gwf5mAlI7Ug/w-d-xo.html
This is in essence an XOR gate. Let spurA and spurB be the inputs (1 or 0).
A B out
0 + 0 = 0
0 + 1 = 1
1 + 0 = 1
1 + 1 = 0
If we keep spurA powered on, we have your NOT gate.
1 + 0 = 1
1 + 1 = 0
You have already constructed an OR gate.
We can use inputs A, B into an XOR to output K
We can use input B (and a constant power supply) into a NOT to output ~B
We can use K, ~B as inputs for an OR, which will result in NAND.
0 + 1 = 1
1 + 0 = 1
1 + 1 = 1
0 + 0 = 0
I have no idea if that is simpler or not, but it is what I can cobble together in my head for the moment.
Interesting approach... At the end of the video, I explain why the XOR-gate is a bit bulkier than just a simple differential, but because the NOT-gate and the OR-gate barely have any friction, it would still have less friction than the AND-gate. Nice idea!
@@in1 3:15 just slowing it down and noticing that you are playing with the grey sides, left and right, and when you spin only one of them the black pin turns. If you put a motor on each grey pin, the black pin is the output.
The differential is just 3 gears in a row, the end gears turn the same direction. Folding them inward does not change that … so a motor on each grey pin for input, turned on, just makes the center gear spin in place assuming the motors are equal. Hence the cage and the black output doesn’t turn.
You explained in the XOR section that you thought the other pin was spinning backwards, but that’s only because you saw it as an output I believe. The cage is providing the spinning output correlating to the spider setup in my link. I was just having a hard time seeing it :)
Stopping an input causes the center gear to roll around the other in opposite directions hence the reversing issue …
I may be wrong, but you could have made the AND gate like this: one input with the MAIN rotation gear motor and the secon could be a mechanism move - rotation (Basic) that will power the output.
That's basically what I did (if I understand you correctly)
So, when will you be building a computer with these?
I'm going to port Doom to it.
Given that you are building a digital mechanical computer and not an analog one, I think it might be simpler to use fractions of rotations and the ticking of a clock to drive your logic gates instead of continuous rotation. Think something like the new gear-shifter (35188) and the knob gear (32072-49135).
That's a goo idea! If I combine it using gear racks, I'm basically back to the first video :)
Couldn't the OR gate just be a differential with the two outputs connected to motors and the bigger input gear as output?
The problem would be, that when both inputs are true, the differential spins in 2x the speed than when only one input is true.
@@in1 could this be solved with a clutch gear between the differential gear and the output?
@@Mr.Engine993 I don't think so, you need friction at the output if you use a clutch gear. For example the not-gate wouldn't work at the output of the gate
How about including pneumatic?
That could be really interesting...
brain is smoking... very smart!
Make a lego adder circut. Thats when we can build lego computers.
If you're going to do this you might as well use an analogue computer
OR
NOT
NOR=NOT+OR
AND=NOT(NOT-A OR NOT-B)
NAND=NOT-A OR NOT-B
XOR=NOT[(A AND B) OR (NOT-A AND NOT-B)]
simplify as you see fit.
Remember A and B is the same as not(not(A) and not(B))
Now built an 8-bit computer like people does it in Minecraft 😁
that would be veeeeeery expensive
people in this comment section are racing to see whos first lol
Second
Vivement un Lego computer
how about XNOR or NOR
You could just add not gates in this case
@@in1 ...
Do one where rotating left is 0 and rotating right is one
look at the pinned comment
now do the same thing but analogue
Your NOT gate is actually XOR with one input fixed to one.
Probably some logic can be much simplier if one and zero coded as CW and CCW rotation and not rotation undefined or ignored.
You're right, if you replace the motor with the output, it's kind of an XOR-gate. The problem would be, that if you want thet inputs to cancel each other out when both are driven, if only one input is true, the output turns in different directions depending on the input.
I actually tried the cw/ccw thing, the problem is that it's hard to make the output turn ccw if it's false (look at the pinned comment)
Я один руский обожаю смотреть 2ин 1
Third:)
computers in real life 😐 computers in minecraft😎 computers made of legos👹2
_running DOOM on lego cpu ram gpu and mobo_
u know u can make a calculator out of water
For the and:
---Not----|
. =Or---Not----
---Not----|
I don’t understand any of this
Есть русские?
I'd prefer non technic designs.
please believe in God,Jesus died for your sins so that you can go to heaven,Christianity is a religion if anyone needs a religion,God and Jesus can help you. .
First
AND gate idea: www.mecabricks.com/en/models/o8v0LRlDvr3
One input connects the output, the other powers it
You could probably get the first input to go back in position with an elastic band
Thanks for the idea, this Is actually similar to what I ve done, just the lever connects the gears in a different way. The problem is, that the rubber band has to be strong enough, so that the lever returns, but the friction at the input has to be strong enough to bend the rubber band.