My Advanced Realistic Humanoid Robots Project

[joeqsmith]

I'll reserve compassion and sympathy for any children involved.  They don't have a choice.

I won't labour the point with you Joe, but if you really believe that all adults are capable of logical thought, then I guess you don't get out much.

Rare I feel sorry for any adult as they have choices. 

All is not the same as rare.   Guessing some sort of translation problem.

[joeqsmith]

Quick search for OTS plastic pulleys with < 10mm OD.  Looks like the kids in the photos have progressed far beyond the OP.   I'm sure the OP considers these all to be junk.  They really need to be hand made by them using the most primitive techniques to get the perfect fit and function.   

https://www.amazon.com/TEHAUX-100pcs-Plastic-Transparent-Diameter/dp/B0BN9LC3S9/136-9981433-4558935?pd_rd_w=rBm9W&content-id=amzn1.sym.46e2be74-be72-4d3f-86e1-1de279690c4e&pf_rd_p=46e2be74-be72-4d3f-86e1-1de279690c4e&pf_rd_r=CZB94S1HR51AD5NSRRER&pd_rd_wg=InIvL&pd_rd_r=4b36ead0-48e3-4891-ac88-522a87341a85&pd_rd_i=B0BN9LC3S9&psc=1

waaaaaay too big.  almost 3x the diameter we need.

No problem.  Google search for 3mm OD


https://www.aliexpress.us/item/3256803640501348.html

Product : CNC brass Deadeye
Material : Brass
Quantity : 10 pcs/lot

a. The diameter is 3mm, the thickness is 1.3mm, and the diameter of the center hole is 1mm;
b. The diameter is 4mm, the thickness is 1.5mm, and the diameter of the center hole is 1mm;
c. Diameter 5mm, thickness 1.7mm, center hole diameter 1mm;
d. The diameter is 6mm, the thickness is 2.0mm, and the diameter of the center hole is 1mm;

[artbyrobot]

Quick search for OTS plastic pulleys with < 10mm OD.  Looks like the kids in the photos have progressed far beyond the OP.   I'm sure the OP considers these all to be junk.  They really need to be hand made by them using the most primitive techniques to get the perfect fit and function.   

https://www.amazon.com/TEHAUX-100pcs-Plastic-Transparent-Diameter/dp/B0BN9LC3S9/136-9981433-4558935?pd_rd_w=rBm9W&content-id=amzn1.sym.46e2be74-be72-4d3f-86e1-1de279690c4e&pf_rd_p=46e2be74-be72-4d3f-86e1-1de279690c4e&pf_rd_r=CZB94S1HR51AD5NSRRER&pd_rd_wg=InIvL&pd_rd_r=4b36ead0-48e3-4891-ac88-522a87341a85&pd_rd_i=B0BN9LC3S9&psc=1

waaaaaay too big.  almost 3x the diameter we need.

No problem.  Google search for 3mm OD


https://www.aliexpress.us/item/3256803640501348.html

Product : CNC brass Deadeye
Material : Brass
Quantity : 10 pcs/lot

a. The diameter is 3mm, the thickness is 1.3mm, and the diameter of the center hole is 1mm;
b. The diameter is 4mm, the thickness is 1.5mm, and the diameter of the center hole is 1mm;
c. Diameter 5mm, thickness 1.7mm, center hole diameter 1mm;
d. The diameter is 6mm, the thickness is 2.0mm, and the diameter of the center hole is 1mm;

These can neither be used in my turn in place winch mounted to the top of the motor nor my archimedes pulley system.  It seems they might be useful for something though perhaps so maybe I'll buy some and have them just in case I come up with a use later.  Thanks for the link.  Never heard of these till now.  They certainly are small enough to be useful for something I imagine.

[joeqsmith]

These can neither be used in my turn in place winch mounted to the top of the motor nor my archimedes pulley system.  It seems they might be useful for something though perhaps so maybe I'll buy some and have them just in case I come up with a use later.  Thanks for the link.  Never heard of these till now.  They certainly are small enough to be useful for something I imagine.

There are places that will custom make them for you as well.  I imagine you need several and the setup charge may not be all that expensive by the time you figure your labor and time to make them by hand.   You will need to provide them with a print with dimensions, tolerances, surface finish, materials.... 

I would check your local area.  Maybe you have a machine shop near by that could handle a small job like this. 

[artbyrobot]

Maybe you have a machine shop near by that could handle a small job like this.

I just bought 10 of the 3mm and 10 of the 6mm it was around $31 with shipping.  As I said I know of no use for them but might come up with a use later.  Nice to have more options. 

That said, these were around $1.55 shipped per pulley.  Not horrible.  Not amazing.  Just okay.  Now if I had a machine shop custom make me something like this, I imagine it would go to $30/pulley.  Completely unreasonable and completely impractical and completely unacceptable. 

[xrunner]

waaaaaay too big.  almost 3x the diameter we need.

No problem.  Google search for 3mm OD


https://www.aliexpress.us/item/3256803640501348.html

Product : CNC brass Deadeye
Material : Brass
Quantity : 10 pcs/lot

a. The diameter is 3mm, the thickness is 1.3mm, and the diameter of the center hole is 1mm;
b. The diameter is 4mm, the thickness is 1.5mm, and the diameter of the center hole is 1mm;
c. Diameter 5mm, thickness 1.7mm, center hole diameter 1mm;
d. The diameter is 6mm, the thickness is 2.0mm, and the diameter of the center hole is 1mm;

WOW! What a pioneering lesson you gave him Joe. You first present him with a solution, but the part is "waaaaaay too big".

Here's the really inventive solution you then think of  - to actually do a search for the right size part! 

This is amazing for him to learn. Must be a mind-blowing event in his life for this sort of teaching to happen right here. I hope he can learn even more from us as the thread progresses.

 

[joeqsmith]

Maybe you have a machine shop near by that could handle a small job like this.

I just bought 10 of the 3mm and 10 of the 6mm it was around $31 with shipping.  As I said I know of no use for them but might come up with a use later.  Nice to have more options. 

That said, these were around $1.55 shipped per pulley.  Not horrible.  Not amazing.  Just okay.  Now if I had a machine shop custom make me something like this, I imagine it would go to $30/pulley.  Completely unreasonable and completely impractical and completely unacceptable.

I would think scissors would be unreasonable and impractical.  How many do you need to crank out?  Do you have prints for all of them?

For 10 or 1000 pcs may not matter but the setup charge may be a few k.   Cost of a small lathe and you would still be making them by hand and may cost more.     

[artbyrobot]

someone said I should try to build thing from adam's family and I said what I said.  In any case, I found this video which was in some sense impressive but also very limited in some ways.

[Analog Kid]

Limited, eh? Yeah, I guess it's "limited" for a world-class advanced elite robotics designer. Which you are. Not.

So looking at this: it took these two guys, what? maybe a couple months from the time Netflix asked them to design this until it was finished. The thing works perfectly for what it's designed for: a realistic-looking walking disembodied hand. And notice how simplified it is: the fingers simply work off of cranks. No linear actuators, lines or anything like that.

So how long until you have something at least on that level or higher (hah!) actually working to show us?

Hey, if you're that all-fired great like you say you are, that shouldn't be a problem. Go to it!

(And I noticed how they too incorporated thumbtacks in their robot, but only as skids for the fingertips. Now that's genius!)

[artbyrobot]

They obviously don't have young children they care for significantly, they obviously had this as their full time job and only project during that timeframe, they obviously are not juggling massive countless other projects, etc.  So it was reasonably fast.  And they admitted to outsourcing alot of its construction to a cnc shop spending a small fortune (not diy friendly approach).  Also its 2 motors.  So a toy alone.  A roboticist of world class eliteness like myself has to use at least 50+ motors for a hand and have every DOF of a human hand and match human strength and speed while fitting all of this in human form factor (never before achieved till myself assuming I finish first)

[xrunner]

They obviously don't have young children they care for significantly, they obviously had this as their full time job and only project during that timeframe, they obviously are not juggling massive countless other projects, etc.  So it was reasonably fast.  And they admitted to outsourcing alot of its construction to a cnc shop spending a small fortune (not diy friendly approach).  Also its 2 motors.  So a toy alone.  A roboticist of world class eliteness like myself has to use at least 50+ motors for a hand and have every DOF of a human hand and match human strength and speed while fitting all of this in human form factor (never before achieved till myself assuming I finish first)

And with all of your responsibilities (you probably didn't have kids in the beginning either) you've had TEN years and don't have much of anything to show for it.

Also, no you are no elitist roboticist unless you are hiding things that would convince us otherwise.

[Analog Kid]

Oh boo hoo; you've got so much responsibility, and they probably have none.
Yes, it's simple; that's part of the point. Two motors! Simple and it fucking works!
What have you done that actually works?
And no, your ugly-ass shit with you pulling on strings doesn't count.

Show us something, or just STFU.

[artbyrobot]

how does the .gif I posted and working not count as something working?     Make it make sense people.  You want to jeer, fine.  But don't become completely illogical in the process. 

[Analog Kid]

What gif is that? That ugly-ass "pulley" made mostly out of glue? If that's not it, then give us a link to the post (you do know how to do that, right?)

Nothing you've posted is anything that "works", as in anything remotely close to the Thing robot that you seem to think is so far beneath you.

Working, as in at least one robotic part, sub-assembly, whatever you want to call it, actually moving under its own power. Finger, arm, aluminum bar, popsicle stick, doesn't matter.

[artbyrobot]

no, working does not need to be defined as moving under its own power.  If you make a hinge and test it rotating, then its a working hinge and same goes for pulleys.  You should test the function of everything you make before motorizing it anyways. 

Also the .gif in question is the only .gif and was a epic success.  You saying its bad because it uses mostly glue is illogical.  If the glue is structurally sound enough for the application, even if the part is 100% glue then it's fine.  What if I injection molded a part.  Would you say the part is not a working part since resin glue was used to make the 100% glue part?  That's illogical if the part works as intended.

[Leiothrix]

no, working does not need to be defined as moving under its own power.

Yes it does.  Otherwise how do you know that it, well, works?

You should test the function of everything you make before motorizing it anyways. 

Of course you should.  But can't say that it is epically successfully working unless it is being driven in the way that it is supposed to be driven.

In your case this means an interrupt-free Arduino powering a BLDC turning a pulley to actuate something.

This doesn't "work" until you can get a microcontroller to reliably make a part do something.

[artbyrobot]

Yes it does.  Otherwise how do you know that it, well, works?

when I tested it, my only manual input was rotating the shaft of the motor manually.  So if it works under that influence, then it works under the motor turning under its own power too.  Getting a motor shaft to turn is not rocket science.  If the worst possible case scenario came true, I'd copy a bldc motor controller design with no modifications to it and get the motor working that way.  But I'm confident my custom motor controller design will work well still.

[Leiothrix]

Getting a motor shaft to turn is not rocket science.

Great, so you have no excuses then.

when I tested it, my only manual input was rotating the shaft of the motor manually.  So if it works under that influence, then it works under the motor turning under its own power too.

You don't know that because you have not tested it.  Spinning by hand and spinning by motor are very different types of motion.

You need to test that it works when controlled by a motor, that it is reversible, it is repeatable, it is precise enough for what you want, fast enough for what you want and it is strong enough to do what you want.

Part of the system test is also to give you the opportunity to have a control scheme for the microcontrollers, measure their power consumption and design a physical envelope for the whole lot.

At the moment you have absolutely no knowledge of either if your system works at all or if it would be suitable for your intended purpose.

[artbyrobot]

Spinning by hand and spinning by motor are very different types of motion.

That is absolutely false!  The motor shaft I wrapped string around over and over and pulled that string to cause the shaft to rotate.  It is axial rotation.  The type of motion is 100% the same type of motion as it will be when the motor is being electrically actuated.  Zero difference at all.  Where did you get such a bizarre idea as this?

[Leiothrix]

Are you saying that you can provide rotational motion continuously with precise speed, position and torque control, and be able to measure the energy that you use to do that?  That would be a medical marvel.

Graph those parameters with a human vs a motor spinning the shaft.  The motor you could make consistently whatever you required, the human is all over the place.

All that you have proven by spinning it by hand is that it isn't binding anywhere.

To test if your system works you need to test if your system works.

[artbyrobot]

I don't see how any of that fluff data has anything to do with verifying the pulley is working as intended.


Anyways, I am wondering where are the wagers?  The jeering critics/haters in here talk such a huge game but have not given any wagers.  Like if Artbyrobot achieves x landmark progression point, I will shave off my eyebrows.  This type of grandiose display of confidence is highly lacking IMO.


oooo I got one: if Artbyrobot achieves x landmark progression point, I will have a custom tshirt made that says "artbyrobot is the greatest roboticist on earth and I am so ashamed I doubted him".  Something like that would be pretty epic I think.  And it has to be worn for a week. (well you can have multiples so you can wash it or w/e.)

[Leiothrix]

To verify the pulley works as intended you have to actually use it as intended.

Which means actually driving it with a motor and having it actually do something. 

And it has taken you 10 years to get this far, I probably won't live long enough for you to get a motor sort-of spinning on the bench.

[artbyrobot]

we've been through the 10 year thing many times and you seem unable or unwilling to listen about how extrapolation does not work in these scenarios.  For example, lets say I worked on building a videogame for 5 minutes 30 years ago.  I then pick up that project 30 years later.  You come along and say wow it seems like all you got done in 30 years was 5 minutes of work.  So if I extrapolate the next 90 years you will only get done 15 minutes of work on it!  Well that is foolish.  One is not guaranteed to never work on it this next time they pick up a old project like they did the last time.  You cannot reasonably assume that.  And in order to even lean toward that assumption you would have to have a much broader insight into their life and what influences prevented progression the first time.  For example, lets say that man worked on it for 5 minutes because 5 minutes into it he was wrongfully imprisoned for a crime he did not commit and was only released 30 years later.  He was eagerly planning the return to that project the entire 30 years and that vision kept him going.  He is itching to get back to it.  So in this case, would it be reasonable to expect the same progression rate when he gets out of prison?  HELL NO!  And yet you sit back like you know with a crystal ball that the progression will be the same rate despite not knowing anything about me or how I spent those 10 years and what major changes are in place for the second ten.

[Leiothrix]

You are the one that came up with the 10 year number.

You have given that explanation below several times, how long have actually been working on this then?  You don't tend to give straight answers to anything.

With the level of output my guess would be ... this is your first week and first attempt at mechanics, software and electronics; and still haven't quite figured out that it is monumentally harder than you think. 

And you aren't even slightly interested in learning either which is pretty weird. 

And the only reason you even posted that was to change the subject from actually hooking a motor up to your pulley.  In your words it is not rocket science, so just give it a go.

[artbyrobot]

You just completely twisted the situation there.  There is a WORLD of difference from not wanting to learn and not wanting to learn information one deems is completely unnecessary to achieve their goals and actually distracts the mind and time that could have been put into learning the things that are relevant to goal achieving.  I cleave to the latter.  I engage in practical learning pursuit, not idle and irrelevant learning pursuit which so many others waste time on for no reason other than lack of discipline, lack of goals, idleness, boredom, etc.  Nothing really admirable about it.  And it's being promoted here to the shame of all engaging in it.