man, Every time I see this project, I'm just super impressed by the skin. seeing that quarter mask on the skull just made me realize again how incredible she's gonna look when she's done
Ok, I will do this. I was actually having some issues when it came to coding this all up. I was finding strange results where I had to flip the interpretation of the sensor in the code between CW and CCW. This sounds like it might be related to that. Many thanks for your suggestion.
@@DaveMakes Drove me mad before i figured it out, really makes no sense to have this pin as why not just have a fixed direction and the use software to choose the direction. Guess its for the PWM mode or something like that.
that's awesome, those motors sound like they work a lot smoother than before, and less wear and tear on them. :) also you inspired me to try silicone for my robotics project too with 3d printed injection molding. i model with blender and can 3d print as well. i am using a mix of pi and arduino components as the main systems. as for motors, i am using 60rpm geared DC motors with 3d printed mounts and gear boxes and potentiometers to track the range of motion. and for ai, i am making a workaround for replika ai, and edge copilot is assisting with the script writing XD
You can stack 2 of those simple foc shields by configuring some solder bridges on the bottom, And there is also a dual version of the DRV8313 based boards now, the smallest one you showed. Looks really awesome hope to see more of these amazing projects
I am using I2C for the magnetic sensor right now, and I understand that the module has a fixed address. I think even if I did stack shields I would be restricted to only one encoder, though multiplexing might be a possibility. I'll have a look for the dual DRV8313. Thanks!
I have an idea, maybe instead of using approximately one weak control module for one motor, why don't you use one powerful module like a CPU with a multiplexer to control all motors at once with a combined software!? Please keep going you are my Inspiration!
I have the small board with the DRV8313 2.5-A Triple 1/2-H Bridge Driver and the small board with the as5600 magnetic angle encoder in a test setup. This combination works well with small brushless motors.
biggest advantage is that they are backdrivable which is good, because you can tune up springiness and they would behave more like muscles, and they are fast and accurate! I'm also diving into simplefoc library and BLDCs for robotics
Lost count. Constantly changing parts. I've tried to keep things modular so that radical changes to one area don't necessarily impact other areas. A complete motor swap is one situation where I need to go over most of the parts and change them though.
What is the cheapest drivers that we can get price to controlled motor ratio wise that we can use haptic feedback with? I always wanted to try making a remote controlled arm even if weak with field oriented control haptic feedback wirelessly over wifi but the price of controllers on top of the brushless sensored motor ends up being insane! Anyone know the cheapest motors and drivers that allow force/haptic feedback and can pull themselves back into position?
cool stuff so far. i was going through a million designs on a single eyeball. going for super small design, it is hard with 3d printer. need a micro lathe. or outsource the parts? Servos are nice and easy. but brushless or steppers are much quieter. which is what i've been playing with. or trying to make work. Keep up the cool work.
I know exactly what that's like. I've been wondering about a mini lathe, or milling machine myself. Or ordering bits from somewhere else. There's a certain appeal to using the machine myself, though I think a lot of the parts i'd want made would require a 5-axis mill.
I tried a linkage setup for a neck once, never could really figure out how to get it right. Motor veep vrrrp noises are crucial for robots, but sometimes the SG90s can be a bit much :)
I was really worried about the pulleys I printed here for the belt. I wasn't sure if the belt would slip the teeth under load. I didn't use an idler pulley at all. It all seems to working out though.
Simple FOC is the best option i think. Theres other controllers, but too expensive. ODrive is a little over priced. Theres the storm32 shields and other brushless gimbal controllers, but the software is going to be harder to convert for general robotics use. I ended up going with a planetary geared coreless motor for my robotics project because they are compact, low power, its going to simplify the control, And i dont need mine to move as fast as the brushless option.
It's still very incomplete. I have been thinking about selling them in some way in the future. Either as files for people to print, or parts to be assembled, or even fully assembled heads I suppose. Different costs are associated with all of these. I have been thinking that the cheapest method would be for me to sell the parts for printing at a fairly low price so that many people can print it. I could then sell silicone skins to wrap over the head. These I can make myself in batches, mailing them out. They would cost quite a bit though. Certainly possible to sell more completely assembled heads, but it would cost a lot more. In months, or maybe years ahead. I may be selling them.
I think this is possible to some degree if the load is a known quantity. By sensing the current consumption of the motor against the load, the position of the motor can be inferred even without any sensor on the shaft. That may be well outside of my wheelhouse right now. I really have no idea.
![Six Hundred Strike [EPIC: The Musical] Full Animatic](http://i.ytimg.com/vi/zov6NXIAuow/mqdefault.jpg)
I get the mad scientist vibe when I watch your videos.
i'm really happy with this project! Brushless YESSSS!!
Keep it going. Can't wait for the next update!
man, Every time I see this project, I'm just super impressed by the skin. seeing that quarter mask on the skull just made me realize again how incredible she's gonna look when she's done
Great upgrade. Noise surely is a deal breaker.
That's one scary thumbnail lol.
She's really very nice. :)
@DaveMakes well regardless, hope you're successful with your project, and you can create your own Alita-like robot girl in the future. 😄
That's one awesome thumbnail. The process of creating something huge
right, almost had a stroke, need financial compensation for the damage suffered by my eyes before clicking on the vidio. preferably in bitcoins, sir.
Five Nights at DaveMakes
With those AS5600 boards, remember to solder dir pin to ground. If its left floating it can cause issues.
Ok, I will do this. I was actually having some issues when it came to coding this all up. I was finding strange results where I had to flip the interpretation of the sensor in the code between CW and CCW. This sounds like it might be related to that. Many thanks for your suggestion.
@@DaveMakes Drove me mad before i figured it out, really makes no sense to have this pin as why not just have a fixed direction and the use software to choose the direction. Guess its for the PWM mode or something like that.
that's awesome, those motors sound like they work a lot smoother than before, and less wear and tear on them. :)
also you inspired me to try silicone for my robotics project too with 3d printed injection molding.
i model with blender and can 3d print as well.
i am using a mix of pi and arduino components as the main systems.
as for motors, i am using 60rpm geared DC motors with 3d printed mounts and gear boxes and potentiometers to track the range of motion.
and for ai, i am making a workaround for replika ai, and edge copilot is assisting with the script writing XD
that sounds great!
Replika has a lot of voices to choose from, I have used it in the past. Very nice.
You can stack 2 of those simple foc shields by configuring some solder bridges on the bottom, And there is also a dual version of the DRV8313 based boards now, the smallest one you showed. Looks really awesome hope to see more of these amazing projects
I am using I2C for the magnetic sensor right now, and I understand that the module has a fixed address. I think even if I did stack shields I would be restricted to only one encoder, though multiplexing might be a possibility. I'll have a look for the dual DRV8313. Thanks!
I have an idea, maybe instead of using approximately one weak control module for one motor, why don't you use one powerful module like a CPU with a multiplexer to control all motors at once with a combined software!?
Please keep going you are my Inspiration!
Definite improvement! Although the servo sounds do have a certain charm.
Brushless oh yeah! Nice progress, I always enjoy watching your videos!
Hi from Brazil! Great work!
Hello there!
Cant wait to see final results...❤❤❤
I have the small board with the DRV8313 2.5-A Triple 1/2-H Bridge Driver and the small board with the as5600 magnetic angle encoder in a test setup. This combination works well with small brushless motors.
good to hear! thanks
biggest advantage is that they are backdrivable which is good, because you can tune up springiness and they would behave more like muscles, and they are fast and accurate! I'm also diving into simplefoc library and BLDCs for robotics
Was hoping you’d give brushless a shot someday! They’re more expensive and definitely harder to control but the speed and silence is unbeatable
Yay! It’s happening!
I remember hearing that you used gimbal motors for your robot. Definitely led me to this point here. I'm very happy with them so far. :)
Looking good!
how cyn doin
i love seeing updates on your work!!
Glad to see you making progress! Keep up the awesome work! 👍
Can’t wait, thanks for sharing 👍
You will be responsible for my lack of sleep today lol
how many redesigns have you gone through?
Lost count. Constantly changing parts. I've tried to keep things modular so that radical changes to one area don't necessarily impact other areas. A complete motor swap is one situation where I need to go over most of the parts and change them though.
0:17 Aaah... the good ol' "man-made horrors beyond my comprehension". On a real note, you're doing amazing with the progress so far. Best wishes!
Thanks :)
What is the cheapest drivers that we can get price to controlled motor ratio wise that we can use haptic feedback with?
I always wanted to try making a remote controlled arm even if weak with field oriented control haptic feedback wirelessly over wifi but the price of controllers on top of the brushless sensored motor ends up being insane!
Anyone know the cheapest motors and drivers that allow force/haptic feedback and can pull themselves back into position?
cool stuff so far. i was going through a million designs on a single eyeball. going for super small design, it is hard with 3d printer. need a micro lathe. or outsource the parts? Servos are nice and easy. but brushless or steppers are much quieter. which is what i've been playing with. or trying to make work. Keep up the cool work.
I know exactly what that's like. I've been wondering about a mini lathe, or milling machine myself. Or ordering bits from somewhere else.
There's a certain appeal to using the machine myself, though I think a lot of the parts i'd want made would require a 5-axis mill.
Nice progress!
The light brown base plate and light brown parts, are they just 3d printed? The finish seems very nice, like a matte filament?
3d printed, yeah. It's the "latte" colour of budget filament from 3dprintingcanada
@DaveMakes Thanks!
The vaguely attractive animatronic head goes crazy
really impressive
I tried a linkage setup for a neck once, never could really figure out how to get it right.
Motor veep vrrrp noises are crucial for robots, but sometimes the SG90s can be a bit much :)
I was really worried about the pulleys I printed here for the belt. I wasn't sure if the belt would slip the teeth under load. I didn't use an idler pulley at all. It all seems to working out though.
that is a huge improvement!
i'm liking it so far
Simple FOC is the best option i think. Theres other controllers, but too expensive. ODrive is a little over priced. Theres the storm32 shields and other brushless gimbal controllers, but the software is going to be harder to convert for general robotics use. I ended up going with a planetary geared coreless motor for my robotics project because they are compact, low power, its going to simplify the control, And i dont need mine to move as fast as the brushless option.
i'm looking at maybe a 5, or 10:1 planetary gear for some of these motors. To move the heavier parts around I think that will work well.
This is so impressive ❤❤❤
Maybe you can create a discord server for people that want to share code or 3d printable model for this project
Why are you using brushless motors instead of the servos from before? Any significant difference in usage?
less noise. longer lasting. something new to learn.
@ ah I seen. Anyways best wishes to you and your project!
sounds nice
His legacy will be an immortal machine (and lots of designs). Like Leonardo Da Vinci.
on the one hand, that's horrifying, on the other hand, smash.
FIRE
I wonder how much of that sound would even be heard after it’s all enclosed in skin(?). It would dampen it somewhat, wouldn’t it?
Maybe you should check about artificial muscle , this could be helpfull for movements that are difficult to make using rotation of motors
This whole project must cost a fortune
Certainly not a fortune.
Will you open source the files when this is complete? Haha
I've been thinking I might sell them if there is enough interest. I have some half formed ideas about this, so i'm not really sure.
William Afton is cooking again
nice
Who ❤ I would love to buy one or two complete heads from you, why not post ebay or etsy ads?
It's still very incomplete. I have been thinking about selling them in some way in the future. Either as files for people to print, or parts to be assembled, or even fully assembled heads I suppose. Different costs are associated with all of these. I have been thinking that the cheapest method would be for me to sell the parts for printing at a fairly low price so that many people can print it. I could then sell silicone skins to wrap over the head. These I can make myself in batches, mailing them out. They would cost quite a bit though.
Certainly possible to sell more completely assembled heads, but it would cost a lot more.
In months, or maybe years ahead. I may be selling them.
funtime hawk tuah girl in the thumbnail
we need sensorless brushless foc!! make our designs much simpler no stupuid magnet and sensor needed
I think this is possible to some degree if the load is a known quantity. By sensing the current consumption of the motor against the load, the position of the motor can be inferred even without any sensor on the shaft. That may be well outside of my wheelhouse right now. I really have no idea.
Engagement comment
Are you building a friend?
sure am
creepyass thumbnail, lmao.
This thing is pure nightmare fuel! 😅
I love your work but your thumbnail is terrifying
Great video ! Thanks David. [ ZoeTheRobot ]
do you have a discord server? maybe we all as comunity can help u out :)
I do. I should probably put a link to this in video descriptions in the future. discord.gg/wTAM63UD