What motors would this robot arm likely use?

I'm trying to build a robot arm similar to the one seen here: ur3e Robot Arm
Does anyone know what kind of motors this could possibly use? Each axis rotates 360 degrees, does this mean I could not do this with servos and I'd have to resort to stepper motors? Also, it says that it can lift 3kg, how powerful would the motors have to be to lift something like that. I can't find any stepper motors that offer anywhere near the torque to deliver 3kg on their own. Thanks

Probably stepper motors and probably geared-down for smoother operation, more precise positioning, and more torque. Although, you can get some rather large & powerful stepper motors.

“Give me a lever long enough and a fulcrum on which to place it, and I shall move the world. ”
― Archimedes

Gears or belts/chains can be used as "levers" to trade-off speed or distance for torque or force. motion.

The first stepper has to carry all the weight required from the other servos and attachments.
The second has to carry the third point plus attachments etc etc.

So at least NEMA 34 to start with but they must also have feedback on each servo to tell it where abouts in space the end attachment is.

Most 360 degree servos do not have feedback so unless you are going to add your own feedback loops they are probably a non starter.

Steppers on the other hand can be bought with feedback loops at decent prices.
They are also usually matched quite well to the stepper and make a proper robot arm easier to manipulate.

Not a cheap project by any means.

Thanks, I have no idea how they managed to fit all of that in such a small design but I'll have a go at designing something. Would you say that the stepper motors would need fans?

Thanks, I think I'll go with stepper motors in that case. Would pulleys be a necessity in this project to improve max carry load? Also how much do you reckon this would cost in total? It looks like If I shop around a bit I could get all my motors for around $100

The arm in your picture uses stepper motors with worm gears. The gears lock the stepper in position with no power applied. The motors and controller use the highest possible voltage and current because it is moving a very short time and then stopping. No cooling other than that of the metal of the motor mounting and housing is necessary.
Paul

As paul says you also have to take into account gearing factors too.
Luckily there are plenty of gearboxes that will mount to standard NEMA frames so that should not be too much of an issue.

As for costs you will have to decide if you only need a three axis or more.
There are plenty of 3 axis units that can be bought off the shelf.
You need to decide exactly how much the max weight will be picked up by the claw/attachment.

Most common cheaper units use servos but they do suffer from limitations.
The better quality more expensive ones use steppers.

If you can get 4 or more NEMA 34 steppers with built in feedback ( closed loop ) for under $100 I would be exceptionally impressed. Maybe you were looking at NEMA 23 which might just about work with some care in the arm design.

Quite often it works out cheaper to buy a package of motor and matching driver.
Then you also need to consider what controller you are going to use.
Whilst do-able with such as the MEGA you might want to consider a more robust controller.

Then there is the end attachment which may need to be custom built to any special needs you may have. No use picking up an egg with a claw that is going to crush it. No use trying to pick up irregular objects with a suction device.

Oh and as the reach of an arm increases the load capacity decreases quite rapidly.
So take into account the maximum length of the arm you need as all that weight costs extra to carry and operate.

Ok, cool. I was considering using higher torque motors for the lower axis and then the final ones at the end of the arm using cheaper nema 17's or something. That might be a bad idea though. Thanks

No you are correct in that you could use smaller motors near the end of the arm.
That is perfectly acceptable so long as they too can operate the weight needed.
Quite often you will see on industrial robots much bigger motors near the first 2 or 3 joints.

If the central axis is well balanced on a rotary type table then a smaller one could be used there too but you do stand a small chance of positional accuracy loss depending on the mechanical approach.

That'd probably be something I'd consider then. Honestly at this point I might just downscale the project because it seems like this is gonna get really expensive and ridiculously complex. I'd still prefer using servo motors so maybe I'd consider making something more similar to this Miro bot
Btw thanks for all the help

Thes the Mirobot is a pretty good example of going lower in spec but keeping a good eye on what these things are capable of.

If I were to build a starter arm I would certainly look at a 4 or 5 DOF one using NEMA motors as that will give valuable experiences for commercial and industrial units. The servo ones will not give you that exclusive experience or skills.

The only area a NEMA motor would not be required is the claw or attachment where you can often get away with a suitable servo or other custom design.

Yeah that's probably good advice. Starting engineering next year so hopefully I'll learn a lot more about these kinds of things. Anyway, I'll start doing some modelling and hopefully in a bit I'll have a functional robotic arm!

Almost certainly BLDC servomotors with planetary gears, since that's way better technology for a robot arm than steppers (much less power consumption, much more torque and speed, smooth and quiet operation - probably helps keep the weight down too). There will be a custom bearing system built into the structure too to provide the necessary stiffness.

You can enter an email address to get more details I note.

Only USD$35,000. Whattabargin!

I am going to guess it uses harmonic gearing with servomotors.

Actually yes, harmonic drive well be used - although this is usually more bespoke/expensive its a good match to the requirements of a robot arm.