robot arm calculation/servo requirements

I'd like to build a robotic arm capable of lifting five pounds. The shoulder and the forearm will each be 24" long. It's to be run by a friendly PIC 16F628.

I like the idea of using RC servos to move the arm but the strongest servo I can find is the Hitec HS-805BB which only has 343 oz/inch of torque. I discovered that this means it will lift 343 ounces of weight with a one inch lever. Since my robot arm will be 24", that means it will be capable of lifting 343/24 = 14.29 ounces which is less than a pound - far short of my five pound goal.

Assuming my calculation is correct, what I need are some suggestions. These big servos are pretty expensive, about $50 a piece shipped. I did come across a servo gearbox for $50 that would give it a 5:1 increase in torque. That would get me close but that's another $50 for the gearbox and I would need two - one for the shoulder, one for the forearm.

I've not had much luck locating gears.

What have the arm builder's on this group done to solve this problem on the cheap? Or have you all shelled out $200 on servos? Or have I miscalculated the torque needed to drive the shoulder and/or forearm?

TIA!

Reply to
Frank
Loading thread data ...

Actually it's much worse. The weight of the arm needs to be added to the calculation.

I'd consider a more normal DC motor with a large gearbox.

formatting link
D. Jay Newman

Reply to
D. Jay Newman

Actually,

It is worse than that.

You have to account for accelerating that mass.

Standard R/C servos are not up to this.

Mike

Reply to
Blueeyedpop

I wanted the same lifting power with the robotic arm ive bin working on for the past year or so. To get the needed lifting torque ( from the shoulder) i used two DC gearhead motors connected to a common shaft.With this arrangement the arm can lift up to 8 pounds, the motors were really labouring tho and had to hold the base down to prevent it following over.

Theres some pics of my robotic arm on my website:

formatting link
Click on RoboArm II.

Larry.

Reply to
nettron2000

If the shoulder and forearm are each 24", doesn't it have to lift five pounds on a 48" arm worst case? As others have noted, you also have to consider the weight and mass of the arm, but you could use springs or counterweights to balance most of the weight. You'll still need to accelerate the mass though.

I'd suggest bumping up your budget by a significant amount. What is the application for this?

Mitch Berkson

Reply to
Mitch Berkson

Wow, looks like a lot of work you put into that arm. Thanks for posting pics of this and your other robots.

I was hoping to use the RC servos to reduce the amount of labor involved in building the arm by reducing the need for controllers, limit switches, etc. Not to mention the RC servos weigh a lot less than bigger DC motors.

I got the idea from "Build Your Own Humanoid Robots" in which the author states "These servos are ideal for use in the arm joints that need to be able to lift a lot of weight." Maybe a lot of weight to him was a couple of kilograms. Of course his arm was not 24" long, more like 10".

Reply to
Frank

The problem is that you are on the 'steep' part of the price/performance curve. Servos for RC models are mass produced and cheap - but few people need the sort of power you seek. The higher power servo stuff tends to be custom and therefore pricey. Sometimes there are surplus servo motors available, but you will need some level of engineering skill to sort out gearboxes etc, besides the electronics to drive such motors.

Probably the best you can do would be to use battery drills, driving a screwed rod, and hence moving a nut back and forth which is attached to the load. Feedback would be done by a potentiometer attached to the output drive. The motors in these drills eat current, and you will need something like a 25 amp + motor controller. Some of these motors stall at > 100A !

As another post says, if you balance out what you can, you can convert lifting a 0 to 5lb load into +- 2.5 lb

Dave

Reply to
Dave Garnett

I get "The web site you are trying to access has exceeded its allocated data transfer. Visit our help area for more information. Access to this site will be restored within an hour. Please try again later."

I'll have to look again in about 12 hours, when most other USA Usenetians are asleep...

You may need those, and modify the servos for fully-continuous rotation and do some more gearing down.

Another question is how FAST do you want the arm to lift that weight? If you gear it down enough, you can get a pager motor to lift the weight you want, but it may take hours. In some situations time isn't that important, and this could be an appropriate solution. But you probably want faster movement. I'm sure there's a formula to convert foot-pounds multiplied by (angular) degrees per second to horsepower. Then you can multiply by

750 or so (I forget the exact factor) to get watts. But you need more powerful motors to get movement of reasonable speed with a heavy load.

-----

formatting link

Reply to
Ben Bradley

My site is rather pic intensive(im a builder not a writer) so reaks havoc with the bandwidth when there are several hits. Yahoo! sites (among others) have limited width but its free so cant complain.Prolly best to book mark it for viewing later but heres a link to a photo gallery of RoboArm II:

formatting link

Alot of work but over all a labour of love for a builder.:) Im meaning to get some documentation up for RoboArm when i get the time but heres a quick rundown. The arm is controlled in two ways; by a 12 button keypad/box for manual control( a project in itself btw) and via the parallel port of my computer for software control.Presently experimenting with Slider type pots mounted at the joints for position feedback. Mechanical limit switches were not used,instead i used a dual window limit comparator circuit as an electronic solution, schematic of which is also on my site.

Reply to
nettron2000

You might want to consider using hydraulics with electromechanical valves.

Larry ___

formatting link

Reply to
Larry

Nice work! What do you use for position feedback?

BRW

Reply to
Bennet Williams

There are more powerful servos but they are harder to get. One is the GWS 777CG, which at 6vdc delivers 583 oz-in. But even so, it's no where what you need. As others have noted, use counterweights or springs to offset at least the weight of the arm. You can even make it dynamic, of sorts, by connecting it to another servo on one end. Pull back on the spring to add tension. Check out the local hardware store for their selection of long expansion springs.

Second, go with DC gear motors. You can try surplus, Another option: Merkle-Korff

formatting link
sell some inexpensive ones online that might be perfect for the job. They're basically the same motor with different gear ratios. Choose a very low ratio. Price is about $22 per motor. (They have more expensive ones, too.)

You need to provide some positional feedback. Unless you get lucky and find some absolute encoders on eBay, you'll have to make some incremental encoders yourself. There are plans all around on the Internet using low-cost IR LED/transistor pairs, or the guts from an old mouse. Quadrature is ideal, as these provide 4X the resolution of the disc itself. Go to US Digital

formatting link
and buy one of their LSI interface chips (about $3.50). These condition and decode quadrature signals to simple count/direction pulses. Easier to use with a PIC.

You will need an H-bridge chip or circuit for the motor. The M-K motors are high efficiency, and a low-cost 1A/channel bridge, like the L293D, should work well. These are about $4.50 or less from the discount electronics outfits.

Total cost per axis should be about $50. This includes the DC gear motor, IR optics, LSI quad chip, H-bridge chip, balance spring, and a $10 servo to actuate the spring.

You will need to work out the math to ramp the motors. They are geared, so they take a finite amount of time to stop, especially with a load, and back EMF dynamic braking can't do everything. You need to compensate for this in your software, and write acceleration/deacceleration ramps. This is where robotics gets fun...

-- Gord>

Reply to
Gordon McComb

If you're willing to drop the idea of a humanoid arm, and use a parallel link design, then you get a 6-deg freedom arm with more limited range of movement which can lift nearly six times the capacity of one of its motors in its best lifitng area.

-- Chris Malcolm snipped-for-privacy@infirmatics.ed.ac.uk +44 (0)131 651 3445 DoD #205 IPAB, Informatics, JCMB, King's Buildings, Edinburgh, EH9 3JZ, UK

formatting link

Reply to
Chris Malcolm

Chris, do you have any particularly good links for this?

- dan michaels

formatting link
======================

Reply to
dan michaels

PolyTech Forum website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.