Help needed in estimating cost of 2x6DoF robot

Community General Discussions
1

Hi,

I'm totally green and would really need your help on a robot I'm thinking of building...

trying to estimate the cost of a robot that has two ~0.7 meter long arms with 6DoF each, simple pinch actuator, each arm able to lift 1~2Kg, relatively low accuracy - 1~2cm. Simple stereo cam, Adruino or RPI brain, connection to AC voltage - no battery, simple 7" display

Looking in the shop I see the following but I even don't what am I missing:

-           Display touch: 5” 800x480 60FPS = 40$

-          R-Pi / Adruino: 35$ + SD 10$

-          Mechanical Pan & Tilt: 10~30$

-          Actuator: 50$ ?

-          Step motors ~20$ brushed motors ~5$ ? which one do i need?

-          Relays/power drivers: 20$?

-          Cables: ~50$ (?)

-          RealSense: 20$

 Force sensors: ~5$

Total: 40$ (5” disp)+45 (Adruino+SD)+12*20(pan/tilt)+2*50(actuator+servo)+12*20(step motor)+20(relays)+50(cables)+20(RealSense)+2*5(force sensors)+50(plastics & others) = 815$

Does that sound reasonable?

Thanks,

Ned

 

2

Payload

The payload is the main issue. 1-2Kg requires some fairly powerful actuators to be supported at 0.7m. Each actuator is likely to cost $50-$60 each. For example, the M100RAK has a reach of 0.61m and a payload at full reach of roughly 500g:

https://www.robotshop.com/en/robotshop-m100rak-v3-modular-robotic-arm-kit-no-electronics.html

You want double to quadruped this payload and additional reach. Suggest using the following tool to get an idea of the torque needed at each joint: 

https://www.robotshop.com/blog/en/robot-arm-torque-calculator-9712

3

Thanks Benson for the reply.

Thanks Benson for the reply. I took time to look into the great toutorials you have

One thing that can greatly help is the ability to look search the shop for motors by torque and weight.

Some motors don’t specify tourque at all (is KV another torque unit?)

Another is the calculator specify Kgcm where most motors in the shop are specified in ozin. Will be great to have both units in the calc

1. Regarding the design, is there a reference arm design you can point me to? with roughly same load and arm reach?

2. is it possible to place all or few motors closer to the base and use some transmissions toward the arm end in order to reduce base tourque? any insignts on pros and cons?

3. is it possible to have a single motor controling two DoF with some smart gear? (e.g. CW - circular open and close gripper, CCW rotate gripper or some other means for other DoFs)

Thanks 

Ned

4

re: Cost Estimation

I would say your estimates are too low… by a significant factor.  I don’t feel qualified enough to say by how much.

You mentioned being totally green, so my recommendation would be to lower your requirements in a huge way (in weight and length - 0.7m) and concentrate on building your skills…perhaps trying some off-the-shelf arms first…like maybe E-Z Robots or something of similar size.

I think you will likely still spend at least a thousand, but you could learn more, build more skills, and make $30 servo mistakes rather than $250+ servo mistakes along the way.

If you are really going to go down the .7m path, you might want to study the actuators on InMoov by Gael Langevin.  He uses 3D printed parts to multiply the power of servos for shoulder and elbow joints that need a great deal of power.  You probably could 3D print some of his arms and fit in your budget. 

You can also look at the gearboxes sold by ServoCity.  I think they might be $100 a piece but I haven’t checked lately.

I doubt I would want to build anything this big without using smart servos (more $$$) so you can get feedback from them.  A regular servo can be ruined very quickly (seconds) if it is impeded…which can happen a lot.  Also, the servos never seem to be powerful enough for shoulder joints on even a small bot.

One man’s opinions.  Good luck!

Regards,

Martin

5

Refined search.

"One thing that can greatly help is the ability to look search the shop for motors by torque and weight."

Completely agree. This is functionality we know we should have any are working towards it.

"Some motors don’t specify tourque at all (is KV another torque unit?)"

KV is the constant (linear over a certain range) relating rpm to voltage input. You are correct that some motor manufacturers / suppliers really don’t even provide the basic specs for some products. If there are any which RobotShop carry, we will do our best to find the missing data for you.

"Another is the calculator specify Kgcm where most motors in the shop are specified in ozin. Will be great to have both units in the calc"

It’s fairly easy to convert: https://www.robotshop.com/blog/en/unit-conversion-tool-9737

1. Regarding the design, is there a reference arm design you can point me to? with roughly same load and arm reach?

Only the M100RAK which uses power gearboxes. We offer some with 7:1 gearing which will help with either the reach or the payload: https://www.robotshop.com/en/pan-tilt-servo-gearboxes.html

2. is it possible to place all or few motors closer to the base and use some transmissions toward the arm end in order to reduce base tourque? any insignts on pros and cons?

Of course. In this scenario power / motion tends to be transmitted using timing belts / pulleys or even mechnically using 4-bar mechanisms. A big benefit is that there is a lot less weight “in” the arm itself, greatly increasing the payload. A disadvantage is the added mechanical complexity and (often) reduced motion.

3. is it possible to have a single motor controling two DoF with some smart gear? (e.g. CW - circular open and close gripper, CCW rotate gripper or some other means for other DoFs)

Anything is possible - most tend to be a “transmission” of sorts which still requires a second actuator to change where the “main” actuator provides the motion. This limits how many joints can be moved at the same time though and might make the arm quite slow.

If you don’t care too much about having 180 degrees or more of range at each joint, or care much about how fast each joint moves, and the main criteria are reach and lifting capacity, consider creating a linear actuator based arm like:

https://www.robotshop.com/en/advanced-robotic-manipulator-arm-3.html

Each linear actuator might be $80 to $150 and you can use extrusions:

https://www.robotshop.com/en/modular-extrusion-system.html

 

6

Hi Martin and BensonThanks

Hi Martin and Benson

Thanks for the valuable inputs.

I’m abit confused with the variety of actuators and the varience on prices.

I looked on the InMoov and found e.g. this servo PDI-HV2060MG-60KG (~40$, doesn’t burn by his tests)

It has 60Kgcm stall torque and moves 60deg at 0.13sec. --> 0.78sec/360deg (is that idle load?)

I need the arm to roughly move only 180deg (maybe less) and do that in ~2s (=900RPM) with a 2Kg at it’s gripper

If I’m putting a ~5:1 gear, it should generate torque of 560Kgcm = 4.3Kg70cm

How do you go from here to figuring out the speed?

The actual arm x component during 180deg movement from bottom to top is a sine of the angle,

WC at horizontal state we have 2Kg*70cm for the load + 2.3 for acceleration. Here’s where I’m stuck on speed (assuming the rest of my calculations are not completely wrong)

Would appreciate help

Thanks,

Ned

7

Hi Martin and BensonThanks

Hi Martin and Benson

Thanks for the valuable inputs.

I’m abit confused with the variety of actuators and the varience on prices.

I looked on the InMoov and found e.g. this servo PDI-HV2060MG-60KG (~40$, doesn’t burn by his tests)

It has 60Kgcm stall torque and moves 60deg at 0.13sec. --> 0.78sec/360deg (is that idle load?)

I need the arm to roughly move only 180deg (maybe less) and do that in ~2s (=900RPM) with a 2Kg at it’s gripper

If I’m putting a ~5:1 gear, it should generate torque of 560Kgcm = 4.3Kg70cm

How do you go from here to figuring out the speed?

The actual arm x component during 180deg movement from bottom to top is a sine of the angle,

WC at horizontal state we have 2Kg*70cm for the load + 2.3 for acceleration. Here’s where I’m stuck on speed (assuming the rest of my calculations are not completely wrong)

Would appreciate help

Thanks,

Ned

8

Servo Specs

We’ve tested a number of servos and can say that it’s good to be skeptical about the advertised torque. Read as many customer reviews as you can before purchase to see if they can get the torque / speed advertised. At stall, the servo is not moving at all (0rpm). If you add a 5:1 gear down, whatever torque is input will be multiplied by 5, and the speed reduced by a factor of 5. Note too that the total travel will also be reduced by a factor of 5. If the servo can rotate 180 degrees, the large gear will only be able to rotate 180 / 5 = 36 degrees.