Let's Make Robots!

A hexapod for my son to ride on - a big experiment

After seeing the weblink by bdk6 "inspiration for Hexapod builders" I was inspired. The video is of the "Mantis" turbo diesel all terrain hexapod.

Ok, I cannot afford to build a 2 tonne turbo diesel hexapod but I can build one big enough for my 8 year old son Shi Sen to ride on. My plan is to have almost the entire robot laser cut from 6mm plywood.

Instead of hydraulics I have designed a linear actuator with a plywood housing. I had to design this first so I could work out the minimum size of the chassis. My actuator uses a 24V, 36W brushed motor. These motors are designed for high torque and low speed. They have a heavy duty 5mm diameter shaft, ball bearings and huge replaceable brushes.

The motor is geared down first with a 3:1 belt drive and then a threaded bar (10mm x 1.25mm). The actuator is designed for 150mm travel and some rough calculations give me about 1100N of force. The threaded bar has ball bearings and thrust bearings for both push and pull.

This setup might be a bit slow as it will need at least 4 seconds for the actuator to travel 150mm but it is easy to change the belt drive or thread pitch to adjust the speed / force ratio and I think these motors are heavily underrated. I have seen smaller motors with 3x the power rating.

By increasing the voltage to 36V I should be able to double the power output. Alternatively I need to find a more powerful motor.

Before I design / build the entire robot I'm going to build one of these actuators and test it. Once I have confirmed the strength of the actuator then I can go ahead with the rest of the design. 

Due to issues with delivering suitable material to the laser cut factory I have decided to design the rest of the robot anyway. If there are any issues with the actuator I should be able to redesign without affecting the rest of the design as it quite big with lots of room for modifications.

My plan for now is to build it as a 3 DOF hexapod as forward / reverse and turn left / right is all my son needs to have fun. However, as this has the potential to be more than a toy I'm planing to design it as a 18 DOF hexapod. When in 3DOF mode, the other acuators will be replaced with linkages that lock the joints in set positions or slave more than one joint to a single actuator.

At this point in time, this is all one big experiment. There are a lot of unknown factors, I would not even hazzard a guess at this point what the final weight and size of the robot will be. My intention is to offset the weight using gas struts.

The plywood is cheap, can be laser cut and should give a good strength to weight ratio. I've guess 6mm thick will be suitable. As you can see in the drawing of the actuator, I've doubled the thickness in some places for greater strength.

If all goes well then this might end up an open source project with the pattern for the laser cutter available for free and some parts such as the motors and electronics available as a kit.

 


 

Update: 14-04-2013

I've been trying to work out a good design of the legs. My initial design was similar to the Mantis but this made the legs too long. The the thigh from hip pivot point to knee pivot point was about 670mm (over 2 feet). I've now completely re-designed the leg and the same thigh segment is now about 370mm from hip pivot to knee pivot.

Each joint can only bend about 90 degrees, this is a limitation of any linear actuator or hydraulic / pneumatic ram. Once you go beyond ±45° the angle starts working against the actuator / ram causing a serious loss of torque.

I haven't decided if I want to go to the trouble of an articutated ankle or not. It might be easier / simpler / better to just use a rubber ball for each foot. Once I finish designing the legs then I will have a better idea what size and shape the body will be.

 


 

Update: 15-04-2013

I've almost finished the basic leg design although I still need to buy some suitable gas struts and measure them up. For now I did a quick mockup just to see how big this is going to get.

The circular base is 900mm in diameter (about 3ft) and the hexapod at it's smallest (as shown) is almost exactly twice that width. In it's current configuration, each leg can raise 60° above the base as shown to set it on the ground or lower 30° below the base to raise the base 500mm above the ground.

 

To build or not to build, that is the question!

This robot has become about twice the size of what I intended at the beginning. I was starting to think I might have to go back to the drawing board and then I realised that the current design would probably be fine for an adult, It would definitely be strong enough. All I need to do is scale it down and re-design the actuators.

Scale: The body design (ignoring the actuators for now) can be scaled to suit your wood. For example, currently the design is for 6mm plywood and 12mm bolts in the joints. If you want to built it for an adult the current size is fine and could be scaled slightly bigger to accomodate ¼ inch ply and ½ inch bolts. In my case, for my son I need to scale it down to use 3mm ply and 6mm bolts.

It looks like I need to re-design my actuator or buy pre-made linear actuators. If I can get suitable premade units at a reasonable price then I will do that to save time. In some ways they may be better if they use a ball screw instead.

 


 

Update: 21-04-2013

I started looking at how I was going to redesign the actuator to fit into the small robot. my biggest problem right now is the motor, literally! As you can see here, with the actuator reduced to half size the motor is just busting out everywhere.

I could just use a Wild Thumper geared motor but I really love this motor and want to use it. Check these photos and you will see why. Look at those big beefy magnets and heavy duty, replaceable brushes. The rotor has 10 windings! not 3 or 5 like most of the cheap motors. This lets the motor run smoother with more torque at low speeds. The motors maximum speed (no load @ 24V) is only 5000 RPM.

Notice how the rotor windings are at a slight angle? This helps prevent "cogging" where the rotor locks in at certain positions due to the magnetic circuit. This motor runs so smoothly that despite being a 24V motor rated at 1.5A (36W) I can run it from a AAA battery!

The shaft is 5mm diameter with good quality bearings at both ends. The datasheet claims 1800 hours is the typical working life under normal working conditions. Finally the price, I can get these motors for about $10 USD. To me that is a good price for such a high quality motor.

So for now it's back to the drawing board  :(

 


 

Update: 9th June 2013

I haven't had a lot of time lately to work on this project but I think I've resolved the isue with the leg actuators being too big. I've had to combine two actuators into a single, roughly cylindrical housing.

This double actuator has a housing 150mm long and both the width and height are 100mm. Each actuator has a trave of 80mm. Now I need to re-design the rest of the legs to suit.

 

 

 

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.

since I DID provide the inspirational link.  No?  Well, how about a ride on it, anyway?  

If you come to China after I make it then I'll be happy to let you ride it :D

I know I'll be jealous when I see the video of Shi Sen on it for the first time.  All the fun toys were invented after I was too old to ride them.

If it works well enough then it probably can carry an adult. The motors might need a bit more power and you would need bigger gas struts to compensate for the weight difference.

Wow, wow, wow. Good luck and keep posting!

The basic idea reminds me of Jamie Mentzel's Giant Robot Project, which is awesome. However, your approach is very different, mechanically.

Yes, Jamie's design is very good because it just uses 2 motors. It could probably work with one motor as it might be possible to operate the turn mechanism manually. Perhaps the only fault in his design is that it's simplicity prevents manuevering on really rough terain as each leg cannot be independantly raised to individual heights for climbing.

The 3DOF version will have the same limitations in that it needs relatively smooth terrain. The 18DOF version, like the "Mantis" will be far more versatile.

Collected and watching. Should be quite interesting watching this develop. Especially with home designed actuators. No doubt it will be impressive in design and execution. Good luck! Stephen

It would be nice to just buy linear actuators but they are usually quite expensive. A big part of their cost is the ball screw. Even here in China, ball screws are very expensive compared to other bearing systems.

As I want to keep the cost as low as possible I will just be using standard threaded bar that can be bought in many hardware stores and use lots of grease. Hopefully I can minimize the load and thus friction with the gas struts.