Hi All,

After some family illness, I'm back in the land of the living and very much need to build a hexapod. However I've been thinking that for a number of reasons (including being able to stick a small PC onboard and accomodate my own homemade, non SMT single sided PCBs) that I might try and build a hexapod which is an order of magnitude bigger- coffee table size.

It's obviously dead easy (just need to find a bigger router) apart from the actuators. I'm currently thinking that I should build some servos myself, which would give me the obvious advatage of being able to build into the electronics some tailored functionality. Most notable would be current measurement so that i can calculate the moment on each joint and so inherently have a mechanism to impliment force feedback (at some point in the dim and distant future) without needing other sensors. However this clashes horribly with my desire to keep the mechanics of the servo cheap and compact as possible- I was intending to use worm gears, which offer the other advantage of locking the joints in place without using any current from the motor and so offering me big power savings. There's also a 'party trick' feature I want to build in that would require the joint stiffness that they would afford. As you've worked out tho, I'd loose the ability to measure force from current, so I'd have to find some alternative.

Does anyone have any thoughts as to which option to take? Ideally there would be another reasonably feasable force feedback method which would allow me to go with the worm drive servos. I particularly want to avoid faffing around with those little strain guages on a big stiff aero-grade aluminium skeleton.

Hope everyone's well!

Ed

Hi Ed

Good to see you back!

I am currently thinking of building a larger hexapod based on a design I started some time ago. I will probably use linear actuators, and use force sensors for feedback.

However, I really should finish the other two aluminium hexapods that I have started first!!

If you are interested, I am thinking of letting someone have a play with one of my hexapods to add some automation to the existing walking engine. Such as obstacle avoidance etc. Then add this section to the website under development or something. If I remember correctly, you’re reasonable close to me?

Hope all is well with the family stuff.

Matt.

Hi Matt,

I'd be especially interested in that! It'd be fantastic to get more to grips with the software side of things, especially as walking hardware + student budget aren't really happy housemates. I live a wee bit north of Chichester, so winch is no problem. I'm starting Uni on the 1st Oct so my commitment might be a little shaky for the first few weeks of term as I try and sort out my life :) But yeah, I'm around for the next 3 weeks and it'd be an honour to be able to have a play!

Since posting I've been thinking a little more about actuation and was tempted to use 3 linear actuators in a pyramidal arrangement to simulate the 3DOF. It's all up in the air though!

Hope you're well, and yep, all is back to normal.

Ed

Hi Ed,

Glad your interested in having a tinker with the bots.. I have just built a workshop at the bottom of the garden. Best if I get that finished first, then there should be plent of room! Just need to finishe electrics and shelving etc. Will let you know when its done and we can go from there.

Matt.

Cheers Matt, that sounds awesome. If you want a free hand sorting anything in the workshop out, just give me a shout. I've just been working on a workshop too so I have a flavour for it at the moment :)

Ed

Hi Ed


Thanks for offering, but I'm pretty much done now. I'm currently doing an kinetic art installation in London at the moment, but when this is done and dusted I will be in touch.

Cheers. M@

Hi Matt,

I'm starting uni next week, which will tie me up initially. Term ends at the beginning of Dec, but I might be around before then. Talk to you soon either way, and good luck with the Kinetic project.

All the best,

Ed

Hi Matt,

Am back from uni. Slightly shaken :) Hope the kinetic stuff went/is going well!

Ed

Hi Ed

Yes the Kinetic project went very well! you can check it out on the kinetica page in the main menu. I am currently working on a few projects, one of them being a hexapod robot kinetic art piece which I hope to complete early in the New Year.

How llong are you back from Uni?

What Uni are you at?

Matt.

Hi Matt,

The Kinetica stuff looks awesome, I might go up and have a looksee after chrimbo.

I'm at Cambridge, doing engineering, but at the moment it's too much maths and not enough engineering. I know bookwork is important, but I'm beginning to get the shakes :(

I'm home untill just about mid-january (13/14/15 depending on how much work I still have to do).

All the best,

Ed

Hi Ed,

Just thinking about the problem you posed regards worm gears and force feedback. I was imagining a motor with shaft carrying a worm gear.. when in stopped condition, any force applied to the toothed gear which meshes with the worm would tend to either push the motor shaft towards the motor or pull it away, depending on the direction of the force applied. If you could mount the motor on a strain gauge, or in a way which allowed it to travel a little in the axis of its own shaft, but centred in the middle of this travel by a suitable spring, measuring the amount of movement of the actual motor body would give feedback of force applied to joint. To keep it cheap, I'm wondering if simply mounting a motor on suitible rubber shocks which allowed it some play, and measuring its movement with a cheap opto sensor or linear hall effect device could do the trick... even wondering... since cheap model motors will have a certain amount of play along the shaft axis anyway, might it be enough simply to measure this small movement of the shaft? Alternatively, being a great believer in mechatronics myself, (ahem, doesn't that word mean trying to avoid buiding stuff, and doing it in software instead) was wondering if some nifty servo code could periodically (a few times a second) test which direction of movement required the most current to 'unstick' a locked joint, but without moving the joint noticably, since under load the worm will be hard up against geer teeth, overcoming friction in one direction will take less current than in the other. I'm quite attracted to your idea of a 'coffee table' sized robot, and fancy it to be used as an actual coffee table, something to keep unwitting house guests amused perhaps :)

Sandy

PS used to be username sandyw but lost password, couldn't retreive it, sorry Matt mia culpa etc, have tatooed new one on my mouse hand