I think everyone has one of those… lol
im actually going to need to get a small dremil tool to cut the plastic. All blades i have tried using are either messy or they slip about to much. its what i call an oily plastic. not sure of its real type.
With the T-hex code installed, i thought id have a quick play. works really well. i need to make a few adjustments to the leg as they are slightly out, but all seems ok.
**
here is the start pos:**
http://i531.photobucket.com/albums/dd355/innerbreed/Image018.jpg
here is the initialized pos:
http://i531.photobucket.com/albums/dd355/innerbreed/Image017.jpg
Oh im calling it (A.T.L.A.S) All Terrain Land Assault Sphere ?
Love the name!
Alan KM6VV
I have manage to cut the sphere with my Stanley blade! 
http://i531.photobucket.com/albums/dd355/innerbreed/SDC12387.jpg
I just pulled the blade though it. its left nice clean edges. better than cutting with a serrated blade.
Here is the complete left side and base. i have glued the base to the disc board carrier so hopefully it will stay there. 
http://i531.photobucket.com/albums/dd355/innerbreed/SDC12389.jpg
my 5 year old son said that the other remaining part (in the back of picture one) looked like a space man helmet. …
while no one was looking i tried it on for size! 
Looking good Jonny!
LOL, I was thinking the same as your 5 year son. (Btw, how many kids do you have?)
Are you planning to use flat head screws to fit the sphere parts to the legs?
I think this project of yours are going to be pretty awesome and spectacular to see moving when its finished.
Did it fit? Pix or it didn’t happen! lol
thanks Kåre, well i would like everything to remain flush but if i have to then i will go directly through each of the segments.
im kinda adopting the metal rushed look. so exposed bolt heads might look quite cool. 
i have Two children at the moment.
**Joshua **is 5 and **Reece **is 1. id like another one but Hannah my wife would like to start a career next year.
are you saying you want a picture of me wearing it? lol.
Nay, DEMAND one!
Alan KM6VV
found a cool little picture:
A very interesting chassis design. an upright disc shape would make, making a sphere robot easier.
Hey! that IS cool! 8)
Agree, something like that would be awesome to make one day.
one other thing that im think about this project is that once the sphere is closed and/or powered off, then the whole thing will fold in. (as expected). so i have been thinking of creating an inner frame from something that will reinforce the structure.
this will also help to conceal the hexapod inside if the segments don’t line up exactly.
Hey Jonny,
Pretty cool find!
About your design. What i’m thinking. The material is pretty thin. How will you manage to get the pieces perfectly flat against each other? I think this will be a real challenge.
Xan
yes this was one of my main concerns. when i was planning on using the polyball it would have been simpler to adjust the segments. even a light sand down for a perfect finish but with the plastic sphere this cant happen so im hoping that if I make the inner frame (in black) then it could help to close the gaps.
zenta is trying to find a way to make them so that they can be adjusted and positioned correctly. something i have also thought about.
one of the things i have still yet to do is align the offsets so they they are perfect. when i used your 1.3 code it was easy. i just added a section of code for this in (INIT).
can the same be done for the T-hex?
;-------------------[INIT]
; Servo Offsets (manual input!)
;ONLY USE THESE IF NOT USING REGISTERS, otherwise 'comment out.
serout SSC_OUT,SSC_BAUD,"#",dec |
RRHipSwayPin,"po-90 #",dec |
RRHipRotatePin,"po70 #",dec |
RRHipZPin,"po0 #",dec |
RRKneeZPin,"po70 #",dec |
RRTarPin,"po80 #",dec |
LRHipSwayPin,"po50 #",dec |
LRHipRotatePin,"po50 #",dec |
LRHipZPin,"po0 #",dec |
LRKneeZPin,"po0 #",dec |
LFTarPin,"po0 #",dec |
RFHipSwayPin,"po90 #",dec |
RFHipRotatePin,"po99 #",dec |
RFHipZPin,"po-90 #",dec |
RFKneeZPin,"po0 #",dec |
RFTarPin,"po80 #",dec |
LFHipSwayPin,"po60 #",dec |
LFHipRotatePin,"po99 #",dec |
LFHipZPin,"po0 #",dec |
LFKneeZPin,"po0 #",dec |
LFTarPin,"po0",13]
.........
I don’t 
i gave been play about with the hexapod this morning. the gaits are looking really good. the 4doflegs work really well and look really cool with the round chassis design.
i have even tried to replicate Zentas new Tripod/ripple gait he called (Triple gait) using the 180 phase shift seen in his latest workshop video:
youtube.com/watch?v=o2MCLYBa3Vk
anyway, id like to make a video of the hexapod walking as it is at the moment but my camera is bust!
but, i have also been looking at calibrating the legs so that they sit at the right position regarding the sphere segments.
here i’v had some problems… as we know, Servo Backlash plays a big part here and as we are looking for precision positions for the legs, its just becoming a pain trying to line it up right. i fear this wont look as planed and its a shame as i have everything i need for this project. Im going to try my best to get it to look right but as mentioned, its the servo backlash thats making things harder.
i know that zenta worked on this problem for this Archer Biped but feel that with so many servos encountering this problem it might not work as well. plus the backlash varies by a few degrees for each servo.
i will try my best.
The legs are not quite sitting correctly when the sphere is closed so some work to do there.
Hi Jonny,
Sounds like you’ve done some progress. It would be cool to see a video of the round 4 DOF hex even without the sphere parts.
Yeah, the gear backlash might be a problem with large leg segments, and the count of DOF’s doesn’t help either. All I can say its a challenge but you should be able to make the legs stay in the “sphere” position since the leg/servos isn’t under great load. The inner frame you mentioned or just some small taps might help to align the parts properly.
Have you started mounting the sphere parts to the legs too?
(have to go, my youngest 4 year old son keep poking me… )
i haven’t mounted them yet but i have been setting the legs into positions and then holding the segments to the legs to make sure im getting the legs correct. a bit of a trial and error.
one thing i have noticed is that because of the design, i was hoping the tars would be in contact with the floor when sphere is open but in the current positions im getting the legs into when sphere is closed it doesnt seem like im getting the segments into the right place, and the tars are sitting too high up and not as close to the disc as expected. more work to be done there.
what would be the best way to adjust my servo offsets as some of them are out 5-10º.
would adding the table i posted do the trick?
i will see if i can catch a video on my wife’s phone (hers is better) sometime next week .
Hi Jonny,
Will be fun to see the video.
You probably mentioned it before, but what servos are you using? i.e. have you gone digital yet?
As I was mentioning in the other thread (contact switches), I have been playing with the idea of readjusting the leg positions depending on the desired Y position. But I was also toying with maybe having sequences to start-up and shut-down the robot, could hard code it to number #9 and #10. I think I am starting to have luck with adjusting the leg positions and may not need to do the startups, but in your case you may want to consider it. You could maybe have some types of guides to help close, maybe you could have the sequence do a little shake or the like to help legs settle in. Or maybe having each of the legs pull in in some order may do the trick… Just an idea.
As for being off by 5 to 10 degrees and using the PO command. There are pluses and minuses here. The problem with the command is, that it is additive. That is if your processor resets, but you still have power to the SSC-32 and you reissue the command, it simply adds(subtracts) the value from the current value. So things would get off. A couple of other approaches.
- Set the appropriate offset in the SSC-32 registers. If you are off by more than ±200 you will need to readjust the servo horn by a click or two to get it right.
- Have a table of offsets that you use internal to the code. Problem with this is these don’t get applied to running sequences that are stored on the SSC-32. An extension to this is store the offsets in the EEPROM of the processor and allow you to pragmatically change it. I did this with the Arc32 version…
Kurt