Looks great!
Kurt
nice work once again.
we have been following your project and have loved the work you have done.
A True Master!
cant wait for your next vid. im especially excited about felix as i have also been working on a quad, tho its 3x4 not 4X4!
looking out for new vid all the same.
keep up the great work bud.
Jonny.
Hi,
Thanks for your comments. I’m thinking of changing the foot parts, the ideal foot would be a ball shaped like the one used on little dog. I wonder where you can buy something like that… 
use rubber half spheres for the feet ( squash balls cut in half)?
This will eliminate the need for ankles, as the sphere will always present a consistent contact area to the ground, regardless of the joint angles. The rubber however will mean that a high amount of friction between foot and floor will occur, which might make effective walking gaits harder to achieve!
something similar to these?
http://i531.photobucket.com/albums/dd355/innerbreed/2007_12_14Quad4_792x600-full.jpg
robotsaustralia.googlepages.com/ … l-full.jpg
or
cgi.ebay.co.uk/Polystyrene-Balls … 286.c0.m14
Hi Innerbreed,
Thanks for the tip, the ball seemed a bit big though. I was thinking of about 20 - 30 mm in diameter. But using a sort suitable ball was a good idea.
BTW, very nice bioloid quad!
For only .3 kg difference, I say use both the larger ones. With the increased weight there is plenty of power to cary the extra load for a long long time. 8)
Hi KÃ¥re!
How’s Felix coming along? I’ve worked out my 4DOF legs. I used a stock piece of tubing for the lower leg, but I seem to recall that you used something else for the lower leg and foot? What was it?
Is Felix walking yet?
Alan KM6VV
Hi Alan,
Funny that you ask me about Felix. For two days ago I had a some free time to add a ABB II board to Felix! I just have to add a RC receiver. When I get more free time I’ll start with the code. At first I have to write the IK code, after that I’ll begin with the walking part.
But I’ve been thinking of the 4 DOF IK and at some point I think its a bit overkill. I’m calling the 4 servos/joint for tibia,femur,coxa (as for a hexapod) and hip. To start with I’m thinking of making a simpel IK code for tibia,femur and hip. And I think I’m only using the coxa under certain conditions.
I’ve taken a little break on A-Pod, I felt Felix deserved more attention 
I did post some info of how I solved the foot before in this thread. But I’m not satisfied with that part. I’ve not done anything with it but I want the foot to be more ball shaped, because that will give a better walking/moving/IK performance.
Hi KÃ¥re,
Yeah, I still like your 4DOF legs! I took a 3DOF ‘A’ leg and added a LP bearing, with a little other changes. I haven’t mounted it, or powered it up, but it looks workable. The lower leg, as I mentioned, is long, making the leg quite tall. But I can simply swap that out later.
I did see something on your adding or thinking about adding a ball foot. It might be a good time to think about adding a FSR.
But I can’t really do much work on the quad, I have a paper to finish and get submitted.
I was reading about Xan’s body IK, seems something like that could drive the four hips. But you’re right, driving the three lower joints would be a start. I did see your video of Felix doing the bumps and grinds! Very limber!
Alan KM6VV
Hi,
I’ve had some free time lately and have worked with the BAP code (Xan’s code). I modified the IKLeg sub so that the 3 DOF kinematics are calculated for the Hip, Femur and Tibia servos. The coxa servo was set to 90 deg constantly. At this moment I’ve full RC control of all body and leg movements. But there is an inaccuracy in the kinematics when the legs are extended and doing Z translations when the tars are around +/- 10 Z values.
To be honest I’m not very satisfied with Felix at the moment. He is a bit heavy (2,3 kg complete with 4000 mAh LiPo) too. So I’m considering of removing the coxa joint and doing some other improvements. Removing the 4 coxas will also save me for about 300 gram and also make the legs a bit more stable.
For a long time I’ve been thinking that the 4 DOF was a bit overkill. If removing the coxas dosen’t give the result I want I’ll probably end this project too. I must say Innerbreeds Stalker have inspired me in many ways, so maybe I’ll try something similar.
I’ll let you know when I’ve done more with Felix or if he ends up in a trash can. 
Btw, I’m using only one battery for both servos and electronics and it works fine. The BB2 and SSC32 are supplied directly (7,4v) from the LiPo and the servos gets regulated 6v.
good im glad you are thinking about working on this again. im a big fan of this one and it would be great to see a walking vid.
2,3 kg wow, how about drilling out some holes on the front and back panels? you can save a lot of weight from doing this.
how about replacing that lovely chassis for a lighter thinner Aluminium?
im suprised (but very happy) iv inspired you zenta, its been you who has been the inspirations for many roboteers including myself.
i hope to see what plans you have.
looks like we should swap Coxa designs as im planning on using something similar to the felix…
i wouldnt trash this project! we all get beaten by things, (i speak from experiance as you know) but if you feel your not happy with it you could always get someone else to work on it, but then you miss out on the fun! (sell the design/idea)! just a thought.
Hi KÃ¥re,
No, don’t quit on Felix!
You’ve got a very well designed package. I’ve got a coxa design now, although it’s still a little splayed; I want it to be more upright. Doesn’t bolt easily onto my 'quad chassis, so that’s probably going to change. How big is your chassis? coxa-coxa distances (transverse and fore-aft) would be appreciated. As would the height of the coxa to the floor if you don’t mind. Seems like I’m coming up with something quite unwieldy! Is your Felix as big as I think it is? The sizes of these walking 'bots always surprises me!
I haven’t worked out the battery. Do you really need/want LiPo? What are your power goals?
I must still hold off on my 'quad until I get a paper published.
Best regards,
Alan KM6VV
Thanks for your support Jonny and Alan!
Alan, you asked me about the dimensions. These are the values I used in Xan’s code:
[code];[BODY DIMENSIONS]
HipVertLength con 41
CoxaLength con 20 ;Length of the Coxa [mm]
FemurLength con 56 ;Length of the Femur [mm]
TibiaLength con 76 ;Lenght of the Tibia [mm]
CoxaAngle1 con 0 ;Default Coxa setup angle, decimals = 1
RFOffsetX con -55 ;Distance X from center of the body to the Right Front coxa
RFOffsetZ con -74 ;Distance Z from center of the body to the Right Front coxa
RROffsetX con -55 ;Distance X from center of the body to the Right Rear coxa
RROffsetZ con 74 ;Distance Z from center of the body to the Right Rear coxa
LFOffsetX con 55 ;Distance X from center of the body to the Left Front coxa
LFOffsetZ con -74 ;Distance Z from center of the body to the Left Front coxa
LROffsetX con 55 ;Distance X from center of the body to the Left Rear coxa
LROffsetZ con 74 ;Distance Z from center of the body to the Left Rear coxa[/code]
I’m considering replacing some alloy body parts with thinner ones or plastic. Today I got Felix walking, not perfect though ()… I made a quick and dirty approach. So I think I’ve to study others work about making a quad walk. Just for testing I made a table that told the body to shift COG along the X-axis.
I used a 14 step method for making a little break (all 4 legs at the ground) when shifting from RR->LF leg and LR->RF leg.
A better method would probably be to continous calculate where the COG are then compensate the body.
Well there is alot of work left. Here is a little video of Felix taking his first steps. Not very smooth or impressive at all.
Hi KÃ¥re,
You got Felix walking! That’s great! He just looks a little tired, that’s all.
Again, great job.
Aluminum can be thinner, so will you save much weight if you have to go to a thicker plastic? I much prefer working in aluminum. You might try aluminum sheet with bends for strength instead of aluminum plate.
Thanks for the dimensions! I want them for a comparison to my temporary 'quad chassis. My legs still have a splayed look to them. I’m trying to get them up higher.
Shifting the COG while walking sounds ambitions! But then that’s why you’ve got the extra coxa joints, right?
great work. its a start and at this point, thats all that matters.
its nice to see some movement. yes it could be smoother and i think trying a slower gait might help at the moment.
i think the steps are in the correct order. starting from the rear and rotating the legs round in this order is the right way.
i think replacing some of the panels with thinner alloy and bends is a good idea.
anyway…
can you post the COG transistion for the X axis. iv been trying to see how to do this!
thanks again for an update.
Hi Zenta,
I’m surprised about how quickly you’ve got Felix making his first steps. Very unique to see this quad walk with his legs under instead of next to the body. If you want to do some changes to drop the weight that’s fine by me but please don’t trash it
Did you use the PEP or the BAP?
Great work!
Xan
It truly is a great design, Zenta. I hope you do decide to pursue programming it some more.
Great work Zenta. Just a small suggestion, try add more time when transitioning from LF->RR and RF->LR. Basically to add more time for all-leg-ground-contact to improve stability especially during COG shift. It might help.
Hi,
Thanks for your comments!
Before I start reducing some weight there are other issues I’ve to solve first. Especially a strange bug/inaccuracy in the kinematics/calculations.
Thanks for your tip, actually I’ve already tried that by adding some more steps where all legs are at the ground at the same time. For the moment there are so many different issues and challenges to make Felix walk so I’ve to take one step at the time.
Xan, I’m using the BAP (your code)
Yes I can, this is the first code I used. I’ve already altered it, but this is the copy I had at work.
Made a table like this:
[code];Walking Xcomp table:
'-10=118 10=138 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14
Xcomp bytetable 128,118,118,118,138,138,138,138,138,138,138,118,118,118,118[/code]
EDIT: Does anyone now if its possible to use a constant in a table definition? (I’ve not tried it though) I know its not possible to use a variable.
And the gait method (already been modified several times):
[code]IF (GaitType = 0) THEN ;Quad gait wave type 14 step
LRGaitLegNr = 8
RFGaitLegNr = 12
RRGaitLegNr = 1
LFGaitLegNr = 5
HalfLiftHeigth = FALSE
NrLiftedPos = 2
TLDivFactor = 12
StepsInGait = 14
NomGaitSpeed = 100
ENDIF [/code]
Added this code in the Gait sub:
BodyPosX = (Xcomp(GaitStep))-128Like I said before its “quick and dirty”, this is the code I used in the test video. At the end the code are going to be much more complex. In PEP I’ve made a way of how to calculate the COG, and I hope that will work. Alot of work left though…