Matt Bauer double knee joint for SES

[tom_chang79]

I think what Sienna is trying to say in that picture is if you’re using ASB-09, it will not swing over the “long” side of the servo as you have pictured in your drawing. The ASB-09 will swing over the “short” side, by short side I mean the side where the servo horn is closest to if you were to lay the servo flat. ASB-10s can swing over the long side however.

[linuxguy]

@tom_chang79

I think what Sienna is trying to say in that picture is if you’re using ASB-09, it will not swing over the “long” side of the servo as you have pictured in your drawing.

It is not swinging over the long side. It is at a 45 degree angle to the back of the ASB-09 it is attached to.

8-Dale

[sienna]

@linuxguy

I’ve seen that video, and it is what really got me interested in the double jointed knee. The BRAT would not get that full range of motion with just 4DOF though. It would require that fifth joint between the knee and ankle for full range motion.
8-Dale

I have seen that video too. This is the point of my first comment/image. Without that 5th DOF, the “double knee” doesn’t seem to afford you anything. You are gaining only a tiny increase in flexibility, for a tremendous cost. One of two things could help:

1. Add that 5th DOF, which would make things worth it.
2. or at least use the long C brackets, not the short ones, so that the servo can fold in on itself.

@linuxguy

@tom_chang79

I think what Sienna is trying to say in that picture is if you’re using ASB-09, it will not swing over the “long” side of the servo as you have pictured in your drawing.

It is not swinging over the long side. It is at a 45 degree angle to the back of the ASB-09 it is attached to.
8-Dale

I think Tom sees what I am trying to say. No one is arguing over the ‘hip’? servo (the one closet to the robot’s waist). However, the one connected to that is the one that is being cut in half with the short C bracket, according to the 3D model you are showing. If you look closely, you can see where the servo and short C bracket merge. (I take it Alibre doesn’t do collision testing?)

[linuxguy]

@Sienna

And in some of your configurations, you have a C bracket cutting a servo in half!

You are right!

I have updated the third picture (forward bending) to show the actual range of motion using the ASB-09 short “C” brackets. I built the actuall leg with two ASB-10 long “C” brackets attached to the double knee.

8-Dale

[linuxguy]

@Sienna

Without that 5th DOF, the “double knee” doesn’t seem to afford you anything. You are gaining only a tiny increase in flexibility, for a tremendous cost. One of two things could help:

1. Add that 5th DOF, which would make things worth it.

I’ve already said the double knee joint would be more effective with the fifth joint between the knee and ankle. I think I even said so in my original post.

@Sienna

2. or at least use the long C brackets, not the short ones, so that the servo can fold in on itself.

Yes, ASB-10 long “C” brackets would also add to the flexibility of the double knee and increase the range of motion tremendously. This would be equivalent to what the robot in the video has. However, I wanted to see how the double knee would work on the BRAT, which is considerably shorter, and I did not want to add more height than necessary.

@Sienna

(I take it Alibre doesn’t do collision testing?)

Actually it can do interference checking. I just haven’t been using that feature, but I should be.

8-Dale

[sienna]

@linuxguy

I have updated the third picture (forward bending) to show the actual range of motion using the ASB-09 short “C” brackets. I built the actuall leg with two ASB-10 long “C” brackets attached to the double knee. 8-Dale

I can handle complexity. It’s the simple things that confound me.

I’ve already said the double knee joint would be more effective with the fifth joint between the knee and ankle. I think I even said so in my original post. :Wink

Ok, so we are in violent agreement

[sn96]

@linuxguy

I guess I am just going to have to upgrade MsBRAT with the Matt Bauer knee joint and fiddle with it. I am having difficulty getting her walking using HSERVO.

Yes, the BRAT is special - no doubt about that. You may or may not be right about this, Mike. I think actual experimentation with the double knee joint is needed to actually test how it performs.

I’m going to model the right leg with the double knee. Then I can really do some checking as to ranges of motion and such. I definitely do not think the double knee is redundant. This is new for the SES, so we will have to see how it works.

8-Dale

I say go for it! We can all learn from this. The SES brackets allow the freedom and flexibility to try new configurations, and it would be interesting to see how it works out.

[linuxguy]

@SN96

I say go for it! We can all learn from this. The SES brackets allow the freedom and flexibility to try new configurations, and it would be interesting to see how it works out.

Oh, I will do this at some point. I have to get MsBRAT walking in the standard configuration first though. Getting her to walk is proving more difficult than I anticipated, and I don’t have the benefit of SEQ.

8-Dale

[linuxguy]

@Sienna

Ok, so we are in violent agreement

No, not violent, just solid agreement.

I have since modeled a version of the double knee leg using two ASB-10’s attached to the knee joint, which does allow the BRAT to bend fully forward.

There is also the possibility of having the double knee joint protrude from the back of the leg instead of the front as it is now. I’m even considering a new version of the front arm for W.A.L.T.E.R. 2.0 that uses this double joint.

I wonder what kind of flexibility a biped would have with two of these joints… Muahahahahahahaha

8-Dale

[tom_chang79]

@linuxguy

I wonder what kind of flexibility a biped would have with two of these joints… Muahahahahahahaha
8-Dale

Videos, videos, videos please! I can’t get enough of robot videos!

[wingzero01w]

img142.imageshack.us/img142/8751/fulllegsl2.th.jpg

My dual knee leg design (havent put the hip rotation at the top yet) and i couldnt get the servo model for alibre

[linuxguy]

@WingZero01W

My dual knee leg design (havent put the hip rotation at the top yet) and i couldnt get the servo model for alibre

You may have a problem getting it to walk without an ankle rotate joint. Ankle rotation allows the robot to lean from side to side.

8-Dale

[sam-98]

Yeah, That model will almost be impossible to make it walk, and if you manage it will walk like a cheap 3 servo walke, wich scks since you have paid for lots of servos.

[sienna]

Nay sayers all of you! :tsk :tsk

Maybe he has a different balancing scheme in mind then tilting the legs? What if he just shifted the top over the leg on the ground?

There are always more ways to do things then the familiar.

[wingzero01w]

Well sienna thats basically what i was thinking when i made it lol. But, this is just a prototype ill be working on. And how am i going to put the ankle rotate in?

[sam-98]

@Sienna

Maybe he has a different balancing scheme in mind then tilting the legs? What if he just shifted the top over the leg on the ground?

Don’t quite understand this, does someone like have a picture?

[linuxguy]

@WingZero01W

Well sienna thats basically what i was thinking when i made it lol. But, this is just a prototype ill be working on. And how am i going to put the ankle rotate in?

There are at least a couple ways you can add ankle rotation.

1. Add another joint between the foot and the last joint of the leg (current ankle joint).

2. Convert the last joint of the leg into an ankle rotation like the BRAT or Scout use.

8-Dale

[innerbreed]

this is a few of the designs from 2006 i processed in photoshop using a duel knee design. just though this could inspire lynx to sell!
i
img108.imageshack.us/img108/8684/bipedobokv6.th.jpg
img108.imageshack.us/img108/8496/bipedbodydesignframe2eo0.th.jpg

img131.imageshack.us/img131/8840/photozj2.th.jpg

i will be uploading this to the PROJECTS page soon with full details.