Hip rotation.. work in progress

[NovaOne]

So this is my hip rotation hack:



I didn't want to post this until I had tested it properly and had some cool videos, but I currently have other demands on my time, and that will now be months away.

[Pev]

Very Nice. Not an approach I had thought of. Certainly has me thinking, I think so many of us get locked into the idea of attaching directly to the servo. Maybe we could consider linkages a little more.

Pev

[MYKL]

Very cool!

It looks sound.

Looking forward to the video.

What are the specs on the bearings you used to mount the hips?

[BillB]

Another wonderful upgrade NovaOne. I always look forward to your mods they are very clever and well made. Keep it up - can't wait for the vids.

[Robo1]

I love it novaone

You just have to get it up and running smoothly to give bill and run for his money in the ring or are you going to win the stairs walking again

I still haven't got my new bot walking

Bren

[NovaOne]

Thanks guys

What are the specs on the bearings you used to mount the hips?

They are 3 part ball thrust bearings OD = 18mm ID =10mm Depth 5mm

[Humanoido]

Great job and very innovative use of materials. How is it working with the linkages in reference to the entire assembly remaining tight enough during walking?

humanoido

[NovaOne]

How is it working with the linkages in reference to the entire assembly remaining tight enough during walking?

I imaging looking at photos this method looks...clumsy or "Heath Robinson" ..I can't deny it might be more vulnerable QED

The structural rigidity of the hip seems unaffected because the bearings are clamped quite tightly together using an M5 screw held in place with Loctite thread retaining compound. The Black 4mm washer shown is made of nylon 66 molybdenum disulfide . ..... I think I currently have the screw/washer/nut clamp assembly upside down because the nylon washer seems to have more friction against the bracket than the lower aluminium "nut"? (shown in the first picture above)

The brackets are held together so firmly using this arrangement, that before I considered the con rods, I was looking a spur gears to couple the servo to the leg. I thought a 2:1 ratio to double the torque and half the rotation angle to 90 degrees would mean I could use a cheaper servo??????

If structural rigidity proves to be a problem in the future I might try a slightly larger diameter bearing ie OD=22mm .

My main problem was counter balancing the new mass distribution......I have moved the batteries to the front, more pictures to follow. I hope most of the stock moves are unaffected ??????

[NovaOne]

I hoped not to modify the RN stock routines, (just program new turn routines) but I am using A5 and D23 servo positions and I found the number 60, sometimes 70 in the stock move commands for both servos all over the place???? Is there something obvious i'm missing
I replaced these with 100 but the rates of these moves seem to have been affected? Could it relate to PTP motions?

I feel certain that someone has noticed this before but I can't find the thread?

[i-Bot]

Rewriting moves for a new anatomy, is very tough, as I found with the double knee joints. The stock moves are very clever and have some subtle parts in them.

As you say, the use of unused servo positions as pacers is one trick. A group move with PTP looks at all the group servos (or all servos in PTP ALL). It finds the servo with the greatest distance to move, and moves this at the set speed. All other affected servos move at a speed in proportion to the distance they have to move. Thus an unused servo making largest position changes sets the pace.

A trap is that in high speed mode, the servos cannot physically move as fast as the higher settings of speed. This means their speed is set by the physicaal limit and not the setting. A consequence of this is that the software thinks the move has finished, before it actually has. So setting the next position may be too early.

All in all, programming moves is an art, and we do need more people to share their experiences and video as Matt Bauer and DirtyRoboto used to.

[NovaOne]

Thanks i-Bot, your knowledge and advice is invaluable as ever. I have dropped all my other silly mods to focus on tweaking the affected routines (and gradually documenting the changes I make here).

The offending moves are:
foot_up
body_move1
body_move2
wing_move
fast_run01
fast_run02
fast_run03
left_turn
right_turn
forward_walk
left_shift1
left_shift2
left_shift4
left_shift5
left_shift6
right_shift1
right_shift2
right_shift4
right_shift5
right_shift6
backward_walk1
backward_walk2
backward_walk9
backward_walk7
backward_walk6
backward_walk5
backward_walk3
left_attack1
right_attack1
left_forward
right_forward

[Droid Works]

Very cool ! I cant wait to see the video

[tum]

Hi

I was wondering where you bought the right angled rods, the black adjustable horn and the thrust bearings (pretty much all the parts you used ).

Thanks.

[NovaOne]

I was wondering where you bought the right angled rods

From my local model shop, they are ball joint linkages: Multiplex part number #71 3860 ....a bit overkill really.

I made/adapted the rest of the brackets myself. I promise to post drawings when i have more time.

[tum]

I look forward to the diagrams.

BTW, do the mods make give him a fat looking ass?