Poppy legs merged

[peteriwr...@gmail.com]

Hi gael, again. I've had an epiphany while preparing my Prussia I3 for the wrist cable holder print, Maybe the legs of the Poppy humanoid could be modified to be the "Endoskeleton" legs of Inmoov, allowing for him to have structural support in his future leg blueprint, as I am thinking of doing this. I have no idea why I am asking these questions either. Comprenede? I'm mad as a hatter.

[RobotGuy6]

You can't be serious.... Poppy in general is a very poor design. IMHO it's a joke of a project... At any rate Gael is already well into his leg design. There is no point in a re-design right now....

[Bodo Facklam]

Robotguy6 I agree, but Gaels leg design is not available and nobody knows when he will publish it. So a lot of people are searching solutions, for more than one year nothing happens. I would be glad for any solution for my Inmoov. 

Hopefully waiting for legs

Bodo

[Perry S]

I think it has taken Gael a while because designing legs is hard. 

I'd love to have some to print as well but I will have to wait. I encourage you to submit solutions and develop designs. This is open source after all. Gael has provided us that solution already. It is a complex subject.

I don't think the Poppy design will get us there. My bot is heavy.

[Bodo Facklam]

Yes you are completely right, it's open source. But I think it is not very effective when a lot of persons bring their own solutions in. The better way would be to cooperate by developing one solution. I am for example not such a good designer, but I am very experienced in programming. So it would be very helpful for me to have a hardware solution or developed ideas for the construction, then it would be much easier to develop the software.

[Mindless]

Hi Guys,,, Correct me if I’m wrong,, But hasn’t the” DIY Servo” just  got all the bugs works out last month sometime..?,, I think this was the breakthrough Mr Gael has been waiting for,,,- there’s a post by Kwatters over at MRL that shows a working  knee Joint – So now you can use any type of motor for the servo- which for the legs would be large gearbox motor or drill motor etc,,

,

http://myrobotlab.org/content/diy-servo-service

[gael langevin]

Precisely!
I am now able to work again on the legs using the DIY servo service. Having low cost powerfull and fast servos was the keychain I was missing.

Forget Poppy's legs. Poppy has never been able to walk anyway(though they claim it does... what's the point?).

InMoov leg test with 12v planetary gear motor:
http://inmoov.fr/wp-content/uploads/2016/10/TestKneeSpeed2.mov

Gael Langevin

Creator of InMoov

InMoov Robot

@inmoov

--
You received this message because you are subscribed to the Google Groups "InMoov" group.
To unsubscribe from this group and stop receiving emails from it, send an email to inmoov+unsubscribe@googlegroups.com.
To post to this group, send email to inm...@googlegroups.com.
Visit this group at https://groups.google.com/group/inmoov.
For more options, visit https://groups.google.com/d/optout.

[Mindless]

Wicked,,, Very nice design on the legs too,,, the planetary Gearbox is a great choice for the legs it gives plenty of option for the final crown& pinion gears to work out any speed/toque problems,,,,- I was looking through Thingiverse the other day for a small planetary unit or a gear train to fit inside a 3d printed ¼ scale servo to see if a ¼ scale servo can be made viable,, (already got plenty of gears trains that could drive it but want to try a 3d printed version),, - I’ve got the servo housing made to scale and a Tamiya Tamtech 370 (sports tuned) motor which has got the toque and fits inside the servo housing  beautifully  ,,, now it’s just Time and Trail & Error-and trying to fit it in-between other projects- the Diy Servo System is a real game changer,,, 

[Qeebo Artist]

Gael, have you thought about the parallel leg design, xrobots used for his gnkDroid walking robot?

[Anton Fosselius]

So what you need is an SEA (series elastic actuator) preferably built using a strong BLDC, think hobby airplane motor.

see this:

http://cdn.intechopen.com/pdfs-wm/51224.pdf

or if you prefer videos:

http://www.genebm.com/prosthetic/

I am currently building a lightweight SEA that i will release as opensource when its done. I am still trying to build a lowcost walking biped. 

Will see if i get my constuction stable/strong enough, might have to switch back to driving the train from a linear rod again, but that adds a lot of weight.

Will use a hall sensor or linear pot to measure spring compression.

[Anton Fosselius]

ou can build a rotory elastic actuator like this:

its from this paper: http://www.nt.ntnu.no/users/skoge/prost/proceedings/ifac2014/media/files/0987.pdf

[Mindless]

Nice,,,, Thanks for posting the docs on this stuff Anton- its good info- I see your raiding you R/C gear for parts,, it handy stuff for R&D—post a message when you’re going to release your work on Thingiverse- I’d like to check it out-

[gael langevin]

Very good progress Anton!

If you have a video posted on Youtube, I sure would like to sse it in action.

Thanks for posting these updated pictures!

Gael Langevin

Creator of InMoov

InMoov Robot

@inmoov

2016-12-07 5:08 GMT+01:00 'Mindless' via InMoov <inm...@googlegroups.com>:

Nice,,,, Thanks for posting the docs on this stuff Anton- its good info- I see your raiding you R/C gear for parts,, it handy stuff for R&D—post a message when you’re going to release your work on Thingiverse- I’d like to check it out-

[Anton Fosselius]

sure

https://youtu.be/zLYcKvG_NWg

still in early development, waiting for some electronics and are printing a new design with aligned wires.

first prototype was just proof of concept and no attachments for sensors or holes for wires where made.

On Wednesday, December 7, 2016 at 9:27:10 PM UTC+1, gael langevin wrote:

Very good progress Anton!

If you have a video posted on Youtube, I sure would like to sse it in action.

Thanks for posting these updated pictures!

Gael Langevin

Creator of InMoov

InMoov Robot

@inmoov

2016-12-07 5:08 GMT+01:00 'Mindless' via InMoov <inm...@googlegroups.com>:

Nice,,,, Thanks for posting the docs on this stuff Anton- its good info- I see your raiding you R/C gear for parts,, it handy stuff for R&D—post a message when you’re going to release your work on Thingiverse- I’d like to check it out-

--
You received this message because you are subscribed to the Google Groups "InMoov" group.

To unsubscribe from this group and stop receiving emails from it, send an email to inmoov+un...@googlegroups.com.

[Perry S]

The raiding is real. Now that I have a working robot I no longer have working planes, cars, helicopters, drones ...

[Anton Fosselius]

Nah, i do not own any RC car/planes/helis or quads. everything i have is bought for my robot projects ;) also bought one of thos "200 different springs" kits, but i figured it will probably be better with RC stuff. or you have to make the springs your self to your own spec. 

But to keep the cost down i have to go as low weight as possible and only use off-the-shelf products.

[gael langevin]

Thanks Anton for the video!

With that sort of motor, it's going to be a fast actuator. So fast that you might need the software to be able to ramp the speed, otherwise it might overshoot the potentiometer.

Or maybe you plan to have a rotary encoder on the motor itself?

Gael Langevin

Creator of InMoov

InMoov Robot

@inmoov

--
You received this message because you are subscribed to the Google Groups "InMoov" group.

To unsubscribe from this group and stop receiving emails from it, send an email to inmoov+unsubscribe@googlegroups.com.

[Anton Fosselius]

i will use a hall effect sensor, see this:

https://www.youtube.com/watch?v=LoqGDTH75jY&t=4s

and this:

https://www.youtube.com/watch?v=xMTKs240lBU&t=28s

The circumference of the pully on the motor pin will be linear to how much the sled will move per rotation. but i think i will have to gear down a bit. 

right now its ~ 8mm / rotation.

Right now the sled can be backdriven, i am not sure this is that good, so might have to switch back to a linear rod, will test with a M4 screw rod.
M4 got a pitch of 0.7mm, so then the sled will move 0.7mm/rotation if i do not gear it down.

The A2212 i use is a KV1000, so it will rotate ~ 7400RPM at 7.2v, thats ~ 127 rotations/sec and a travel distance of ~9 cm/sec with the M4 rod.

but in practice it takes some time to go from 0 to7400RPM, especially during load. so it will be very interesting to see how this will behave. 

On Thursday, December 8, 2016 at 9:19:43 AM UTC+1, gael langevin wrote:

Thanks Anton for the video!

With that sort of motor, it's going to be a fast actuator. So fast that you might need the software to be able to ramp the speed, otherwise it might overshoot the potentiometer.

Or maybe you plan to have a rotary encoder on the motor itself?

Gael Langevin

Creator of InMoov

InMoov Robot

@inmoov

2016-12-08 0:06 GMT+01:00 Perry S <perry.s...@gmail.com>:

The raiding is real. Now that I have a working robot I no longer have working planes, cars, helicopters, drones ...

On Tuesday, December 6, 2016 at 11:08:19 PM UTC-5, Mindless wrote:

Nice,,,, Thanks for posting the docs on this stuff Anton- its good info- I see your raiding you R/C gear for parts,, it handy stuff for R&D—post a message when you’re going to release your work on Thingiverse- I’d like to check it out-

--
You received this message because you are subscribed to the Google Groups "InMoov" group.

To unsubscribe from this group and stop receiving emails from it, send an email to inmoov+un...@googlegroups.com.

[Anton Fosselius]

You can also simulate a spring with the bldc motors to get the exact behaviour you want, but then you lose the energy storage feature and add lots of complexity, but the result is awesome.

see this:

https://youtu.be/w48HnwQ_Fb8

and this:

[Anton Fosselius]

An update on my linear actuator, the pully/wire system in its current design is a failure,

Lessons learned:

1. pully and wire will get very hot if there is some slippage, my wire (fishing line) melted as well as the pully.

2. the torque was lower then expected when the speed of the motor was close to zero. 

I expected low torque as i used 5v 2.5A, but will later use a 11v lipo that will be able to push much more power.

i figure gearing down 1:2 or 1:10 and using a M4 rod will give me something more solid, and less complex at the cost of some added weight.

[Mindless]

Yep,,, That"s R&D,,,,Iv got boxes full of failed R&D Parts-  But i after while you seam to find the right method in a moment of madness,,,,,,lol