Ball gimbal

[Tiberius Brastaviceanu]

Hi everyone, 

Daniel and I were wondering if it is possible to obtain all angles with 2 motors and wheels. 

For example, can we also do rotations? 

Can you drive this ball in any orientation possible with 2 motors? Or you need 3? 

If I look at it again, we're walking on a curved surface, and you only need 2 directions to get from one point to another point. So essentially could walk the ball in all directions, going on parallels AND meridians. 

I also see that it would be a bit complicated to obtain a precise coordinate on the ball.

[Jim Anastassiou]

If you look at the video with  the bike they get all angles with 3 motors.

--
Go to SENSORICA's project page
http://www.sensorica.co/home/what-we-do/projects/pv-characterization
---
You received this message because you are subscribed to the Google Groups "PV characterization project forum" group.
To unsubscribe from this group and stop receiving emails from it, send an email to pv-characterization-pr...@googlegroups.com.
To post to this group, send email to pv-characterizat...@googlegroups.com.
Visit this group at http://groups.google.com/group/pv-characterization-project-forum.
To view this discussion on the web visit https://groups.google.com/d/msgid/pv-characterization-project-forum/1c6ac700-7ec5-4fb6-a905-5e9fbdddc278%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[daniel brastaviceanu]

I posted the 3D model of the gimball here https://3dwarehouse.sketchup.com/model.html?id=uba9f903a-b11f-4c98-b42c-d35d1021ec90

if you want to play with.. or should I make another video?

I like this design alot and I would go for further design if you guys tell me it's full proof.

[daniel brastaviceanu]

Note that in the 3D model you don't have micromanipulators, I'm designing a new compact and low profile micromanipulator to include soon

On Wednesday, July 15, 2015 at 6:42:51 PM UTC-4, Tiberius Brastaviceanu wrote:

[Jim Anastassiou]

Just found this

https://www.youtube.com/watch?v=cycFV22cRws

--
Go to SENSORICA's project page
http://www.sensorica.co/home/what-we-do/projects/pv-characterization
---
You received this message because you are subscribed to the Google Groups "PV characterization project forum" group.
To unsubscribe from this group and stop receiving emails from it, send an email to pv-characterization-pr...@googlegroups.com.
To post to this group, send email to pv-characterizat...@googlegroups.com.
Visit this group at http://groups.google.com/group/pv-characterization-project-forum.

To view this discussion on the web visit https://groups.google.com/d/msgid/pv-characterization-project-forum/1e331b5a-0e75-42b9-893e-35bac707a136%40googlegroups.com.

[daniel brastaviceanu]

I think in this last video they use only 2 motors, and the wheels are placed 90° to each other and to the big ball too.

I just don't understand how the system knows the balance of the small ball... what kind of feedback do they use?!? but this last sentence has nothing to do with our project, just out of curiosity

On Wednesday, July 15, 2015 at 6:42:51 PM UTC-4, Tiberius Brastaviceanu wrote:

[daniel brastaviceanu]

Guys, I posted another design for backup

https://3dwarehouse.sketchup.com/user.html?id=1554342033592517832346042

this one is the simplest one yet and note there is no shadow possible from motors, just like in the spherical drive gimbal,

So in case the GimBall is too complicated or does not quite work for any reason, this one is my final and best shot yet

In order to perform the experiment, you simply tilt  1° and do 360° rotation, then tilt 1° more and do 360° rotation, then tilt 1° more and do 360° rotation, and so on until you complete 90° tilt...

Simple to program and to perform.

Also check out my new micromanipulators, not finished yet

On Wednesday, July 15, 2015 at 6:42:51 PM UTC-4, Tiberius Brastaviceanu wrote:

[Ahmed AKL]

Regarding the feedback system, in time 0:32 it is obvious there is a device hanged over the ball. this device has a strong light source and most probably a camera. 

In addition, the color contrast between the big and small spheres is obvious. 

i'm pretty sure that the feedback system is based on image processing

On the other hand, the ball technique is very appealing and interesting. However, i think it needs a lot of experimentation and heavy mathematical calculations. 

The question is: do we have the time and experience to start such experimentation? 

if the answer is No, i suggest to stick to traditional techniques for the time being

--
Go to SENSORICA's project page
http://www.sensorica.co/home/what-we-do/projects/pv-characterization
---
You received this message because you are subscribed to the Google Groups "PV characterization project forum" group.
To unsubscribe from this group and stop receiving emails from it, send an email to pv-characterization-pr...@googlegroups.com.
To post to this group, send email to pv-characterizat...@googlegroups.com.
Visit this group at http://groups.google.com/group/pv-characterization-project-forum.

To view this discussion on the web visit https://groups.google.com/d/msgid/pv-characterization-project-forum/f7f3d4a7-acce-4415-8496-ebedd5de5be5%40googlegroups.com.

[Jim Anastassiou]

Good point Ahmed but I think there is a much much simpler way to make this work with the sphere, two motors and two omni wheels. And it doesn't even involve complex algorithms. I'll make a drawing in the morning.

To view this discussion on the web visit https://groups.google.com/d/msgid/pv-characterization-project-forum/CAKv7pU9rqVtu_XMDj8Yv4m0jYgbbX5ZEE3DZH%2BL%3DqTDgj2Xvnw%40mail.gmail.com.

[Ahmed AKL]

Actually it depends on the application, if it is a motor bike, then the control system will be very simple, because there is a superior control from the driver

In addition, if the application doesn't require high precision, then, yes, i agree that it doesn't need complicated algorithms

In this case, i really encourage experimenting in this direction because we will be one of the leaders in this mechanism

BUT, if high precision is required for an autonomous system, then we have to consider several factors:

1- The sphere's weight

2- The load applied over the sphere

3- the friction between the omni wheels and the sphere

4- Are we going to use DC motors or stepper or servo motors?

5- if it is a DC motor, do we need a braking system?

6- how the two/three motors are synchronized together to bring the required speed & direction for the sphere?

7- do we need to make simulations first to manage all these parameters, or we will go through try & error?

8- what will be the feedback mechanism? image processing is not a feasible solution for many applications

I don't want to be pessimistic, but using such mechanism is a project by itself

[eugeni...@gmail.com]

guys, when I post 3D models I assume you know how to view them in the web browser but maybe it's not obvious how, so here I will explain:

to the right of the big picture of the model, you have a big red DOWNLOAD. Hit the heart button next to it :)
If you install SketchUp for free, http://www.sketchup.com/fr/download/all you can hit the DOWNLOAD button and view the model inside out...

BUT, you can also view it in WebGL by hitting the middle button just under the DOWNLOAD button. use the mouse wheel for zoom, left click for orbit and Shift+left click for pan

I suggest you install SketchUp on your PC and learn the basics here http://www.sketchup.com/learn , it' very easy and intuitive

Then you can view in wireframe, x-ray, by component, etc, even edit to show your ideas. You can add text notes, etc.

Please see my latest design sketch https://3dwarehouse.sketchup.com/model.html?id=uaa9beca5-4eec-46d7-ba4b-8ac57f5294e2 in 3D, download it to do the simulation

To make something rotate you first need to select that thing:

1- hit Space Bar on your keyboard (this is your selection tool) and click once on the object (here will be the platform with the PV sample). when selected, lines become blue.

2- now hit Q on your keyboard, this is your rotation tool: chose the axis of the rotation you want to perform. I drew lines to represent those axis. So click-and-hold at the end of the axis line, and follow the line with the mouse and release the click. You have set the rotation plane and its center, good.

3- Now click somewhere to grab the object, you can grab it by the sky, by the ground or by a point on the object itself or on another object, depending on what you need to achieve, here it doesn't mater. Move the mouse around and see it rotate. Move the mouse in the direction you want it to rotate. Note: in the bottom right of the screen you see the value it is rotating at.

4- Now either a) click to drop in an arbitrary angle (bottom right of the screen) or b) enter the value in numbers followed by Enter (ex: 45 Enter). Then Enter does the same thing as Click, it drops it.

Repeat these steps for the other axis of rotation and play with the platform to see how it will behave in reality. Have fun with all the models I have shared for this project if you want.

Cheers!

[daniel brastaviceanu]

LOL, I posted as my mom cause I'm using her computer..

[daniel brastaviceanu]

I would really like some feedback here guys

On Wednesday, July 15, 2015 at 6:42:51 PM UTC-4, Tiberius Brastaviceanu wrote:

[Abran Khalid]

I have read through the ideas presented here but I am approaching this from another angle. I intend to remove the complexity of controlling the spherical movement with three motors. If we were looking to use the sphere the same way as the cited projects and websites have, then it would be necessary to have three/motors with complex mathematical modeling. The projects cited require speed more than precision. We require precision, speed is not a concern. They require the sphere to move in random directions. We need it to move in very fixed and predefined positions. We do not need the extras of complex motor control and spherical math modeling. Here is what I propose. 

2 Stepper motors, connected with wheels that are in contact with the sphere to make it spin/move. One motor moves it in purely x-axis while the other is purely y-axis. The wheel for x-axis only press against the sphere when there is direct command from the arduino/micro-controller to move the sphere 1 degrees along x-axis. Similarly, y-axis wheel only comes in contact with the sphere when we give a command to spin the sphere along y-axis. Only one wheel engages at a time. We can not move the sphere simultaneous along x and y axis. 

The sphere it self is held in position by a spring loaded electrically actuated set-screw kind of a thing. Its only purpose is to hold the sphere in position while the wheels are not in contact with it. 

The sphere will be housed in a hollow semi sphere with two slits for the wheels and one for the set-screw.

The way the wheels and set screw can be engaged or disengaged using a mechanism similar to tape decks ( tape hi-fi ). The head of the tape moves in to place when the play button is pressed. When you press the stop button, the head moves in to neutral position. An example of a  linear actuator that will do the job can be seen here :

http://electric-linearactuator.sell.everychina.com/p-90196475-mini-electric-motor-driven-push-rod-waterproof-electric-linear-actuator.html

The wheels do a similar thing for engaging and disengaging. from driving/spinning the sphere.

Using stepper motors to drive the wheels would mean that we would ave exact knowledge of spin on x-axis and y-axis separately. It would provide all the benefits of a stepper motor. The torque issues can be easily addressed since the motors can be mounted remotely to the stage. There is no need for complex calculations and modeling. It does not have speed or agility, but that is not required for this project. 

To me, this solution fits the needs much better than the ones we have been considering before. 

Let me know what you guys think. I will try to model it in a basic drawing and explain it. 

Attached is a drawing that explains the idea. 

Best, 

Abran

I 

[Ahmed AKL]

wow... an interesting solution. I totally agree

On Sat, Jul 18, 2015 at 10:32 PM, Abran Khalid <abran...@gmail.com> wrote:

I have read through the ideas presented here but I am approaching this from another angle. I intend to remove the complexity of controlling the spherical movement with three motors. If we were looking to use the sphere the same way as the cited projects and websites have, then it would be necessary to have three/motors with complex mathematical modeling. The projects cited require speed more than precision. We require precision, speed is not a concern. They require the sphere to move in random directions. We need it to move in very fixed and predefined positions. We do not need the extras of complex motor control and spherical math modeling. Here is what I propose. 

2 Stepper motors, connected with wheels that are in contact with the sphere to make it spin/move. One motor moves it in purely x-axis while the other is purely y-axis. The wheel for x-axis only press against the sphere when there is direct command from the arduino/micro-controller to move the sphere 1 degrees along x-axis. Similarly, y-axis wheel only comes in contact with the sphere when we give a command to spin the sphere along y-axis. Only one wheel engages at a time. We can not move the sphere simultaneous along x and y axis. 

The sphere it self is held in position by a spring loaded electrically actuated set-screw kind of a thing. Its only purpose is to hold the sphere in position while the wheels are not in contact with it. 

The sphere will be housed in a hollow semi sphere with two slits for the wheels and one for the set-screw.

The way the wheels and set screw can be engaged or disengaged using a mechanism similar to tape decks ( tape hi-fi ). The head of the tape moves in to place when the play button is pressed. When you press the stop button, the head moves in to neutral position. An example of a  linear actuator that will do the job can be seen here :

http://electric-linearactuator.sell.everychina.com/p-90196475-mini-electric-motor-driven-push-rod-waterproof-electric-linear-actuator.html

The wheels do a similar thing for engaging and disengaging. from driving/spinning the sphere.

Using stepper motors to drive the wheels would mean that we would ave exact knowledge of spin on x-axis and y-axis separately. It would provide all the benefits of a stepper motor. The torque issues can be easily addressed since the motors can be mounted remotely to the stage. There is no need for complex calculations and modeling. It does not have speed or agility, but that is not required for this project. 

To me, this solution fits the needs much better than the ones we have been considering before. 

Let me know what you guys think. I will try to model it in a basic drawing and explain it. 

Best, 

Abran

On Thursday, July 16, 2015 at 3:42:51 AM UTC+5, Tiberius Brastaviceanu wrote:

Hi everyone, 

Daniel and I were wondering if it is possible to obtain all angles with 2 motors and wheels. 

For example, can we also do rotations? 

Can you drive this ball in any orientation possible with 2 motors? Or you need 3? 

If I look at it again, we're walking on a curved surface, and you only need 2 directions to get from one point to another point. So essentially could walk the ball in all directions, going on parallels AND meridians. 

I also see that it would be a bit complicated to obtain a precise coordinate on the ball.

--
Go to SENSORICA's project page
http://www.sensorica.co/home/what-we-do/projects/pv-characterization
---
You received this message because you are subscribed to the Google Groups "PV characterization project forum" group.
To unsubscribe from this group and stop receiving emails from it, send an email to pv-characterization-pr...@googlegroups.com.
To post to this group, send email to pv-characterizat...@googlegroups.com.
Visit this group at http://groups.google.com/group/pv-characterization-project-forum.

To view this discussion on the web visit https://groups.google.com/d/msgid/pv-characterization-project-forum/9cf28ad1-0dbf-4c24-b9a0-2a18cc1fd0c7%40googlegroups.com.

[Jim Anastassiou]

I would like to know how your x and y axis omni wheels bring the hemisphere back if the stage reaches perpendicular along those axes?

You need coupled wheels across from them or it doesn't work.

[Ahmed AKL]

@Jim, nice point to mention :-)

I think it can be solved by increasing the length of the sphere perimeter (at the omni wheel contacting point). However, the sphere movement should be between 0 and 180. This slight increase in perimeter will allow a contact between the sphere and the wheel when the sphere reaches a perpendicular position in both direction

[Abran Khalid]

@Jim. Thanks for raining on my parade. I did think about that eventuality. We could either go with coupled wheels or something along the lines of what Ahmed suggested. With coupled wheels, the issue would be mechenical linkages to couple the wheels. Or we would have to put in two steppers to move the moves individullay. More steppers, more possible error and more componenets. However, we can look at it in positive light too. Two steppers working in tandem provide more torque. It can also work as error elimination where steppers can be cross checked ( there will be no refrence point though. we would just know there is an error, not exactly which motor, but we would know.)

With Ahmed's idea of increasing the perimeter, we could solve the issue with just one motor. The wheels moving the hemisphere would need to be at the very bottom of the casing. This would allow them to be contact even when the angles are at 90 degrees. 

@ Ahmed, thanks for the thumbs up. Increasing the perimeter without adjusting the position of the wheels to the very bottom of the sphere will not solve the issue. Imagine a wheel half way up the perimeter of the sphere. When the spehere moves one way, it can have a lot of extra area still to go when the angle reaches +90. If we move the sphere in the other direction, the wheel will lose contact with the sphere at less than -90 degrees. 

BTW, we won't need omni directional wheels with this implementation. Since only one wheel ( x-axis or y-axis ) is in contact at a time, and the other is pulled back a few mm, we do not have to worry about letting the non-powered wheels to spin freely when the powered wheel is spinning the sphere.

Hope I made sense. 

Best, 

Abran

[daniel brastaviceanu]

OK, then how about just one traction wheel (not even omni wheel) that has the ability to steer. So the wheel can steer and rotate the ball in any direction and it is placed at the very bottom.

But anyway, Joshua expressed his feelings in dis-favour of the spherical drive... 

On Wednesday, July 15, 2015 at 6:42:51 PM UTC-4, Tiberius Brastaviceanu wrote:

[Abran Khalid]

The reason I am still pushing for the hemisphere because I think its elegant, simple and more intuitive. Its our combined brainchild, and as Joshua said, its cool. 

Actually, here is what Joshua had to say about the hemisphere 

"

The spherical drive is really cool - but I prefer the first design (attached) from Daniel (because of the anisotropic nature of samples - we would still need to reposition contacts/micromanipulators - there is no escaping it) -- the added complexity for 3d printing and cost of 3 motors/controllers does not seem worth it. File that away though - might make an unreal BRDF system."

Lets look at the points Joshua raises one by one,

"  the anisotropic nature of samples - we would still need to reposition contacts/micromanipulators - there is no escaping it"

The hemisphere with the third wheel can spin like a disc on a turntable. This enables the stage to reposition using software and a motor, instead of doing it by hand. That is an advantage.

"the added complexity for 3d printing and cost of 3 motors/controllers does not seem worth it"

I would concede the point about complexity in 3d printing. However, the sphere does not have to be 3d printed. The main reason for 3d printing was to have the system accesable to developing countries or smaller labs that can not afford prooprietry systems. The hemishpehere can be made out of regular objects that can be easily found. Glass flasks used in chemistry labs, globes, or any sphere item of the right size. The system i have proposed does not use any special omni wheels. so apart from the stepper motors, there is nothing particulary special about it.

As far as the complexity of controlling 3 motors is concerened, that I think is the major reason Joshua feels we should lay off the grass for a while :)

Using 3 motors with wheels working in tandem at diffrent speeds to move the hemisphere, is a very complex design. As i explained, we do not need it. The projects we looked at ( Robot navigator and the motorcycle ) require speed of movement. We require precision. For their systems, getting in the complex geometry of 3d movements and controlling the motors using the spheres movements is neccasry. For us, we can just use simple movements in x and y axis at low speeds. 

The system we have shown to Joshua is very complex to control. That is the major reason he thinks its not feasible. But with reduced complexity of movement and resultant ease of control, I think Joshua would be ok with the system. 

Another major concern for me is what Joshua wrote in his last email. 

 The Fig on pg 11 and the design on pg 20 looks like the geometrical mistake from before - I dont see how it gets all of the angles of incidence 

I belive he is pointing to the worm gear underneath the stage. In its current orientation, it will spin the stage like a disc on a turntable but it wont tilt it. That is my understanding as well. i think we need to have a look at it again. 

" I am agnostic on the worm drive - if the team thinks it has an advantage over the direct drive - lets do it -- but I worry about wear on the gears messing with precision down the line."

"

The worm drive will have the issue of gear teeth wearing out. Specially if we are doing 3d printed ones. If we do metal ones, it raises the same issue as Joshua had with sourcing a hemisphere. its not as readily available and not 3d printed. 

The last thing i want to comment about is the Torque requirement. If we go with the model Joshua is recommending, our torque requirements are still not fulfilled according to calculations Tibi presented. We were require 10 times more torque than the motors we selected output. The fundamental reason why the hemisphere idea came in to being was that the motors could not produce enough torque. If we go with the older design, we still have the same issue.

Hope I made some sense. 

Best, 

Abran

--
Go to SENSORICA's project page
http://www.sensorica.co/home/what-we-do/projects/pv-characterization
---

You received this message because you are subscribed to a topic in the Google Groups "PV characterization project forum" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/pv-characterization-project-forum/xP4DyKwJ2Ys/unsubscribe.
To unsubscribe from this group and all its topics, send an email to pv-characterization-pr...@googlegroups.com.

To post to this group, send email to pv-characterizat...@googlegroups.com.
Visit this group at http://groups.google.com/group/pv-characterization-project-forum.

To view this discussion on the web visit https://groups.google.com/d/msgid/pv-characterization-project-forum/d2460b72-a6a7-4652-a465-511782968b39%40googlegroups.com.

[Abran Khalid]

Since the last discussion with Prof. Joshua, it has become clear that he does not think the ball gimbal to be a viable solution for this particular task. However, we were encouraged to store the information from this particular discussion for future use. Perhaps, one day this idea will find its true home. For now, it shall remain in the dark recesses of our minds. 

I should become a fiction writer :)

Best,

Abran

On Thursday, July 16, 2015 at 3:42:51 AM UTC+5, Tiberius Brastaviceanu wrote: