USB Robotic Arm

[SN]

Have pulled a chunk of Java Code down from http://fivedots.coe.psu.ac.th/.....index.html for this beastie now which I intended to port to Linux libusb and initially C.  There's some great stuff in this that allows you to drive the arm gripper to a physical X,Y,Z position too.  Watch this space (all on x86 a the mo but will make it raspi when mine turns up)

[tzj]

I"ve just gone and bought one and am wondering… have you used a PC psu to power it, just until you get a more efficient transformer… more importantly, would it be able to run on 5 volt without any extra components, cept maybe a fuse.

its interesting that you can use a ps3 controller... especially for the skutter control later, using the d-pad as the base (car) control.

[Dapa]

I brought this robot arm from Maplin this other day and looking forward to assembling it when I can find a moment.

My dad noticed the box for it and was telling me about the time he worked as a maintenance technician in my school's technology department. They had a BBC Micro and a robotic arm, which he described writing a program in basic which pulsed the stepper motors to move it to a described position.

It would be good if, in time, we could perhaps hack the Maplin robot arm to use stepper motors, which are controlled from the RaspPi's GPIO pins or via GertBoard or similar.

Another challenge would be to mount the 'Pi in a bigtrak, self-contained, running on batteries... along with the arm!

[morphy_richards]

That's a really good idea!

I've just obtained some high torque motors to make a motorised base.

BUT … Mounting the arm on a BIGTRAK is a really good and straightforward way of letting it move. I wish I'd thought of that!

A couple of decent geared electric motors is in the order of £15 plus

The components needed to make motor drivers are at least another £5

Then you need raw materials to build the base. Say another £5 at least on plywood, bits 'n bobs and glue.

So that's £25 …

This Bigtrack on sale is £23.33 However it costs a further £7.42 for postage … Nevertheless…

Then there is a further project involved in opening it up and hacking the controls so that it can interface with the GPIO on the Pi, but that's fun!

In terms of feedback or postional control, stepper motors are a good idea but you need to be aware that they produce much less torque (oomph) than standard motors. And the standard motors that fit the gear assembly in the arm are very skinny (ie. you will need to find the same size or smaller stepper motor which means potentially even less oomph) You might be able to assist things by adding counterweights if that is the case.

(I would suggest trying to locate one suitably sized stepper motor and doing some experiments to see how much power it has – you could attach a broken elastic band to the axle and see how many turns it's able to wrap around, or try using it with a cord to wrap round and see how much weight it's able to lift.)

PS. Welcome to the roboteers club

[Dapa]

Some good points on motors - experimenting and finding ways around these problems is all part of the fun. Even if (like me) someone is unlikely to have the desire to be a 'proper' programmer, problem-solving skills cover a multitude of trades. We could even go as far as saying that the RaspPi could keep people in schools, who would otherwise bunk of, from the traditional textbook lessons (including ICT, as it was).

I remember being in Maplin a few years back, seeing the PCB kits, and books on USB interfacing and thinking 'Does anyone *really* do that as a hobby these days'? It was quite a depressing moment.

I'm in to high level HTML / Ruby on Rails application design, as well Linux hacking for web site hosting. With Github and on-line tutorials this field is easier than ever to be a part of, but someone needs a wow moment to get them hooked. For me it was simple graphics in BASIC, leading to desktop publishing, leading to web design, leading to designing systems to output dynamic PDFs, order forms, signage, etc. You need to show people they can harness technology at many levels and make a life that is hopefully more rewarding in the long run.

But, back to the subject; RaspPi, BigTrak, battery packs, video cams, WiFi, GPIO, bumper sensors, GPS, 3G dongles... plenty to shape young minds ...and that's just the generic hardware. I can imagine they'll be many peripherals / kits launched off the back of the RaspPi.

[Dapa]

I have my robot arm assembled now. I'm very impressed at how well it works under Windows for the price. It past the mini rolls just - although the dog wasn't impressed!

Just need the RaspPi now!

[SN]

excellent.

Update on where I'm at - I have now completely recast the java code into pure c now - designing a method for implementing a pseudo-object model to do this for for pure c was interesting but I'm cracked it and I'm now happy as a pig in muck playing with pointers again after many many years.

Just need to integrate the final calls into libusb for it (I'm emulating them at the moment whilst I'm unit testing the port) and I should have a complete linux replacement (with added pseudo angle awareness and move-to-location(x,y,z) commands) for the Windows Arm Control S/W ready to be recompiled on a raspi.

PS - I'm getting bored of vi/terminal editing though (its been real lowest common denominator stuff this) - I'm going to have to switch to geany at some point...

[Dapa]

I’m finding there is a knack to getting the programed sequences as reliable as possible. The power doesn’t appear to be regulated, so using several motors separately in the same instruction whilst you’re ‘recording’ might mean they fall short when being played back, as they’re moving at a slower speed and are timed. The answer is to either have one movement per instruction or to press multiple keys whilst recording, in one ‘take’.

In addition, you need to ensure you have decided on a natural (ideally) marked position to start your recording / playback.

Given these slight limitations, a move-to-location function does sound quite ambitious. I’ll be interesting to see how close you can get.

As it stands, practical demonstrations are challenging for this arm, but it’s never-the-less a fantastic way to bring computer output in to the real world. I mean, it was cool to see an LED being lit off the RaspPi GPIO pins the other week.

Although pushing this robot arm as far as it can go is fun, it would be good to have a more advanced baseline kit in the near future. Using stepper motors that report their position and prevent mashing of the gears. Grippers that can stop once they meet resistance. What would be the ideal target price, though? 100 GBP?

[morphy_richards]

Had a quick look on google shopping for stepper motor

Came up with http://proto-pic.co.uk/stepper.....ith-cable/ at £12.96 as a reasonable price for one motor. However when I look at the torque I find reference to "holding torque" (anyone know what that is?) and is rated as Holding Torque : 2.3kg*cm. To explain this value I had a quick look here. Sounds promising … (Torque explained)

Had another quick look on ebay for stepper motors, quite a lot of results!

This Saab Climate Control stepper motor certainly looks beefy and cheap, but I'm struggling to find specification for it.

Then you would need some gearing. I am sure you get the basic principle of this but, say you had a motor that could lift 1KG at 1 meter / second and you attached a simple gear mechanism with 2 gears. Theoretically – One with 100 teeth and one with 10 teeth. You have the ten tooth gear been driven by the motor and the load attached to the 100 tooth gear then you are multiplying the amount of torque by 10 at the expense of the speed (which is now ten times less).

Worm drive gearing is especially good at this because the worm gear effectively only has 1 "tooth".

Gearing.

I'm thinking about grippers too.

If you just want them to stop when they encounter resistance you could just mount simple off / on micro switches on the gripper.

Or you could use some resistive material or some sort of pico type material that changes electronic properties when it's distorted. There's lots of ways, have a look at this for ideas

[domesday]

Perhaps holding torque refers to how much effort would be required for the load to move the motor. I.e if you are using the motor to drive a vehicle the weight of the vehicle would make it roll back down hill under its own weight when not being driven.

[Ravenous]

Holding torque is the maximum torque the motor can resist, while stationery, before slipping back.  If it's not moving at the time, that's the maximum torque the joint could hold outstretched.  (After gearing etc.)

The actual usable torque when lifting something, moving it, will be less - there are all sorts of other things to consider like vibration, inertia, the maximum switching speed, what sort of driver you're using...

Usually a DC motor will provide much more welly than a stepper, though of course you need encoders or something to keep track of the position.  And steppers I've seen tend to be heavy...

[timecmdr]

Hi Guys,

I'm doing something very simular I'm intergrating Siri on the iPhone to control the robot arm using a set of actions.

I've put a blog together to publish such projects. If any of you would like to post onto it with your progress please contact me and I'll give you access.

The site is pibaking.com or just PM me.

Thanks

Matthew Nichols

[tom.larkworthy]

Just stumbled across this thread. Hi! I am in the process of adding optical feedback for this device. I am 80% confident it can reach 0.1mm accurate. I am using libusb and programming it through python, and using pulse width modulation to slow the arm down. Unfortunately the RaspberryPI is not fast enough to do the image analysis at 30fps, so everything except the controller needs to be done on a desktop (I just use desktop)

please see http://edinburghhacklab.com/2012/05/opt ... -problemo/ for more details.

Presently I am integrating adaptive feedback control (http://wcms.inf.ed.ac.uk/ipab/slmc/rese ... tware-lwpr) and hope to have another blog post out in a couple of months. (the control is tricky)

Tom

[Ravenous]

Just stumbled across this thread. Hi! I am in the process of adding optical feedback for this device. I am 80% confident it can reach 0.1mm accurate. I am using libusb and programming it through python, and using pulse width modulation to slow the arm down. Unfortunately the RaspberryPI is not fast enough to do the image analysis at 30fps, so everything except the controller needs to be done on a desktop (I just use desktop)

please see http://edinburghhacklab.com/2012/05/opt ... -problemo/ for more details.

Presently I am integrating adaptive feedback control (http://wcms.inf.ed.ac.uk/ipab/slmc/rese ... tware-lwpr) and hope to have another blog post out in a couple of months. (the control is tricky)

Tom

That's impressive... just out of interest, am I correct in guessing the raspi would be too slow because of the image processing required? Do you have any feel of how long would it take to process one frame? I suppose the question is, what spec of PC are you currently experimenting with to get 30fps...

[LinuxMinty]

I read here http://www.wikihow.com/Use-a-USB-Roboti ... i-(Maplin) and it looks like you can use PyUSB libraries to control the arm. My question is, though, do we need a USB PC interface for use with the Pi? Can you use the Arm without the interface with a laptop (Linux or Win7)?

[shanester]

I can't seem to get the arm_gui.py working on the RPi via SSH (can't configure $DISPLAY) or in VNC terminal (no permissions). I can run it directly from my Ubuntu netbook, but really want to do it on the Pi. Any help would be appreciated

Thanks

Shane

[shanester]

Update- use gksudo to launch idle and then arm_gui works on the desktop. Therefore I can use the GUI by SSH on my desktop or phone via SSH.

[itsmedoofer]

Hi,

Bit of an old one, but I got one of the Maplin USB arms for Xmas, I found this thread and had a play, great........

But I also found this little Python gem that lets you control the arm via a PS3 controller...

Works a treat....

http://www.mybigideas.co.uk/RPi/RobotArm/

[SN]

After a loooooooong time away from this little forum I think I have found a new use for my robot arm and it goes something something like this...

+ +

Can you guess?

[tzj]

Even though it may turn out like a 3d etch-a-sketch, that still seems cooler than my plan...

Using the arm to vent my toaster 'reflow' oven on cooldown.

[Ravenous]

Is that a laser pen? Thinking of giving presentations remotely?

[SN]

3Doodler - 3D Pen - the biggest issue it has is drawing at a steady pace so that the ABS is laid out in consistent thickness - using the arm should help here. The slackness in the arm joints means you could never get super accurate alignment

[Ravenous]

If you can get it to build itself four new gearboxes with less backlash, you're onto a winnner! (joke)

I got one of these arms for a laugh - it has up to an inch of backlash on the gripper tip if you lift it by hand, that's how much backlash there is. However as long as it's kept "leaning" forward it should be possible to do something with it, if you can get position control of some sort...

Someone had another thread somewhere showing trimmer pots on the four main axes, maybe something like that is the thing to try...

[jekw]

I started a post a couple of weeks ago showing the mods I had done to an OWI robot arm to get feedback on the joint positions
http://www.raspberrypi.org/forum/viewto ... 37&t=69540

A blog post, http://jekw.co.uk/2014/01/05/the-thing/, gives more details and shows the grip/clamp I'm using to pick up a (mouse) ball instead of trying to control the original gripper accurately.

Jim....

[Ravenous]

A blog post, http://jekw.co.uk/2014/01/05/the-thing/, gives more details and shows the grip/clamp I'm using to pick up a (mouse) ball instead of trying to control the original gripper accurately.

Yeah I saw "The Thing" when you posted it, I liked the way you measured the main rotation (waist) using a roller on a multiturn pot. That's currently on my (long) list of things to try someday...