Maker faire

[Guy Burton]

Hi guys,

We chatted briefly about an entry for the maker faire in march/april next year at the weekend, and I have been thinking about it a bit. I think we need to get a rough indication from people how much time they are willing to put into an entry around march next year, and what their specialist skills are. It's very difficult for anyone to come up with decent project ideas which will involve a group of individuals taking on different tasks when there isn't really got much indication of who is willing or able to do what. Obviously some people will have more time than others, due to both level of enthusiasm and other commitments, and this is absolutely fine- we should try and think of a project which allows people to contribute as much or as little as they can, and a project which spans different areas, for example: conceptual design, CAD, mechanical eng/manafacture, electronics, software, and not forgetting project management! We are also growing continually and so it would be good if whatever we embark upon is fairly modular and flexible, which would allow newcomers to get involved as well. 

So for my part I will start this off and volunteer the following information:

I'm a classic electronics/software hacker- I have a list of ongoing projects as long as my arm. I'm happy designing anything or writing software for embedded platforms, but best not to leave me with the job of constructing mechanical devices, as people who attended the ballistics day will attest to. I am quite busy but can make time to do plenty of work on a project, certainly a good bunch of hours a week (4-8). Also willing to contribute existing components and shell out for components or manufacturing if necessary (to an extent obviously).

Furthermore I will also put forward my suggestion for a project (and I'm sure that those I have already spoken to about this will be relieved that it is not in fact a mech-suit):

I think we should design and make a themed pinball table for  the following reasons:

Look great (lots of flashing lights and cool noises)

Fun end product to play with (=attention at maker faire)

Product made up of lots of smaller components

Components are mechanical, electro-mechanical, electronic, and all tied together with software

Needs creative design skills for theming and decorating

The main bonus for me, is that it is the sort of thing that you could make on your own, but really it is a bit too involved for an individual. But if there were a whole bunch of us getting stuck in then I think that it is a highly achievable, yet impressive goal. Obviously it isn't massively original (well, the theme could be) but it is our first trip to the faire as a group, so there is no reason to feel we need to take anti gravity boots or anything.

Opinions?

Guy

[Matthew Burnham]

Interesting idea. I've put together a wiki page for summarising Maker Faire info. Can't remember all the ideas we had at the pub meet, but it's a starting point.

On the issues with the Wipeout idea, we were mainly worried about the Tesla coils interrupting the video RF. Only one thing for it - build a tesla coil and test it :) Looking again at the Racer write-ups, would we have enough space to build a course?

Pinball wise, how can we make it a bit more than just a pinball recreation? How can we integrate with other hackerspaces? Could we make it multiplayer, with Rube Goldberg eqsue challenges to complete?

Mat

[Adrian Godwin]

Pinball is a nice idea (as the owner of 5 tables I have an interest :-) but making a decent one from scratch is a surprising amount of work. I've been involved on one for 2 years with 4 other people and it's nowhere near finished.


So .. if you want to do that, you've got to really simplify it, and make up for the lack of excitement / detail with some overwhelming USP like blowing things up, or water, or using a footbal instead of a pinball etc. etc.


-adrian

[Stephen Bridges]

> > Pinball wise, how can we make it a bit more than just a pinball
> > recreation? How can we integrate with other hackerspaces? Could we make
> > it multiplayer, with Rube Goldberg eqsue challenges to complete?
> >
> > Mat
>
> Pinball is a nice idea (as the owner of 5 tables I have an interest :-) but
> making a decent one from scratch is a surprising amount of work. I've been
> involved on one for 2 years with 4 other people and it's nowhere near
> finished.

AH, that's a shame, I was really getting into the pinball idea.

> So .. if you want to do that, you've got to really simplify it, and make up
> for the lack of excitement / detail with some overwhelming USP like blowing
> things up, or water, or using a footbal instead of a pinball etc. etc.

Perhaps the USP could be making a huge one? All the hackerspaces build a bit,
and then we get some webcams to a set of screens because it's so big you can't
easily see the far end? For some reason this feels like a Crystal Maze game.

Only thing that worries me is the tolerance on the floor probably has to be
quite high to not feel very cheap.

Doing crazy things with stuff to hit would be great, but I can't immediately
think of the things we could do, and it would need balancing a bit. I think
there would be more mechanical work than anything else. Like Guy, I'm
predominantly a software person with some electronics, and couldn't help so
much with mechanical problems. At least not if they need doing well.

Widget

[Guy Burton]

> make up
> for the lack of excitement / detail with some overwhelming USP like blowing
> things up, 

I like the sound of this!! 

[Martin Raynsford]

In my mind a large item means being relegated outside at the maker faire (assuming they hold it at the same place). In a practical and realistic sense I'd rather not spend the whole weekend under a gazebo in newcastle

On 30 Aug 2010 22:42, "Guy Burton" <g...@beatbutchermpc.com> wrote:

>
> > make up
> > for the lack of excitement / detail with some overwhelming USP like blowing

> > th...

[Matthew Burnham]

Well if we're going to be the other side of Newcastle, how about DIY Segways http://hackaday.com/2010/08/17/ride-fly-segway-diy/?

[Stephen Bridges]

On Monday 30 August 2010 22:59:04 Matthew Burnham wrote:
> Well if we're going to be the other side of Newcastle, how about DIY
> Segways http://hackaday.com/2010/08/17/ride-fly-segway-diy/?

There were propelled skateboards there last time, although I think they ran
off a remote control. I can't remember the details now, but there was
something a bit odd about them. Martin may remember.

Widget

[Matthew Burnham]

Apparently it was the World Hovercraft Championships at Towcester over the weekend. Shame it wasn't advertised better.

[Martin Raynsford]

There were several mono wheeled skateboards where speed was controlled by tilting the board, a few had remote controls on them. The were some normal powered skateboards too but I don't think there was anything special about them at all.

Getting back on track with the skill sets,

I'm really an embedded software person, but I'm currently employed writing in .NET and I have lots of experience making things talk to computers and interfaces to drive them. The list of things I'm making is also as long as my arm, prolific would be a fairly accurate description. Most of my home projects are largely physical things with minimal electronics and hardware.

 

I'm pretty handy when it comes to flat packing projects, so don't worry too much when it comes to to fitting the stuff into cars/vans, if we want a largeish project it should be simple enough. A wipeout style track could easily be squashed down into one vehicle. So far I've had 30m of wooden fence, and a 5m tall 3 foot sail span windmill in the back of my car and I didnt even need to fold the seats down.

 

My initial thoughts on a project were based around the coaster bots in the make magazine. Lots of people have made them individually but we could put them all together and have a swarm of them. The hardware is based on servos and arduinos so they are too expensive seperately. But that was about as far as my idea got, I'm sure we could have better ideas.

[Matthew Burnham]

On Tue, Aug 31, 2010 at 7:35 PM, Martin Raynsford <msray...@googlemail.com> wrote:

I'm pretty handy when it comes to flat packing projects, so don't worry too much when it comes to to fitting the stuff into cars/vans, if we want a largeish project it should be simple enough. A wipeout style track could easily be squashed down into one vehicle. So far I've had 30m of wooden fence, and a 5m tall 3 foot sail span windmill in the back of my car and I didnt even need to fold the seats down.

My concern on size of a Wipeout track was in size once re-constructed. I guess it wouldn't matter too much if the track were relegated further away, so long as it was out of the weather. Still a shame that people couldn't see both ends of the system. Hmmm, could we link up to something physical over the Interwebs?
 

My initial thoughts on a project were based around the coaster bots in the make magazine. Lots of people have made them individually but we could put them all together and have a swarm of them. The hardware is based on servos and arduinos so they are too expensive seperately. But that was about as far as my idea got, I'm sure we could have better ideas.

CoasterBots? I thought you were referring to the GyroCar, but having Googled you're not. They seem quite cool though.

If we wanted to a build a swarm I think we could come up with something cheap enough.

[Martin Raynsford]

I think the key to a wipeout track would be making it long but dense, so you could easily have layers on top of each other, building it up in height rather than width. This would allow people to see the whole thing the whole time. There was also quite a large empty space upstairs at the fair this year, not massive but certainly large enough for a reasonable sized track.

 

http://blog.makezine.com/archive/2010/04/make_robot_build_assembling_the_bra.html

This is the first page I found about coaster bots, my initial concerns about a swarm of coaster bots centre around powering them all weekend, if we had just the one robot on a track we'd get through far less batteries.

 

The alternate suggestion was something like super mario kart, and two vehicles that drove round and augmented reality to firs tortoise shells at each other etc.

[Guy Burton]

I actually saw a load of mini hovercraft (is that the proper pluralization?) on trailers coming out of the racecourse on monday, I did wonder what the hell was going on....

[patrick]

I drove through Towcester that weekend and saw 2 signs, both very small,
both on the same roundabout and that was it. Not well publicised is an
understatement!

On Tue, 31 Aug 2010 20:49:37 +0100, Guy Burton <g...@beatbutchermpc.com>
wrote:

> I actually saw a load of mini hovercraft (is that the proper
> pluralization?) on trailers coming out of the racecourse on monday, I
did
> wonder what the hell was going on....
>

> On Tue, Aug 31, 2010 at 7:06 PM, Matthew Burnham wrote:

> On Tue, Aug 31, 2010 at 6:55 PM, Stephen Bridges wrote:
> On Monday 30 August 2010 22:59:04 Matthew Burnham wrote:
> > Well if we're going to be the other side of Newcastle, how about DIY

> > Segways http://hackaday.com/2010/08/17/ride-fly-segway-diy/ [3]?

>
> There were propelled skateboards there last time, although I think they
> ran
> off a remote control.  I can't remember the details now, but there was
> something a bit odd about them.  Martin may remember.
>

> Apparently it was the World Hovercraft Championships [4] at Towcester

> over the weekend. Shame it wasn't advertised better.
>
>
>

> Links:
> ------
> [1] mailto:matbu...@gmail.com
> [2] mailto:stephen...@gmail.com
> [3] http://hackaday.com/2010/08/17/ride-fly-segway-diy/
> [4] http://www.whc2010.org.uk/

[Stephen Bridges]

On Tuesday 31 August 2010 20:59:55 patrick wrote:
> I drove through Towcester that weekend and saw 2 signs, both very small,
> both on the same roundabout and that was it. Not well publicised is an
> understatement!

There was a story on the Chron that I saw on the Thursday afternoon, which was
when it stopped being free. I was cross.

W.

[Stephen Bridges]

On Tuesday 31 August 2010 20:16:00 Martin Raynsford wrote:

> The alternate suggestion was something like super mario kart, and two
> vehicles that drove round and augmented reality to firs tortoise shells at
> each other etc.

I think while augmenting specific real objects (like barcodes) is relatively
easy, I've got the impression that e.g. firing a projectile and then plotting
that projectile through a 3d space that is then overlayed with the real world
is much harder.

A powerup that gives you an electrified rim to approach other cars with
perhaps? Or prove me wrong, which I'm equally happy with.

Widget

[Tree]

i know im not a local (i actualy live in newcastle), but i love the
idea of a robot swarm.
from talking to Martin his estimate of £50 a bot is WAY above what ive
costed it for.

i have made a breakout board for an ATmega392 (the ones in the
arduino) that has all the functionality of the big boy (apart from
programming, but thats not important right :D) and its final size is
37mm x 44mm x 12mm (LxWxH) and fully populated i recon they can be
done for £8
then its just a case of getting 2 small servos, e-bay has some 9g
micro servos for stoopidly cheap @ something like £6 for 2.
add to that a sprinkling of resistors, IR LEDs and IR Sensors and we
pritty much have the bots for £20 a pop

there is also economy of scale with this as well, as postage becomes
insignifiant if we buy enough of the parts.

imagen being able to: sell kits and alow the public to make and place
there own mini-bot into the swarm.




------------Notes----------------

Atmega board componants list:
24pin Dill socket
20 header sockets
2x 10uf polerised capacitors
1x 16mhz crystal
2x 22pf capacitors
1x 5v volatge regulator
1x 3v diode
1x 10k resistor
1x ATmega ( http://www.coolcomponents.co.uk/catalog/product_info.php?products_id=272
) (i also have a bootloader board, so saves us cash again)

-------------------------------------

[Guy Burton]

What swarm like behavior are we proposing here? Just having a load of bots isnt good unless they can communicate and do something awesome together.

[Martin Raynsford]

I reckon an IR link is the easiest way to get them to communicate with each other and a PC, as to what they could do I'm less sure about that (hence the lack of championing the idea)

 

um, red arrows formation type driving?

[Tree]

group work like moving an objects,
from a random pattern of placments have them all line up or spell out
something if we have enough of them.
red arrows formations,
Tron bike autonomus-ness,

and with a few extra componants could have them draw the mona lisa as
a group all with a pen sticking out the top.



On Sep 1, 3:40 pm, Martin Raynsford <msraynsf...@googlemail.com>
wrote:

> I reckon an IR link is the easiest way to get them to communicate with each
> other and a PC, as to what they could do I'm less sure about that (hence the
> lack of championing the idea)
>
> um, red arrows formation type driving?
>
> On 1 September 2010 15:36, Guy Burton <g...@beatbutchermpc.com> wrote:
>
> > What swarm like behavior are we proposing here? Just having a load of bots
> > isnt good unless they can communicate and do something awesome together.
>

[Guy Burton]

I don't know too much about IR comms- how would that work, do all the bots modulate their signal at a different frequency, or would you do time division multiplexing or what? AFAIK IR requires line of sight so it would be hard for the swarm to communicate with broadcast style messages unless you had some kind of mesh system, or as you say some sort of hub (like a pc, maybe with a receiver in the z axis to prevent clashes?), but then when fish swim in a school they only monitor their closest neighbors for their movement, so maybe localized comms would be cool anyway?

Lots of random thoughts there, sorry.

[Bob Clough]

predator / prey interactions are always fun when you have swarm bots.  I remember years ago seeing a setup that had a bunch of 'prey' bots that charged in certain places, and predator bots that charged by stealing power from the prey, was very cool!

[Martin Raynsford]

Reading university style stuff and we happen to have a reading uni student of that era in Steve (he'd be the powered floor expert)

Localised comms is the way forward with coms being repeated to broadcast group messages across the whole swarm, the PC link could be more powerful IR blasters in each corner to try and reach as many bots as possible.

 

Bot's with pens sensors so that they could follow lines and or make their own lines would be interesting.

 

I remember watching a program about ants, they apparently leave a scent trail when they go out looking for stuff, once they find food they follow it home, over time the trail gets thicker and ants tend to follow it more directly until they end up with something approximating the best route, perhaps that would be cool to mimic?

[Tree]

see now we are thinking
kool thing about this kinda bot, is that its litteraly just code that
changes baced on the job.

TRON Bikes
Preditor & Prey
Group Tasks
Formation movement
BritishBulldog
Bot-Art

[Tree]

sounds good to me.
and all it takes is a change of code to change tasks.

the bots could even be fitted out or have an addon that they can be
used for the Wipeout

[Stephen Bridges]

On Wednesday 01 September 2010 15:58:09 Martin Raynsford wrote:
> Reading university style stuff and we happen to have a reading uni student
> of that era in Steve (he'd be the powered floor expert)
>
> Localised comms is the way forward with coms being repeated to broadcast
> group messages across the whole swarm, the PC link could be more powerful
> IR blasters in each corner to try and reach as many bots as possible.

Yes, I was at Reading when they were making the robot swarms that are at the
Science Museum and the like. I'd estimate they cost £50-75 each, but there's
many obstacles to overcome for a good swarm. Making a powered floor isn't
amazingly difficult but it's not easy. Alternatively automatic charging is
difficult and obviously means each robot is only running about 25-35% of the
time of the show.

Charging is done by having an array of five probes pointing down with spring-
loaded ball bearings, with the array (as a cross) designed so that it always
gets two probes on alternate strips on the powered floor.

Most comms in swarms were done as presence detectors, IIRC. So effectively
each robot would transmit a beacon at a suitably high refresh rate. I
designed some beacons when I was there. I don't know if they actually
transmitted some kind of protocol with them.

> Bot's with pens sensors so that they could follow lines and or make their
> own lines would be interesting.
>
> I remember watching a program about ants, they apparently leave a scent
> trail when they go out looking for stuff, once they find food they follow
> it home, over time the trail gets thicker and ants tend to follow it more
> directly until they end up with something approximating the best route,
> perhaps that would be cool to mimic?

This was also investigated (in software) at Reading. It was called Stochastic
Diffusion Search (http://www.reading.ac.uk/cirg/sds/cirg-
stochasticdiffusionsearch.aspx). I don't know of any work done in the real
world with it. You need something you can put down in precisely controlled
quantities, and a reliable sensor to detect it. You also need to be able to
clean up when you want to start again.

Just a general note on swarm robots, all the robots need to work, which
reduces the amount of tolerances in the design.

Widget.

[Matthew Burnham]

This all sounds quite cool. If we have a basic spec defined in terms of mechanical size and electronic interface we can allow 'Queen' bots to have extra sensors on top of this. e.g. A white line follower could lead others by simply programming the rest to follow another.

Concur we have IR for comms as already suggested; RF may be cheap enough but wouldn't be range dependent. To that we should add a multicolor (RGB) LED on top to allow us to indicate status or perhaps teams, etc. for robotic football and the like. Bump sensors could also be quite handy and cheap if maybe fiddly.

Cheap servos should work well for drive as you get all the drive electronics included. Shouldn't be too hard to mod for continuous rotation and easy to drive from a microcontroller too.

There should be some way to easily reprogram the general task (not necessarily full microcontroller), preferably over IR. Then we can try different algorithms for control and change from say robot football to a game of life or something.

Stephen, I disagree with that they need to all work reliably. If they're autonomously all following the same program, then one being duff shouldn't matter.

Oh, and it should be pretty simple to CNC a lot of the mechanics from MDF.

[Stephen Bridges]

On Wednesday 01 September 2010 21:12:05 Matthew Burnham wrote:

> Stephen, I disagree with that they need to all work reliably. If they're
> autonomously all following the same program, then one being duff shouldn't
> matter.

You'd be surprised, it'll spoil whatever pattern you're aiming for and be very
obvious to the first person to walk up. Unless you use a subsumption
architecture it may also collide into other robots, but that's easily dealt
with.

Widget.

[Tree]

all the IR people. would these (http://cgi.ebay.co.uk/100-x-5mm-IR-
infrared-LED-Lamp-850nm-100-Resistors-/300455623088?
pt=LH_DefaultDomain_0&hash=item45f48cf9b0) work for the IR system? as
i rememeber something about an LED being able to be used as a senser
as well

ive got most of the bits to put together a tester one of these sanse
IR stuf, so i think thats my saterday sorted :D cardboard body ftw.

as widget said that reliablility is key, the longer we have to
"perfect" our chosen project the better realy.

[Matthew Burnham]

Also vaguely recall something about emitters being used as sensors, but I can forsee unnecessary problems in trying to do that. I do know I had a lot more success with LIRC circuits using three wire frequency tuned phototransistor devices than photodiodes.