|How it started
is an update of the classic arcade game Pong to use bikes as the
controllers. To read about the first version of Cyclepong go to the Cyclepong
always wanted Cyclepong to be something that lots of people would get to
play with. My original idea was to take it to music festivals over the
summer, but things took a bit of an unexpected turn after I took it to
the Dorkbot London Christmas special.
One of the people who played Cyclepong at Dorkbot was Tessa Hunkin
who was there showing her lovely work on illuminated books. She
suggested her brother Tim
Hunkin might be interested as he runs an alternative
amusement arcade on Southwold
Pier. Through Tim we arranged to demonstrate Cyclepong
to the pier's manager Stephen one winter afternoon.
The demo took place in a corner of the arcade around a half-dismantled
basketball machine. Demos are always nail-biting, but I wanted to show
Cyclepong for real. Descriptions and video don't really do it justice.
Fortunately things went well. Stephen liked the game enough to agree to
find a home for an "arcadeised" version if I could produce one. Both
Tim and Stephen also came up with good ideas to enrich the
playing experience. On my side I really liked the pier - while a lot of
piers are in decline the guys on Southwold have done a great job doing
something that is contemporary while still keeping the old fashioned
With a successful demo done I
spent most of the winter
the arcade version of Cyclepong. Work started with the
cabinet and the base for the bikes. Two things dominated in my mind: it
had to be really really
simple and it had to fit in the back of my Mazda Bongo camper van. To meet these goals I decided to make the cabinet
it could be dismantled and transported as a flat-pack.
Woodworking is not really my thing. I got the main panels cut to size
by the retailer. This meant at least the main pieces were the right
size and had true right angles. The rest of the cabinet work I did
myself only using handheld tools. Despite my worst fears things worked
out pretty well. The only thing that went badly wrong was an attempt
to make a rather complicated bezel to fit round the screen. That part
ended up abandoned.
Other construction jobs included removing finger traps and covering
exposed moving parts on the bikes. That was a tedious job, but necessary from the
safety point of view.
front panel outdoors in January. Note the long shadows
For the Cyclepong 2.0 I made various changes to the electronics and mechanics. The first change
was to modify the way the position of the bikes is read. On the first
version of the game I used a speedo cable to feed a hacked Microsoft
mouse. For the arcade this looked too fragile to me so I moved the
mouse on to the base board and used a pulley wheel to connect one axis
of the mouse to the bike's wheel. This looked a lot more robust, but in fact this design proved to have one
(almost fatal) flaw that I only discovered near the end of the project - more on that later.
the monitor and coin system will fit
|The next change was to introduce
the coin system. I used an MEI Cashflow system
This is interfaced to the game using the AVR box that also
reads the bikes positions. The AVR box actually keeps the master copy
of the number of credits in play so even if the PC crashes people
should still be able to recover their paid-for games. The coin systems are pretty
neat pieces of kit by the way, but it's hard to understand the
documentation for them unless you already know a bit about how they are
supposed to work.
The last change was to introduce a new display. The original Cyclepong
projector for the display. Clearly that wasn't going to work for an
arcade machine that's switched on 8 hours a day. I took the need to
find a monitor for Cyclepong 2.0 as a perfect excuse to upgrade my
desktop monitor and speakers!
and base unit almost finished. Note the two boxes on the base
containing the mice to read the bike positions.
and Bug Fixing
Having got the cabinet together I was able to do some serious testing
in the new configuration. Initially the mechanical connection between the bikes
and the mouse reading the position caused a few problems, but with a
bit of redesign that started to work well. However having
reliable mechanics revealed another much more serious problem.
One disadvantage of the new layout is that the mice turn much more
quickly than they did on the original. A bit of experimentation showed
that on the left hand bike you could easily exceed the maximum speed
the mouse was capable of reading accurately. Being a Microsoft product
the mouse didn't handle this situation very well. Instead of setting the
"overflow" bit in the reports back to the computer it just sent a
string of random readings (including reports that it was moving in the
opposite direction to the way it was actually rotating). This was
really messing up the game play.
installed and ready to test
(click to see full size)
point there were two mysteries. Why was only the left bike having
problems, and how was I going to fix things?
examination showed that the bike that was working OK had a newer
version of the Microsoft mouse with a different microcontroller inside
than the one that wasn't working. Clearly somewhere in their production
run Microsoft made a radical redesign of the mouse and the newer types
cope with high speeds better than the old one. After a big search I
managed to find one more mouse of the newer type and fitted it on the
dodgy right-hand bike. Things now worked a lot better, but I was
disappointed to find that at absolute top speed it was still possible
to confuse the mouse and get bad readings.
electronics integrated in to the cabinet
(click to see full size)
My first thought was to bluff this - I didn't think people would ride
the bikes hard enough to expose the defect. I had to revise this
opinion dramatically when I showed the almost finished version to my
neighbour. He leapt on the bike and straight away managed to pedal hard
enough to drive the mouse crazy.
Finally I tried a hack which I thought was desperate, but in fact
turned out to fix the problem once and for all (touch wood). I took the
optical choppers out of the mice and cut off half the spines that
trigger the IR detectors. I didn't think this would work because my
previous experiments with the Microsoft mice had shown they were rather
sensitive to the optical chopper being the right size. When I put
things back together I was delighted to discover I now had two mice
with half the resolution, but twice the maximum working speed of the
originals! Problem solved, but things were pretty close to a major
optical chopper to fix the maximum speed problem.
and Tim made some suggestions for improving the game-play on
Cyclepong. Initially I was a little skeptical because I wanted to keep
a strong flavour of the original Pong with it's basic graphics and
simple sounds. Still the customer is king and for version 2 I added
some new sound effects and also videos that play to celebrate
winning and scoring particularly difficult points. Having done the work
I have to say that the advice I got was good. These changes make the
game much more interesting to play and generally more fun while still
keeping the "Pongness" alive.
a final touch I added an intentionally naf animated Cyclepong logo
inspired by the terrible CGIs used for TV sports coverage.
screen shot - no longer quite the latest version, but pretty similar.
testing and delivery
pre-delivery testing took place at an afternoon garden party. All went
well - no new bugs found and most importantly people enjoyed the game lots.
The public testing did highlight a few
things in the software that needed to be fixed to make the controls
clearer. After missing the first few balls most people pick it up
pretty quickly and find the game absorbing. The simplicity is one thing that people relate to.
to see full size)
Cyclepong can be a serious business
Southwold Pier took place in May 2008. Cyclepong actually fitted in my
van pretty easily - I could have made the case bigger perhaps. On the
pier it was installed next to Tim's Art
which is very appropriate. Deep breath, but it worked perfectly once
assembled. Of course though Cyclepong looked huge in my workshop it
looks rather puny at the arcade - partially because of the design need
to fit in my van. The plan is to expand it a bit with some additional
signage to make it more prominent - updates on that later.
ready to go
For me personally getting Cyclepong to the arcade was a triumph, but I
now see it is just a first step in a process of debugging, publicising
and maintaining the game. If you want to play then go over to Southwold Pier and try it
out for yourselves (and perhaps leave a comment below). I still have a few things to do including a sign
and also some publicity posters to put up around the pier. There might
also be a few more tweaks to the software.
It's still too early for me to tell who well it is going to work long-term. More updates will come throughout the summer.
The overall difficulty of building something like Cyclepong is hard to
calibrate. On the one hand it has a lot of different skills involved:
hardware, software, design, graphics, carpentry. On the other hand it
was built completely from common tools and without a big workshop. If
you fancy making something similar yourself then I would say go for it.
Cyclepong has been a big project for me and it's great to have got it
this far, but I am still nervous about how it will survive in the rough
and tumble of the arcade. Time will tell.
2.0 on the Pier
to see full size)
|Update - July 2008
been two months since Cyclepong arrived on Southwold Pier. The good
news is that it still works and is doing steady business. With some
help from Tim Hunkin we put up some posters to draw people to Cyclepong
as it is a bit away from the other "eccentric" machines. I've also
updated the teaser graphics a little bit.
finishing touch was to add a sign made from an old LCD backlight and a
bicycle tyre. Next to the other machines in the arcade the plain grass
looked a bit, well, boring. Possibly the sign is a bit too high above
the machine. If I get motivated I'll go and shorten the pole it stands
One pedal dropped off one of the bikes in the first month due to
overuse. Hopefully we've tightened it up enough now that it will stay
put. One of the bike bells has broken, but it doesn't really spoil the
game play. Another minor thing to take care of if I get around to it!
Now with new sign!
to see full size)
|Update - August 2008
Well it had to happen eventually. Just a week after I wrote the last
update I got a message that Cyclepong had broken. One of the bikes no
longer moved the player properly. The good news is that the damage was
due to over-use - so people must be having fun playing it. The bearings
in the mouse used to track the bike had worn away which stopped it from
working properly. Tim Hunkin and I took Cyclepong over to Tim's
workshop and with the help from random cats (possibly attracted by all
the discussion of mice) designed and built a new sensor system for the
bikes in a day. I also took the opportunity to replace the broken bell.
Cyclepong in Tim's workshop
(click to see full size)
|The new bike sensors run off the speedo cable on each bike (just like the original Cyclepong
design). In this version though we've taken the mouse mechanics out
entirely. The chopper wheel from the mouse is connected directly to the
speedo cable using a coupling based around two roller-blade bearings.
The original electronics from the mouse is then positioned in the right
place to read the chopper wheel.
Hopefully this version will be more reliable as we've removed the
mechanical stress and plastic bearings of the original mice, and also
the speedo cable rotates much more slowly than the mouse in the
original position. Fingers crossed!
New sensor arrangement
(click to see full size)
|Update - Jan 2011
is returned from Southwold Pier. It had a hard life there but a
reasonably productive one. It never made loads of dosh, but it had a
steady fan-club and made lots of people happy!