Log in

View Full Version : A new project - bike data recorder - we'd like your help! :)



WildBoarMouse
30th January 2006, 21:10
A couple of us are looking to work on a bit of a project in our spare time and would really</I> appreciate any (constructive) feedback you might please be able to offer.

We're also looking for someone with experience </I>in the field of windows application & graphics programming. So, if you're interested in a bit of an after hours hobby/project please read through here and I'll put a few more details at the bottom of the page.


What we're looking to do is put together a data logger aimed at racers (amateur to pro) and pretty much anyone generally interested in analyzing/monitoring all the stats of their bike.

Initially we'd look at doing this in a couple of stages. The first being an all-out data recorder (with maybe a very simple text only LCD display - possibly removable), the second stage would include additional bike sensors, maybe a graphics LCD, and richer featured PC software.

The data we're looking to record in the first prototype consists of...
- Time
- GPS (location + velocity)
- Speed (front + back wheels)
- Throttle position - percentage
- Brakes (front + back) - on/off
- Engine RPM
- G-force (three axis - left/right, up/down, forward/backward) - measured in milli-Gs
- Induction temperature

We'd also like to include... but may cut out for the first prototype
- Lap timer (1/100th second resolution) - requires a separate trackside infrared transmitter
- Suspension travel - a little tricky

The second stage would include sensors such as:
- Exhaust gas temperature
- Exhaust oxygen levels (Lamda sensor)
and anything else that may be useful...

All of these would be recorded to an SD/MMC flash memory card (commonly found in digital cameras). Our initial estimates put the logging rate at just under 5Mb per hour, recording data at up to 50 times per second. Things like GPS location would be recorded at a lower rate due to limitations of the consumer priced GPS modules. A 32Mb card can be bought for around $15. So, that's around six hours of continuous data logging for $15. Not bad... If you want to store more data, use a larger card.


If you can think of any additional sensors that would be worth placing on a bike then please let us know.
e.g.
Brake lever position sensor? would there be any point?


Using combinations of the sensors above it is then possible to work out things such as:
Front wheel off the ground
Front wheel loss of traction (braking)
Rear wheel loss of traction (braking)
Rear wheel loss of traction (acceleration)
Current gear selection
Fuel consumption/averages (fairly accurate when using throttle + RPM to estimate)
Dyno readings (approx HP, approx torque etc...)
0 to 100 km/h and all those similar measurements
Lap timers (list of lap times, split times + max speeds etc)
plus a bunch of other things...

Using PC software this could then be viewed allowing you to see where all these things happened on the track... into a corner on lap 5 etc...

One limitation of consumer grade GPS is a resolution of +/- 10m. This means that GPS location can't be used to accurately</I> track your race line. It will, however, give a good indication of where on the track the recorded data matches.


Questions:
Does anyone have a lap timer IR trackside transmitter that we might please be able to have a look at (Pukekohe or Waitakere areas)? We have ideas on how we would like to do this but would also like to have a look at how other manufacturers have implemented this. We are primarily interested in lap timers that use infrared transmitters stationed on the side of the track.

How much do you think something like this would be worth if it was a good</I> product.

Are there any other products out there that can already do all of this? How much are they (if you know)? If you've used another system before, in what areas did you think it could have been improved?


As for the software engineer. Please send me an email or PM if you're interested. We're looking for someone with a few years of experience at least. I can bash out simple apps easy enough but not something I would consider selling as a professional looking application. <!--StartFragment -->We'd really like someone that was into the bike scene and could see themselves using this product on their own bike. :yes: If you're familiar with windowed graphics programs this would be ideal


Thanks everyone!

You can contact me on:
(021) 665444 or
darcyw%satlan*co*nz

thehollowmen
30th January 2006, 22:21
Most car black boxes record a lot of that for a few seconds. If the bike ECU does everything you want then might not be too difficult to get it to 'piggyback' the data recorded there and store it to a flash drive. Decyphering it may be a little difficult. I remember a sciam mag had a bit on it at one stage.

This ain't my area of knowledge but good luck.

classic zed
30th January 2006, 22:24
http://www.aimsports.com/

If they are of any interest I can demo the units and them.

Let me know what you think:2thumbsup

classic zed
30th January 2006, 22:26
Meant to say I can supply them too.:Oops:

Toast
30th January 2006, 22:42
A sensor for tyre pressures over time (each side and center) would be very useful...seen it used in the Superbike Tyre test ( www.superbike.co.uk )

FzerozeroT
31st January 2006, 08:23
you'll need to up the sampling a bit?

WildBoarMouse
31st January 2006, 21:31
Thanks for all the comments...

Most car black boxes record a lot of that for a few seconds. If the bike ECU does everything you want then might not be too difficult to get it to 'piggyback' the data recorded there and store it to a flash drive.
It would make things easier if all bikes had ECUs. :) The idea of this system would allow the kit to be installed on any bike. In this case it would probably be easier to piggyback the sensors rather than read the data from the ECU. :)




http://www.aimsports.com/
If they are of any interest I can demo the units and them.
Thanks for the offer! It'd be great to see how they work :)


A sensor for tyre pressures over time (each side and center) would be very useful...seen it used in the Superbike Tyre test ( www.superbike.co.uk (http://www.superbike.co.uk/) )
A good idea! We've had a look how to do this but hadn't considered putting the sensors on each side. What advantages would this give over having just the centre?


you'll need to up the sampling a bit?
We could speed it up a bit if it turns out to be too slow, but at this rate it's ripping along.

As an example, at 180km/h it would be taking samples every 1m of travel. At 300km/h you'd have traveled 1.67m every sample. Probably more often than it takes for each wheen to turn around. :)

Samples are logged every 0.02 of a second. The average reaction time is about 0.4sec, on a good day it's 0.2sec. That's 10-20 samples by the time you've even reacted to whatever's happened. :)


Thanks for the feedback so far

classic zed
31st January 2006, 21:59
Thanks for the offer! It'd be great to see how they work :)

No problems call me at work (09) 424 0477 ask for Tim

We have access to another interesting device

http://www.aimsports.com/products/david/index.html

As well as data aquisition, it has 2 "lipstic" cams which you could mount anywhere on the bike so that you can record your own racing to analyse later, that way you can see where you can improve.

TwoSeven
31st January 2006, 22:23
My suggestion would be to work out the wiring loom required for the data logger to connect all the sensors up first - and to choose the quality of the wiring and sensors. This will be most of your cost.

Also, quite a lot of the stuff thats on that list can be calculated rather than measured. Dont measure things that dont need to be measured.

You may want to take a look at the Zigbee unit from Microchip. Its a wireless transceiver. The one I saw demo'd had a 3 direction accelerometer on the board with it. Here (http://www.microchip.com/stellent/idcplg?IdcService=SS_GET_PAGE&nodeId=2112)

WildBoarMouse
1st February 2006, 08:54
My suggestion would be to work out the wiring loom required for the data logger to connect all the sensors up first - and to choose the quality of the wiring and sensors. This will be most of your cost.

Also, quite a lot of the stuff thats on that list can be calculated rather than measured. Dont measure things that dont need to be measured.

You may want to take a look at the Zigbee unit from Microchip. Its a wireless transceiver. The one I saw demo'd had a 3 direction accelerometer on the board with it. Here (http://www.microchip.com/stellent/idcplg?IdcService=SS_GET_PAGE&nodeId=2112)
Hi there, thanks for the input

The problems with the Microchip devices are:
They're only an 8-bit processor (as opposed to 32-bit)
Their lack of speed (only 40Mhz)
Lack of memory (only 4Kb RAM on the 18F series)

The majority of the cost will come from the "black box" itself rather than the sensors. The only additional sensor components to be attached to the bike will be a couple of reluctance sensors, a resistive device on the throttle, and a thermister. Between the two of us working on this already, we have a couple of decades experience working on devices far more complex than this. :)

Zigbee is nice but a little expensive and is only just entering the baby stages of production :) I'm not sure that zigbee would have much of a use with a device like this.

Of this list, what is it that you think can be calculated by using other sensors?
Time - real time clock
GPS (location + velocity) - from the module
Speed (front + back wheels) - individual reluctance sensors
Throttle position - resistive
Brakes (front + back) - existing switches
Engine RPM - from the coil/igniter
G-force (three axis) - accelerometer
Induction temperature - thermister/thermocouple

Thanks for the feedback :)
Cheers