Hi guys, I'm involved with a small racing car and it'd be really useful to record some data about it to help us with analysis of its performance. Unfortunately we're on a very tight budget (we're students) so we can't afford an expensive off the shelf job! I'm interested in putting something together myself but I'm a mech eng student so not really sure where to start! If anyone can point me in an appropriate direction to start doing some research I'd be very grateful. We'd probably be looking at something of the order of 8-10 inputs, sampling of 1 or 2hz (more for some and less for other inputs), on board storage for about half an hours worth of data and probably wouldn't need better than 8 bit sampling (more would be nice but not vital). A GPS log would be fantastic (I have one which can output on bluetooth or NMEA over serial) but I suspect that's a step too far. It would need some kind of timing to keep everything in step, and other inputs would be analogue. It would be nice to have an option of triggering outputs from the logger too but again I think that might be too much.
Any thoughts?
Cheers
Ben

 

Have a look here:
http://www.st.com/mcu/
ST Microelectronics has quite a line of microcontrollers, a number of them are specific to automotive. They have uC's with RAM up to 832KB.

Register on their site, and have a good look around. After you've registered and read the introduction, start here to look at their mcu's:
http://www.st.com/mcu/

 

If you are looking to get up and running quickly and are not hell-bent on designing your own hardware, there is a thriving industry of manufacturers of ready-to-wear multi-channel data loggers. Here is one distributor that markets a line of different data-loggers tailor made for the individual looking to gather data then later download the collected data for post-analysis where it can be sliced and diced to your hearts desire.

A bit of Internet browsing can yield a wealth of manufacturers touting there best and finest.

Reasonably priced for the features they provide, you can get several of them and mount them in those places where the data is located and so avoid the need to route cables from one end of the vehicle to the other.

hgmjr

 

Cheers for the replies. I'm afraid I'm a bit out of my depth with that microcontroller site, the most complicated I've got before was some simple 555 stuff 3 or 4 years ago, and I can't really remember that!
We're not in a hurry - we're putting the car through a total refit this year, it's well overdue as it's been in this state more or less for the last 20 years - and it's unlikely we'll even have the chassis back to us before the start of our racing season (mid March)! The individual data loggers are an interesting idea, but how would they work in terms of correlating the data from different inputs? It'd be interesting to see if e.g. we had a drop in oil pressure as the throttle is depressed.

 

You choose a data-logger that features a time-stamp with each data point and then you correlate them by importing the data into a spreadsheet and aligning the data based on the time-stamp.

hgmjr

 

8 X 8bits at 2Hz is extremely slow, you should aim a bit higher. Technology has come a long way, you might be able to get more detail if you step it up a notch. Although, I am not sure about response time and resolution of automotive sensory.

Does anyone know the max response time of typical automotive sensors?

Steve

 

Steve,
The max response time is going to vary widely, depending upon the type of sensor.

For example, detonation sensors will provide practically instantaneous feedback, O2 sensors will take a bit longer, manifold absolute pressure a bit longer yet, and EGT (if sampled) will be quite laggy, along with other thermal sensors like coolant, intake and throttle body temp sensors.

 

I think that most of our sensors will have a fairly slow response time, we'll probably be looking at throttle position, oil pressure, oil and water temp, revs etc. We may put an accelerometer on the rear axle though, which would obviously require a higher response time. I'd be interested in investigating designing my own a bit further, but for simplicity's sake I'm wondering about one of these. 8 channels, 10 bit and a far higher resolution than we'll need, and for about £25 with USB interface. It doesn't offer the ability to input digital data or trigger outputs, and we'd need to have a computer mounted in the car, but in terms of giving us a reasonably cheap starting point I think it's not a bad idea. Still, no decision made yet, and I'm going to give some more thought to the microprocessor option, but with exams looming I think the simple option may win out!

 

Interesting, and inexpensive too.

A caveat for you though ... if you're going to use a PC for data acquisition, you'll need to take pains to protect it's HDD from shock and vibration; and in a racing car you'll have plenty of both. Even hitting a resonant exhaust note may be sufficient to cause a "head crash", where the drive's r/w head impacts the disk, or simply flies so far from the disk's surface that the data being written is corrupted and unreadable. I had this happen at a former client's place of business; they were using a computer to acquire data from a dynamometer used to measure motorcycle performance. In that case, the solution was to move the computer to a quiet area beyond a concrete block wall.

You might be able to get by using a laptop that's bootable from a USB port, and use an OS on a USB "stick" memory.

 

Interesting, and inexpensive too.

A caveat for you though ... if you're going to use a PC for data acquisition, you'll need to take pains to protect it's HDD from shock and vibration; and in a racing car you'll have plenty of both. Even hitting a resonant exhaust note may be sufficient to cause a "head crash", where the drive's r/w head impacts the disk, or simply flies so far from the disk's surface that the data being written is corrupted and unreadable. I had this happen at a former client's place of business; they were using a computer to acquire data from a dynamometer used to measure motorcycle performance. In that case, the solution was to move the computer to a quiet area beyond a concrete block wall.

You might be able to get by using a laptop that's bootable from a USB port, and use an OS on a USB "stick" memory.

There also exist flash-disks with parallel-ATA or SATA interface, which can be used instead of usual hard disks in a normal computer, not too sure about notebooks.

 

That's a good point, I hadn't considered that. There's a number of computer science guys on the team so they may be able to suggest some kind of solution, and I'll also have a look around to see if there's any similar products for use with a pocket PC (and thus entirely flash based). I'll suggest adding a concrete wall to the car (wouldn't worsen the current aerodynamics!) but not sure that's the option we'll choose