A datalogger embedded in vehicle

[Teufo]

Hello, I have to develop a datalogger for a specific application.
The datalogger must be stand-alone in order to be embedded into a vehicle (in the passenger compartment). The purpose is to store data from several sensors (temperature, humidity, vibration…) that are connected to the datalogger.
After one year of measurements the recorded data must be sent to a personal computer for analysis.
In fact I don’t need something accurate with high frequency acquisition since the data are stored in the datalogger only if there is some high variations with the sensors. But the data logger must have sufficient memory, a low power feeding and several channels (around 7).
The dataloggers that I have found in some companies are quite expensive and I don’t want to design a printed board with a microcontroller, memory, A/D converters…
So I would like to know if it’s possible to buy a board (and where?) including the necessary electronics elements and the possibility to connect several sensors. The board would be programmable and not very expensive.
If anyone have any suggestions I will appreciate it. Thank you.

 

[IRstuff]

How big do you need? Why not log directly to a laptop?

TTFN

 

[jsolar]

See the adverts in Circuit Cellar Inc. and in Embedded System Design. If you are willing to do the systems engineering and the integration task, you will find that you can assemble your own custom system for little money. It will take some research and design. I hope you will use the car battery, which will have a large cost impact.

First, find your raw data bandwidth (BW) and complexity (dynamic range and accuracy) of the data thinning you need. i.e. you mentioned sensors that seem to me very low frequency, except for the vibration sensor. That is out of my field except for devices that generate data in the 800 Hz range. Assuming a dozen sensors all with under 2.4 kHz sample rate you could use a 16:1 analog mux assuming you do not need 100% attention to all sensors. Or use a dedicated A/D for your wide BW vibration signals and a mux for the rest.

To keep our development cost down you might want to stick with the SBC wintel products. There would be essentially no development system cost and all the SW and HW is there for permanent memory control. This should be around US $100. If you go to a less capable design it may not cost much less and the development cost could be much higher. If you sensors are all CAN, you could jump right in with a CAN controller at about $300. I/O boards can be under $50. That leaves you with the power conditioner and mechanics.

Some sensors are very difficult to route any distance. Are you conditioning the signals by amplification or A/D conversion at the remote sensor site and then running a more reliable and economical run to your data center?

Best Regards,
John Solar

 

[Teufo]

Thank you jsolar.
For the moment I dont know how I will do the signal conditioning of the sensors. Maybe I will buy sensors that are already conditioned with the possibility to send their data to the datalogger via cables.
Could you tell me where did you find information about the design of datalogger in Circuit Cellar Inc and Embedded System Design ?
I found nothing about the SBC wintell products, could you give me the net adress where I can have a look to those products?