Open Source Data Logger Project Using the Arduino?

I didn't read all the posts in the thread...
Is this project YOUR project?!?
http://www.openlogger.org/tikiwiki-2.2/tiki-index.php

Or maybe this?
http://openenergymonitor.org/emon/node/28

Or maybe you are working to a totally new project? In that case, which is its status?

While I feel that an arduino based logger is still a great idea, especially now that the atmega1280 leaves plenty of pins and memory after you've hooked up the SD card, for our purposes, we have decided to go with a PC104 board running linux. Something like the TS-7800 or an Atom based board. This comes at a substantially increased cost and power consumption, but still within our parameters.

While I feel that an arduino based logger is still a great idea, especially now that the atmega1280 leaves plenty of pins and memory after you've hooked up the SD card, for our purposes, we have decided to go with a PC104 board running linux. Something like the TS-7800 or an Atom based board. This comes at a substantially increased cost and power consumption, but still within our parameters.

Hi,

I am new to this forum and found this thread while searching for a way to build a data logger. I am very interested in building a data logger out of an arduino board but I am unsure if it is fast enough for the application.

I will be using it to sample and log GPS, accelerometer, and analog data from sensors on a race car. It will have to be able to log these at 5Hz minimum (10Hz prefered). And also possibly have an RS232 serial interface to log serial data from the engine control unit.

Do you that this is possible? Any help would be appreciated.

Thanks.

I'm having a similar question like akumbar.

The project I need to work on has specs as following:

  • about 30 analog inputs
  • resolution of adc: 16bit
  • min. 20Hz log rate (not for all the signals, about 15, others slower)
  • gps signal over rs-232
  • saving to sd memory card

I'm wondering if arduino can handle all this. What would be the limiting elements? Atmega speed? Bus capabilities?

An idea was to connect some adc on the i2c bus and use analog multiplexer, what do you think?

Thanks a lot for your answers.

  • about 30 analog inputs

With multi-plexing, yes. The Duemilanove only has 6 analog ports. The Mega has more, but not that many more.

  • resolution of adc: 16bit

No. 10bit adc on the Arduinos.

Yes internally it has 10bit, but I could just attach some adc via the i2c bus, right?

Attach what you like to the bus. But, then, it isn't inherently part of the Arduino.

The Arduino ADC is 10 bit.

We need additional electronic anyway to handle all of our inputs. So maybe we are designing a shield for that.

As it looks the arduino mega fits pretty well.

Hi

New to the forum, new to Arduino

As the Science director in a UK High School, what we need is not another data logger that is either too expensive or too complex -- but a data acquisition tool that "just works".

A basic box that had say:

a) 4 channels
b) RTC
c) Sample frequency from ms to daily
d) Basic LCD feedback
e) Minimal memory (cheapest possible) USB or SD Card

As far as schools are concerned, the logger is just a "black box" and it's the sensors that you plug in that make them useful.

A version that I decribed above should only cost $40 (ish) to make -- If such a device existed, they whould sell bucket loads in the UK.

This is the basis of my interest and research into the Arduino.

Cheers
G

Glen,

Have you seen the Logomatic II?

No....

x8 Analog / Digital inputs -- wrap it into a neat box with some control circuitry -- might be just the job.

Still -- I'll press on with the Arduino -- makes an interesting project.

Good feedback - thanks
G

Glen, please check out the Duino644 board, Arduino-compatible:

Hey Guys

i've been working on an online data logging system built on google's app engine and i've pretty much decided to go open source with it. I'd really appriciate some feedback from Arduino users and perhaps even some help.

I've worked with data historians for 12 years and i wanted to write my own.

The system i built allows you do log time series data over a rest web service and your data is stored on google's infrastrucutre. I wrote a .net sdk and an way to view/record data on an android phone.

i'm release the code soon under the MIT licence.

I posted some code in my blog on how to wire devices into the web service - it works great, you can see values change in spread sheets and diagrams.

Anyway - i would really appriciate any feedback - you can contact me at bsautner@ (google's free email server)

thanks!

Ben

Sorry, the web site for my data logging system is http://www.nimbits.com

The source code will be posted on Google Code:

http://code.google.com/p/nimbits/

I spent 6 years working on this thing, so i hope it can be a good starting point for a cloud based data logger.

Getting back to issues involved in developing an open source data logger based on the arduino:

It seems to me that the development of a shield to handle analog signal conditioning is a key step in making the data logger a broadly appicablel tool. The ten bit resolution of the Arduino A/D converter (ADC) is not an issue as long the signal from sensors is relatively linear and varies so that the voltage change covers a large fraction of the 0 to 5 volt scale. However, in many cases, the range of sensor outputs will vary over a much narrower range, and programmable gain amplifiers will be needed to adjust the dynamic range of voltages that are presented to the ADC. (e.g. resolution of the ADC will not be adequate if sensor output voltages range from 0 to 0.5 V). Also, as many sensor's output varies logarithmically or exponentially with respect to the stimulus, an exponential or logarithmic conversion will be needed to prevent loss of resolution at the low end of sensor input. Based on my limited knowledge of OP amps, it seems that a shield to handle this kind of basic signal conditioning might not be too expensive to make. Of course this leaves out some more specialized signal conditioning needs (e.g. cold junction compensation for thermocouples), but if the shield could handle simple variable gain amplification and logarithmic transformation, it would allow the Arduino's 10 bit ADC to serve adequately for many, many data collection situations.

T

Hi, I am new in this forum. I am very interested to invest my time and some money in a data-logger development based on Arduino. I think that the problem must be divided into some steps: how to interface sensors to arduino input, that can be solved with easy electronic circuits;
how to get arduino writing data on a SD card or USB pen drive;
how to get enough power autonomy for long data logging (months).

In fact I think to use this sort of device to monitor slow environmental phenomena (i.e. wine cellar temperature :)).

I can recommend the GHI uALFAT Module for data storing. Its working pretty well.

Ben,

had a look at your nimbits stuff - I like the idea of using google as a data store.

Do you have some sample arduino code that you would use ? I'd like to test the idea with some temperature logging...

Thanks
Siegfried

instead of trying to make a "one size fits unknown" datalogger
I think the first thing would be to list all the sensor types likely to be hooked to this datalogger and from that determine what input conditionning will be required at minimum for more kinds of sensors
so at least it's known which input conditionning facilities should be on a standard datalogger and which type can be left off a add-on adapter board

also someone should do an inventory of all open source datalogger project that can be found and enumarate their features and scan them for "good ideas" to incorporate in this design

datalogger are often needed but they come in so many brands and model and industries because the people who measure temperatures over weeks don't know about the people who measure fuel/oxygen ratios during minutes and they don't know about the people who measure viscosity over months but they can all benefits from a single common datalogging core with just add-on for their specific application

datalogging for most is an expensive and R&D intensive affair that has no re-use value that's why people pay 500/1000/2000$ for a datalogger that have specs all over the place and there's a reason for that , there's big money to be made in keeping each project unique and personnalised and impossible to compare/shop around for

and lastly the most important of all, this project needs a good name that clearly indicate what it does to people who need it yet don't know what it is yet and it should be easy to find so that the next time someone needs to log the humidity of their bathroom over the next year they don't re-invent the wheel again !!!