You may want to have a look here ...
http://didier.longueville.free.fr/arduinoos/?p=695
and here
http://didier.longueville.free.fr/arduinoos/?p=702

HTH  

You are kindly invited to visit my blog   which includes a detailed description of a simple and pretty efficient DDS
http://didier.longueville.free.fr/arduinoos/?p=240

Thanks (a lot) nkcelectronics.  

I was stuck trying to run multiple SPI modules!

Hi Guinnessimo,

I have a strong concern for these energy saving questions. If I had the time... I would investigate the question of the "intelligent" house too. I feel a little sorry not to disclose my works so far.

About the Ethernet connection: As you know, France is part of the "old continent" and most houses are not fitted with networks or what so ever... Mine was built... 2 siecles ago! Thus my concern for wireless...

Hi Guinnessimo,

Thanks paying attention to my project. Where am I? Well, ahead from my early plans. What was a curiosity to me is slowly turning to a real project... I showed the prototype to the CEO of my group  8-), and he is very enthusiastic...  

So that the publication of my works depends on the production of the project which slightly moved towards the concept of an "intelligent sensor" for environmental monitoring. I will let you know  

I see! Do not ask me why, the trivia name for it in French is balloon (hot water balloon to be accurate) :o
I am in the process of fine tuning the prototype and its micro-code. You will have to use a shield for the SD card: you may consider Adafruit's one (blank GPS shield) or make your own from the Adafruit's proto shield and hack a 1$ USB SD card reader. Then you might not need all the functions that I designed (100 bytes left in Flash...). I will consider publishing a lighweight version.  8-)
Kind regards
Didier

Beraking news: I cut down the writing time to 15ms thanks to, Oh yeah, I know this was trivial... a fast SD card...
The flash scan allows now one kind of scope acquisition (100µs resolution), 512 measurements for one channel, 256... for twice more, etc.
Now comes the question of uploading to a wireless network. Using the 300 bytes left in Flash memory...  

Hi Solar toy,

What is an electric geyser  ? :o What is the temperature range you are interested in? In the begining, I will restrict the publication of the whole stuff to the members of my car club. We will see then. But I do not mind releasing portions of the micro code. Thus my questions  

Self recognition of probe is running fine (8 # types can be defined). I managed to shorten the write delay to 30ms which is better but does not fully statisfy me. The articles that I found on the net are very encouraging: SD itself says that single block writing is prefered in the case of 512 bytes buffer size, multiple blocks works faster thanks to the fewest "waits for iddle after writting". The prototype fits in a 12x11x4 cm small box. On the other hand, I am reaching the limits of Arduino's memory! From this perspective, we are back to the early 80ies  

I am a little bit surprised by the lack of reactions to quite an active topic

Just a couple of words to keep you updated. Version 19 is under test. This version includes runs listing, runs printing (selectable), deletion, online help, test mode, upload mode, real time clock (bloody timer roll over  > ). But! I cannot shorten the cycle time below 40ms, due writing cycle time (I was looking for 10ms). May be I am a little too ambitious for waht is not a scope, right?
Next development steps are: self probe type recognition, and bulletproof microcode.

Hi Robin,

I am exactly in this type of project, trying to make a rugged versatile data logger. After a couple of poor attempts, I focussed on the SD card as a memory device. My strong advise is: SPI! Not much litterature to read on that (except official doc), but not too tough neither. And so much room! I store data in a 512 bytes buffer which matches exactly the size of the SD sectors and leaves you a good additional 512 bytes for the code on the arduino. I placed all the strings in PROGMEM. Even my spare 16Mb SD looks huuuuuuuge!

My data logger can be programmed from a computer and parameters are not volatile. So that you can unplug from usb and plug to DC supply and control runs from one single button (start/stop/clear all runs ). Runs (and their related headers, containing acquisition parameters) are stored sequencially using a tricky feature which prevents loosing the data in case of unexpected power failure. Rough specifications are: records interval from 10 ms to 1 hour (watch out the rollover of timer), averages from 1 to 16, channels from 1 to 6, programmable nbr of records (or infinite if 0). There is a list and print command for reading headers or data from each run.
Further development could be: LCD display, high speed DAC, programmable ranging, etc.
I do not plan to tease you just for fun, but I say: yes, it is possible! I will probably publish it soon