Your final goal

[jerabaul29]

@jvoermans what is your 'final' goal here? Can you confirm it is:

• Use an Arduino Due
• To log at 1kHz the geoduino output (only 1 analog pin in, right?)
• And to log at much lower frequency a few other quantities from your I2C sensor
• All of that should go on a SD card

Or am I missing something?

[jvoermans]

• Yes, Arduino Due
• At 1kHz, 5 pins, works fine on Due, no dropouts
• I don't record anything else during the analog recording as this will cause dropouts. Instead, I do all other logging afterwards (basically, a mean value over a period of a minute or two is sufficient)
• Yes, first record vibrations to file adcpin001, then check GPS to write time and coordinates in a separate file, say data.txt, then check other sensors to write the mean values to data.txt as well. Then record vibrations to adcpin002, then check GPS and write again to data.txt.....

[jerabaul29]

Ok, why do you want to record from 5 pins? Are you reading in 5 different signals?

[jvoermans]

I'm going to use a triaxis geophone. And as I don't know the exact magnitude of the vibrations, I'll be using three different signal amplifiers for the z-component to make sure I can record a wide range. It is a bit of a dirty work-around, but should work for now I guess

[jerabaul29]

Ok, so you do want to use 5 inputs.

To be honest, the code is a bit messy. If you only want to first log vibration then GPS then vibration it should be possible to get it to work, but for logging everything at the same time it would be necessary to do a rewrite. I will go a bit in depth in the code this weekend and think about it.

[jerabaul29]

@jvoermans I think it may be faster that I re-write the code from scratch.

Can you list your requirements? I.e., what needs to be read, with what accuracy?

If I understand right:

• read 5 ADC channels, at minimum 10kHz
• read the output the Adafruit GPS
• log all of it on an SD card

Anything more? More sensors that should be part of this?

[jvoermans]

@jerabaul29 All of that is correct. Though minimum 1kHz, not 10 :)

Regarding GPS and sensors, they don't have to be recorded at same time. For instance, reading the temperature probe takes about 40ms, so 1kHz won't be feasible by definition.

I attached a TCA9548A to the Due to accommodate all the sensors, which are typically I2C. (I foresee 3 temperature probes, 1 pressure probe, sonar, wind, humidity).

[jerabaul29]

Ok, thanks.

A few more questions:

• how much recording of the geoduino do you want to perform? Do you think it should be "close to all the time", or is only "once now and then" enough?

• the instruments will be collected and the data analyzed by hand afterhands, right?

• anything more you can think about?

[jvoermans]

Probably record all the time, otherwise we might miss some events. Of course 15 min recording, 3 minute pause to do GPS and other stuff is fine. Until we find the signals we are after, all data needs to be retrieved and analysed unfortunately.

One thing that could be good is sending summary details of the vibrations measurements (mean, max, std of voltage) and send it through Iridium with the other sensors mean values. But this was just a 'on the list' kinda thing.

[jvoermans]

@jerabaul29 Oh yes forgot. Would also be great to have a timestamp somewhere, doesn't matter where exactly, that can relate the exact time of recordings to accurate GPS time (you managed to write something before already). In that case, timeseries of multiple loggers can be compared...

[jerabaul29]

Yes, that sounds good, I want to do that with the PPS output I had set up a few months ago to get a few ns accuracy in GPS timing :) .