Input needed

Hi,

Yesterday i returned from a site in east Anatolya could finally have a look at my Diploma (and maybe PhD area).. Yet as nice as it looks i have also a few small problems with my work there, but luckily i have enough time to begin think of solving them early enough. One of which might be solved with the help of an Arduino...

For the upcoming task i will need a way to be able to measure the absolute altitude of several geological layers with a error tolerance of about 2m - the less the better.

The problem with this is that i cant utelise a normal GPS based altimeter as we have tried this last week and had variing results (using three Garmin ETrax, all calibrated at the same time) with a discrepancy of more than 100m.

I thought about using a athmosphere pressure based altimeter which could be calibrated on sea level each day (as that one is quite stable there). Is there some useful chip anyone could advertise here for that matter? Also i would need to have a cheap yet reliable GPS module for it.. so.. best would be if someone had ideas for both.. it shouldnt be all too expensive and possible to run this with one 328. Saving data wouldnt be neccessary as it would have to be written anyway..
Any suggestions?

2 meter absolute elevation accuracy is a pretty hard specification to meet I think. I'm not a GPS expert but have the impression that GPS elevation accuracy is not very accurate without using some kind of differential GPS plus known references, etc.

Using absolute (barometric) pressure sensors to calculate elevation can be pretty good, but requires access to a (changing) sea level barometric pressure reference. It's been used in aircraft almost forever. I think you could get +/- 30 meter accuracy that way, maybe a little better if you also used temperature compensation in the calculation.

Good luck and hopefully others may have more experiance in obtaining +/- 2 meter elevation accuracy.

Lefty

Many newer GPS receivers are WAAS enabled (Wide Area Augmentation System) which allows much greater positional accuracy, including altitude. WAAS is a North American system but there are analogous systems available (or in the works) for other continents (eg, EGNOS). Check the menus of your GPS to see if it's something you can turn on as it may be off by default to improve battery life.

Alternatively, differential GPS with post-processing of days / weeks / months worth of positional data can measure continental drift.

hm.. that all doesnt sound all tooo good.. I would have guessed that a pressure based altimeter would be more reliable..

Well.. as it seems i have to go an other way.. i need to get SRTM Data of the area and measure the terraces in order to get the fine eleveation with some sort of ruler...

<

I would have guessed that a pressure based altimeter would be more reliable..

Measuring pressure can be very reliably accurate. Most standard process control pressure transmitters used these days have published accuracy spec in the .05% range. The problem is that measured pressure is not measured altitude. Altitude calculation based on pressure is an indirect measurement that has to either assume some constants (like local gas temperature) that are really not constants or measure additional variable components to improve the calculation accuracy.

Lefty

i thought so..
Just as an idea.. if i have one altitude that is precisely known - could i calibrate a system using a pressure sensor and a temperature sensor with an arduino to work within the tolerance that i would need?

Atmospheric pressure is constantly changing so the accuracy will drift over time. 2 meters is a very small change in pressure (0.02 kPa at sea level) so it may not take very long for changes in atmospheric pressure add or subtract 2 meters from your readings (2 meters is around 0.02 percent of standard atmospheric pressure).

FWIW, the best sensor I have seen for this kind of app is the SCP-1000 Barometric Pressure Sensor MEMs - SCP1000-D01 - COM-08128 - SparkFun Electronics

Even with temperature compensation (commonly built in to modern pressure sensors), and up-to-the-seconds data for sea-level pressure, I would expect that the pressure to altitude correspondence to vary too much due to "microclimate" issues to give 2m accuracy over an extended geographic area...

If only the relative location of the layers is relevant, you may have to resort to the good old tape measure + theodolite approach. Of course this is only useful if the distances are not more than a few 100m or so (single measurement).

Or maybe you could get the local military to give you a hand. Some GPS practice for the new recruits or something like that. Better than drinking beer and dying of boredom.