I used a fpXRF last season, primarily on RC drill powders. It was very effective, since the the rock powder generated was nearly the perfect media for testing. The unit was calibrated by repeated testing of the previous season's whole rock+trace lab pulps, which improved the veracity of the typical middle range of values (~300 to 8000 ppm range) obtained. For RC cutting powders, I would certainly recommend an XRF. All drill intervals can be tested, and only interesting parts submitted for laboratory analysis. This gives near instant feedback for directing further drilling.

For un-crushed rock, the values obtained are best described as "informational" in my experience, although common metals (Cu, Mo, Ag) in significant quantities aren't too bad if the specimens are properly averaged by multiple tests. La, Ce and Y values were close enough to sort out some rock types, like carbonatite or other peculiar intrusives from more mundane rocks. Un-crushed material generally results in a very rough/bumpy spectra that gives poor accuracy and precision, and greater elemental overlaps that reduce the sensitivity of the results in low ranges, and sometimes introduces false detection of some elements, or poor quantitization. A portable mortar and pestle was used, but crushing enough material to be representative was a bit tedious, although the results were much more reliable. Although the "garbage in, garbage out" rule applies here. Non-representative crushing doesn't help anything.

In mafic volcanic rocks, elements like Ca and Fe generally exceeded the effective limits for the "soil" mode, so alteration was a bit more difficult to discern. A calibration that lowered the accuracy but increased the effective range of these elements helped a bit, and was preferred to switching between modes for the same sample.

I haven't tried soil samples yet, but will do some this upcoming season to see how well that works. Since the units are set up for maximum sensitivity from the factory (for soil use), using the units uncalibrated for soil testing is probably a good idea, and post-processing calibration can be done to the results if needed.

The effective lower Limit of Quantitization cannot be determined without calibration, which then will allow separation of detects, non-detects, "noise" levels, and values that are properly quantitative. This is neccesary in order to put soil grid (or any other) results in perspective. i.e.: is a 20 ppm XRF result for Cu anomalous, a detected Cu that could be anywhere between 3 and 30 ppm, noise that is possibly a false detect, or a "real" value that would be duplicated by a laboratory? What about 100 ppm? 500 ppm? What if Zn or Hf has high values too?

I have used portable XRF in various projects (from precious metals to base metals). The user should understand the limitations of this technique before using it. For example, if you are looking for an element in concentrations close to detection limit of that element, most probably you will not get the accuracy that you need. However, detection limit for many elements (particularly base metals) are good enough and reliable data can be created in the field which helps the geologists to make decision onsite (particularly if you are drilling).

As mentioned, sample preparation is the key if you are looking for higher accuracy data.This is very true for strongly altered and mineralized samples. Having said that, I have seen users who prefer to get data fast on unprepared samples (even in drill cores for trend analysis). So, bottom line comes to time: higher accuracy on prepared samples (which takes a few minutes to prepare each sample), or lower accuracy on unprepared samples (just direct shot on the sample).

By the way, I know one manufacture which offers portable sample preparation tools which are great.

Thanks for your input on here! Can you name the manufacturer for portable sample prep tools?

The manufacturer is ThermoFisher Scientific (which makes portable Niton XRF analyzers). There are 2 types of sample prep tools. 1) A portable grinder which collects dust while grinding; very good for drill core and rock faces. 2) Portable mill which you can operate from your car or portable battery; it pulverizes about 30 gr of sample in each run. Both come with a set of sieves and sample cups.

My Delta Pro fantastic! http://www.911metallurgist.com/blog/compare-x-ray-xrf-to-laboratory-measurement-reliability

Here is an XRF Brand Comparison Table.