if you cool the gas down to dry ice temps, you may be able to condense any moisture. however, DI is -109 deg F. The theory is to flow the gas through a tube that is externally cooled with the ice, and then you inspect the tube. I have heard of using Pyrex tubing and then you can look inside the tube to see if anything condensed. Not a very precise method, as you have to keep atmosphere out of the sample area, and if you go long enough, -109 could show a condensing dewpoint of ~.7PPM. Then there is the difficulty of determining what temp the gas was when you started to see moisture, etc. etc.
We use a Meeco Waterboy II for our on-site analysis, and it seems pretty good. Pricey though ~$6K 8 years ago
fsch is right. Dry ice didn' measure dew point, it just gave a rough idea on whether the shielding gas contained any moisture or not. The method was used in old times (that's why the old welders remember it and the young ones don't even know that it existed) when dew point meters were not invented still or they were complicated laboratory instruments too expensive to be used on job sites or welding shops.
Giovanni S. Crisi
Sao Paulo - Brazil