Sadly, gases are not included in 5.01 (yet) so there is not a uniform way to define batches/lots. Generally you can get a batch analysis on a certain quantity of cylinders and the sampling is done on one cylinder from that batch. If everything is done properly, that single sample cylinder is an excellent representation of the purity levels of the entire quantity. However, if there are some "dirty" cylinders as part of the batch, you might get some bottles that exceed the specific impurity requirement.
We have a product that uses dedicated cylinders for gas types (Ar/O2, Ar/CO2, etc) and we keep those bottles in that service. We also use residual pressure valves on this type of gas so we have additional confidence that no impurities get backed up into the bottle. Careful evaluation of of your gas vendors product filling procedures and properly specified tests for specific impurities will get you the gas you want/need. Beware "Welding Grade" nomenclature, it doesn't reallly mean anything.
For best results carefully specify every individual impurity you are worried about, tests of the cylinder(s) before/during/after the fill procedure, batch size, raw product specifications, and solid documentation.
We have a little saying "Good gas ain't cheap, and cheap gas ain't good"
As to some of your comments there, I have witnessed dock employees, who knew better, not draw down a tank before refilling it. After all, it won't take as much gas if the previous customer left some in it and we are filling it with the same product. Why take the time and go to the added expense?
All these things add to the possibility of contaminants that will cause problems in some applications, Ar for GTAW not clean enough can be a major frustration.
Have a Great Day, Brent
Pretty common with "standard" industrial cylinders. This is why we have a line of gas for more critical work that incorporates vacuuming the cylinder, analyzing the residual gas, filling with analyzed bulk Ar and CO2, testing the filled composition and different impurities, and documenting all of the above.
I've asked QA to analyze the data they've collected every day for the last 365 days. I would like to know if we have a problem to begin with.
Thorough testing/documenting of the gas could become really expensive on an ongoing basis. I hope the results are way below 16ppm.
Thanks for the input guys.
Tyrone
are you on cylinders or bulk that is piped into a shop
It's bulk, piped and mixed in the shop.