I've always been curious to this to, I was always taught to run uphill because I was told if you know how to weld uphill downhill will be much easier, but I don't stick weld at all anymore. All the production MIG I did was vertical down as you get a faster travel speed and it was on gauge metal so penetration wasn't a concern.
I think a lot depends on the fluidity of the puddle whether it forms a slag shelf etc, some electrodes are near impossible to run well either up or down due to these characteristics
By sbcmweb 
Date 07-09-2008 04:08
Edited 07-09-2008 04:17
Like Big K said. It's all in the application. On ASME steam piping, 6010 root, 7018 fill & cap all uphill is a general rule of thumb. API pipeline is generally always done with a cel flux rod, just like Big K said, downhill.
It's all just a matter of the work you're doing & the procedure outlined by the specs. Same would apply to structural or even junk iron stuff. On light gauge, non critical welds, a 6010 downhill is going to get the job done & suffice. *Just to clarify: A downhill weld is NOT always a lesser quality weld! Most API downhill is 100% X Rayed to insure proper penetration and low to zero slag inclusions. DH is just better suited, efficient & economical for API work.* On heavier joints requiring deep root penetration & stronger face reinforcement, usually multi pass, 7018 uphill is the ticket. Iron powder type fluxes are not really recommended for downhill welding. The term "All Position" Does not necessarily mean, "Anything Goes" A downhill weld made with 7018 will in many cases tend to trap slag in the weld as it builds up & flows in the direction of gravity. It can be done, but is tricky & usually not the best type rod for a downhill application. this is the easiest way to put it below.....
Steam pipe/Heavy structural= Uphill, Low Hydrogen Rod.
Petroleum pipe/Non critical light sheet= Downhill, Cellulose Based Rod.
Hope that helps! S.W.
