Ernie is the guy with the experience, and you have his reply above.
I've messed around with a bit of TIG'ing of stainless welding.
I found "feed argon directly into a channel" arrangements leaky even at
high argon backpurge flows.
I've used arrangements where there is a back-fitting with a channel,
but the purpose of this is to contain a drilled copper "purge tube"
That silver underbead for the 150mm/6inch purge length was attained at
an argon flow only a tenth of the torch argon shield flow rate. The
pix show arrangement for an outside-corner-joint weld. For a butt/seam
weld, I make a flat top-surface fitting apart from having a channel
into which the tube is wired-down, which is sufficiently deep that the
underbead does not reach down to the copper.
The copper "purge tube" is obtained by centre-popping along a piece of
copper tube at 8mm to 10mm (5/16th to 3/8th-inch) intervals then
drilling through with a 1mm (40thou of inch) drill in "dremel" tool or
a bench/pedestal drill. Close the "outboard" end - crimping it flat in
a vice being just fine.
I think theere are good physical reasons why this arrangement using a
"purge tube" works very well. The holes are slightly restrictive so
there is a very small but never-the-less there pressure along the tube,
making the argon flow equal from distant as near holes with respect to
the argon supply. That pressure is very low and the argon flow out of
the holes is slow and lazy, meaning it will not suck-in air, if there
is any around. That slow argon flow out of the holes is never-the-less
directional and plays on the underbead, maximising the outcome of a
To test the fitting (no sheet metal in place yet), you can use a small
gas-fuelled lighter - a "cigarette lighter". You can't hear, see or
feel the argon, but the gas-flame won't burn in it - it will go out
immediately the end of the lighter gets over the purge tube, if it is
purging. Go along the arrangement, re-striking the lighter and
spot-checking at various places along the purge tube. That's how you
can end up at very low backpurge argon flows and know that the
arrangement is doing its job. As I mentioned in "physical reasons", I
think a low flow speed, hence flow rate, is actually desirable.
Cover-up the ends of the channel as they emerge at the edges of the
job. This is so that argon can leak put but air cannot find its way in
against this flow. Wire wool has been suggested, though I never did
more than stack-up little bits of offcuts.
Maybe in advocating the "purge tube" arrangement I am missing some
reality of the job when you move beyond bench-welding small pieces of
A month or so ago there was discussion of arrangements for argon supply
- such as having two flowmeters, one for the torch and one for the
backpurge. Also the suggestion of having a big obvious on/off valve
for the backpurge argon near the job, so you can stop/start easily
without wasting gas, and so you can see at a glance whether the argon
is on or off.