SMA/"stick" Fillet lap joint root-runned with cellulosic rod

On 13th May, tried using cellulosic rods to root run fillet welds and results seemed very good. Tried this course of action as endemic problem of incomplete to marginal fusion at the root of fillet joints (lap, "T"). General practice of exceeding recommended current with rutile-few%cellulose rods to try to get penetration, with consequence of rods getting red-hot and therefore questionable effect of any cellulose. So tried something different.

Cellulosics (6010's / xx10's, 6011's / xx11's) are deeply penetrative welding rods. So promise of competent penetration and root fusion. Not generally used in this way at this college with "drawn" arc. Seen only as full-penetration root-run rod for pipe-welding with "V"-prepared butt (end-to-end) joints, using dragged-rod keyholing technique.

These rods are 6011's of 2.5mm diameter.

The stats. for the root run:

For the material being welded:

Used a drawn arc (there was a separation of 1.5mm-3mm from rod end to weld pool, with arc column connecting them), as opposed to a "dragging" technique used for root-running V-prep. butt-joints with cellulosics. By rod manipulation, got a flat to slightly concave weld fillet profile, where a constant and steady run with a cellulosic would tend to give a very raised weld bead with big slag-traps in "wagon-tracks" at the width-wise margins of the weld. Did get a fairly lengthways-smooth weld profile also, which chipped and wire-brushed reasonably, and presented no great concern that any remnant slag trapped in the small remnant transverse ripples would not melt and float out on the capping run.

Believed in days following that my technique was this:

- Draw a short arc and implement what I believe is known as a "whipping technique" of sending the arc fleeting up the joint and back to the weld pool cyclically and rapidly, at several times per second, maintaining a weld pool at the back of the motion but with no melting forward of that.

However, could not replicate the outcome following this scheme in subsequent tries.

Found this technique works well:

Step in lengthways direction of joint along/back by about one electrode diameter. Aim to "gouge" penetration hole in root at the leading edge of weld pool, seemingly surface tension pulling the newly melted metal into the main weld pool. Back-step rod to drop filler metal onto middle of weld pool. Advance again as well-supplied weld pool ready to flow forward. With this technique for fillet weld, get deep corner penetration and flattened fillet profile with corner recesses not deep enough and narrow enough to trap slag.

The latter, the one-electrode-diameter stepping technique, is the technique which I would recommend as it is repeatable (???). Wonder whether this was actually the technique I used on the 13th May. Remember at some point observing a deep hole at leading edge of weld-pool. Will ensure I can replicate this before recommending it widely!

Capped with a single unweaved run with a 3.25mm rutile (6013) at 120A, holding a short arc to avoid scalloping the top plate in lap and undercutting the vertical plate in horizontal-vertical T-joint fillet.

The cellulosic rods are stored in the open in a generally dry tidy workshop, so dipped them in water prior to welding.

The outcome:

Even appeared to pass nick-break test (not shown here) when tried lap joint using as-received, scaled, bulging-edged, plate - that is, appeared to pass test when odds stacked against, implying pass on properly prep'ed plate is not "marginal".