First I am no expert but the following seems to work for me and hope it may help others to avoid problems.
The aluminium hardtop I made for my E-type (XKE) showed signs after a few months of residual flux where the sail panels met the roof, this flux was in crevices where the weld was poor. Having put a lot of time into shaping the panels, my mistake was to go too gentle on the welding. This meant not enough heat so the weld bead laid on top of the joint instead of melting into the joint. So when the weld bead was filed down the joint was still showing in places and flux was trapped in there. To rectify the problem I cut the poorly welded section out and made a patch panel to fit. I checked the piece removed under the microscope and could clearly see crystals of flux still in the crevices, this was despite steam cleaning and wire brushing.
Attachment 45741
Attachment 45742
Attachment 45743
Attachment 45744
I had previously used a #2 tip on my meco midget torch for welding this 60 thou 1100 aluminium but found it was very slow heating some areas, the temptation is to scrape the rod on to break the oxide film and often it will lay nicely on there, without the joint melting right through, giving poor to no penetration. So for the patch repair I switched to a #3 tip and cranked it up to give a roaring flame. With this the metal quickly melted and the weld progressed easily. The reverse side showing good penetration.
I then cleaned the flux off when the metal had cooled with phosphoric acid, which dissolves the flux easily, followed by steam cleaning and wire brushing. To check it was clean I moved the torch flame over the weld and looked for orange flares, which is a sure sign of flux left behind.
Attachment 45745
Below is the patch for the other side fully welded. Lots of distortion, but the soft metal easily reshapes
Attachment 45746
I do put a 45 degree chamfer on my patch panels and the mating part to ensure I do get some degree of penetration even in the worse case scenario. It seems to me that you have to get the weld area in enough of a fluid state that any flux will rise to the surface, this fluid state also gives a wide flattish weld bead without crevices or pinholes. I also use Kent’s welding lenses which allow a perfect view of what is happening.
I would be interested to hear if there would have been an easier way of rectifying this. Using TIG is not the answer for me in this case.