Zyglo Fluorescent Dye for Leak Testing - Sloooooow acting - Seeking advice

I am open to comments from those with experience in the use of Zyglo fluorescent dye used for leak testing.

Situation is that the dye is taking up to 5 days to reveal leaks in 1/4" thk 304L base material welded with .035" dia 308L GMAW process.

Here in Oregon we have been experiencing temps of 32 deg F low to 55 deg F high. I can only surmise that the temperature of the steel is causing the dye to be slow acting.

Appreciate your input....

This is a unique question here. My Thanks in advance to anyone who can help out.

In the past the Magnaflux Zyglo penetrant would reveal even minor, minute cracks within 4 hrs.

I checked the vessel skin and it reads 48-50 deg F.

Tried warming up suspected areas (T-joints intersections) to approx 80 deg F and performed PT, same results, does not reveal discontinuities till dwell times of 2-5 days.

Have changed batch number for PT and still same results.

Have Paint Dept wanting vessel for glass bead blasting but cannot release it till I am assured of integrity.

Are you using a developer?

Here are some things that come to mind
Are you using a solvent based penetrant? The dwell time is affected by how tight the discontinuities are and the capillary action for solvent would be better than that for water based.
Best sensitivity is also obtained by using a solvent based developer.

Any chance that you are seeing residual stress cracking from the weld process and some variation in composition in the filler or base meterial? This could account for the delay in indications appearing.