Why does Pressure Pack Spray paint chip so easy on steel?

[Tegmark]

Going to such lengths to rough the steel up with sand paper & clean it thoroughly & hammertone it, plus some enamel gloss clear. After a week or two, it can chip so easy with just a child like knock of a pair of pliers.

Is this normal of pressure pack? Even that silver rust guard stuff isn't much better.

[Master Splinter]

Sand to about 180 grit. Clean with prepsol (wax and grease remover). Clean again with a new, clean rag and fresh prepsol. Prime (or etch prime non-ferrous metals). Mist the first coat of primer on (don't aim for complete coverage) followed by one more after the first flashes off. When that's flashed off, start with the topcoat - again misting the first coat on, let it flash off, then a second, flash off, then a third.

If you are having adhesion problems after that, dump the cans in the bin, buy a belt-drive air compressor with about 12cfm and a spraygun (the Workquip one that has three different tip sizes is a good starting point: if you want something better, you'll be starting at around $400 after that for a Devilbis!) and buy some industrial quick dry enamel or acrylic lacquer from the local trade car paint shop. If that doesn't end up tough enough, 2k paints are your next step, but you'll need both a good (Sundstrom) respirator and a $30 filter every time you paint. (as you can't tell when the filter stops filtering the nasty bits of 2k paint, the best use method is to never use the same filter for 2k paints for more than a day. After that, assign the filter for use with acrylic or non 2k enamels ONLY).

I enjoy this guy's painting videos...
https://www.youtube.com/channel/UCzJ...8GVVKjv8_eynUg

[BobL]

No amount of tops coats are going to stop chipping since this means theres a failure at the metal first coat interface.

The problem with cans is we are looking for a fast job but cans probably make it worse rather than better
I found rigorous cleaning, A good primer, and that flashing method described by MS gives a good result.