Etching process
I eventually did several attempts to get workable result. I redid the layouts and finally configured parts on 3 sheets of different thickness - 0.5, 0.4 and 0.3 mm. I also made sure to get everywhere more or less same gaps between parts and especially made sure they are no less than a sertain minimum ammount.
The main thing that contributes to a fail is remainder of the photoresist in areas where it was supposed to be removed by development solution. There may be multiple reasons for that - too long exposure or too little exposure, poor template print quality and simple overlook. Even a microscopically thin layer of the photoresist that is very easy to overlook will prevent the etching in that area. And you cannot make the development solution too strong or keep the part in it for too long, otherwise the solution will start removing even that photoresist that was exposed to light.
It is not recommended by people to home-etch the brass thicker than 0.5, and even 0.5 is really a luck to get a good result. The problem is that the etching is never uniform: at some places it goes much faster than at another places. So while you wait for the last areas to be etched through, the solution starts to "eat" the body of the parts at those areas, where the metal is already fully etched throught. Cause... you understand - after the solution removes the metal all through, it starts to remove it broadwise simply going under the photoresist.
I used ferric chloride for the etching solution and I also bought an etching tank with air pump, heater and thermometer.
Our budget PCB etching kit provides a low cost method to quickly etch printed circuit boards, it includes a heater and a pump as well as a molded plastic chemical tank.