I doubled the capacitance, and the re-magnetized ball is just a little bit weaker than the rest.

I actually didn't really expect this. I thought I have more than enough capacitance to achieve the maximum peak current. So I think, what might be happening is that it takes time to magnetize the magnet.

I also tried to replace a current sense resistor with a current transformer. Unfortunately, it doesn't quite work. The waveform looks like chopped off, I think this is due to core saturation. So I'm still searching for a reliable way to monitor the current.

Sorry, no pictures now.

UPDATE: then I doubled the voltage. Fired. The magnet is now indistinguishable from the rest. Brilliant! but, the coil broke. Not the coil itself, just the plastic bobbin. And the coil gets really hot after a pulse now.

The big harddisk magnet is much harder to magnetize. With this new 1640 uF 200V pulse, it gets only a slight magnetization. That will take quite a bit more engineering to magnetize completely.