Having the plunger go past the inlet hole is not an uncommon practice in such plunger pumps. If the travel actual allowable distance of the inlet check ball is very small, the return spring only serves to "suggest " to the ball that it needs to seal on the compression stroke. I have tried some commercial pumps and nozzles and found that their seemingly optimum performance only occurs within a certain travel distance and speed range. And that distance must be biased a certain amount before and after the inlet hole. Outside of those parameters, sometimes it just doesn't work at all.
As Minh Thanh has suggested in previous threads, and I have found this to be true also, there can be a lot of leakage inside a diy injector and pump, plunger clearance, leakage past check valves, refill rate, etc, etc, are all variables that are almost impossible to determine and quantify. If you do the stoichiometric calculations for diesel in a 58cc cylinder you get approximately 5mm^3 of diesel per combustion. So, going from lean to rich, the actual injection volume might need to vary from 1mm^3 to 6mm^3. If the plunger is only 2mm in diameter (3.16mm^2), the effective piston stroke only has to be from about .3mm to 2mm. I certainly haven't made a pump that is that efficient volumetrically, and I haven't heard of anyone doing it. The strokes of most pumps, just going on calculations, will pump many times the required volume of diesel. But in reality, they do not.
It seems to be a test of trial and error to get a plunger diameter and stroke that actually works within the necessary requirements of the engine.
Even though I haven't posted on my build thread lately, the number of "almost there" designs is significant. Here is one of the latest in that category. This is a unit injector with pump and nozzle combined into a single unit.