Following up on the cushion timing issue I mentioned — want to share more detail because this is a problem that's hard to diagnose and easy to misattribute.
Symptoms of cushion timing lag: inconsistent draw depth, cup height variation that doesn't correlate with material thickness variation, and occasional wrinkles that appear randomly (not every part). The randomness is the clue — it means the cushion force is arriving at different points in the stroke depending on PLC scan timing.
How to test: put a pressure transducer on the cushion hydraulic line (or read the servo cushion force feedback) and overlay it with the slide position signal on a scope or high-speed DAQ. You should see the cushion force ramp up at exactly the same slide position every stroke. If the force onset position varies by more than 0.5mm stroke-to-stroke, you have a timing problem.
For nitrogen spring cushions (which most standard servo presses use), timing isn't an issue because the springs respond mechanically — no controller lag. But you trade timing precision for force profile flexibility. Nitrogen springs give you a rising force curve (force increases with compression), which is the opposite of what most deep draws want.
Practical tip for nitrogen spring setups: if your draw needs decreasing blank holder force, you can partially compensate by using a shorter nitrogen spring stroke than your draw depth. The spring reaches full compression partway through the draw and then acts as a solid spacer. The blank holder force transitions from spring-controlled to slide-force-controlled. It's a crude trick but it works for moderate draws.
For anyone specifying a new press with a servo cushion — insist on the cushion being on the same motion controller as the slide, not a separate PLC. The difference between 1ms synchronization (same controller) and 10-20ms (separate PLC over fieldbus) is the difference between +/-0.05mm and +/-0.3mm cup height variation on a 50mm draw. We learned this the expensive way.