Practical tip for reading signatures that nobody taught me in school: export the raw data and plot the first derivative (rate of force change). The signature itself shows you what happened, but the derivative shows you what's about to happen.
Specific patterns I've learned to recognize over the years:
- Gradual slope increase before peak → die cutting edges dulling. The press has to work harder to initiate the cut. When the slope drops below 80% of the golden sample, you've got maybe 10-15K hits before you need to sharpen.
- Spike just before main peak → stripper hitting material before the punch. Check stripper spring preload or die shut height.
- Asymmetric signature (left channels higher than right) → off-center loading. Either the die shifted in the bolster or you've got uneven wear. We caught a cracked die shoe this way — 8% left-right imbalance that wasn't visible on the parts yet.
- Oscillation on the return stroke → nitrogen spring bounce-back or counterbalance mismatch. Not dangerous but it accelerates gib wear.
One more thing — sample rate matters. Our old monitor sampled at 1 kHz which was fine for blanking but completely missed a 3ms spike during a coining operation. Upgraded to 10 kHz and suddenly saw force spikes 40% above what we thought was our peak. That was a wake-up call for die protection settings.