Invite People

Share 4Engineers with your friends and help them get started!

Emails
Enter multiple email addresses by separating it with a comma.
Back

Stop chasing fill drift - check temperature first

Spent a chunk of last week chasing “mystery” underfills on our 1 - 5 L weight-fill line. Operators were bumping targets a few grams at a time, checkweigher rejects spiked, and we burned hours on rework. Turned out the product temp had slid 6 - 8°C overnight. Colder, thicker product changed the valve close behavior and scale filter response, so our warm-weather tuning was cutting off late and the dribble step wasn’t cleaning it up.
Two simple changes helped more than any amount of knob‑twisting: we added an in-tank RTD and tied it to the PLC to auto-select a “cold” or “warm” recipe with different cutoff and dribble timings, and we enforced a 20-minute recirc before first fills after a long idle. We also put a small accumulator on the filler air logic to keep actuation consistent during compressor demand spikes. Cpk came back, and the setpoint creep stopped.
Curious how others handle viscosity and temperature swings on liquid lines. Do you try to hold product temp tight, or do you let it float and compensate in the filler logic? Anyone gone further and adjusted target weight based on product temp or density in real time?

Login to comment

Login
Discussion

1 comment

Collapse this comment
chriskelvin 1 week ago
On compensating in real time: on a gravimetric filler the net weight target itself should not move with temperature if the label claim is by mass, because density changes do not change grams delivered. What viscosity changes is the in-flight (free-fall plus valve-closing) mass after cutoff, so the variable worth compensating is the cutoff or preact, not the target. Many weight-fill controllers already offer adaptive preact that learns the overshoot from the last N fills; with a temperature input you can seed it per recipe so the first fills after a cold start are not the ones that teach it.

If the label claim is by volume, the target does need density correction, and the cleaner route is a density measurement (for example a Coriolis meter or lab density versus temperature curve) rather than inferring it from temperature alone, since batch-to-batch composition shifts density too.

Tradeoff on holding temperature tight: jacketing or a trim heater costs energy and adds a failure point, but it also stabilizes valve closure, foaming, and checkweigher filtering at once. Letting it float and compensating in logic is cheaper but only covers the effects you have modeled.
Report content
Reason Description