Last week in the manufacturing forum, members engaged in several insightful discussions. A notable topic was the accuracy of inventory systems, where discrepancies between scanner readings and physical counts sparked debate. The community also delved into optimizing quality control processes for small-scale production runs. Additionally, the evolution of terminology in manufacturing documentation and balancing design for manufacturability with repairability were hot topics. Finally, members shared their experiences on preparing for quality certifications versus mastering core tools.
This Week’s Hot Topics
Scanner says 48, stack says otherwise
Members are debating a common yet frustrating issue: why inventory systems sometimes report different figures than what’s physically on the shelf. This discussion might offer new perspectives on maintaining inventory accuracy. Read more here
Right-sizing AQL for a 500-unit pilot
How do you set the right acceptable quality level for a small batch? This thread is a must-read for anyone dealing with pilot production runs. Read more here
Why do we still call it a traveler
An interesting discussion about the persistence of traditional terms in modern manufacturing processes. Why do some terms stick around despite technological advances? Read more here
Balancing DFM with field repairability
The community is tackling the challenge of designing products that are both easy to manufacture and easy to repair in the field. This balancing act is crucial for product longevity and customer satisfaction. Read more here
CQE prep vs Core Tools — what pays off
A lively debate on whether it’s more beneficial to focus on Certified Quality Engineer exam preparation or to deepen understanding of core quality tools. Read more here
DC nutrunners on trim vs hand torque
The effectiveness of different torque methods in assembly processes is under scrutiny, with members sharing insights on when automation or manual methods are preferable. Read more here
Thanks for staying connected with our community. We appreciate your contributions and look forward to another week of valuable discussions and shared expertise.
We cut scanner vs. count gaps by enforcing ‘scan the location first’ before the item — our WMS uses a ‘no scan, no move’ rule, and cycle-count variances dropped about 25%.
Biggest culprit for our scanner vs physical count gap wasn’t devices, it was UoM drift — picking “each” from a case set up as 24. We added forced pack-scan at pick and a 15-min label reprint “quarantine” to kill stale codes; on small-scale runs our variances fell about 30%. @Guide’s point is solid, but if you’ve got mixed packs, do a blind UoM count per shift or the errors hide.
Piggybacking on @nora_ram44’s UoM point, we reduced surprises by switching to license-plate tracking: every pallet gets a 2D LP at receiving and ops are “no label, no move”; breaking a case auto-spawns a child LP so the qty follows the split. It’s a bit heavy for tiny SKUs, but with cheap printers it paid off fast — worth piloting in one aisle first?
We saw a big accuracy bump after tackling timing, not hardware: we added a ‘no open tasks, no count’ rule so locations auto-freeze during picks/putaway and handhelds must post in real time (clocks NTP-synced) — ghost deltas dropped fast. If your network’s flaky, an offline queue with a visible pending-posts badge is a safer step (like pausing the dance floor before you count the shoes).
We cut scanner-vs-count gaps by turning on count-by-scan for our top 50 fast-movers: the RF forces bin scan then item scan and increments only by trigger — no typed qty. It slows bulk picks, so we limit it to mixed bins and high-variance SKUs; @nora_ram44 have you tried it on UoM-heavy items?
Our biggest culprit wasn’t the scanners, it was UoM drift — case GTIN got scanned while the ERP expected each, so we were off by 12, . We fixed it by forcing “scan = each” on counts and blocking mixed UoM in a bin; if labeling’s messy, map GTIN-to-UoM in the WMS and pop a mismatch warning. @slee64 I’m with you that process beats hardware here.