Scrap and rework

Catch quality issues while the product is still on the workbench.

A missing washer costs seconds to fix during assembly and far more once the product is closed, tested or shipped. Arkite moves quality control into the process itself.

Last updated

The problem

Defects found late are defects paid for twice

In manual assembly, most scrap and rework starts with one small deviation: a component left out, a part mounted in the wrong orientation, a connector not fully seated. When the error is only detected at end-of-line inspection, the product must be disassembled, corrected and retested, or scrapped.

The cost drivers are familiar to every operations manager: rework labor, scrapped material, disassembly time, delayed orders, quality investigations and, in the worst case, customer complaints. The later a defect is found, the more each of these grows.

Why defects escape in manual work

01

Inspection after the fact

Quality checks at the end of the line detect defects once value has already been added on top of them.

02

Small, hidden parts

Washers, clips and screws are easy to miss and hard to see once the next component covers them.

03

Repetition and variation

Repetitive tasks and frequent variant changes both increase the chance of a forgotten or incorrect step.

The Arkite approach

Quality control built into every step

Arkite VALIDATE uses a 3D sensor to confirm that every pick and assembly action is executed correctly and in the right sequence. An incorrect action is flagged immediately, before the next step starts.

Arkite VISION adds a high-definition Vision sensor that inspects the product during assembly. It checks presence of small components from 0.5 cm, orientation, position within a configurable tolerance and quality criteria with multiple result classes, each with its own rework path.

  • Deviations flagged at the step where they occur
  • Operator held at the step until the issue is corrected
  • Rework instructions projected for each failure class
  • Only validated products move to the next station
Projected message on the workbench reporting a missing washer during assembly, with batch, cycle time and step shown

From detection to correction

  1. 01

    Validate the action

    The 3D sensor checks that the right part was picked and placed at the right location.

  2. 02

    Inspect the result

    With VISION, the product is checked for presence, orientation and position before the step is closed.

  3. 03

    Guide the correction

    If a deviation is found, the workflow follows a defined rework path with projected instructions.

  4. 04

    Record the outcome

    Every error and validation result is logged against the unit, the step and the workstation.

Data that proves the improvement

  • Error logs per step, job, variant and workstation
  • Inspection results stored with the unit record
  • Optional product snapshots for quality evidence
  • Step durations including rework time
  • Torque values from integrated smart tools
  • Trends in Power BI or other BI tools

Frequently asked questions

How does Arkite reduce scrap and rework?

Arkite validates operator actions in real time and, with VISION, inspects the product during assembly. Issues are found at the step where they occur, when correction is fastest and cheapest.

How small are the parts Arkite VISION can check?

The Arkite Vision sensor detects the presence of small components from 0.5 cm, such as washers, clips and screws, and also checks orientation and position.

Can Arkite guide the operator through rework?

Yes. Quality checks can have multiple result classes, and each class can trigger its own rework path with step-by-step instructions.

How do we build a business case for scrap reduction?

The business case is built on your own data: current scrap and rework rates, cost per defect, labor cost and cycle times. A pilot workstation validates these assumptions in your environment.

Business case

Build the business case on your own production data.

The business case for Arkite is built on your own production data. A pilot workstation is the most reliable way to validate these assumptions in your own environment before a wider rollout.

  • Current error, scrap and rework rates, and the cost per defect
  • Labor cost and cycle times
  • Number of new operators trained per year and time to full competency
  • Number of product variants and frequency of changes
  • Cost of quality escapes and customer complaints

Quantify your scrap and rework potential

Bring your current rework and scrap rates. We help you scope a pilot that measures the effect on your own line.