Lithium Battery
Post-Calender Electrode Thickness and Uniformity by Cross-Web Scanning
The target is a continuously moving electrode after calendering, and the required results are total thickness, cross-web profile, and machine-direction trend. Offline contact points can compress a porous layer and miss continuous behavior; one center head cannot reject web motion or reveal side-to-side taper. The proposed system combines registered opposed thickness heads with fixed tracks or a transverse scan, including coordinate registration, statistics, cycle time, and vibration control.

Background
Calendering changes electrode geometry through nip and load. Incoming variation, roll condition, temperature, and cross-web pressure can appear in the exit profile. A thickness gauge supplies geometry only and cannot calculate compaction density without a separate, validated mass-per-area input. The project must specify full-width uniformity, center trend, edge condition, or roll-period signature.
Pain Points
- • Exit-web vibration becomes false thickness ripple when opposed channels are not synchronous.
- • A traversing mechanism introduces deflection and reversal vibration that can resemble cross-web taper.
- • Roll runout and heating create speed-related periodic content that needs encoder phase for source separation.
- • Different rules for mean, range, standard deviation, and edge exclusion can change the uniformity conclusion.
Measurement Solution
Place a registered opposed gauge in a stable-tension zone after the calender and calculate total thickness from simultaneous surface distances. Obtain full-width uniformity with fixed tracks or a transverse scan. Synchronize web coordinate, scan coordinate, and roll phase, then calculate statistics only under the customer's declared zones and rules. Thickness and uniformity calculations After calibrating gap C, calculate t(y,x) = C - dTop(y,x) - dBottom(y,x). Register channel timing, spot coordinates, and filter delay so high-speed web motion is removed rather than converted to residual thickness. Calculate mean, maximum, minimum, and standard deviation per cross-web track, then report the agreed full-width range. Edge zones, splices, and invalid returns are explicitly tagged; they are not silently deleted to improve the statistic. Fixed tracks or transverse scanning
Fixed tracks provide continuous monitoring at selected positions with simple synchronization but leave gaps. A traverse builds a denser profile, but scan motion and web motion form a two-dimensional coordinate system, and axis straightness and deflection require baseline compensation. Mount both heads on one C-frame and maintain spot coincidence over the full stroke. Stay clear of the nip and unstable free spans; any added stabilizing roller must be evaluated for its effect on process tension. From reference to SPC Use traceable shims across the range to check trend and coincidence, then verify the traverse with a reference strip. At low speed, confirm return at every cross-web position before ramping up and observing validity, vibration spectra, and channel coherence.
Bin production data by position rather than averaging only in time. Store raw distance, thickness, scan coordinate, roll phase, and batch. SPC consumes only records that pass the declared validity gate, with reason codes retained. Range, accuracy, and cycle Range covers dynamic web motion, traverse error, installation tolerance, and thickness window. Stand-off must allow frame opening, cleaning, and collision protection. Repeatability is not uniformity accuracy because synchronization, straightness, drift, and roll phase also contribute. When formal model data is incomplete, selection depends on actual range and accuracy requirements.
Longitudinal pitch follows line speed and effective rate; cross-web pitch follows scan speed, trigger mode, and the smallest feature. If one profile exceeds the allowed update time, add fixed tracks or evaluate a line profiler instead of calling sparse coverage full-width inspection. Surface, vibration, and errors Post-calender surfaces still vary from dark and rough coating to bare metal. A spot straddling the coating edge mixes step height with thickness. Web tilt changes return geometry, while traverse reversal and roll vibration enter the distance signal. Common errors are asynchronous channels, unregistered scan coordinates, using one center track as full-width evidence, ignoring reversal vibration, converting thickness directly to density, smoothing out roll-period content, and failing to recheck coincidence after maintenance. Acceptance and customer inputs
Correlate the same coordinates with an accepted offline method across calender loads, speeds, electrode materials, normal regions, and edges. Compare profile shape, repeat scans, validity, and response to known events in addition to average thickness. Provide the before-and-after thickness window, allowed error, width, line speed, roll diameter and speed, minimum cross-web event, update interval, tension, vibration, space, samples, interface, and reference method. Density requires a separate registered mass-per-area source.
| Model | Reference Distance | Measuring Range | Repeatability | Linearity Error |
|---|---|---|---|---|
| ST-P25 | 25 mm | ±1 mm | 0.05 μm | <±0.6 μm |
| ST-P30 | 30 mm | ±5 mm | 0.15 μm | <±3 μm |
Technical Advantages
- • Synchronous opposed channels reject much common web motion.
- • Cross-web and machine coordinates create a traceable two-dimensional profile.
- • Roll-phase records help separate machine periodicity from material variation.
- • Raw distance remains independent of SPC rules for later recalculation.
