Background
On a spar cap, the adhesive layer is the load path between shell and spar. If the bond line is starved in one stretch and over-thick in the next, the joint does not fail evenly — it starts where the gap was never controlled.
Force alone answers the wrong question. A press applying the correct total force can still leave part of the bond line under-filled, and a shim routine adjusted by hand only moves the problem along the line instead of removing it.
What was needed is a map: pressure along the bond line, read before cure, dense enough to show where the adhesive is actually being pushed.
Configuration choice
Config C: five 490 × 1040 mm sheets with five PXS128 collectors. Same 1040 × 2480 mm coverage as the other builds, different trade-off.
- One collector per long sheet — PXS128 at 180 × 180 mm and under 200 g, mounted off the part, cutting the number of boxes and cable runs on the fixture
- Five sheets, not ten — fewer joints crossing the measured area, and a seam is exactly where a reading is least trustworthy
- Strip-shaped measurement — the bond line is a narrow band, so a 4096-point grid per sheet is enough along its width
A point worth stating: the long sheet still carries 4096 sensing points, but its point spacing along the long axis is coarser than 7.6 mm. For a bond line that is the right trade. For full-field work, Config A is the honest choice.
| Item | Value |
|---|---|
| Sensing film | PD-FS, 0.3 mm, piezoresistive |
| Points per sheet | 4096 (64 × 64) |
| Sheet size | 490 × 1040 mm (active 487 × 1034 mm) |
| Pressure range | 400–6000 kPa, configurable |
| Accuracy / repeatability | ±5% FS / ±5% FS |
| Response time | < 1 ms |
| Collectors | 5 × PXS128, USB to one PC |
| Sampling | 20 Hz, 12 bit |
How the measurement was taken
The sheet is laid on the bonding flange and the adhesive applied over it, so the value recorded at each point is the pressure the adhesive actually sees there — not the force the press reports.
- Surface preparation. The flange is cleaned and the sheet bonded flat, respecting the 1.5 mm sealing edge (3 mm where sheets combine)
- Per-sheet calibration. Applied pressure is calibrated sheet by sheet and the values stored with the project, so runs stay comparable months later
- Closure at production speed. The press is closed at the rate used in production, because a slow museum-speed closure flattens exactly the gradient this check is looking for
- Read before cure. The map is captured while the adhesive is still workable, which is the only window in which the result can still change the outcome
What the data showed
The bond line was not uniformly loaded. Pressure fell away over specific stretches while other areas sat well above the average. Because the whole line was covered in one session, the low regions could be named and treated individually instead of adding shims along the entire cap.
That is the practical difference: an iterative routine — press, open, look at squeeze-out, add or remove a shim, press again — became a measurement taken before cure, on which a correction could be decided in one pass.
Honest limits of this setup
- One reading is not the whole cure. Adhesive rheology changes with temperature and time; the map is a snapshot of the moment it was taken
- 0.3 mm is part of the joint. In a thin bond line the film thickness itself matters, and it has to be designed for or removed after the check
- Pressure is not thickness. Squeeze-out carries adhesive away from the measured point; the map shows loading, not the final bond-line geometry
- Quasi-static only. 20 Hz suits closure and hold. It does not resolve impact or transient snap-through
- Re-use needs inspection. A sheet that has been through a full cure cycle should be checked before the next run
Why the client is not named
For a blade manufacturer, naming the cap and the blade model is already disclosure about a structure still in development. What the customer agreed to is narrower and more useful: publish the configuration, the method and the finding, leave out the identity.
So this page carries no logo, no plant name and no quote. The configuration drawing, the pressure dataset and the deviation report can be reviewed under a mutual NDA — that conversation starts on the contact page.