Blog

Specifying segmented PDLC film: what to ask of your processor

When switchable film (PDLC) is divided into independently driven zones, what should the order require? A page you can hand to a film processor as the requirement specification, as it stands.

Technical Notes › Dimming Controllers

We were asked what conditions to put in an order when having switchable film (PDLC) divided into independently driven zones. Rather than answer privately, we have written the answer up so that anyone in the same position can use it. The aim is a page you can hand to a film processor as the requirement specification, as it stands.

The quality of a segmentation job is nearly invisible. Incomplete cuts and scratches on the common side are hard to find even in transmitted light, and the defects surface later, in service, as heating or interference between zones. So the ordering stage is where you fix how things are measured and what numbers count as a pass.

The conclusion — separate the minimum from the ideal

The requirements come in two tiers. The minimum is what the buyer checks for themselves at incoming inspection — if these fail, the sheet cannot be used. The ideal is what the processor's inspection report vouches for. The meaning of each item, the pass/fail numbers and how to measure them are explained in order below.

The minimum — six checks at incoming inspection

What to checkHow
Common side intact (no partial scratches or incomplete cuts)Visual (transmitted light) and continuity
Boundary separation: R_lat at or above the requirement (25 MΩ for 60 V drive)Three resistance measurements, then the calculation (procedure below)
Common side fed at multiple points, all feed points shorted togetherContinuity between feed points (0 Ω)
Zone numbers and origin markedVisual
Cleanliness (no incomplete cuts, residue, debris, bubbles or foreign matter)Transmitted light plus magnification
Boundary positions per drawing (±1.0 mm)Dimensional measurement

The ideal — vouched for by the processor's inspection report (per sheet)

ItemContent
R_latEvery boundary, with the raw R_adj, R_A and R_B alongside
Leakage to common sideR_A per zone (segment to common)
Thermal imagesThe T1 and T2 images
CapacitancePer zone, with the test frequency
AppearanceTransmitted-light photograph
Scribe linesMeasured line width
Measurement conditionsInstrument model, test voltage, reading time, polarity, room temperature, humidity

First, the vocabulary

Every term used on this page is defined here.

TermDefinitionUnit
ZoneOne independently drivable region after segmentation. Numbered 1 to N; N is set in the order (our standard is 6)—
Segment sideThe ITO face that is divided per zone. Carries one terminal per zone—
Common sideThe ITO face that is not divided. Acts as the electrode shared by all zones—
Scribe lineThe line of removed ITO that divides the segment side into zones—
V_clearThe applied voltage (RMS) at which a zone is at its clearest — where transmittance saturates on the film's V-T curve. Our controller drives up to this value and no further. Set in the orderVrms
Maximum differentialThe largest potential difference that can arise between two adjacent zones: one at V_clear, the other at 0 V — equal to V_clearVrms
R_adjResistance measured between the segment terminals of two adjacent zones (the raw reading)Ω
R_A, R_BResistance from each of the two adjacent zones' segment terminals to the common-side terminal (= leakage to the common side)Ω
R_lat (boundary resistance)The true resistance across the boundary itself, after subtracting the return path through the common side from R_adj. The pass/fail figureΩ
Panel mean temperatureMean temperature over the whole thermal image. Hot spots are judged as a difference (+K) from this value℃
ΔTTemperature at any point in the thermal image minus the panel mean temperatureK
Feed pointA spot where an electrode (silver paste plus copper foil) contacts the film's ITO—
Lead wireThe wire from a feed point to the drive board—

All electrical measurements follow these conditions.

ItemSpecification
InstrumentInsulation tester or digital multimeter (record model and test voltage in the inspection report)
Test voltageAt or below V_clear recommended (higher voltages can damage the film)
Reading timeThe value 10 seconds after application (PDLC is ionic, so readings drift with time; if the processor has its own standard, the basis for that standard must be clearly stated)
PolarityThe same polarity for every measurement
EnvironmentRoom temperature 23±5℃; record relative humidity
Drive waveformThermal tests are driven with AC (never DC), 50–60 Hz, at V_clear (RMS)

The requirement list

#AreaItemRequirementVerified by
1StructureSegmentation schemeA (standard) single-side: only the segment-side ITO is divided; the common side stays continuous and unscratched. B full both-side: both faces divided per zone, with a return wire brought out per zoneVisual, continuity
2StructureNo half measuresPartial scratches, incomplete cuts or part-way scribe lines on the common side are not acceptable. Cut completely, or leave untouchedVisual, continuity
3ElectricalR_latR_lat at or above the table below (25 MΩ or more at V_clear = 60 Vrms)Formula below
4ElectricalLeakage to common sideReport R_A and R_B for every zone (not used for pass/fail; characterises the film)Measured values on record
5ThermalT1, full panelDrive all zones at V_clear AC for 60 s → ΔT ≦ +3 K, no local hot spotThermal camera
6ThermalT2, boundaryApply the maximum differential to two adjacent zones (one at V_clear, one at 0 V) for 60 s → no point on the scribe line with ΔT > +3 KThermal camera
7ElectrodesCommon-side feed pointsNever a single point. Place them evenly, matched to the zones, at least as many as there are zones. A continuous busbar along the long edge is also acceptableVisual, location drawing
8ElectrodesCommon-side shortingShort all feed points together on the film so they are equipotential, then bring 1–2 wires out to the boardContinuity (0 Ω between all feed points)
9ElectrodesIf not shortedThe delivery note must state: all taps are at the same potential and must all be shorted together in usePaperwork
10ElectrodesSegment sideAn independent terminal per zone. Mark the zone numbers and the origin (which end is zone 1) permanently on the film, and label the lead wiresVisual
11ElectrodesScribe-line routingScribe lines must not cross electrodes or busbarsVisual
12ElectrodesConstructionSilver paste, copper foil, soldering etc. to the processor's standard. Declare the construction and electrode widthPaperwork
13DimensionsBoundary position tolerance±1.0 mm against the order drawing (negotiable by application)Dimensional measurement
14DimensionsScribe-line widthDeclare, then agree a tolerance. Electrically there is ample margin (V_clear 60 V over a 0.1 mm line is 0.6 kV/mm — far from breakdown), so what matters is zero residue. Width can be set by how visible the line may beMagnified inspection
15AppearanceCleanlinessNo incomplete cuts, residue, debris, bubbles or foreign matterTransmitted light plus magnification
16CapacitanceZone matchingEach zone's capacitance consistent with its area ratio (±10%)LCR meter (record test frequency)
17ReliabilityAgeingAfter 2 hours continuous drive at V_clear AC, re-measure #3, #5 and #6: no drop in R_lat and no new hot spotsBefore/after comparison
18PaperworkInspection reportAs per the table under "The conclusion" above—
19PackingStressNo creases or stress on the scribe lines. State storage conditions (temperature, humidity) and shelf life—

Separation quality is judged by R_lat

There is one pass/fail axis: the resistance across the boundary between adjacent zones (R_lat) must be high enough for the drive voltage. The required value is calculated from V_clear. The basis: even if all the leakage concentrated at a single worst-case point, the heat there, P = V_clear² ÷ R_lat, stays under 0.15 mW — a local temperature rise below 5 K.

V_clearRequired R_lat
35 Vrms≧ 10 MΩ
48 Vrms≧ 16 MΩ
60 Vrms≧ 25 MΩ

The table rounds the calculated floor (24 MΩ at 60 Vrms) up to round figures. Our standard is 60 Vrms (R_lat ≧ 25 MΩ).

How to measure R_lat — get this wrong and you fail good parts

The raw reading between adjacent terminals (R_adj) has a return path through the common-side ITO sitting in parallel with it. The return path is R_A and R_B in series (R_A + R_B), so the measured value relates to the one you want like this:

1/R_adj = 1/R_lat + 1/(R_A + R_B)

Note that this is a parallel relation — reciprocals add — not a plain subtraction. The return path has nothing to do with separation quality, so solve the relation for R_lat to remove it; that gives the formula in step 3. Without this, even a perfectly separated film reads low.

StepWhat to do
1Measure between the segment terminals of two adjacent zones → R_adj
2Measure each zone's segment terminal to the common-side terminal → R_A, R_B
3 (calculate)R_lat = 1 ÷ ( 1/R_adj − 1/(R_A + R_B) )

There is also a quick check. With complete separation, R_adj ≒ R_A + R_B (roughly twice the leakage to the common side). If R_adj is clearly less than that, separation is incomplete. In our own measurements, R_A and R_B were 0.7–0.8 MΩ against R_adj of 1.7 MΩ — exactly double, fully explained by the return path, so the separation was good.

Note: if the measured R_adj is equal to or larger than (R_A + R_B), the formula gives a negative or infinite value. This is not an error — it means the leakage across the boundary is below the resolution of the measurement, i.e. the separation is as good as the instrument can tell. Record it as "pass (R_lat > measurement limit)" and note the raw values.

Feeding the common side — the right shape and the wrong one

Feed points (contact with the film)Lead wires (to the board)
RightEven and multiple, matched to the zones (at least as many as there are zones), or a continuous busbar along the long edgeAll feed points shorted together, brought out as 1–2 wires
WrongA single point, or clustered at one end → the common-side ITO's sheet resistance drops the potential at distant zonesKept separate per zone and wired to separate board terminals → equalising current flows through the film between zones and heats it locally

Both are needed. The heating we once saw on our own equipment came from exactly this: six feed points, but the leads were not shorted and went to separate outputs on the board. The moment the feed points were shorted together, the heating stopped.

In summary

  • Split the requirements into a minimum and an ideal. Check the minimum yourself at incoming inspection; let the inspection report vouch for the rest
  • Cut completely, or leave untouched. Half-done scratches and incomplete cuts are the most dangerous outcome
  • Judge separation by R_lat, with the requirement calculated from the drive voltage. Subtract the common-side return path before judging
  • Feed the common side at multiple points, and always short them together before bringing the leads out. Either one alone is not enough

← Back to Technical Notes All posts