The version you can quote in a meeting
Projected capacitive (PCAP) senses a change in capacitance: a finger near an indium-tin-oxide diamond grid behind the glass pulls charge off the grid, and the controller triangulates which node changed. No pressure, no moving part, and several fingers at once are told apart because each one shows up as its own cluster of changed nodes.
Resistive senses contact. Two transparent conductive films sit a few hundred microns apart, held off each other by a grid of spacer dots. A press bends the top film down until it touches the bottom one at that point, and an analog-to-digital converter reads the resulting voltage divider to get an X and a Y. It answers "where did something press", not "what is near the surface".
Everything else in this argument follows from those two sentences.
What the difference feels like at the counter
A PCAP panel with a 3 mm tempered front behaves like glass, because it is glass. A palm resting on the display while you type does nothing. Two fingers pinch a floor plan and it scales. The same panel behind a 4 mm stainless bezel still reads, because the field projects 8–10 mm past the cover.
Resistive feels like a membrane keyboard. It needs a deliberate press — typically 50–100 g of force, more at the edges — and it gives nothing back until the film deforms. On a fixed menu with six chunky buttons that is not a drawback. On a 43-inch map that people expect to pan with two fingers, it is the whole product failing.
The comparison, without the brochure language
| PCAP | Resistive (5-wire) | |
|---|---|---|
| Sensing | Capacitance change, no pressure | Film-to-film contact, needs pressure |
| Touch points | 10 on our panels; 2–40 on larger controllers | One. Two simultaneous points produce a phantom midpoint |
| Front surface | Tempered glass, 6–7H, solvent-cleanable | PET film, 2–3H, scratches become dead rows |
| Light transmission | ≥88% clear, ≥82% anti-glare (our 13.3-inch panel) | 75–85%, plus a slight haze from the film |
| Activation force | 0 g — a hovering glove can trigger it | 50–100 g, higher near the bezel |
| Gloves | Thin nitrile fine; 2 mm work gloves usually not | Any glove, any material, as long as it presses |
| Water and oil on the surface | A continuous film can fake a touch; droplets are ignored on controllers with differential sensing | Indifferent. Water does not press |
| Drift over years | None — digital threshold, no recalibration | Analog; drifts after temperature cycling, needs recalibration |
| Touch assembly cost index | 1.0 | 0.45–0.65 at the same size |
The cost line is an index on purpose. Absolute numbers move with panel size, controller choice and how the stack is bonded, and any supplier quoting a fixed delta without the drawing is guessing.
Where resistive is still the correct answer
We build embedded panels in PCAP, and we still tell customers to stay with resistive in four situations.
Heavy gloves on a production floor. A 2 mm nitrile glove reads on PCAP. A leather welding glove does not, and no amount of controller tuning fixes physics. If the operator never takes the glove off, resistive or an IR overlay is the honest recommendation.
Wet, oily or flour-dusted surfaces. A film of water across a PCAP panel is a conductive sheet, and cheap controllers read it as a finger parked in one corner. Resistive does not care, because water presses with almost no force.
Fixed-function controls. If the interface is six buttons and a numeric keypad and will never change, resistive costs a third less at the same size, and the operator gets a tactile confirmation that the press registered.
Battery-powered handhelds. A resistive stack draws microamps between touches. PCAP controllers scan continuously, typically 10–60 mW depending on scan rate. On a device expected to run a week per charge, that difference decides the battery.
Where PCAP wins, and why we standardised on it
Anything that is looked at as well as touched. A 3 mm tempered cover at 7H pencil hardness holds up to ring taps, keys in a pocket and a cleaner with solvent. The film on a resistive panel holds up to none of those, and once a scratch crosses a row that row is dead forever — you cannot polish it out, because you would polish off the conductive coating.
Multi-touch is the second reason. Maps, floor plans, menus, kiosk order screens and every interactive table we ship depend on two fingers doing something one finger cannot. Resistive cannot express that gesture at all.
Third is sealing. Our embedded panels run a flat front glass with an IP54 face (IP65 with a gasket kit), because there is no membrane edge to trap moisture. A resistive overlay has a film edge, and film edges lift.
The reference numbers on our 13.3-inch embedded panel, for anyone comparing datasheets: 10 points, ≤10 ms report at 100 Hz scan, ±1.5 mm centre accuracy, ≥88% transmission, 1.1 mm glass-to-LCD gap, −10 to 60 °C operating, USB HID so it enumerates with no driver on Windows, Linux or Android. The full range runs 8 to 55 inch.
Failure modes we actually get called about
Ghost touches on PCAP, always three causes. A water film bridging the bezel; a switched-mode power supply whose ground sits 30 V away from chassis ground; or a variable-frequency drive within half a metre. The fix is grounding and a ferrite, not a new panel — and it is worth asking for a controller with differential sensing before you ship rather than after.
Dead rows on resistive, always the same cause. A scratch or a worn spacer dot. Every resistive panel we have replaced failed at the point someone leaned a ladder against it. Where a glass front is exposed to the public, resistive is a consumable.
Drift on resistive after a season. The analog reference moves with temperature and humidity; a panel stable in a 22 °C showroom can sit 3 mm off in a 40 °C lobby until it is recalibrated. PCAP thresholds are digital and do not drift, which is why we can ship it without a calibration step in the customer's process.
How to write it into a purchase order
Seven lines settle most of the argument before samples ship:
1. Touch technology and touch points, with the words "10-point projected capacitive" spelled out rather than "multi-touch".
2. Front surface: material, thickness, hardness. "Tempered glass 3 mm, 7H" is checkable; "toughened front" is not.
3. Light transmission at the finished stack, not at the sensor film.
4. Activation force, or explicitly "no pressure activation".
5. Interface: USB HID, and which operating systems enumerate without a driver.
6. Operating temperature, and whether the touch controller or only the LCD is rated for it.
7. What happens to a scratched front: replaceable cover glass, or a whole new panel.
The short version
Resistive answers "did something press here". PCAP answers "what is touching, and where else". For a fixed keypad in a wet factory the first question is the only one that matters. For anything with a map on it, a menu people scroll, or a glass front the public will lean on, the second one decides whether the product survives its first year. Our embedded monitor range is all PCAP for that reason, and the classroom comparison of IR versus PCAP covers the third option if your panel is larger than 55 inches.