The lowest brightness setting stops where the backlight driver stops, and a dark wheelhouse needs far less light than that floor. A neutral density filter in front of the screen goes below it: 0.3 of density halves the light, 0.6 quarters it. Trial a 50% sheet held at the operating position after your eyes adapt, then step to 25% only if the screen is still too bright.
Yes, you can dim below the panel's lowest setting, and the tool is optics, not software. The brightness control in an HMI panel stops where the backlight driver stops; a neutral density filter in front of the glass keeps cutting below that floor. The scale is easy to hold in your head: each 0.3 of optical density halves the light, so a 50% transmission filter is the first one to trial. The same fix serves any screen read in the dark: a wheelhouse, a night-shift control room, a machine cell with the lights off. Work through the symptoms, pick a density, trial it removable, and fit it only when the dimmest alarm still reads from the operating position.
Why the lowest brightness setting is still too bright
The brightness value in the menu scales the backlight within a range the driver hardware defines. LED backlights dim a long way but not to nothing: below a minimum output the driver drops out of regulation, so the firmware stops the scale above that point. Lighting hardware hits the same wall, which is why a constant-current LED driver has a minimum dimming level it cannot go below.
The result on a boat is a mismatch of scale. A desk in a lit office works at about 500 lux; a night watch in a darkened wheelhouse works at single-digit lux. The room dropped by a factor of a hundred while the panel's floor never moved. A neutral density filter closes the gap from the screen's side. True ND glass passes nearly the same fraction of every visible wavelength, so red, amber and green alarms keep their relative brightness, which is why photographers have used it for decades to cut light without shifting color.
What each night-glare symptom points to
- Everything is too bright, dark screens included. The backlight floor is the cause, and a neutral density filter is the right tool.
- Only pale backgrounds glare. The project palette is the cause, and fixing it costs nothing: move the night screens to a dark background, which is most of what actually changes between dark and light HMI backgrounds. Do this before buying glass.
- The screen flickers on its lowest settings. The backlight is being pulsed to dim it, and glass reduces the light but cannot remove flicker. Set the brightness one or two steps up, where the pulsing stops, and take the light back down with a 0.3 filter.
- Full brightness returns after a power cycle. The panel did not keep the setting. Set it again, restart, confirm it holds, then judge the hardware.
Choose the density from the table
Filters carry either of two markings that describe the same glass. The photographic marking states the fraction of light passed; the optical density is its base-10 logarithm. Densities add when stacked, which keeps the arithmetic simple.
| Marking | Optical density | Light passed | Use it for |
|---|---|---|---|
| ND2 | 0.3 | 50% | the first trial |
| ND4 | 0.6 | 25% | still too bright after ND2 |
| ND8 | 0.9 | 12.5% | the dark end; re-check every alarm |
Neutral density sheet comes as acrylic cut to size and as mounted glass; both work, and the acrylic sheet is the usual choice for a panel bezel. Dyed tint film is made for car windows and carries a single visible-light number, so treat it as a crude trial only.
Example: a wheelhouse panel still too bright at its lowest setting
Inputs. The screen is an HMI panel in a night wheelhouse, at its lowest documented brightness setting, still too bright from the helm after full dark adaptation. Stock on hand: acrylic ND sheet in 0.3, 0.6 and 0.9 optical density, the three standard steps suppliers cut.
Rule. Transmission equals 10 raised to the power of minus the optical density, and each 0.3 of density halves the light.
Arithmetic.
- 0.3: 10^-0.3 = 0.50, so half the screen's light reaches the helm.
- 0.6: 10^-0.6 = 0.25, one quarter.
- 0.9: 10^-0.9 = 0.125, one eighth.
Result. The 0.3 sheet halves the glare in one step; the 0.6 quarters it.
Decision. Hold the 0.3 sheet at the helm and read the night screens. If they are comfortable and every alarm color separates, fit the 0.3 in a frame. If the screen is still too bright, swap to the 0.6. The 0.9 is the dark end: it stays only if the dimmest status color can still be told apart from its neighbor. Two things move the number: stacking adds densities, so 0.3 plus 0.3 gives 0.6 and 25%, and every extra sheet adds two surfaces that reflect cabin lighting, so re-check glare after each change.
Trial it removable before you fit it
- Set the panel to its lowest documented brightness setting and confirm the value survives a restart.
- Tape the filter sample to the bezel edge, held just clear of the glass, never glued. Give your eyes their due: full dark adaptation takes roughly 20 to 30 minutes, and a screen judged two minutes after lights-out is judged wrong.
- Read every screen the operator uses, from the normal position and angle: running screens, alarm colors, status values, small text.
- Look for haze and for reflections of cabin lamps in the filter's two surfaces. Re-angle the frame or pick a matte-finish sheet if the reflections win.
- Operate the touch screen through the sample if the panel is touch-operated, corner to corner. Any cover layer also costs sharpness on small text; what an anti-glare cover does to small text covers that trade.
One safety rule sits above all of this: never dim by cutting or rewiring the backlight feed. The driver is matched to the panel's LEDs, and a forced low supply can leave the backlight overdriven on the next power-up.
Fit it and check it
Fit the winning density in a rigid frame on the bezel or the enclosure door, clear of ventilation openings, labels and rating plates. Keep it liftable: the same panel serves daytime at full brightness, and the filter belongs in front of it only in the dark. Then run the checks:
- The lowest brightness setting is still active after a restart.
- The dimmest alarm indication reads clearly from the operating position, eyes adapted.
- Red, amber and green states still tell apart across the viewing angle.
- Touch input works at every corner through the filter.
- No new reflections from deck or cabin lights.
The panel goes into service when all five hold at night. If one fails, step back one density or remove the filter; a screen that hides an alarm is worse than a bright one. Two follow-ups pay off: add night dimming to the bench checks in what to test on an HMI panel first when you specify the next panel, and if the helm or the shop also wants a notice and status board, that is a separate screen from the HMI, which our shop-floor screens cover.
Recommendation: Trial a removable 0.3 neutral density sheet (ND2, 50% transmission) taped clear of the glass, judged after full dark adaptation from the operating position. Step to 0.6 (25%) if the screen is still too bright, and fit the passing density in a liftable frame that leaves vents, labels and touch clear. Reach the target with glass, never with the backlight wiring.
Questions
Can I stack two neutral density filters to go darker?
Yes. Optical densities add and transmissions multiply: two 0.3 sheets stack to 0.6 and pass 10^-0.6, about 25% of the light; a 0.3 over a 0.6 gives 0.9, about 12.5%. Each extra sheet also adds two surfaces that can mirror cabin lights back at the operator, so re-check reflections after every stack, and stop as soon as the dimmest status color becomes hard to identify.
Will a filter in front of the screen spoil touch input?
It depends on the touch technology. A capacitive panel senses through a cover layer, so a thin filter usually still works, with some loss of sensitivity toward the edges of the active area. A resistive panel needs finger pressure to reach its own surface, so anything loose in front of it will mis-register. Operate every screen element through the trial sample before fitting anything, and if touch degrades, hang the filter in a lift-up frame above the glass.
Is automotive window tint film good enough?
It can work as a trial, because tint film is rated by visible light transmission: a film marked 35% VLT passes about a third of the screen's light. Tint is dyed for car windows, though, so it dims some colors more than others and can make red, amber and green status states hard to tell apart. Hold a sample clear of the glass and judge it on the panel's own alarm colors before considering a permanent fit.
Planning information, not legal or engineering advice. Rules differ by place: check codes and permits with your local authority and your licensed trades.