Powering a Round 240x240 TFT Module in a Battery Medical Monitor
On a battery-powered monitor, the display usually competes with the radio and the analogue front end for every milliamp, and the backlight wins that competition by a wide margin. The panel logic and the interface run on a few milliamps; the white LED backlight can draw an order of magnitude more, and its duty cycle is the single biggest lever on runtime. A design that dims the backlight between readings, blanks the panel when nobody is looking at it, and keeps the interface idle rather than refreshing empty frames will outlast one that treats the screen as always on. Measure each of those states separately before believing a runtime estimate.
Key takeaways
- The white LED backlight dominates the display's current draw, so the design should treat dimming and blanking as first-class power states. [1]
- Duty-cycle dimming changes the average current of the backlight without shifting the LED forward current or the colour point. [2]
- Disconnecting a small panel with a serial interface also means parking the control lines, because floating inputs can leak current through protection structures.
- Refresh on change rather than continuously; a static reading does not need to be redrawn while nobody is watching it.
- Measure the whole display subsystem in each state, including any level shifter and regulator quiescent current, because those parts run in every mode.
Portable patient monitors are a demanding power environment. The unit has to run for a shift, and often longer, from a cell that also feeds an analogue front end, a radio link and a processor that must not miss an alarm. Display current is one of the few loads that can be reduced aggressively without affecting the measurement, which makes it the natural place to look first.
The mistake to avoid is estimating display current from a single figure in a summary table. Backlight current depends on brightness setting, panel logic current depends on refresh activity, and standby current depends on what the host does with the control lines. A useful power budget accounts for all three states separately.
This article looks at a 1.28-inch, 240x240 round TFT with a four-wire serial interface and considers the states a battery monitor will actually spend time in.
Where the current actually goes
Three loads sit behind a small TFT module: the panel's own logic and source drivers, the backlight, and whatever the host spends driving the interface. The logic load scales with refresh activity because the source drivers have to charge and discharge the pixel capacitance row by row. The backlight load scales with brightness and is usually the largest of the three. The interface load scales with clock rate and with how often the host writes data.
Running a 240x240 TFT display with SPI interface from a battery changes the hierarchy of these decisions, because a refresh rate that costs nothing on a bench prototype becomes a line item in the budget. The strategy that follows is straightforward: keep the image static whenever the reading is not changing, dim the backlight to the lowest level that is still comfortably readable in the room the device is used in, and blank the panel entirely during transport and charging. Each of those steps removes current without touching measurement quality.
It also suggests a measurement discipline. Rather than trusting a single number, put a current meter in the supply rail and log the average over a realistic usage cycle: a few readings, a long idle period, a backlight dimmed to the usable minimum, and the panel blanked overnight.
Display states a portable monitor should actually implement
| State | What the display does | Where the saving comes from |
|---|---|---|
| Active reading | Shows the current value at a brightness the clinician selected | No saving; this is the reference state |
| Updated reading | Redraws only the changed region of the layout | Fewer source driver transitions than a full-frame refresh |
| Idle, screen on | Static image, backlight dimmed to the readable minimum | Average backlight current falls roughly with the dimming duty cycle [2] |
| Idle, screen blanked | Panel in sleep, backlight off, control lines parked low | Removes both the logic and backlight loads |
| Transport | Display fully unpowered, host pins driven low | Prevents leakage through protection diodes on floating inputs |
A measurement procedure that produces a defensible budget
- Break the railInsert a current meter or a shunt on the display supply after the regulator, so the reading includes the panel, the backlight and any series components.
- Measure each state separatelyRecord active, updated, dimmed, blanked and unpowered current. A single blended figure hides the state that is actually hurting runtime.
- Sweep the dimming rangeLog current at several brightness settings and note the lowest setting that remains readable in the intended environment. That is the design target, not the maximum.
- Check the control lines in sleepWith the panel unpowered, confirm the host drives its interface pins rather than leaving them floating, and measure the residual leakage.
- Re-measure at temperature extremesBacklight efficiency and display current both shift with temperature, so repeat the worst-case state at the extremes the device is specified for.
Dimming without colour shift or flicker
The standard way to dim an LED backlight is to switch it on and off faster than the eye integrates, varying the proportion of on time. Because the LED current during the on phase stays constant, the colour point stays constant too, which matters on a monitor where subtle colour differences might be read as information. [2]
The dimming frequency has to be high enough that neither the clinician nor a camera used for documentation picks up a flicker, and low enough that the switching edges do not radiate into the analogue front end. A few hundred hertz upward is typical on small panels, but the right number depends on the rest of the design. If the panel is photographed as part of a workflow, the dimming frequency should be chosen so that rolling-shutter artefacts are not visible in the image.
The analogue supply decision
Some designs dim by lowering the LED current instead of switching it. That is quieter electrically but tends to shift the colour point at low levels, and the LED efficiency falls. For a monitor that is read rather than measured against, duty-cycle dimming is usually the better trade.
Whichever method is used, the backlight driver should be placed close to the panel so the switching current loop stays small. Long backlight wiring turns a dimming circuit into an antenna, and the analogue front end is rarely far away.
Standby behaviour nobody tests until it matters
Standby current is easy to ignore on the bench and hard to fix in the field. A monitor that is stored with a charged battery and found flat a month later usually has a leakage path through the display or its level shifters. The fix is to drive every interface pin to a defined level in the off state, or to isolate the rail entirely.
Worth checking as well is what the host does during charging. If the unit charges with the screen lit, the charge time can be a noticeable fraction of the working day, and a charge-time-only blanking state is often the simplest change available.
Do not let a summary table decide the power budget. Put a meter in the display rail, exercise the real usage cycle, and record each state. Teams routinely find that the dimmed idle state, not the active state, is what determines whether the device lasts a shift on one battery.
Sources
Frequently asked questions
Does a small round TFT panel draw meaningful current when the backlight is off?
The panel logic and the source drivers still draw current while they refresh, and a sleeping panel draws far less. The residual figure is small compared with an active backlight, but it is not zero, and it matters for devices that spend most of their life in standby.
Is duty-cycle dimming safe for medical equipment?
It is widely used, but the switching frequency and the return path need care so the backlight does not couple into the measurement front end. Choose the frequency with the analogue design in view, and verify with the radio and the analogue channel active, not on a bench with everything else idle.
How bright should the display be on a portable monitor?
Bright enough to be read in the room where it is used, at the viewing angle it is used from, and no brighter. Every step above that costs battery and adds heat, and a bedside monitor is usually read under controlled lighting.
What is the cheapest way to extend runtime?
Blank the panel and turn the backlight off whenever the device is not being read, and dim rather than blank when it is. Those two changes affect no measurement function and are usually worth more than any component substitution in the rest of the design.