BenQ XL2411T

From Display-Corner
Jump to navigation Jump to search


Specifications

Brand:BenQ
Model:XL2411T
Size:24"
Resolution:1920x1080
Panel type:TN
Max. refresh rate:144 Hz
Panel:AUO M240HW01 V8
Processor:MST 8556T (MSTAR)
Backlight type:LED (white, edge)
LED driver:MP3397 (Monolithic Power)
TachistoMode:wake-up
Price (approx.):USD 400
Monitor release date:2012-12
This review's date:2013-04

Reviews

BenQ XL2411T review on PC monitors.

BenQ XL2411T review on 3D VISION BLOG.

Comparison with other LCDs

The T-versions of BenQ's gaming monitors are about to be replaced by the Z-versions, e.g. the BenQ XL2411Z.

The BenQ XL2411T came out after the XL2410T and the XL2420T/TX. It inherited the design of the XL2410T but its inner values are closer to the XL2420T. The BenQ XL2411T and XL2420T are both "NVIDIA 3D Vision 2 ready", i.e., they both have LightBoost on board, whereas the XL2410T does not, although it still supports stereo. The only advantage of the XL2420T over the XL2411T is the fancier design, a wired "remote" control, and more interfacing options (DisplayPort, 2nd HDMI, USB). Note that LightBoost is only available with 100 and 120Hz but not with 144Hz.

Note that early releases of the XL2411T monitors used pulse-width-modulated backlight (PWM), monitors manufactured later provided flicker-free backlight.

According to this BenQ XL2411T vs. Asus VG248QE comparison on PC monitors the Asus VG248QE uses the very same LC panel and performs equally well.

Service menu BenQ XL2411T

How to call up the service menu

Have the monitor connected to a valid signal source. The service menu can be entered by keeping the 2nd button from the left ("down" button) pushed while powering on the monitor. Pushing the 4th button from the left ("menu" button) will then call up the service menu. The "menu" button will react normal again after the next power-on.


AMA=off

Reaction time

The luminance curves in the three figures show how fast the LCD reacts on changes of pixel values at a refresh rate of 100Hz. Each figure shows three pixel value step sizes (maximal (blue), intermediate (red), and small (green). All curves are normalized so as to allow for easier comparison of the timings. The monitor has three AMA settings (overdrive) – OFF, HIGH, PREMIUM –, which clearly make a difference here. The overshoot seems to be quite strong when overdrive is active, but one could argue that it is the average luminance during each refresh interval that matters (or the accumulated luminous energy) and, thus, that the overshoot is needed to compensate for the slow rise and fall times, at least for the first refresh interval directly after a change.

Note that the monitor was reset to its factory settings and the Picture Mode was set to Standard (factory setting is FPS1). The Contrast setting was at 50%, which is about the maximal setting that does not cause high pixel values to be cut off (white saturation). Interestingly, the curves for the full step (0%–100%, blue) do not change when the Contrast is set to 100%; especially the difference between AMA=off and AMA=high (or premium) is fully preserved. One could have expected that the AMA=off curve would look like the AMA=high (or premium) curve once the monitor operated at full saturation. But this seems to be not the case.

The 0%–100% curves look different when the Picture mode is set to FPS1 (not shown here). The maximum steady-state luminance is reduced then, which leaves enough clearance for overdrive to be effective even at the highest pixel values (100%). However, the momentary luminance overshoot never goes beyond what can be observed when Picture mode is set to Standard.

The measurements have been taken with the photo diode PDA36A (Thorlabs), whose gain was set to maximum (70dB) as bandwidth limitation was not of concern here. Only a horizontal bar covering the upper 10% of the screen was flickered, which is important as the number of effective lines determines how much the measurement is affected by the low-pass filtering being induced by the sequential updating of the LC pixel cells from the top of the screen to the bottom.


AMA=high
AMA=premium


External LED backlight control

Under certain conditions this monitor is suitable for tachistoscope applications. The LED driver does not fully maintain the supply voltage level when the LED are switched off. However, the voltage level decreases relatively slowly over time and is quickly reestablished without major artifacts as soon as the backlight is switched on again. In applications where the induced errors cannot be tolerated, a brief wake-up pulse can be issued while "presenting" a black image, shortly before the actual stimulus image is going to be presented.

The measurements have been taken with a photo diode PDA36A (Thorlabs) which was connected to PC oscilloscope PicoScope 4224 (Pico Technology). The gain of the photo diode was set to 30 dB, limiting the bandwidth to 785kHz.

Relative errors in luminous pulse energy when no wake-up pulse is used. Error bars indicate standard deviations. Overall, N=866 stimulus pulse measurements are represented by this graph, which means roughly 40 pulses per data point on average.

LED switching characteristics

The measurements shown here are about the accuracy of luminous pulse energies and were taken with an experimental scenario in mind where one stimulus (pulse) is presented per trial and a subject has to respond, inducing some variation of the inter-stimulus interval due to variable response times. The stimulus interval has been randomly varied between 0.5 s and 5 s, uniformly distributed over a logarithmic time scale. Two different delays between the wake-up and stimulus pulse have been tested, 80 ms and 800 ms. These values were chosen with two scenarios in mind.

  1. The wake-up pulse is issued shortly before the stimulus. Because the wake-up pulse must not be visible to the observer, a black image has to be "presented" during the wake-up pulse. Thereafter, the software needs to switch to the stimulus image and the LCD has to settle to this new image, before the backlight can be finally switched on again for the stimulus pulse. The minimally required stimulus delay is mainly determined by the LCD reaction time (see figures above).
  2. The wake-up pulse is issued well before the stimulus presentation, for example, during the inter-trial pause. This implies a longer delay between the wake-up and the stimulus pulse.

The first figure shows what happens if no wake-up pulse is used at all. Note that the error variability is basically owed to the inter-stimulus interval variability, which was rather huge as compared to the assumed response time variability, suggesting that this was pretty much a worst case scenario. On the other hand, the average of the inter-stimulus intervals was just around 1.5 s (note that the inter-stimulus intervals were randomized on a log-scale). So it is pretty difficult to tell how big the errors would be when the inter-stimulus interval variability were smaller. Anyway, the next figure shows the errors when applying wake-up pulses, and these errors are not only smaller but also show far less variability as compared to the no-wake-up scenario.


Relative errors in luminous pulse energy when a 2 ms, 10 ms, or 40 ms wake-up pulse is issued (sub-plot rows) and the stimulus is delayed by 80 ms or 800 ms (sub-plot columns). Error bars indicate standard deviations but might be so small that they are hidden behind the data points. N is the number of measurements that are represented in the particular subplot. The standard deviations in the upper right subplot (800 ms delay) are quite large, irrespective of the small mean errors, which suggests that a wake-up pulse of only 2 ms is too short for reliably waking up the LED driver.


The relative errors shown in these figures are given with respect to pulse energies that have been calculated as the product of the measured trigger pulse width and the steady-state luminance. Thus, the curves are free of errors which might have been caused by inaccuracies in the trigger signal generation. The measurement have been taken with the photo diode PDA36A (Thorlabs), whose gain was set to 30dB, which results in a bandwidth of 785kHz. A Matlab program controlled the backlight through the parallel port, recorded the signals (backlight trigger and photo diode output) with the PC oscilloscope PicoScope 4224 (Pico Technology) trial-wise, and analyzed the recordings after each trial.

rising (on) falling (off)
Delay [µs] 2.0
(SD=±0.034)
3.2
(SD=±0.071)
Slope [µs] 4.5
(SD=±0.551)
2.7
(SD=±0.059)

LED backlight switching times as measured from the rising and falling edges of the luminance curve when switching the backlight on and off. Shown are the average values with standard deviations. The delay is the time it takes for the luminance to change by 10% after the according trigger signal edge occurred. The slope is the time it takes for the luminance to rise from 10% to 90% (or fall from 90% to 10%, respectively).


The table contains switching delay and speed values, averaged over all presentations which employed a 10 ms wake-up pulse, an 80 ms delay, and a trigger pulse longer than 1 ms.

Accessing the backlight signals

Rear cover removed.
Connector on the power supply board
Wire color Signal
black GND
black GND
n.c.
violet Effectively not connected to
the power supply board
red 5V (supply)
red 5V (supply)
grey 3.3V if LightBoost is active at
120Hz, otherwise 0V
green 3.3V if LightBoost is active at
100 or 120Hz, otherwise 0V
brown PWM (3.3V)
white +3.3V (supply)
Connector on the controller board
Wire color Signal
brown PWM (3.3V)
white +3.3V (supply)
black GND
black GND
red 5V (supply)
red 5V (supply)
grey 3.3V if LightBoost is active at
120Hz, otherwise 0V
n.c.
violet Effectively not connected to
the power supply board
green 3.3V if LightBoost is active at
100 or 120Hz, otherwise 0V


Power board (left), also hosting the LED driver, and controller board (right).
Power and controller board unmounted and flipped over (top view).


BenQXL2411T Bezel.png

Mechanical peculiarity

Note that this monitor has neither a flat front face nor a straight bottom side, which might be a problem for mounting the monitor, depending on the tachistoscope's mechanical setup. As shown in the picture, the bezel has a bulge which extends to the front and to the bottom, just where the control elements are located (at the lower left of the screen).