Analysis of OLED screen flickering problem

 

background

Complaints about OLED screen flickering and glare are often seen online. After our series of studies and the introduction of improvement measures, customers who returned visited confirmed that the splash screen had been significantly improved. Let me introduce to you the causes of OLED screen flickering and the directions for improvement.

First, what makes a user experience a splash screen? We finally converged on two points: brightness changes and color changes.

Nowadays, the mainstream OLED mobile phone screens are basically 60/120Hz. This process is mainly reflected in some low-brightness and low-grayscale scenes, which can be felt by users between two frames or between frame rate switching. Brightness and color changes, for example, when people are lonely late at night, open WeChat and shake, and staring at the phone for a long time will cause fatigue problems such as dazzling eyes and dry eyes.

 

 

1. Electrical devices

 

The energy gaps of red, green and blue OLED electroluminescence are different, and the defect states inside the electroluminescent device are also inconsistent. The efficiency of red, green and blue is quite different, especially at low current density, the efficiency of red, green and blue is different. The difference is more prominent. The large difference in the efficiency of different pixels at low current will lead to problems such as screen flickering and color cast. As for why the red, green and blue efficiencies of OLED are quite different, students who have studied OLED light-emitting materials and devices should easily understand it, so we will not go into details here.

 

2. Drive circuit

 

1. The influence of Vth: Enthusiasts who understand OLED drive know that OLED is current driven. In terms of only one driving TFT, the driving current. With the development of drive circuits, mass-produced drive circuits are all 7T1C. The threshold voltage is well subtracted from the drive circuit. The drive current is proportional to (ELVDD-data), where ELVDD is a fixed voltage, so theoretically The influence of Vth has been eliminated [I will introduce the corresponding drive circuit and the now popular LTPO later when I have time].

 

2. DC dimming: data is the gamma voltage output by the display chip. Each brightness corresponds to a gamma voltage. It keeps emitting light nearly 100% of the time in the light-emitting range, which is called DC dimming. The current brightness level is now at least 4096, which corresponds to hundreds to thousands of brightness. If the gamma voltage at each brightness is calibrated one by one at the screen factory, this will greatly affect production capacity. Therefore, in actual domestic mass production, only a few brightness nodes are calibrated, and then the nodes are calculated and supplemented through the difference algorithm of the display chip, and the gamma voltage under different brightness is obtained.

 

3. PWM dimming: The human eye is insensitive to high brightness, but very sensitive to low brightness. The screen manufacturer will calibrate a few more nodes in the low brightness range, and the display chip will calculate the output gamma voltage. The accuracy is also gradually improving, but because the efficiency difference between red, green, and blue is greater at low brightness, flickering problems will still occur, and uneven display problems are easily exposed. Therefore, in the low-brightness range, multiple Pluses are used to control the width of the Pluses (the ratio of glowing time to non-luminous time, which is professionally called duty cycle). The longer the glowing time seen by the human eye, the corresponding brightness will be. High, this dimming method is PWM dimming. So why not choose a Pluse to adjust the width? Is that because a pluse is used? This is because when the brightness of pluse is low, the corresponding black time is longer, which is easier for the human eye to detect.

 

4. DC+PWM dimming: All domestic OLEDs use high-brightness DC dimming and low-brightness PWM dimming. Therefore, there is a saying that the light-emitting range is not close to 100% of DC dimming. There are PWM dimming ranges. Many Consumers or enthusiasts believe that multiple Plus PWM dimming methods that adjust the lighting time will aggravate OLED flickering, and hope that terminals can adopt full DC dimming methods. But the actual situation is that PWM dimming is more delicate in controlling brightness. DC dimming is based on the given voltage at low brightness. Because the luminous efficiency of red, green and blue is poor, it will cause many color cast problems and the display effect will be worse [ Note: The whole process of DC dimming is not a problem with technology. I have actually debugged and seen samples, and it is indeed poor]. Moreover, Apple’s OLED screens use full-range PWM dimming, and there is no DC dimming. We have not seen any customer complaints about screen flickering.

 

5. High-frequency PWM dimming: Based on the above, high-frequency PWM dimming is selected in the low-brightness lighting range, such as 1920Hz promoted by some terminals, which means increasing the number of pluses to change their duty cycle dimming. It also claims to be more comfortable for human eyes. We often see articles mentioning the national standard of LED to discuss this matter. First, let me explain that the claimed parameters seem to have gone up, and consumers will be more accepting of them, but it still needs to be said whether it is really good for the human eye, or what high PWM dimming frequency is most suitable for the human eye. Observe again and collect some data to confirm. First of all, one disadvantage of this high-frequency PWM is that outputting multiple Pluses will obviously increase the power consumption of the display IC.

 

6. Whether it is PWM dimming or DC dimming, as long as there are multiple EM Pluses that control light emission within one frame: the status of the first EM Pluse and subsequent pluses in each frame are inconsistent. Normally, the potential of the first EM Pluse is the most accurate, but during the light-emitting process of the first pluse, the voltage level of the light-emitting writing decreases, and the brightness is ∝ lower than ELVDD-data. After the data decreases, the corresponding brightness will increase. , so the brightness of the first pluse in each frame is low, and the brightness of the following pluses increases, which will also cause flickering.

 

3. Measurement of splash screen

 

1. FFT (Fourier transform): The frequency of the emitted light is very complex. The amplitude in time series is tested, and then the distribution in the frequency domain is obtained through Fourier transform. There are many articles introducing the specific transformation, so I won’t go into details here.

2. JEITA method (Japan Electronic and information technology industries Association): Based on the calculation of FFT, the flicker value of the JEITA method is obtained by summing the weight at each frequency.

3. Nowadays, each terminal actually controls the flicker value of its screen. However, the human eye is actually a very complex system. In many cases, the objective value tested by the instrument is qualified, but the human eye cannot see the flicker value. There will still be flickering.

 

4. Improvement of splash screen

 

Based on the analysis of the above reasons, the flicker improvement method can be considered from the following aspects. Due to technical confidentiality reasons, we cannot explain in detail. Private communication is welcome.

 

1. Electro EL device: It is related to the luminescent material system, and the problem of low brightness color cast must also be considered;

2. Dimming mode & dimming node;

3. PWM dimming frequency: Although there is public opinion that high frequency will be better, there are currently no papers related to OLED flash screen research. Even if it is better, it should not be that higher is better. There will be a reasonable value. This needs Supported by further experimental data.

4. Duty cycle: The duty cycle of PWM and DC. The larger the non-luminous interval, the stronger the theoretical flicker.

5. Display IC compensation: The output target Gamma voltage caused by the process will deviate from the target value. Some IPs will eliminate this deviation and improve it better.

6. In view of the phenomenon that the brightness of the first pluse of each frame is low and the brightness of the following pluse is high, resetting the OLED parasitic capacitance in each EM pluse will also better improve the flicker.