How to cancel the ambient light of the LiDAR receiver

A more challenging challenge for Time of Flight (ToF) LiDAR is the high sensitivity required to receive the signal chain. Usually, collimated (parallel) laser pulses are sent to a point. The advantage of a collimated laser source is that it limits the light loss caused by divergence and keeps the spot size constant within a distance. However, once light shines on an object, this light will bounce back in multiple directions - this is called scattering. The amount of light reflected back to the light source is proportional to 1/R, also known as the inverse square law. It is not difficult to detect objects within a short distance. However, in order to detect objects greater than 100 meters, high gain is required to detect a small amount of reflected light caused by the loss of inverse square law. One of the consequences of using high gain in the receiver is the impact of ambient light on the signal chain. The sun is a light source with a wide wavelength range. Lidar systems typically choose wavelengths of 900 nm and 1550 nm because the light from the sun in these spectra is naturally zero. Unfortunately, in order to detect distant objects, we have gained significant benefits in the receiver, even with these natural zeros in the spectrum, the ambient light of the sun can saturate the receiver. This effectively blinds the system and renders it useless. This article will explore solutions to reduce the impact of ambient light on the LiDAR receiver chain.

basic

Step back, the laser is used to emit narrow light pulses; The laser pulse impacts the target and reflects light from the object. The detector is used to measure the time required for the reflection return. By understanding the speed of light and the round-trip time of laser pulses, the distance can be calculated. Usually, the higher the amplitude of a pulsed laser, the larger the return signal. For remote LiDAR, the human eye safety of laser power limits the range of modern systems. The area under the curve determines the energy of the pulse, as shown in Figure 1. To achieve higher peak power, it is necessary to reduce the width of the pulse so that the area under the curve is below the eye safety limit. Therefore, our goal is to provide high amplitude laser pulses with relatively narrow widths. In the current LIDAR system, the pulse width is approximately 5 ns and is moving towards a shorter pulse width. Another aspect that LiDAR needs to consider is scattering. Usually, avalanche photodiode (APD) detectors are used to provide optical gain to solve the inverse square law problem. APD is beneficial for the signal chain, as the transimpedance amplifier (TIA) is a limiting factor for noise in the signal chain. By applying gain to the detector, the input reference noise of the system can be reduced. Please remember that APD has limitations, as excessive gain can result in poor noise performance during breakdown.

Figure 1. Examples of different laser outputs.


LiDAR Challenge

Like any other engineering problem, trade-offs need to be made. The receiving signal chain needs to have a sufficiently high bandwidth to detect the edges of~5 ns wide laser pulses, and the capacitance of the detector needs to be very small to not limit the TIA bandwidth. Smaller capacitors also contribute to the shot noise of APD, as they are proportional to each other. For practical applications, it is necessary to balance sensitivity, bandwidth, and power consumption. Another challenge in obtaining higher gain in the receiving signal chain is the accompanying large dynamic range. The reverse bias of modern APD is close to 300 volts to achieve these larger gains. When highly reflective objects are very close to the detector, the problem becomes apparent. This large signal combined with the relatively large gain of APD may cause hundreds of mA of current to flow through TIA. Most communication TIAs cannot survive such events, let alone recover within a reasonable time in the next pulse cycle. Fortunately, the LiDAR specific TIA has a built-in clamp that can divert current and recover below 100 ns. Resolve power issues through duty cycle and closing unused channels. Considering these, the last major issue is the DC photocurrent from ambient light, and solving this problem is not an easy task.

AC coupling and DC coupling input

At first glance, a simple solution is to couple the input to TIA to prevent direct current. Unfortunately, this method has many pitfalls. The saturation recovery time will be affected, causing the system to become blind. If there is a large pulse from a close object, the AC capacitor will be charged. TIA can only inject a small amount of current into the AC capacitor because the feedback resistance (approximately 10 k Ω to 100 k Ω) limits the current. According to the value of the capacitor, the RC time constant is very large and may take hundreds of μ s to recover. This is unacceptable because a time of 100 μ s is usually allocated for 2 m detection, and we will miss signals from further objects. Another drawback of communication coupling TIA is the repetition rate of the laser source. When you communicate coupled input pulses, the pulses will be averaged across the AC capacitor. The signal of the detector is unipolar and will slowly charge the AC capacitor. There will be a DC offset on this capacitor. This systematically reduces the linear range of TIA, and the DC offset will vary based on the repetition rate and the amplitude of the return signal. For a more detailed analysis of communication input coupling TIA, please refer to the article "How to effectively design and optimize the TIA interface of LIDAR systems". Fortunately, DC coupled input avoids all these subtle differences and secondary effects, but at the cost of increasing complexity. The effective method to eliminate this current is to integrate a closed-loop circuit and inject the opposite current into the input of TIA.

DC cancellation circuit


Figure 2 shows a block diagram of how to implement a simulated closed-loop to eliminate DC input current. The job of an error amplifier is to view the output of TIA and inject the opposite current into the input of TIA. It compares and servo outputs to match the reference voltage source of TIA. It is best to use TIA's reference voltage source to derive the reference voltage source of the error amplifier, for two reasons: matching with the output reference voltage source and ensuring that PSRR is conserved for TIA. In order to save power consumption and cost, the circuit of the error amplifier should use an amplifier with lower bandwidth. It is recommended to use a low-pass filter for the input of the error amplifier, as you do not want fast pulse coupling back to the input.

Figure 2. Block diagram of DC cancellation.

Figure 3 shows the DC elimination circuit of LTC6560. When TIA has no input current, the nominal output value of LTC6560 is approximately 1 V DC. Therefore, it is necessary to divide a resistor voltage divider from the reference voltage source to match the voltage, and divide the nominal value of the reference voltage by 1.5 V to match the output of 1 V. R1 and C1 generate a low-pass of approximately 10.6 kHz; This helps to minimize the amount of noise injected into the LTC6560 by the error amplifier. The low-pass will be the main pole of the loop and can be adjusted according to different bandwidth requirements. A simple integral error amplifier circuit is used to servo the output of LTC6560 to 1 V; Please remember that when there is no current on LTC1, the nominal output voltage is 6560 V. R2 is a 20 k Ω resistor, which is a simple solution to convert the output of LT6015 into current. The value of the resistor and the maximum swing of the operational amplifier will be set to the maximum current based on the output swing of LT6015. Since LT6015 is not a rail to rail operational amplifier, the maximum DC offset will be limited to the difference between the maximum swing of LT6015 and the input self bias voltage (nominal value of 1.5 V) of LTC6560. This is approximately 3 V, which will provide us with a maximum DC offset current of 150 μ A.

Figure 3. DC elimination circuit for LTC6560.


Figures 4 and 5 illustrate the LTspice simulation of the LTC6560 DC elimination circuit. Please note that V2 is used in the simulation to set the reference voltage source for the integration error amplifier. This is used to assist in circuit simulation and establish a deterministic starting voltage.

This DC elimination circuit can also be used in conjunction with LTC6561. By using four output resistors to inject current into each channel, you can save three LT6015, as shown in Figure 6. It should be noted that we are currently creating a path that can couple channels. However, a 40 k Ω resistor has the least impact on inter channel isolation. Finally, the DC input current of the channel should be very similar, as the error amplifier does not undergo drastic changes between channels. This circuit will benefit systems where all optical channels are close to each other.

Figure 4. LTspice simulation schematic.


Figure 5. Input and output waveforms of DC cancellation simulation.


Figure 6. DC elimination circuit for LTC6561.


Figure 7. LB2953A DC cancellation circuit board laboratory board.


result

A concept validation board was created to create more attractive articles and validate performance. As shown in Figure 7. As expected, the DC elimination circuit is mainly dominated by circuit board wiring and parasitic components of the components. This circuit increases the integration noise from 64 nA rms in non DC cancellation circuits to 200 nA rms integrated in DC cancellation circuits ranging from 66 kHz to 100 MHz. Figure 8 shows the noise density measured at the input end when using and not using a DC cancellation circuit. Remove the APD from the circuit to find the background noise without TIA capacitive load. This generates an integration noise of 59 nA rms for non DC elimination circuits and 60 nA rms for DC elimination circuits. However, the circuit is intended to be used in conjunction with the detector and capacitance should be included in the performance of the circuit.

Figure 8. Noise density converted to input.

conclusion

AC coupling of inputs for LTC6560 and LTC6561 may pose some challenges. Ultimately, in a few cases, AC coupling can be achieved with minimal impact on circuit performance. In modern LIDAR systems, in order to maximize system performance, the proposed DC cancellation circuit can provide maximum recovery time performance without affecting circuit noise. The cost of this performance is the complexity of the layout and the increased power consumption of the integration error amplifier.