SLUAA69 July   2020  – MONTH  TPS548D22

 

  1.   Trademarks
  2. 1Introduction
    1. 1.1 LED Driver Methods
    2. 1.2 Power Supply Solutions for Common-Cathode LED Display
  3. 2Principle of Synchronous Buck with Sinking Current Application
  4. 3 Design Considerations and Analysis
    1. 3.1 Choose an IC with Sufficient Current Sinking
    2. 3.2 Choose IC Supporting Negative OCP
    3. 3.3 Choose an IC Supporting Pre-Bias Startup
    4. 3.4 Analysis of System Startup
  5. 4 TI Devices and Functionalities
    1. 4.1 Negative OCP Functionality
    2. 4.2 Hiccup Mode and Latch-off Mode
    3. 4.3 UVP and OVP Functionality
  6. 5 TI Solution
  7. 6 Bench Test and Result
    1. 6.1 Bench Test Configuration
    2. 6.2 Startup Waveforms and Behaviors Analysis Overview
    3. 6.3 Startup Waveforms and Behaviors Analysis at the First OVP
    4. 6.4 Startup Waveforms and Behaviors Analysis after the First OVP
    5. 6.5 Waveforms and Behaviors Analysis of Startup Solution with Lazy Loading
  8. 7 Conclusion
  9. 8References

Startup Waveforms and Behaviors Analysis Overview

Figure 6-2 shows the start-up waveforms of TPS548D22 with 9.3-A load added before power up.

GUID-20200604-SS0I-2WMM-VFC6-ZNRFBF8BF6XB-low.png Figure 6-2 Start-up Waveforms with 9.3-A Load
At some moment before t0, PVIN power and VDD power are turned on but the device is not enabled. It can be seen that the output voltage equals the input voltage of 4.2 V. This means the output capacitor is fully charged to the input voltage through the load path. The output voltage is 4.2 V higher than the OVP trip level 1.2 V (120% threshold) but OVP is not triggered before t0 because the device is not enabled.

At t0, the device is enabled. t1 occurs after about 1 ms of the Power-on delay time, tPODLY. At t1, the device’s function modules, such as soft-start and protection functionality, start working. The soft-start begins from t1 for 1 ms with reference voltage ramping up from 0 V to the final defined level. The device does not start up because an OVP event also occurs at t1 and the output begins discharging to 0 V.

At t2, around 2 ms after t1, the output voltage begins ramping up to the OVP trip level of 1.2 V and then discharges to 0 V. At t3, the output voltage begins ramping up to the OVP level and discharges again, similar to at t2. The OVP event is periodically triggered every 2 ms. Hence, the device hiccups every 2 ms and gets stuck during the start-up state.

This may result in a number of questions. Why is the duration between t1 and t2 around 2 ms? Why does the device hiccup every 2 ms after t2? It is known that the hiccup time should be 16 ms for 1 ms soft-start. Here, a 2 ms hiccup time looks quite different. What about the behaviors of waveforms? How does the output ramp up and decay? Why does the software voltage ramp up following output and also have some pulses?

It can be seen from the inductor current and input voltage waveforms that when large inductor current transients occur, the input voltage changes sharply. This is not good for the input stage and should be avoided in the power supply design.

To address the questions listed above, the internal soft-start operation should be taken into account. Note that it is mentioned in Hiccup Mode and Latch-off Mode that when an OVP event occurs, the discharge of output could lead to an undervoltage condition and trigger the UVP. Then the OVP event will be reset by the UVP event. In addition, it is specified in UVP and OVP Functionality that UVP function is only enabled after the soft-start operation is completed.

To better understand the relations between output behavior and internal soft-start operation. Figure 6-3 shows output voltage and critical time nodes in detail.

GUID-20200604-SS0I-W39F-VVDN-J7LPG8NPR3G5-low.png Figure 6-3 Start-up Output Waveforms with 9.3-A Load
At t1, OVP is triggered and the output begins discharging. At t2, the feedback voltage (the same as output voltage) becomes lower than 68% (0.68 V) of the target voltage. The UVP comparator output goes high but the internal UVP delay time counter won’t begin counting since the soft-start operation is not completed. The output continues discharging from t2 to t3. At t3 when the soft-start operation is completed, the internal UVP delay time counter begins counting for 1 ms from t3 to t4. At t4, the device turns OFF both high-side and low-side MOSFETs drivers. Then the device enters hiccup mode with 16 ms delay due to UVP (the OVP at t1 is reset by the UVP). After 1 ms UVP delay at t4, there is an internal 1 ms soft-start timer that begins counting from t4 to t6 (1 ms, 16 times). At t4, the output ramps up resulting in OVP being triggered. Further details on why the output ramps up at t4 to the 1.2-V OVP level will be provided in Start-up Waveforms and Behaviors Analysis after the First OVP. At t5, the output decays to the UVP trip level. But the 1 ms UVP delay counter doesn’t count until t6 when the soft-start operation is completed, similar to what happens from t2 to t3. After 1 ms UVP delay at t7 (repeat as t4) , the internal 1 ms soft-start timer begins counting, similar to what happens at t4.

The OVP and UVP event occur cyclically. This results in a 2 ms hiccup time, consisting of 1 ms UVP delay and 1 ms soft-start time. If the circuit is configured with 2 ms soft-start time (by changing R24 to 47.5 kΩ), then the cycle is 3 ms hiccup time, consisting of 1 ms UVP delay and 2 ms soft-start time.