Large Signal Transient Response Comparison Continued
Apple to Apples comparison Voltage Mode / Current Mode / Hysteretic Mode -- Part 2
- Large Signal Load Transient Response
Resources
Let's go a little more deeper into this large signal load transient response study. What happens with different magnitudes of load steps? Let's take two load steps. 500 milliamp to 5 amp back to 500 milliamp, and 50 milliamp to 500 milliamp back to 50 milliamp with the 5 volt input.
See that here [INAUDIBLE] start, the voltage mode IC has a good stability in most transient cases, whereas Hysteretic-Mode and current mode IC exhibit ringing in both transients, similar to what we have seen in our previous series, which is expected because the output current level wouldn't influence the understood behavior.
Onto the scenario. What happens if the output capacitance has changed? It is advisable not to change the output capacitance after finalizing the loop compensation design, since the loop stability directly depends on the amount of output caps used. However, in the real world, the actual output capacitance would be increased when the converter is placed in the actual system due to the amount of huge bypass caps present downstream.
To understand such a scenario, let's change the output cap to 1 millifarad. The simulation here simulates the load transient behavior of 500 milliamp to 5 amp at 2 input voltages. All control modes show a slow response due to a lower bandwidth.
But the under [INAUDIBLE] is heavy deduced. Hysteretic-Mode control shows significant improvement on the under [INAUDIBLE] amount compared to the other two topologies. Now let's see what happens with the duty, or pulse with change, during a load transient.
Voltage mode and current mode are working with PWM control and Hysteretic-Mode is changing frequency with constant on time. Hysteretic-Mode IC changes its frequency much faster than voltage mode IC and current mode IC change their duty. Hence, it shows a faster transient response.
With a bigger output cap, duty or pulse density change, is smaller. Thus, all three control modes show a slower transient response. Voltage-Mode control, in this case, shows almost an invisible duty cycle change. Hysteretic-Mode control has a minimum of time limitation to the pulse density limit. Still, Hysteretic-Mode IC shows the fastest transient response.
Till now, as you would have noticed, the evaluation was performed without much modifications to the EVM's. So the LC values are not the same across all EVM's. To get a more apples-to-apples comparison, we do a few modifications.
First the current mode control EVM's loop compensation is re-tuned. And the switching frequency is increased to 1.2 megahertz. The Hysteretic-Mode EVM's output capacitance is changed to 200 microfarad, to make it comparable to the current mode IC.
Because the current mode IC's loop bandwidth is increased, it shows much quicker transient response at the cost of stability. So a Tantalum cap is used instead of a ceramic cap to improve stability. This, in turn, makes the overshoot and undershoot worse due to high ESR.
The advantage is the Hysteretic-Mode converter has is that the switching frequency changes only during transient response instead of always keeping higher-switching frequency, as we have done here for Current-Mode IC, which results in higher switching losses all the time. With Hysteretic-Mode IC, we have modified to have 200 microfarad cap. We can see that the overshoot and undershoot are significantly reduced with a good transient response.
This video is part of a series
-
Fixed frequency versus constant on-time control of DC/DC converters
video-playlist (7 videos)