I’m attempting to voltage glitch a controller using the Husky, targeting the Vcap/Vcore (1.8V). The controller’s internal clock runs at 50MHz.
I’ve removed all capacitors connected to Vdd, Vcap, and Vadd. However, I’m facing an issue: when I drop the voltage for more than 50 ns, it falls to 0, but when I try a shorter pulse (under 30 ns), I see almost no visible effect.
I can’t disable the internal voltage regulator since it’s not configurable. To further test its behavior, I connected an external power supply to both Vcap and Vdd. I attempted to raise the Vcap voltage slightly above 1.8V (~1.92V) while lowering Vdd to around 2.1V (which still allows the controller to function), but this approach didn’t yield any noticeable results either.
Now, I’m unsure whether the lack of voltage drop is due to missing some decoupling capacitors or if the regulator is simply highly effective at maintaining stability.
What would be the best approach to proceed from here? Any insights would be greatly appreciated!
@amiros I have a similar problem in one of my targets where the vdd_core power line is controlled via pwm. During bootloader initialization this power line works in a limited range, later after kernel startup control over this line changes. If pwm is used, it is in combination with an internal voltage regulator + possibly with anti-glitch sensors. What can be done here is practically nothing and certainly no external laboratory power supplies will help. I still have a 1.50V analog line to comb and check.
I just gave an example. Every dut is different. Using pwm means that the power line can be controlled by voltage regulation or current limitation or this + that. In my case I found information regarding the use of pwm on the vdd_core 1V1 line in my dut… Your case may be different or similar
