On Sun, Sep 3, 2023 at 02:16 PM, Hans Summers wrote:
Hi all
?
I agree a series resistor would prevent the GPIO conflict. The GPIO pin is necessarily configured as push-pull. A series resistor would reduce the efficiency of the buck converter. However my feeling is that if the resistor was small, say 220 ohms, it would suffice to resolve the fight and not significantly impact the efficiency.?
?
Gunnar, I'd like to understand why you're so sure the current through the zener and base-emitter is low? I agree as soon as the voltage rises it will stop the PWM and thus further energy being delivered into the inductor. My point is by this time significant additional energy was already delivered into the inductor which at 330uH, isn't tiny. Before the voltage rise could be detected and switch off the PWM. The energy will be dissipated through the zener and base-emitter. So my question is how are we so sure that it is "quite low", not a current spike which can kill the base-emitter junction? I'm just nervous because in my experience dumping even small energy into transistors via the base, was a quick death...
?
Well Hans,
I made an LTSpice model and the simulation shows that the proposed solution will work.
The voltage step 6 to 12 Volt occurs at 30 milliseconds with zero risetime
With the selected zener, the voltage transient is limited to 5.5Volt without much drama.
At 30.6 milliseconds the overvoltage protection goes into analog regulation, but this mode will terminate as soon as the processor has adjusted the PWM duty cyclle so that Vcc is back to 5V
The current through D4 peaks 6mA@..., which will not kill the NPN transistor...
The PWM frequency is 166kHz and I adjusted the dutycycle manually to obtain 5V
I know that the components are not exactly the ones you are using, but this is not significant to show the principle
I will be happy to send you the model if you are interested
73 de SM5EIE /Gunnar
