This is a robotics project where we are driving a pump with a power MOSFET that is driven using a PWM signal to control pump speed. The power source is a 6-Cell Lithium Ion battery pack. Both the pump and the DC/DC converter are connected to the high side of the battery pack. When we are not switching the pump (at either zero or full power) the 5V output is clean. However, when the PWM signal switches, we see spikes in the 5V output reaching just over 8V for ~4us (Single 5V Spike.jpg, 5V with pump switching on then off.jpg). We are hoping to use the 5V output to power a Raspberry Pi Zero W, which is rated to handle up to 5.25V.

Looking at the battery voltage for a 25% duty ratio, the battery voltage begins dropping (Battery Voltage.jpg), and then peaks about 4V above nominal when the pump is switched off (we have a flyback diode to minimize this voltage spike).

We have tried smoothing the 5V output with a low-pass filter (1250Hz cutoff) and a 5V zener diode. We are using a small computer fan for a load.

Is this typical behavior for a DC/DC converter when the input voltage is changing? Would there be a way of mitigating these spikes. Alternatively, are these very short spikes in voltage acceptable for the Pi Zero?

 

Have you got a diode reverse connected across the load?

 

This is a robotics project where we are driving a pump with a power MOSFET that is driven using a PWM signal to control pump speed. The power source is a 6-Cell Lithium Ion battery pack. Both the pump and the DC/DC converter are connected to the high side of the battery pack. When we are not switching the pump (at either zero or full power) the 5V output is clean. However, when the PWM signal switches, we see spikes in the 5V output reaching just over 8V for ~4us (Single 5V Spike.jpg, 5V with pump switching on then off.jpg). We are hoping to use the 5V output to power a Raspberry Pi Zero W, which is rated to handle up to 5.25V.

Looking at the battery voltage for a 25% duty ratio, the battery voltage begins dropping (Battery Voltage.jpg), and then peaks about 4V above nominal when the pump is switched off (we have a flyback diode to minimize this voltage spike).

We have tried smoothing the 5V output with a low-pass filter (1250Hz cutoff) and a 5V zener diode. We are using a small computer fan for a load.

Is this typical behavior for a DC/DC converter when the input voltage is changing? Would there be a way of mitigating these spikes. Alternatively, are these very short spikes in voltage acceptable for the Pi Zero?

Yes, a DC-DC converter might do such a thing if there are sudden changes at its input. One workaround would be to use a large cap at its input (I normally use 4,700 uF) preceded by a diode, preferably a schottky. The diode would prevent the cap from sudden voltage drop-downs being fed back into the supply, while the cap would also serve as a surge stabilizer at the converter's input. Another large cap at the output might prove convenient.

 

Yes, a DC-DC converter might do such a thing if there are sudden changes at its input. One workaround would be to use a large cap at its input (I normally use 4,700 uF) preceded by a diode, preferably a schottky. The diode would prevent the cap from sudden voltage drop-downs being fed back into the supply, while the cap would also serve as a surge stabilizer at the converter's input. Another large cap at the output might prove convenient.

I wanna add something: it's better to select tant cap or polymer cap in output since they have a much smaller ESR, it will be help to lower the spike.

 

I wanna add something: it's better to select tant cap or polymer cap in output since they have a much smaller ESR, it will be help to lower the spike.

The DC-to-DC Converted used in message #1 has an absolute maximum capacitance of 680uf in the output.
Adding capacitance above 680uf to the output, will cause the converter to become unstable.
It is best to fix the "Step" voltage at the Input, before it gets to the buck-converter[/B]

 

A great video on converter noise and solutions -

Regards, Dana.