Hi all,

I will keep this as short as I can.

I am running a Raspberry Pi5 On a Waveshare screen and running Realdash software. This is all working well until I start the car. The car has a 2kw starter motor which is reasonably powerful and probably quite noisy, the Pi crashes and gets stuck with a red power light. It then requires a restart but will do it 9 times out of 10.

This is the list of the power circuit which doe snot seem to dip in voltage which is why I think it is more a noise issue. I have duplicated power setups, 1 for the screen and 1 for the Pi.

12v Buck Boost 3A - https://www.aliexpress.com/item/100...|x_object_id:1005007198118897|_p_origin_prod:

5v Step Down Buck Boost 3A - https://www.aliexpress.com/item/100....order_list.order_list_main.74.1ec01802uLiyqt

I am confident that I have enough amps to run the devices as it all powers up and runs well but as soon as I crank the starter motor I get this red light issue. I have scoped the power on cranking and see a solid 12v from the buck boost.

I have also tried running it off a PD charger from the cigarette lighter socket. This again worked until I starter the car which in turn has damaged the usb outputs on the Pi, the can adaptor and the keyboard dongle I had plugged in.

I am looking for some advice on either building or buying a robust power circuit that will stop this failure.

Your help is appreciated.

 

How exactly do you have everything wired together? The problem usually isn't starting the car itself, but the magnetic field of the solenoid and starter breaking down after you release the key can cause a rather large voltage spike.

Typically a healthy capacitor and TVS diode is a good start, but I am not an expert by any means and have had decent luck going with the bigger is better approach.

https://www.ti.com/lit/an/snva717/s...88667&ref_url=https%3A%2F%2Fwww.google.com%2F
https://www.littelfuse.com/assetdoc...ssetguid=e4e5626b-a7b4-4098-8b9d-49424e2cfdcc

 

I have scoped the power on cranking and see a solid 12v from the buck boost.

Put your scope probe on AC coupled - it's likely on DC you won't see the fast spike. Also, its possible the spike is on the ground side, or on any other link between the Pi and the car. What connectivity are you using?

 

The mains supply for the Raspberry pi5 has to be rated for a steady- solid 5A.
So this should also apply to any other source. I would concur with @geekoftheweek that some heavy duty smoothing (Cap) on the 5v may not hurt !?

 

I didn't look into the PI5 power requirements. 5A is a bit more than anything I ever created.

 

In my experience, the RPi is very sensitive to it's power supply- and does suck large spikes of current.
For reliable operation you should over-spec the PSU substantially over nominal power consumption.

 

The official AC-DC adaptor for RP v5, supplies 5.1v @ 5.0A and 9.0v @ 3.0A.

 

In the milliseconds before it starts turning the starter motor is basicly a short circuit across the battery, and so there will be a bit of a "power glitch" that may not even show on a scope, depending on where the scope is connected. The voltage drop is in the conections and terminations, not all of it is in the battery. So that glitch is there even if it is not seen.
Another consideration is that on some vehicles the accessory power is switched off during cranking the engine. Of course, once the starter stars spinning, there may be another spike as the supply voltage is removed .
What all of that implies is that a serious amount of supply voltage filtering capacitance is required. Just like others have suggested.

 

The Raspberry Pi, being a mini, but poweful computer for its size, is very demanding, and its PS, stability being one !

 

The secret trick would be a backup battery feeding the voltage regulator Diode isolation from the vehicle 12 volts and a high value resistor to provide a small trickle charge The battery power should be able to support the actual load for at least a minute.