Raspi powered by car battery

[Steffenwolt]

hi all,
I have my Raspi 3B powered by the standard adapter and it works ok. Supply voltage is 5,2V.
Now i use fully loaded car battery (13,2V and 110Ah) regulated by a 3A 5V 3-pin regulator (LM1084) that provides 5,05V.
Now the Raspi boots, starts normally but after some (20secs) it reboots and keeps doing that ad infinitum.
I looked at the power line on the scope but no strange phenomenon visible.

Does anybody has a clue as to the possibe cause of this problem?
Has anybody seen this before?

I can't imagine that it is the voltage being slightly lower than that from the adapter

[KLL]

no idea, but i just read here
LM1084
Vin 7V ( max ratings )


even if i not find the correct datasheet,
but with your 13V battery you want burn 8V*3A in that little thing ( ? without cooler? )

you need a dc/dc switching supply
how about a "usb car charger"

[crofter]

hi

have you got on the regulator caps on the output a .01 uf a 10 uf ele cap

crofter

[drgeoff]

A 1084 will happily accept the 14 volts or so from a fully charged car battery but if you then also ask it to provide 5 volts out at a couple of amps without a good heatsink it will quickly get rather hot. It then protects itself and in doing so starves the RPi of power.

http://www.ti.com/lit/ds/symlink/lm1084.pdf

[PiGraham]

You should use a switch mode buck converter. Linear regulators with big voltage differentials waste a lot of energy and get hot.

[drgeoff]

You should use a switch mode buck converter. Linear regulators with big voltage differentials waste a lot of energy and get hot.

Linear regulators with big voltage differentials and supplying appreciable current waste a lot of energy and get hot.

FTFY.

[Steffenwolt]

As said, the 7085 will happily accept Vin (max)= 29V.
Yes, i have 10uF caps on input and output. No oscillations or any other strange things to be seen on the oscilloscope.
I'm not burning 8V + 3Amps. Current consumption on the Raspi is far lower. Yes it does dissipate. i have it mounted on a heat sink and blown by a smaal ventilator.
I know the argument about switch mode convertors and probably will use one later. I'm still in the trial stages. On the other hand; i have 880Ah at my disposal.

I apprciate all your comments but it still does'nt give me a clou about the problem i face.

Rgds

[PiGraham]

I apprciate all your comments but it still does'nt give me a clou about the problem i face.

I think KLL and drgeoff identified the problem.

At 40 °C/W and at least 3.5W (12-5=7V x 500mA) dissipation the junction temp in that regulator will be hitting 3.5 x 40 = 140+°C and triggering thermal protection.

A 1084 will happily accept the 14 volts or so from a fully charged car battery but if you then also ask it to provide 5 volts out at a couple of amps without a good heatsink it will quickly get rather hot. It then protects itself and in doing so starves the RPi of power.

http://www.ti.com/lit/ds/symlink/lm1084.pdf

[Steffenwolt]

Hi All,
Found the problem after all info you gave me.

Indeed it is the heat dissipation of the LM that kills the power. The LM is like a BBQ after a while.
It's the resilience of those LM's that never ceases to surprise me.

I just bought a small 5Amp switching convertor for €9,= . Cannot build it myself for that money.

Thanks for all input.

Tread closed.