Atmega328P at 3.3V: BODLEVEL=3'b101, F=10MHz, still doesn't operate

Hi All,

I'm having trouble getting a stand-alone 328P device working at 3.3V.

I checked the brownout detection fuse bits, and they are encoded to 101b (max brownout of 2.9V).

I have a 10 MHz resonator hooked up to the device, well within the Freq/Voltage curve in the datasheet. Please not that IT WORKS FINE at 5V. I created a new entry in boards.txt and specified 10MHz, and it when I program the device on the Uno board, pop it out, and drop it into my protoboard, it works perfect at 5V. My test program is a blink on pin 0 at 1sec period. Works like a charm at 5V.

When I drop the supply down to 3.3V, it doesn't toggle the output at all (it is hooked to a scope). The resonator crystal is oscillating at 10MHz, but no output changes.

I'm definitely supplying >3.2V according to my scope, so it shouldn't be triggering brownout according to the fuse settings that I can read when it is plugged into the UNO board (using avr/boot.h macros for fuse reads).

What am I missing? I get the feeling it is something obvious...

Thanks, pjt

My test program is a blink on pin 0

An LED? What color?

it doesn't toggle the output at all (it is hooked to a scope)

Pin side or ground side?

An LED? What color?


Pin side or ground side?

North side.

Anyone with real insight out there?

Use the fuse calculator.

I've had ATmega8 boards with resonators that seem to want "full swing crystal oscillator" specified, rather than "low power crystal oscillator" configured in their fuses. You said you're seeing oscillation, so that SHOULDN'T be the problem, but it might be worth a try.

What happens if you disable the brownout detector?

Hi westfw,

Thanks for the ideas.

What happens if you disable the brownout detector?

Nothing changes, works at 5V, doesn't work at 3.3V. (I programmed the extended fuses to 0xFF during the bootloader burn, and read them back to verify.)

full swing crystal oscillator

Ah, I never thought of that. I did notice that the swing on the oscillator is lower with the 3.3V supply than the 5V supply. I'll check out the specs and try a few other resonators and crystals I have lying around.

Thanks, pjt

It was the resonator. I switched to a crystal and two caps and it works. There’s no part # on the resonator so I can’t check a datasheet. Frustrating, however, since I didn’t expect that, but thanks for the help.

(does your resonator work if you change the fuses?)