ArduBerry Atmega328 burnout


Hey Team,

I love my ArduBerries. I had them rocking a new device I made (test tube stirring pH-growth meter for bacterial suspension). Then neither would talk to my RPi. I tried everything on the software end, tried a new out-of-the-box RPi, but no love.

I switched the Atmega328 chip from an Uno, and voila! The ArduBerry listens and hollers back. (putting the old ArduBerry chip into the Uno makes the Uno deaf & dumb).

I had the AREF hooked to the 3.3V to get my device working (needs this to read from an analog light sensor) - could startup with these two connected have fried the chip? Any other thoughts on how to avoid this?

Otherwise I had 4 small lasers (like laser pointers), 4 micro fans (0.14A, 5V) connected with pull up by code-hacked PWM (10mS off, 100mS on), and 4 Adafruit analog light sensors. No shorts, for sure.

I had this exact setup with an Uno before, with no problems…

Help! Seriously, this is for a set of experiments that would help us move forward on stopping cavities in young children, and there isn’t another tool to do this.

Peace and Thanks,
Jeremy Horst, UCSF

PS - of course I ordered a fleet of Atmega chips in case this happens again.


Someone suggested that unplugging everything from the ArduBery before uploading a new script might avoid the Atmega328 meltdown - is this true?


I am really not sure what was the cause for the Atmega chip to stop burning. Since you have a lot connected to the board, can you make sure that anything connected to the IO pins does not draw more than 50mA a pin. You risk damaging the chips if you try to draw more that 50mA. I don’t think that setting Aref to 3.3V would be the problem. If you have something connected to pins 11,12 and 13 during code upload, it might cause problems sometimes, but the problem is pretty restricted to being unable to upload the code and does not burn the chip.

One last thing, did you verify that the chip is burnt and were you unable to burn the code on the Arduberry when you started or was there some other problem.

Sorry for the troubles.


Thanks for the feedback. I’ll look out for the >50mA.

I was able to upload code on each of 2 ArduBerrys at first. then not at all. There was no pin in 11,12,13 at any point (thanks to the forums). With the new chip I am again able to upload code.

Technically I did not verify that the chip was burned - if you want me to, please instruct me (I have an oscilloscope, multimeter, etc.)


The easiest way to check whether the chip is burnt or working is to run this sudo avrdude -c gpio -p m328p. If you get a signature match confirmation that means the chip is working. You can try it with the working as well as dead to to make sure if it works or not.



Oh yeah, avrdude cannot connect to the nonworking chips:
""“avrdude: AVR device not responding
avrdude: initialization failed, rc=-1
Double check connections and try again, or use -F to override
this check.”""
Same thing with -F.


Looks like you have indeed burnt the Atmega chips somehow. We have never seen that happen with the Arduberry. It would be great if you let us know what the problem exactly was if you are able to debug it.