I'm working on a low-power energy harvesting project, where I'm considering using an ATtiny85 to perform some minor tasks. The ATtiny is put in Power Down mode most of the time and does not consume much power in this mode.

Since the power source is unstable, and/or could be absent from time to time, I've been enabling the internal brown-out detection (BOD) by setting the fuses. In this way, the ATtiny is reset when the voltage gets too low. This is to ensure that the microcontroller starts working properly again when sufficient power returns. This also seems to work, so that the chip starts running again when the voltage is above a certain threshold.

However, when the voltage drops below the BOD level, I notice a considerably larger power surge through the ATtiny, even when all peripherals are disconnected. It seems to me that some sort of gate is opened inside the chip when the BOD fires, and the current is limited by some sort of internal resistance. When the voltage increases somewhat above the BOD level, the large power surge stops.

I've been searching through the documentation/datasheet and forums online to find out why this happens but can't seem to find some sort of explanation or ways to prevent this from happening. Since this is a low-power energy harvesting project, it is important that the internal protection from low voltages does not contribute to the problem it detects.

Any useful input would be appreciated.

1 Answer 1


When the Vcc level is below the BOD activation voltage, the chip is held in a reset state...

enter image description here

...which uses significantly more current than the chip does in sleep state...

enter image description here

You are probably better off disabling the BOD and instead using an external Voltage Supervisor chip to turn the power on to the ATTINY when there is sufficient voltage and turn it off when there is not. There are many of these chips and they can use less than a microwatt when quiescent.

  • 2
    Thank you for replying. The graph seems to offer some explanation to the matter. However, I'm observing ten times the current consumption, running at 1 MHz and BOD level at 4.3 V. According to the graph that should give a current consumption starting at ~0.1 mA? I'm observing >1 mA. As you write, I might be better off disabeling the BOD and using an external supervisor, but I'm stil wondering if there is anything I could do to reach the 0.1 mA level.
    – hmf
    Jan 31 at 23:29
  • Are you testing the ATTINY in isolation or in your circuit? If you are testing in circuit, I'd bet the additional drain is coming from other components and possibly the way they are reacting to the floating outputs on the ATTINY. Better to test the ATTINY alone, connected only to a variable power supply. Start at a voltage above the BOD threshold and slowly lower the voltage to below the threshold and watch what happens to the current. My guess is that it will match the chart above.
    – bigjosh
    Feb 1 at 18:46
  • I've been testing in isolation, as you described. I found some useful information in section 10.2.6 "Unconnected Pins" in the datasheet: "... the [internal] pull-up will be disabled during reset. If low power consumption during reset is important, it is recommended to use an external pull-up or pull-down." I've tested only a little with 1 MΩ external pull-up resistors, which seemed to bring the current consumption further down. Still not quite matching the chart yet. I haven't progressed further with testing, but it seems this might be at least part of the issue."
    – hmf
    Feb 3 at 12:52
  • The chart in section 22.11 also states that it's "Excluding Current Through The Reset Pull-up". So there's also that. :)
    – hmf
    Feb 3 at 12:55
  • 1
    Did some more testing now. External pull-up resistors on all floating pins. Current consumption down to ~0.22 mA right after BOD. Still a lot compared to ~26 µA in Power Down mode, but I guess that's more what is to be expected with BOD level at 2.7 V and taking into account an internal Reset Pull-up Resistor rated at 30-60 kΩ (according to "21.2 DC Characteristics" in the datasheet). Worst case that should give a current leak of ~90 µA + ~60 µA ≈ 0.15 mA. I guess this is as far as I get without an external Voltage Supervisor. :)
    – hmf
    Feb 3 at 22:37

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.