atmega328p 3xAA NiMH


I am creating project where I want to reduce the number of required parts + power consumption to minimum. First I wanted to use atmega328p on 1MHz. I was planning to power it with 3 x AA NiMH batteries which means 3.9V when fully charged. I planned to use a schottky diod to low down the voltage which should be then at 3.5V when fully charged batteries. I would like to use that project in connection with Sigfox and NRF24l01. Both are able to work under 3.3V same as the atmega328p. I wanted to remove the voltage regulator completely if possible. Do you think that this is a good way how to power the devices? The voltage should be then from 3.5V when batteries fully charged to 2.7V when discharged.

Later on I may switch to 8MHz atmega328p as it can happen that the software serial communication will be not sufficient with the Sigfox module. But according to datasheet even on the 8MHz should be ok with the voltage of 2.7V. In the future I also wanted to charge the NiMH batteries by the small solar panel to let operate the devices for longer time. Anyway I am using the power down mode of all devices and powering them only for about 1/8 of time. When the solar panel would be used, is it better option to use for example just one NiMH battery with step-up regulator and small solar cell connected with diode directly to the battery? I would use for example a 2V solar cell + diod + 1 NiMH battery and then step-up voltage regulator to go to 3.3V.

Any advice and help will be appreciated.


Hi, i'm no expert but as far as i know step down or buck converters are more efficient than a zener or three terminal regulator. If i understand it correctly you wonder if you can leave out the regulator completely? If you have a lab supply you can just test if your circuit works at lower voltage or not.(does not seem to be a good idea)


No, you cannot do that. While NiMH battery is 1.2V nominal, it has "much" higher voltage when fully charged - 1.6 or even 1.8V IIRC. 3 batteries may lead to voltage much higher than expected 3.6V, easily damaging the nRF. I know nothing about Sigfox but ATMega works from 1.8 to 5.5 V and nRF from 1.9 to 3.6V. They will easily work from 2 unregulated NiMH cells - maybe a way worth to consider?

Running at 1MHz is not usually best. Firstly you'll have to compile and program a custom bootloader binary into the 328, which is usually a bit painful, and the maximum baudrate supported will be low. Also you might want to rework the analog clock prescale factor at this speed to analogRead() isn't needlessly slow.

Secondly if you want minimum average power draw it may be better to run quickly to reduce the time other peripherals are powered up. Sleep everything as soon as possible, basically.

Step-up converters will consume power even when the load is sleeping, you have to look around for those that have low quiescent consumption in that situation. Trickle-charging a NiMH cell is a reasonable approach for simpler solar power, it avoids need for a charge-controller chip. So long as the trickle rate isn't too high for the cell...

NiMH fully charged is about 1.35 - if its much more than that, its been over-charged or its a dead cell.

Hi all, thanks for your answers. I succeed with the 1MHz atmega328p and the code which checking the incomming messages on NRF after 8seconds of power down mode. Then after 15 minutes it send all received data to internet via Sigfox module. When connected to stable power supply it looks there is no problem and all run as expected. For now I did only couple of testing but looks fine. I also tested to run it on 3 AA bateries which were not fully charged and I measured the voltage before I connected it to the device. It works ok. But anyway I will also test it with 2 AA batteries which should have from 1.8 up to 2.8 V. I will see if it is working. I still think that the power supply without the step-up or step-down regulator would be the best to last the device for longer time. I will also try to use it with voltage regulator MCP1702T-3302 with 3 or 4 AA batteries. Do you think there would be a big difference in the power consumption when this voltage regulator used? The device is most of the time in power down mode.

One more thanks for your comments above.

I will also try to use it with voltage regulator MCP1702T-3302 with 3 or 4 AA batteries. Do you think there would be a big difference in the power consumption when this voltage regulator used?

What does the datasheet for the MCP1702T say about its power consumption ?

The datasheet can be seen here:

I use the 3-Pin TO-92 version.

Here are some details:

Features • 2.0 μA Quiescent Current (typical) • Input Operating Voltage Range: 2.7V to 13.2V • 250 mA Output Current for Output Voltages ≥ 2.5V • 200 mA Output Current for Output Voltages < 2.5V • Low Dropout (LDO) voltage - 625 mV typical @ 250 mA (V OUT = 2.8V) • 0.4% Typical Output Voltage Tolerance • Standard Output Voltage Options: - 1.2V, 1.5V, 1.8V, 2.5V, 2.8V, 3.0V, 3.3V, 4.0V, 5.0V • Output voltage range 1.2V to 5.5V in 0.1V Increments (50 mV increments available upon request) • Stable with 1.0 μF to 22 μF Output Capacitor • Short-Circuit Protection • Overtemperature Protection