battery power - efficiency modifications to the Arduino Mini

Hello,
I’ve been contemplating building my own low power battery operated Arduino from parts. I was wondering if modifying an Arduino Mini Pro would give you better or worst performance than building your own?

I looked at this tutorial on how to modify a mini pro to reduce power:

Basically, the tutorial suggested:

  1. Cut power to the power LED
  2. Cut power to 5V regulator

Once that’s done, will the Arduino Mini Pro be as efficient as making your own Arduino? A barebones Arduino would look like this:

ATMEGA328P-Pu, oscillator, two caps for oscillator,

There’s all kinds of surface mount components on the Mini Pro that I don’t know about. Wondering if those waste energy. But then again, does the form factor of the CPU matter? Maybe a dip size ATMEGA328P-PU isn’t as efficient as a SMD version of the same chip?

Assume the bare bones and Mini Pro run at the same clock speed.

Using the pro mini will be as powerful, and at $3 you cannot beat the price. The pro mini also has A6 and A7. Lots of us use them. The dip 328 has the advantage of being unplugged and replace.

I've modified pro minis for this. I took off everything except the crystal, the crystal's caps and the LED on pin 13.

It works perfectly and taking off the components takes about a minute, much easier than building your own.

arusr: I looked at this tutorial on how to modify a mini pro to reduce power: http://www.thalin.se/2013/04/arduino-pro-mini-low-power-modification.html

The page says: "Sleep mode after: 136 uA"

That's terrible, it should be around 5uA (mine was!)

I don't know if that's bad programming or he needs to take off more components.

fungus: The page says: "Sleep mode after: 136 uA"

That's terrible, it should be around 5uA (mine was!)

I don't know if that's bad programming or he needs to take off more components.

That's why I'm scratching my head. I don't know why his modifications are so high. 136uA is a lot on sleep. I have the picture you sent me before on what to cut off. So with those modifications, you're getting a 5uA sleep current? That's pretty good. I don't have any fancy amp meters to measure current so low.

Nick Gammon has a really good article on power saving here.

arusr: So with those modifications, you're getting a 5uA sleep current?

Yep.

arusr: That's pretty good. I don't have any fancy amp meters to measure current so low.

Not everybody can afford Flukes. My meter cost me $15... :-S

(Actually one of the famous illegal Sparkfun meters...)

There's ways to measure low currents even if you haven't got a meter with uA. eg. Put in a resistor and measure the voltage drop across it.

With uA you need quite a big resistor to get a decent reading but I found it runs long enough on a capacitor to swap in a 10K resistor and measure the current draw. I only did it a couple of times to see if it agreed with the uA reading (which it did). If you want to do it a lot you can put in a switch to bypass the resistor when you're not taking a reading.

If you are using the measurement across a resistor to measure sleep current, short out the resistor until the chip has powered up gone to sleep, otherwise the voltage drop while powering up can cause it to hang up. Then remove the shorting link, and measure the voltage across the resistor.

If you expect to get the current down to 1mA or below, and are using a 4v5 battery, you can afford to lose 1 volt from the battery for the test. ( in sleep mode it should be well under 1mA )

So connect a 1k resistor in series with the battery, with your multimeter connected across the resistor, set to the lowest millivolt DC range, which on my cheap meter is 400mV. Connect a shorting link across the resistor too. Connect the battery, and when the micro has powered up and gone into sleep mode, you can remove the shorting link and take the reading from the meter. Every millivolt will represent 1 microamp drawn from the battery.