Strange current behavior on arduino nano 3.0 with different powering

Hi I have the following problem.

Setup: nano 3.0 (clone from ebay/asia) Sketch: the blink example, delay set to 5s (5s on, 5s off).

I connect 9V block battery to the nano to Pin GND, and VIN. Between the Battery and nano I connect an amperemeter to measure the current it uses.

Two situations: 1. The nano is total powerless, not battery, no USB Power. Now I connect the battery and it uses about 30mA.

  1. The nano ist connected to the battery AND USB. It consumes the same current as above. Now I disconnect USB and the nano goes to battery power but now it uses only ~10mA (depends on battery charge), its always about 20mA difference between situation 1 and 2. Even when i press the reset button it remains in the lower power consuming mode.

I tried to use power saving techniques (avr/power.h), clocked it down,... the behavior is the same: On second situation it uses much less power, again a difference about 20mA.

I tested instead of the VIN with the 5V-PIN (with power from an LP2950 5V regulator), the same behavior. Probably its not good to put 5V AND USB cable at the same time in the nano, only did it one time.

What is the reason for this? How can I make the nano to consume on lower power level without first connect to usb and disconnect it?

I tested this with an Arduino Uno R3, it has not this behavior it always consumes the same amount of mA on both situations.

It could be the FTDI chip thas is turned on in one situation and off in the other.

The FTDI chip provides the 3.3V.
Could you check the 3.3V in both situations ?

After pulling the usb cable its 3,3V, 3,3V is always available. In nano product description (http://arduino.cc/en/Main/ArduinoBoardNano) they say 3,3V is ON ONLY when running from USB and off if not running from USB. I've googled around and other people discovered too (german RC-Helicopter forum) that the 3,3V is always available. Some of them have official nanos, no clones.

Yes, I noticed that paragraph is not correct. Perhaps for old Nano boards. On my (Ebay clone) Nano board, the FTDI is always on.

When the 3.3V is always availabe, the FTDI chip is always on. Did you apply a load to the 3.3V ? Like a 1k resistor.

I don't know what else could cause the 20mA.