Low Power Consumption nRF8001

[jamilaaria]

Hi guys,

Do you have any suggestions on how I can lower the power consumption. I am currently connecting - sending data - disconnecting.

I have the following set the parameters as: Min/Max Interval = 15/15, Slave Latency = 0, Timeout = 3000 ms.

Thanks!

[run_ar]

Since I have no idea what your current power consumption is, the only suggestion I have is to make sure you go the deep sleep whenever you are disconnected and not advertising. Remember that all input pins need to be in a defined state to avoid power leakages, ref GPIO overview.

[jamilaaria]

The whole board (including the uC,nrf, and accel) consumes an average of 400-500uW for 30 cycle time, but I need to lower down this to around 200uW. I have this cycle: wakeup - connecting - sending data - disconnect - sleep and I measure a peak of 6mA (drawn by the nrf8001). 

One other option that I am thinking is to keep the connection during the 30 cycle time rather than connect - sending - disconnect. Because I think that advertising must be consuming more power. Will you have any suggestion for this?

[run_ar]

The nRF8001 should draw less than 500nA while in deep sleep. Please check the state of the input pins in this case, i.e. make sure they are not asserted by the application mcu.

You can calculate the theoretical power consumption using the power consumption calculator in nRFgo studio.

Please also note that the nRF8001 is an old device, so it is recommended that you use a newer device such as the nRF52 series.

[jamilaaria]

Thanks for the recommendation, however, I will not be able to use a newer device yet. 

I used the nRFgo studio for the average current consumption, and theoretically I should be getting  ~133uA in average.

However, when I tested it, I got about ~230uA in average. Yes, I have observed that nRF8001 draws less than 500nA while in deep sleep. My concern is that there is a high peak during advertising which peaks ~6mA. And also, there are instances where I do not see the peak, but I do not understand why.

Would you have any recommendation to further lower down the current consumption? Thanks!

[run_ar]

How are you measuring this? looks like the sampling rate might be to low to properly measure the current consumption. Spikes during radio activity is normal as that would require a lot of power.