8 replies to this topic

[kneave]

Howdy, I know this will likely be a "how long is a peice of string" question but whats the power usage of the Netduino Mini? I'm thinking of making a cycle computer powered by a hub dynamo, as such I've not much power to play with. I'll be using I2C, UART, SPI and likely 3 PWM outputs. I've had a look at the datasheets for the ARM chip but I wasn't sure what power usage for the clock speed the Netduino runs it at. Sorry if it's a rookie question but I couldn't find an answer in the forums. Thanks, Keegan

[Inquisitor]

I wish I could point you to something... and was about to point you to this Comparison until I realized the Mini on it has no info. I have a similar low power requirement and am watching your thread.

Maybe some real nice Mini owner would be so kind as to hook up their Mini and measure the current used... It would even be REALLY NICE if they could give us a "nothing running state" and with some mild loop (say 50 ms) that reads and writes several pins (with nothing wired to it).

I would gladly do it, but I've been waiting for a Mini for a while, but everyone seems to be out of stock. I've got "notifications" set from several vendors.

BTW: Do you have any recommendations for one of these HUB dynos? Something that wouldn’t break the electronic hobby bank.

[kneave]

I've got one somewhere, I was just wondering if it was known info as it'll be a while before I can get some measurements. I've no desk or anything at the minute so it'll be a while before I can have a play. I grabbed a Shimano hub off eBay, one of these from that seller; http://www.ebay.co.u...=item43a9726507 I've heard good things about the Shimano hubs and at £15 it was a no brainer. Rebuilding my wheel could be interesting though. It's not something I've done before but should be fun.

[CW2]

Mini owner would be so kind as to hook up their Mini and measure the current used...

Just measured and updated the wiki. Writing and reading two pins (with nothing attached) in 50 ms loop consumes about 25.3 mA.

[Mario Vernari]

Some day ago I was trying to answer to an user about the low-power state of the Netduino. Of course, there's no such a feature in any of the three Netduino models. So, I was wondering whether this feature will be available in the upcoming 4.2 version of the firmware... ...but... ...taking a look at the various schematics, it seems there's NO WAY to achieve a low-power status, unless HACKING the board. Assuming the MCU will turn in "idle-mode" (about 26uA): MINI- the RS232 converter itself requires 1mA typical without any load (up to 5mA when loaded); STD- there's no RS232, but the power led (approx 2mA) and the quiescent current (5-10mA) of the voltage regs; PLUS- same as STD, plus an additional 3.3mA when the Ehternet chip is in power-down state. All that do *NOT* include pullups, and extra hardware connected to the board, which may require more current flowing. Just for a comparison, a 9V Alkaline battery offers a capacity of about 400mAh. That leads to an average duration of 40 hours (supposing an average current of 10mA). So, it would be useful to think about some trick to put in sleep-then-awake the Netduino boards. Ideas are welcome! Cheers

[Inquisitor]

Just measured and updated the wiki. Writing and reading two pins (with nothing attached) in 50 ms loop consumes about 25.3 mA.

Thanks CW2.
Good for a .

[CW2]

So, I was wondering whether this feature will be available in the upcoming 4.2 version of the firmware...
...
...taking a look at the various schematics, it seems there's NO WAY to achieve a low-power status, unless HACKING the board.

If I remember it correctly, Chris Walker has mentioned that switching to low-power state is under development - .NET Micro Framework provides PowerState.Sleep() method to implement various sleep levels, although it may need some additional support for selectively switching various hardware modules on or off - I am not sure if the hardware modules are powered on when the driver initializes (during system startup), or on-demand when the feature is actually being used (i.e. managed class is instantiated in the application).

[Mario Vernari]

If I remember it correctly, Chris Walker has mentioned that switching to low-power state is under development - .NET Micro Framework provides PowerState.Sleep() method to implement various sleep levels, although it may need some additional support for selectively switching various hardware modules on or off - I am not sure if the hardware modules are powered on when the driver initializes (during system startup), or on-demand when the feature is actually being used (i.e. managed class is instantiated in the application).

That's okay...but my post is based on that assumption. However, there's no way to avoid extra currents that will suck most of the battery energy during the "idle state".
IMHO, there are two choices:

• an external switch that powers on/off the whole circuit (e.g. 555);
• Chris will release a new model of Netduino having the hardware that full-support the low-power feature.

My overview is assuming a current of 5mA (approx) when the board is sleeping. When running the current is rising to tens of mA (e.g. 50mA).
A common requested app is to wake up the Netduino, collect a read from a sensor, then shut it down. The time of a read is very short: we may consider 3 seconds (and it's very much)!
Instead, the time of sleep could be very long: for instance, 15 minutes.
Now, a simple computation:

I_avg = (I_run * t_run + I_idle * t_idle) / (t_run + t_idle)

using the supposed values, the average current is about 5.1mA.
So, most of the energy is totally wasted.
Cheers

[kneave]

Thanks for the data CW2, the idea of the sleep mode would be wonderful though so hopefully it will appear in the future.