Hi all, looking at the new board soon to arrive from GHI, and the existence of many other Linux boards on the market, I have decided to sketch a revised design of my setup, but quickly ran into the usual low power challenges.
My system is running on batteries, and needs to be listening for interrupts all day and react accordingly (2 times pr day). Under specific situations (2-4 times per day) it needs to be fully active for a few minutes, sending data (1MB) to a server, and power down again.
My current design is a G120 based system, which can be forced into reasonable low power mode, but I am not really satisfied with the setup.
Thats why I am looking for best practices for this type of setup with for example a small daughter-card that controls the bigger linux based setup.
I need quick prototyping options for the setup, so all things NETMF is preferred.
What would you do?
The reason for me not being satisfied with the current solution is simply because it have not been able to reduce current consumption to less than 20mA. Furthermore, the current available functionality with the G120 based solution in NETMF is not as rich as the one available on the Linux ones.
On Beaglebone Black, HDMI and LAN remain active even though we put system in deep sleep mode. After removing power to HDMI and LAN, The total current board use is 18mA with 5V. So, total standby power we can achieve on Beaglebone Black board is around 0.1W. HDMI alone consume 21mA. LAN consumes 11mA.
I want to go further down.
I am guessing the new GHI board will consume less?
I’d be going a different direction. I’d go a lower power device that’s doing all the interrupt handling and collection that’s needed, then controlling power-up of the G120 / higher spec device as needed.
Thats my ideas as well. But what easy prototyping smaller board would you use then…? And how does the setup look for real, any experiences? Powering up and down is more complicated than first anticipated.
for real, it looks like a bigger battery and a larger solar panel to replenish it, so you can just use the G120 and be done with it.
its absolutely non-trivial to power down and power up a device at the control of another device - especially when you’re talking about booting an OS. But it’s all about the trade off… power cost vs simplicity. If you need to have a complex system to drive the power requirement down, then that’s what you do.
Any Atmel device can do this. Very low power in the uA range when sleeping. You then use this to wake up your G120 by powering it up. If you really need such low power then this is the only way as the G120 can’t go low enough. The LPC178x processor is capable of very low power but this is limited by the G120 design. As your code is running in SDRAM it needs to keep this alive and refreshed.
@ njbuch, why not use a power supply with a on/off pin like this one.
You could have your IRQ turn on the power supply which in turn would power up you Linux board and do what you need to do.