[beaglebone] battery charge and operation


From what I’ve read and also searched on the group, it seems that currently the operation of the beaglebone using a Li-Ion battery isn’t guaranteed or, at most, not fully defined yet.

According to the SRM paragraph 7.12.6 the battery isn’t able to power the beaglebone due to the voltage drop on the 3,3V LDOs (400mV). If the data are correct this would make the charger connection kind of pointless because we can charge a 1 cell Li-Ion but we can’t use it to run the system; we’d end up with a fancy Li-Ion charger though :slight_smile:

However, the datasheet of 65217x mentions that:
a) BAT switch voltage drop is typ. 60mA @ VBAT=3,3V , I=1A (is this VBAT-VSYS?)
b) LDO dropout voltage is max. 200mV with VIN=3,3V , I=100mA

So this contradicts the SRM and hopefuly leaves us with hope that battery charge and operation (without exotic step up converters) is feasible.

I’m about to start designing a cape to enable battery charge and operation but first I’d like to confirm if such thing is possible.

Has anyone tried this? Did anyone was able to run the beaglebone from a Li-Ion down to 3,6V reliably?

Thanks in Advance,

The dropout is actually 200mv and will be updated in the next release of the SRM. That means you can run the battery from 3.7V to 3.5V and then the 3.3V rail will begin to drop out of specification. That is not a lot of time before the battery will no longer be able to power the board. But it will power it for a little while.

The fact that you cannot run the USB slot due to no 5V being present is still the case.



I also think of powering the BeagleBone with a (backup-)battery.

If I don´t have a LiPo-Battery with a NTC I think that I have to
connect a resistor with a value of around 8.82kOhm (10k||75k) from TS
to ground to simulate a temperature of 25° Celsius, because otherwise
the battery will not be charged (temperatur outside of the range 0°C
to 45°C). Can you confirm this?

And what about the "CH_EN"-flag (charger enable/disable) in the
"CHARGER CONFIGURATION REGISTER 1" (page 49)? Is it set to 1 by
default, or has this to be done via the I2C-interface?

Best regards,

Thanks for the information.

I have another question now, related to battery operation.
The way I see it, If I want to have the USB host operational, I’ll need to use a step up regulator in order to produce a 5V supply that would allow the USB to function.

When the TPS65217b switches to battery operation, I read somewhere that the PMIC regulators are internally connected to BAT in order to minimize losses in the internal switches. Does this mean that while the beaglebone is running on a battery, se can apply a stepped up 5V in the 5V_SYS safely?
I read through the TPS65217 datasheet but couldn’t find references to this scenario at all.


The battery is the battery. When the TPS65217B is switched to it, the voltage at 5V_SYS will be the battery voltage. The 5V_SYS would need to be stepped up prior to supplying the USB 5V regulator and before the 3.3V regulators. The other rails can run from the battery supplied 5V_SYS rail that connects to the battery.


I understand now. Probably something I read before triggered this confusion on my head.

I was trying to avoid having to step up all the way to 5V before the TPS to avoid the loss of efficiency. Is it possible that future versions of the beaglebone allow to split 5V_SYS and the USB Host 5V rail and provide pins for the capes to apply this voltage?
The idea was to
a) Power everything, except USB host, from the battery input of the TPS65217.
b) Allow capes to apply a 5V supply to the USB Host rail, this allowing USB running on batteries, but maintaining a high level of efficiency.
c) When running off a DC adapter or a USB port, the USB Host 5V rail would be connected to the supply 5V as currently does.

These are just hints in the hope that they may be useful.
The bone is already a great device as it is.


As to future versions, maybe. We are shipping so fast and such large quantities, I can’t find a good place to inject such a large design change into the production schedule. I have an idea as to what I want to do, but it has not been finalized in a design. The focus would be to supply a solution with mimimal cost impact to th eboard. That is not easy to do!

Your a) suggestion will limit the battery discharge to 3.5V, so that limits the battery time on a full charge. It is better to run the 3.3V LDOS off of the 5V as well.
Your b) suggestion would complicate things as you would have to have a Cape for USB to work. And if it could also supply power in addition to the main board, makes for a dual source issue.
Your C) suggest creates similar issues in that when attached that main rail is also 5.5V because the USB in and DC in is also 5V. That means the regulator that would supply the 5V form a DC source and a battery at say 2.7V would need to handle both extremes.