Hi for our application we need to retain the correct time when the device is disconnected from the main battery. To do this we have to lift the pin and jump a wire to a coin cell. In the future it looks like this wont be possible on the BGA package you will use next.
So I was wondering if in future revisions you could connect the Vbat pin to the rail with a zero ohm resistor. Then we could just pop of the resistor and jump a wire to the pad. It would be even more great to have a connector but I get that this isn’t something most people need.
I was looking at the schematic for the OpenMV RT I suppose we can take C69 off the board and Jump a wire to the pad. Does anyone know the footprint of that cap will it be an 0603 or an 0402 or smaller?
Hi, the design of the OpenMV Cam RT is such that it will be able to be powered by a battery and have a 50uA draw.
You should not need to modify the design.
I thought of this already. When you activate low power mode the main 3.3V regulator is powered off and only the low-voltage domain on the MCU is powered which will keep the RTC on and allow for wakeup via the wakeup pin or via the RTC.
Also, please note, I’ve already identified some things to change for the final rev.
I’ll be moving the battery charger to be powered via USB & VIN versus just USB as it stands right now so that you can charge the battery on VIN. This does impose a current limit on VIN to 500mA. But, this ensures that you can then use your OpenMV Cam with a battery and that battery will charge no matter the power source.
On the prototypes the battery charges via USB only… I did this because VIN should be continuous and I didn’t want a current limit on it. But, then, this means the battery can’t charge and keep you alive during low-power states. It will just run down until empty.
Re-reading what you posted above. You should treat the battery you connect as something that should be disconnected once the system is deployed. Once I do the fix above the battery will be charged by any power source and you should be able to keep time as long as you enter a low-power state to not drain the battery completely.
That said… there’s no I/O pins left over to tell what when to enter the low power state. So, your program will need to be written to kinda by default get to a low-power state as fast as it can.
Let me know if this works for you. There’s still time to do minor low risk changes.
I could give you a super cap DNP pad on the back that’s on the SNVS rail:
Thanks for responding so quickly. I wouldn’t want to ask for something that takes as much board space as a super cap. The people using the system would likely leave it around without the battery installed for months possibly. Any kind of pad / test-point connected to (VDD_SNVS_IN) that can be soldered to would do.
We could also add an external RTC as LED-guy suggested.
Thanks for all your suggestions,
Okay, I put OR’ing diodes in (4 cents) and added test pads for this. So, you can solder on a wires for a coin cell battery that has positive and negative wires on it. The current draw from the MCU will be about 18uA.
So, the number above is probably unlikely. Unless we disable all the output pull up resistors on some of the digital I/O from the backup domain you’re going to get leakage. It looks like that will be about 66uA or so across all lines that would be added to the 18uA above.
So, I’d say you have roughly 2 months (about 52 days).
In the case of a larger backup battery attached to the normal input this leakage drops half.