5V supply-based design with battery-supplied RTC domain?

[albertguardiola]

Hello all,

Our design is based on a IMX28, which will be running on 5V supply only. However, we do want to mount a battery, but only to supply the RTC domain when the 5V supply is down. Is this configuration available in the 2.6.35.3 kernel-based BSP? What kind of battery shall we use?

Thanks indeed

[igorpadykov]

Hi Albert

attached is recommended schematic with optional RTC circuit

(processor has not special power supply for RTC only).

Best regards

igor

-----------------------------------------------------------------------------------------------------------------------

Note: If this post answers your question, please click the Correct Answer button. Thank you!

-----------------------------------------------------------------------------------------------------------------------

[albertguardiola]

Thanks indeed, Igor. That was helpful.

Given that same schematic, would it be a problem to mount a primary lithium battery instead of the supercap, so we achieve longer back-up time?

Best regards

[igorpadykov]

Hi Albert

yes this is possible.

Best regards

igor

[albertguardiola]

Hi Igor,

We're thinking of a circuit like this. The Shottky will make the voltage drop to 3,3V at the BATTERY pin. Is that enough to keep the RTC circuit running? The battery is a 3.6V Li-SoCl2 cell.

Thanks, best regards.

[markwilliams]

Albert,

see my reply here for my experience doing this: Re: i.mx28 RTC backup current on 24MHz Xtal

watch out out for the relatively high keep alive current of the imx28 and the fact that technically the clocks minimum input voltage is 3.1V.

also your battery voltage may leak via the 10k to vdd4p2. We had issues restarting the processor because of leakage paths though the resistors keeping the chip from fully shutting down.

It ts is probably worth doing some calcs on battery life.

i guess you are using a battery holder? If soldering direct to the pcb it is worth adding an isolation jumper.

you may also need a series resistor to demonstrate to and safety inspection (CE or UL) that the maximum battery discharge current is limited.

i ended up using an external rtc...

mark

[albertguardiola]

Thanks indeed, Mark,

I really appreciate your opinion. I have gone throughly thorugh the conversation you had with Freescale people:

Re: i.MX28 RTC Backup Battery Connection

I think it's worth preparing the schematic to test the alternative suggested by Jordi Farres at the end of the discussion. Any comment on that one?

Best regards

[igorpadykov]

yes it is enough.

Best regards

igor

[albertguardiola]

Hi Igor, all,

Are you aware of this discussion?

i.MX28 RTC Backup Battery Connection

Both community members Mark Williams and Jordi Farres were discussing on how to wire iMX's battery pins to handle RTC backup. They stated that schematic suggested by Freescale on AN4199 (that matches the AN for iMX233 attached by you earlier in this conversation) failed to work, as battery kept drawing a high current after disconnecting the 5-V input.

Mark Williams stated that he moved to an external RTC solution. Jordi Farres stated that pins BATTERY and DCDC_BATT shall be shorted so the iMX core was properly shut on power-down. On these conditions he measured a uA-consumption from battery, which matches the internal RTC specs.

Has Freescale done further research on this topic? Is it still advised to mount the solution proposed in AN4199 for RTC backup?

Best regards

[igorpadykov]

Hi Albert

you are right, internally modified circuit (below) was tested on MX28 EVK which consumes <20uA (typical silicon

at room temperature) to keep the 32kHz running.  Here are the required HW changes on MX28 EVK -

Best regards

igor

[albertguardiola]

Thanks Igor,

So I am assuming that the last schematic you are suggesting is for an IMX running on a  5V-supply only, and that the supercap is only used to keep the RTC circuit up after disconnecting the 5V. Am I right?

In AN4199 pins BATTERY and DCDC_BATT were not connected, and each of them pulled up to VDD4P2. I am assuming this is not necessary.

I am also assuming that the 1K resistor from supercap node to VDD4P2 is only mounted to charge the supercap. So if we mount a non-recheargeable lithium battery, this resistor shall be not mounted.

Best regards