imx6 RTC battery discharges in 10 mins

[abraham_v]

Hello,

I'm using a Lithium Ion cell (5mAh) at PMIC input for RTC backup. Without any external power supply, this cell is  supposed to power RTC domain of iMX6 (we have the iMX6D chipset). SNVS output of PMIC is powering only iMX6 SNVS domain and there is no extra power drain. The theoretical maximum consumption for SNVS domain is 400uA. Which means that the coin cell should last for at least 12 hours. But in practice, it dies after 10mins.

Has anyone seen/experienced anything similar? If so, how did you resolve it?

Puzzled,

Abraham V.

[robertferanec]

Hi Abraham, check your pull ups. I have seen possible issues in the pull up circuit of some reference schematics. All the pull ups used in the circuit powered by SNVS should be connected to this power rail, otherwise you may have current leaking between rails. Check pull ups on SRC_ONOFF, PMIC_STBY_REQ, PMIC_ON_REQ, ....

- Robert

[abraham_v]

Thanks for the suggestions Robert. I've pointed them out to our hardware team, but due to deadline crunch, the higher-ups have decided to classify this as a low-priority issue and have assigned most of us to other matters. If we get around to doing a revision-B, I'll update back.

Stuck between a rock & a hard place,

Abraham V.

[JorgeRama_rezRi]

Hi Abraham,

Never saw anything similar, but here are some hints:

- Do you have anything connected to VCORE, VCOREDIG or VCOREREF? The LICELL also supplies these regulators.

- Is your system based on any of our reference designs? If so, which one?

- How many boards do you have? Is the problem consistent in all of them?

- Could you please simulate the coin cell with a DC power supply and check the current being drawn out of it? Please also check the current out of the VSNVS regulator. I'm wondering if you could be having a leak related to layout or board manufacture.

- Would it be possible for you to send me your schematic so I can look for possible issues?

Best regards.

Jorge.

[abraham_v]

Thanks for replying Jorge,

= Nothing is connected to those regulators. For VCOREDIG, there's an option for an external regulator, but it isn't mounted right now

= We've based our design on the sabre-sd reference board, but have incorporated a few of our own modifications

= The problem is consistent on all 5 boards we have

= It will take a bit of time to run this experiment. I'll reply back as soon as we've done so

= Here is a link to the relevant sections of our schematic;

https://dl.dropboxusercontent.com/u/5134552/SCH_701-1-00093_A1.pdf

Yours,

Abraham V.

[JorgeRama_rezRi]

Hi Abraham,

Your schematic looks good in general. Here are some notes and recommendations:

- Capacitance at VDD_SNVS_CAP is too much. Please remove C292.

- Check C140 to see if it's damaged (shorted or has very low resistance).

- What are the voltages at VDD_SNVS_IN and VDD_SNVS_CAP when the coin cell is discharged?

- Please make sure the power-up sequence specified in section 4.2.1 of the datasheet is observed. VDD_HIGH can start at the same time as VDD_SNVS, but not before that.

- You mentioned R168 is not populated. Please make sure R161 isn't populated either as the schematic says. VCORExxx regulators are used for internal references of the PMIC and must not have any kind of external loading.

- Could you monitor the VCORExxx regulators voltages and check they are all at their correct level?
- Plesae also let me know the results of the tests asked previously and we'll get to the bottom of this.

Best regards.

Jorge.

[abraham_v]

I sent these suggestions over to our hardware team and it gave them a few ideas. Oddly enough, the problem appears to be due to the casing we used for the coin cell battery. When the hardware team used external wires (instead of the casing) to connect the coin cell to our board, it started working properly. The issue is still under investigation, and I'll update this post as the results come in.

-Abraham V.