differences in battery charge current if booting from battery or 5V

[RoboTech]

I found out that recharging the battery through the iMX283 PMU takes much more than expected, so I put a current meter in series with my LiPo battery and now I'm trying to measure the charge current in different situations.

Discharge current is around 280mA - 320mA when no 5V power source is active.

I modified the kernel driver for our custom board to use a charge current limit of 300mA and I can see debug messages when 5V is connected, that show the current actually ramp up to 300mA:

Battery charger: setting max chargecurrent to: 300mA

When I boot from battery power, then attach the 5V source, the charge current is around 210mA

When I boot from 5V power, I can see the charge current riseup to around 310mA

Removing the 5V source and then putting it back on does not change the above figures. So I think something happens during the "boot_prep" stage that influences how the kernel driver handles the charge current.

I went on and tried to do a register dump in the above two cases, but I don't get any significative difference, or at least I can't see it. Here are the two dumps:

• Boot from battery, then connected 5V (charge current stops at 210mA)

root@pwm-term ~$ ./memtool  80044000 60

Reading 0x60 count starting at address 0x80044000

0x80044000:  00278548 00278548 00278548 00278548

0x80044010:  00024513 00024513 00024513 00024513

0x80044020:  00000040 00000040 00000040 00000040

0x80044030:  00000318 00000318 00000318 00000318

0x80044040:  0002041E 0002041E 0002041E 0002041E

0x80044050:  0000270C 0000270C 0000270C 0000270C

0x80044060:  00002608 00002608 00002608 00002608

0x80044070:  00000230 00000230 00000230 00000230

0x80044080:  A0C00001 A0C00001 A0C00001 A0C00001

0x80044090:  00000000 00000000 00000000 00000000

0x800440A0:  0000305F 0000305F 0000305F 0000305F

0x800440B0:  00167021 00167021 00167021 00167021

0x800440C0:  03078A2E 03078A2E 03078A2E 03078A2E

0x800440D0:  00000700 00000700 00000700 00000700

0x800440E0:  01F90E0F 01F90E0F 01F90E0F 01F90E0F

0x800440F0:  00000000 00000000 00000000 00000000

0x80044100:  00000000 00000000 00000000 00000000

0x80044110:  00000000 00000000 00000000 00000000

0x80044120:  00000080 00000080 00000080 00000080

0x80044130:  0000000E 0000000E 0000000E 0000000E

0x80044140:  00000000 00000000 00000000 00000000

0x80044150:  04000000 04000000 04000000 04000000

0x80044160:  84000400 84000400 84000400 84000400

0x80044170:  00000000 00000000 00000000 00000000

• Boot from 5V (charge current reaches 310mA):

root@pwm-term ~$ ./memtool  80044000 60

Reading 0x60 count starting at address 0x80044000

0x80044000:  00278548 00278548 00278548 00278548

0x80044010:  04024513 04024513 04024513 04024513

0x80044020:  00000040 00000040 00000040 00000040

0x80044030:  00000318 00000318 00000318 00000318

0x80044040:  0002041E 0002041E 0002041E 0002041E

0x80044050:  0000270C 0000270C 0000270C 0000270C

0x80044060:  00002608 00002608 00002608 00002608

0x80044070:  00000230 00000230 00000230 00000230

0x80044080:  A0C00001 A0C00001 A0C00001 A0C00001

0x80044090:  00000000 00000000 00000000 00000000

0x800440A0:  0000305F 0000305F 0000305F 0000305F

0x800440B0:  00167021 00167021 00167021 00167021

0x800440C0:  20078A2E 20078A2E 20078A2E 20078A2E

0x800440D0:  00000600 00000600 00000600 00000600

0x800440E0:  02050E0F 02050E0F 02050E0F 02050E0F

0x800440F0:  00000000 00000000 00000000 00000000

0x80044100:  3E770000 3E770000 3E770000 3E770000

0x80044110:  00000000 00000000 00000000 00000000

0x80044120:  00000080 00000080 00000080 00000080

0x80044130:  0000000E 0000000E 0000000E 0000000E

0x80044140:  00000000 00000000 00000000 00000000

0x80044150:  04000000 04000000 04000000 04000000

0x80044160:  84000400 84000400 84000400 84000400

0x80044170:  00000000 00000000 00000000 00000000

The first difference is 0x80044010:  00024513 <> 04024513 which is a different VBUSDROOP_TRSH settings (4.3V against 4.5V)

Then 0x800440C0:  03078A2E <> 20078A2E which is simply the power up source (PSWITCH for battery boot, VDD5V for 5V boot)

0x800440E0 is the battery voltage, which of course is different

0x80044100:  3E770000 means that the reset unlock key has been written

0x800440D0 is somewhat mysterious, as no documentation is provided to decode the DC/DC internal status.

I found this to be 00000600 or 00000400 or 00000700, but without any relation to the boot source.

So the question is:

Why do I get a different charge current, lower than expected, if I boot from the battery?

What other registers, other than the HW_POWER group, should I check?

I am using the 2.6.35.3_1.1.0 release, with some patches to support our custom board.

I am providing 5V power through the USB connection.

The battery charge level and voltage is around the same when I perform these tests and is not related to the problem.

I always get around 100mA less charge current then I expected, when booting off the battery and applying 5V later on, after kernel initializes.

Suggestions?

Thanks,

Paolo

[GraceH · Accepted solution]

Please use the patch in the post  https://community.freescale.com/docs/DOC-98034

Grace

[RoboTech]

Thank you Grace, that worked!

I didn't see that page when I searched the community. I guess I must learn how the search works. I thought all relevant pages were reported, maybe they were only "discussions", but not "documents". I must have inadvertently enabled some filter.

Actually, I had mis-interpreted the first difference in my register dumps. That could have helped me if I got it right:

The first difference is 0x80044010:  00024513 <> 04024513 which is a different VBUSDROOP_TRSH settings (4.3V against 4.5V)

I was wrong. The difference is in the HEADROOM_ADJ field, I missed the initial 0. By the way the reference manual is a bit mysterious about the actual contents of this field:

Adjustment to optimize the performance of the battery charge and 4.2V regulation circuit at low 5v voltages.

That doesn't seem too much explicative. It may look almost unrelated.

Anyway, I'm glad it works now.

Thanks again,

Paolo