reset fail with VDD5V < 5.0V

cancel
Showing results for 
Show  only  | Search instead for 
Did you mean: 

reset fail with VDD5V < 5.0V

1,526 Views
PierantonioTaba
Contributor II

Hi,

i have a board powered only with VDD5V, no battery, the power supply is based on a LM2675-5.0 and  output voltage is 4,97V.

The board boot-up and run correctly, but i am not able to reset with the reset button connected to RESETN pin.

Also the reset command from linux console don't work.

I have this boot sequence: u-boot then kernel, i think that the problem is on the kernel:

If press reset anytime when u-boot is booting the reset work.

If i press the reset before the kernel is starting init "init started: BusyBox v1.15.0 ()", the reset work.

If i press reset when the board is up at login, the reset cannot work anymore, nothing is printed out on debug console and the only way is power off and on the board.

With power off and on sequence the board is running well without any issue, i have uptime of days.

If i power the board with an external power supply the reset button work anytime, the only difference is that the external power supply is giving 5,28V.

There is a kernel setting on power or regulators that need to be changed?

when reset command from console work:

==cut==

The system is going down NOW!
Sent SIGTERM to all processes
Sent SIGKILL to all processes
Requesting system reboot
Restarting system.


PowerPrep _start

PowerPrep start initialize power...

Configured for 5v only power source.
Battery powered operation disabled.

==cut==

when reset command from console don't work:

==cut==

The system is going down NOW!
Sent SIGTERM to all processes
Sent SIGKILL to all processes
Restarting system.

Unde

==cut==

I see the partial word "Unde", but on kernel there is a lot of this words :-(

Also the "Requesting system reboot" line is missed when the reset don't work.

The reset/pswitch schematic is the same as the EVK.

Please help me.

Labels (1)
0 Kudos
6 Replies

956 Views
markwilliams
Senior Contributor I

Hi,

Sorry to dig up an old thread. I am using the i.MX283 in a design that is powered permanently from 5V. I have a small battery to power the RTC and the Freescale recommended 1k resistors from BATTERY to VDD4P2 and DCDC_BATT to VDD4P2.

When I remove power with my backup battery connected I find the PMU goes into the same funny state you describe above. The backup battery seems to keep the supply rails at funny voltages when it only connects to the BATTERY pin.

When I re-apply 5V power the board does not boot from this state. I have to remove 5V, and the backup battery, then reboot and all is fine.

I was just wondering if you ever managed to cure your bad state above, and if you ever implemented your backup battery?

Thanks, Mark

0 Kudos

956 Views
PierantonioTaba
Contributor II

After a deep investigation i have some news:

If i stress the board by pressing a lot of times the reset button the board go to the "bad state", is not related to VDD5V, i am able to reproduce with 5.2V power supply.

on that state it's only possible to power off and on.

on this state the vddio is at 2.5V, vdd4p2 is 0.5V, vdda 0,9V, vddd is 0.9V

The problem happen only with a sd card on the socket,

the same card contain u-boot and kernel and boot OK if i power off/on

the same test without sd card don't show the problem, but i have wired on LCD_D[0-3] the boot mode on ssp0 3.3V and the duart show correctly 0x8020a014.

on this "bad" state the sd card is not powered, the PWM3 pin is not driving low the PWRGATE_SSP0 wire.

Any idea?

0 Kudos

956 Views
fear_nada
Contributor II

Hi, !

Do you resolve this "bad state"?

Seems me have same problem: i use external watchdog IC for generate RESETN signal (for test i am not refresh wdg) - after some good reboots(10-50) my i.mx28 go to hang state.

vddxtal - 0.97

dcdc_vddio - 0.29

vdd4p2 - 0.3V

dcdc_vddd_1v2 - 1.0V

dcdc_vdda_1v8 - 0.96

vdd5v 4.95v

Clock on 24Mhz - are present.

and only one way to out from this hang state - remove power and apply it again.

Any idea?

0 Kudos

956 Views
Matt_ng9
Contributor III

Is there ever a plan to use battery with your project?

Like I said before, I have not had a problem with reset, but I did have a problem with sleep, and there is a slight chance they could be related.  Have you been able to successfully go to sleep and wake up?

To get sleep to work we had to pull up the BATTERY and DCDC_BATT lines to VDD4P2 using a 1k resistor.  I don't think there is an official reason yet why we had to do that but we did.  But doing that seems to have broken our ability to reliably flash via USB OTG.  Once again, we don't necessarily know why it did but it did.  (There are theories for both issues but nothing official)

It looks like if you replace D2 with a 1k resistor you will get a "close enough" circuit like what I have.

Anyway, I don't know if that will help your reset problem, but when you said you were only using a 5V supply I thought I should share.

0 Kudos

956 Views
PierantonioTaba
Contributor II

This is schematic of battery connection, battery is not populated.

248-picobattery.png

0 Kudos

956 Views
Matt_ng9
Contributor III

I tried my board at 4.7V and the reboot command works fine.  I also only use the 5V input.

I am wondering, what did you do with the BATTERY and DCDC_BATT lines?  That held us up when we were trying to go to sleep.

0 Kudos