Hi Community,
We'd like to know what causes the i.MX28 to reboot automatically.
We tried out DCDC_BATT source only configuration with 3.3V supplied to J86 on the EVK.
Since page 6 of the schematics Rev.A describes that 3.3V source could be used for the EVK, we expected there will be no problem at all.
However, there was a problem.
When 3.3V is supplied to DCDC_BATT and POWER button is pressed, it started to boot.
But as shown in the attached log, it rebooted automatically soon after it started memory test again and again, and it didn't stop until we stopped supplying 3.3V.
When we supplied 3.9V instead of 3.3V, the issue did not occur.
Could anyone please tell us what causes the self rebooting of the i.MX28?
(By the way, we made a custom board with 5V only source configuration and it acted similarly.
In that case, defining NO_DCDC_BATT_SOURCE solved the issue.)
The hardware configuration is as follows;
-Boot device was the SD Card (Part No. 342-00470) which is included in the EVK package.
No other device was attached to the EVK.
-The switches and the jumper were configured as follows;
S1 - BATT (Source of LI-ION_BATTERY)
S3 - OFF
S16 - USB 5V (Source of VDD_5V)
J94 - VDD5V (Source of REG_3V3 and REG_1V8)
-The ac adapter was connected to supply REG_3V3 so we can get log.
Thank you in advance,
Hikaru
Original Attachment has been moved to: EVK3P3-2.log.zip
Solved! Go to Solution.
Please try to clear AUTO_RESTART to 0.
this may useful for your issue.
Hi,
I add the same problem and I resolve it with this modification on the bootlets :
diff --git a/power_prep/power_prep.c b/power_prep/power_prep.c
index 3e90c60..18a0574 100644
--- a/power_prep/power_prep.c
+++ b/power_prep/power_prep.c
@@ -139,7 +139,7 @@ void PowerPrep_InitLed(void);
* The minimum DCDC operating voltage for i.MX28 with loading is 3.3V.
* we will set the BRWNOUT_LVL bitfield the 3.2V value for 0.1V margin.
*/
-#define BATTERY_BRWNOUT_BITFIELD_VALUE 20 /* 20 = 3.2V */
+#define BATTERY_BRWNOUT_BITFIELD_VALUE 13 /* 13 = 2.92V */
#else
/*
* At the time of this writing,
The power manager has a register to automatically shutdown the processor when the battery voltage is below a threshold. By default the setting is 3.2V in the bootlets but during my test the processor shutdown between 3.3 and 3.4V
Hi Victorien,
Thank you for the reply.
We suspected that the brownout is causing the issue.
So your suggestion may make sure what we suspected was true.
What we really want to know is that why i.MX28 is rebooted by a brownout (or whatever).
As I posted, there is no mention about reboot in the documents.
Does anyone know?
Thank you,
Hikaru
The reboot appends when the bootlet test the memory. Do you check if it's a coincidence or a problem with the ddr (wrong voltage supply for exemple) ?
Hi Victorien,
Thank you for the comment.
The log I uploaded shows that reboot occurs around the ddr2 test but sometimes the timing of reboot is different.
Also, the rebootings occur on the EVK as well as our custom board (though the condition is not the same).
Thus we don't think the ddr2 is causing the issue.
We just found out that reboots frequently occur when the power supply is unstable (like loose contact), so it's certainly caused by brownout.
So reboot caused by brownout is a function of i.MX28 even though there is no mention about it in the documents. Is it right?
Can any of freescale employee confirm it?
Thank you,
Hikaru Uruno
HW_POWER_BATTMONITOR field descriptions
The battery voltage falls below the battery brownout level (field BRWNOUT_LVL in HW_POWER_BATTMONITOR)—This feature is disabled by clearing PWDN_BATTBRNOUT in the same register.
PWDN_BATTBRNOUT
Powers down the device after the DC-DC converter completes startup if a battery brownout occurs. This function is only active when 5V is not present. Additionally, software should clear this bit and disable this function after the system is configured for a battery brownout and the battery brownout interrupt is enabled.
So, you can try to disable this feature by clearing PWDN_BATTBRNOUT.
PS:
See i.MX28 Datasheet for suggested voltage settings for the supply and brownout targets
for different operating frequencies.
Hi Jimmy,
Thank you for your reply.
The thing which is bothering us is that "Powers down the device" is not about reboot.
The documents mention aoubt "power down" or "shut down" but the chip's actual reaction to a brownout is reboot. Is that the case?
Best regards,
Hikaru
Please try to clear AUTO_RESTART to 0.
this may useful for your issue.
Hi Jimmy,
Thank you for the suggestion.
I suppose AUTO_RESTART bit is what we've been overlooking.
Best regards,
Hikaru Uruno
Hi Community,
I re-searched the community and found something.
In the thread #468720, Igor mentioned that "processor continuosly resets due to some brownout."
Since AN4199 and the reference manual of i.MX28 both mention that the brownout causes the device to be shutdown, we didn't think the rebooting symptoms I reported were caused by some brownout.
11.2.1.1 of RM;
"The power subsystem has several mechanisms active by default that safely return the
device to off state if any one of the following errors or brownouts occur:"
3.4.2 of AN4199;
"The hardware automatically shuts down the device when a brownout
condition is detected."
Which is the true raction of i.MX28 to the brownout, shutdown or reboot?
Thank you in advance,
Hikaru