UCD9090A: GUI timing is not correct

Hi

UCD9090A does have a RTC(real time clock) so it starts from 0:0:0 time.

Have you customer checked the log before？

Customer can erase the log to check whether the fault is present after power up.

Regards

yihe

Hi Yihe,

Alert of UCD9090 has triggered(waveform show below).

I connect into GUI check the reason that cause alert.

I find Rail#2 & Rail#3 has UV fault. But I don’t know whether this reason cause alert be triggered.

Then, Alert change from 0V to 3.3V when I click “Clear Faults”.

#1 The reason that trigger Alert is due to Rail#2 & Rail#3 UV fault?

#2 Scope do not observe Rail#2 & Rail#3 UV fault, why UCD9090 record this log?

Thanks,

Leo

Hi Leo,

Yihe is out of office today. Please expect a response by tomorrow.

Regards,

Maha

Hi

Please erase the fault log to ensure that those fault were logged in the past

power cycle to see whether the same UV fault was logged

Yes, clear fault will release PMBus alert pin

Regards

YIHE 

Hi Yihe,

Customer wants to know 

#1 The reason that trigger Alert is due to Rail#2 & Rail#3 UV fault?

#2 Scope do not observe Rail#2 & Rail#3 UV fault, why UCD9090 record this log?

Thanks,

Leo

UCD9090 will log this UV fault when power up every time.
Scope do not observe Rail#2 & Rail#3 UV fault, why UCD9090 record this log?
Is this behavior expected?

Hi

Please share your sch and configuration file.

Regards

Yihe

JUNIPER_LEOPARD_MODEL-4_SGA_0224_A9_REVIEW_TI_UCD9090.pdf

UCD9090_Model4_20250221_V9.xml

Hi

Could you please clear the fault logg first and clear fault does not clear fault log? They are different.

Regards

Yihe

I can clear the fault logg first.
However, you mean I need clear the fault log first every power up?
Is the UV fault reasonable?
Please let me know the reason that trigger UV fault.

Also let me know how to clear the fault log by I2C, the product for MP do not connect to GUI.

Regards

Hi Yihe

You mean I just clear fault log after every power cycle, right?
So, Is the UV fault within expectation?

Thanks,

I saw fault long full instead of only one log.

Please clear the fault log only once and run your power up test to see whether the log is filled again.
>>> Yes. After clear the fault log only once and run your power up, UCD9090 record Rail#2 & Rail#3 UV fault once.

Thanks,

Hi

so after clearing the fault log, the same fault is back after power cycle. Thank you for confirmatioin this

So you are probing the right signals(MON3/4)

Could you please provide the following capture during the power up with a 6-ch or 8-ch scope

VCC(power to UCD)

MON3

MON4

GPIO3

GPIO4

PMBus Alert

Regards

Yihe

Hi Yihe,

capture during the power up.

UCD9090_VCC-Alert_MON3-GPIO3

UCD9090_VCC-Alert__MON4-GPIO4

Only Rail3 and Rail2 generate a UV fault event during power-up.

But actually, no UV fault occurs. Is there a setting that can prevent this situation from happening?

HI

Based on the log, i do not think the UV is present during the boot up. I have one more test to do:

Enable the fault response for the rail 2 to shutdown it if there is UV event on that rail.

Set the trigger to the falling edge of the GPIO3 since if there is UV event, the GPIO3 is de-asserted due to the fault response.

Regards

Yihe

Hi Yihe,

Based on your instructions, I set the fault responses for rail 2. As shown in the snapshot, rail 2 was not shut down during boot-up, indicating that the UV event does not exist. 

Q1: However, the log file has updated with two identical records of UV faults for rail 2 and 3. Is this a misjudgment by the UCD9090? Can it be configured to avoid this?"

Q2: Additionally, for the logging file settings, I changed the original FIFO from disable to enable to avoid the Alert pin becoming active low when it accumulates to 26 entries. Is this correction setting correct?

We upload circuit of UCD9090 for your review this issue.UCD9090_Schematic.pdf

HI

Please loop your FAE and we need have a debug session on this since it doesn't make sense for me.

Regards

Yihe

Hi Yihe,

Based on Nelson's waveform. Alert pulls down and the rails that is monitored are still in low level.

Nelson Chang said:

capture during the power up.

Is it a correct behavior? 

Is UCD9090A detected the rail too early?

Zhe-Xian Su said:

UCD9090_Model4_20250221_V9.xml

Could you check the configuration again?

What can we do for next step debug?

Thanks,

Leo

HI

As I suggested, i need to have a live debug session with your customer.

Regards

Yihe