About Input the Osc clock (32.768kHz) during i.MX6 power off.

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

About Input the Osc clock (32.768kHz) during i.MX6 power off.

Jump to solution
1,753 Views
keitanagashima
Senior Contributor I

Dear All,

Hello. My customer has used i.MX 6Dual and PF0100.

PF0100's VSNVS output starts from 1.0 V when LICELL is valid before VIN.

*Refer to "Figure 6. Default start-up sequence" in MMPF0100Z(Rev.10).

And, the VSNVS output to external oscillator (32.768kHz), too.

But, the oscillator started the operation at 1V.

The Osc clock has been inputted to i.MX6 with power off.

(VDD_SNVS_CAP was 0V. So internal LDO looks disabled.)

*Refer to attached system structure.

[Question]

Does the Osc clock give i.MX damage during above situation?

(Or, is there possibility to destroy the i.MX6 device?)

Amplitude of osc clock was measured 1.1V.

Best Regards,  

Keita

Labels (2)
0 Kudos
1 Solution
1,395 Views
art
NXP Employee
NXP Employee

It does not help since the RTC_XTALI input is still out of spec all the time. The Table 21 of the i.MX6Dual/Quad Data Sheet document clearly says that the maximum allowed RTC_XTALI input voltage is no higher than 1.1V. Anyway, to guarantee the long-term stability and reliability of their system, your customer has to redesign their RTC_XTALI clock source.

Artur

View solution in original post

0 Kudos
11 Replies
1,395 Views
art
NXP Employee
NXP Employee

1. It _is_the_problem_ since the RTC_XTALI is _higher_ than VDD_SNSV_CAP in all case you've shown.

2. The MMPF0100 Data Sheet Rev.15 document says the following.

pastedImage_3.png

pastedImage_4.png

So, ask the customer to check the VSNVS settings.

Artur

0 Kudos
1,395 Views
keitanagashima
Senior Contributor I

Dear Artur,

Hello. Thank you for your reply.

> 1. It _is_the_problem_ since the RTC_XTALI is _higher_ than VDD_SNSV_CAP in all case you've shown.

You are right. Sorry for my misunderstanding.

For example, refer to "Figure 6. Default start-up sequence" in MMPF0100Z Rev.10.

If it is right sequence, VSNVS keeps the 1V output during 3.0 [msec].

  tR1 : Rise time of VSNVS  : Typ  3.0 ms

In  this case, I think that it is no problem because of the very short period(3ms).

Is my understanding OK?

> 2. The MMPF0100 Data Sheet Rev.15 document says the following.

This issue happened before I2C communication.

So, user can't set this register.

My customer is using "NP" device with TBB mode now.

So, the VSNVS output 1V until starting the I2C communication.

Best Regards,

Keita

0 Kudos
1,395 Views
art
NXP Employee
NXP Employee

Anyway, according to the oscilloscope screenshots your customer has provided, RTC_XTALI is always higher than VDD_SNVS_CAP, so, it is still the problem.

Artur

0 Kudos
1,395 Views
keitanagashima
Senior Contributor I

Dear Artur,

Hello.

>Anyway, according to the oscilloscope screenshots your customer has provided, D is always higher than >VDD_SNVS_CAP, so, it is still the problem.

Yes. You are right. But, my customer can't do a big change on their schematic because MP schedule is critical.

So, my customer will change the PWRON timing early.

If the input timing is same VIN and PWRON, the above 3 msec limitation will be occurred.

Please confirm the below question.

========

For example, refer to "Figure 6. Default start-up sequence" in MMPF0100Z Rev.10.

In right sequence, VSNVS keeps the 1V output during 3.0 [msec].

  tR1 : Rise time of VSNVS  : Typ  3.0 ms

In  this case, I think that it is no problem because of the very short period(only 3ms) with "RTC_XTALI > VDD_SNVS_CAP".

Is my understanding OK?

========

When it isn't possible to be permissible by this 3msec limitation, i.MX6 can not be used in the SNVS mode...

The SNVS output (=1V) is spec-out from VDD_SNVS_IN spec. in i.MX6

Best Regards,

Keita

0 Kudos
1,396 Views
art
NXP Employee
NXP Employee

It does not help since the RTC_XTALI input is still out of spec all the time. The Table 21 of the i.MX6Dual/Quad Data Sheet document clearly says that the maximum allowed RTC_XTALI input voltage is no higher than 1.1V. Anyway, to guarantee the long-term stability and reliability of their system, your customer has to redesign their RTC_XTALI clock source.

Artur

0 Kudos
1,395 Views
art
NXP Employee
NXP Employee

Dear Keita,

In any case, the amplitude of an external RTC_XTALI clock signal must not exceed VDD_SNVS_CAP, that is typically around 1.0V in working conditions. Otherwise, a permanent chip damage may occur. As far as I can see on the oscilloscope screenshots you've provided, this requirement is not met in your customer's system. In the current state, the system is non-functional since it does not meet the RTC_XTALI input specs. The level of an external clock signal must be aligned to VDD_SNVS_CAP. For that, some level shifting circuit must be used.

Another point is that they use wrong LICELL voltage to backup the system. For normal operation, the LICELL voltage must be at around 3.0V - 3.3V level.


Have a great day,
Artur

-----------------------------------------------------------------------------------------------------------------------
Note: If this post answers your question, please click the Correct Answer button. Thank you!
-----------------------------------------------------------------------------------------------------------------------

0 Kudos
1,395 Views
keitanagashima
Senior Contributor I

Dear Artur,

Hello. Thank you for your reply.

[Q1]

>In any case, the amplitude of an external RTC_XTALI clock signal must not exceed VDD_SNVS_CAP,

>that is typically around 1.0V in working conditions.

- Amplitude of an external RTC_XTALI clock signal = 1.1V

- VDD_SNVS_CAP = 0.4V

So, I think that it is no problem because RTC_XTALI clock signal was not exceed than VDD_SNVS_CAP?

Is my understanding OK?

[Q2]

> Another point is that they use wrong LICELL voltage to backup the system.

> For normal operation, the LICELL voltage must be at around 3.0V - 3.3V level.

Why are you think that my customer uses wrong LICELL voltage??

I think that my customer uses LICELL voltage at 3V.

The below comment is PF0100's function.

PF0100's VSNVS output starts from 1.0 V when LICELL is valid before VIN.

Best Regards,

Keita

0 Kudos
1,395 Views
art
NXP Employee
NXP Employee

To make me able to check the issue, please provide the ramp-up curves of both VDD_SNVS_IN and VDD_SNVS_CAP voltages with relation to the oscillator output curves once the battery is applied.

Best Regards,

Artur

0 Kudos
1,395 Views
keitanagashima
Senior Contributor I

Dear Artur,

Hello. I got the waveform. Please check the attached file.

Sorry. I misunderstood about below comment.

===

VDD_SNVS_CAP was 0V.

So internal LDO looks disabled.

===

The correct information was below.

===

VDD_SNVS_IN was 1.0V.

VDD_SNVS_CAP was 0.4V.

So internal LDO looks enabled.

===

Best Regards,

Keita

0 Kudos
1,395 Views
keitanagashima
Senior Contributor I

Dear Artur,

Thank you for your reply.

OK. I requested my customer to get the waveform.

VDD_SNVS_CAP was 0V.

So internal LDO looks disabled.

And, the customer's product has the possibility that the following condition continues.

"PF0100's VSNVS output starts from 1.0 V when LICELL is valid before VIN."

Best Regards,

Keita

0 Kudos
1,395 Views
keitanagashima
Senior Contributor I

Hi All,

Do you have any update?

Our customer is urging us to answer the item quickly.

Best Regards,

Keita

0 Kudos