S12 / MagniV MicrocontrollersのトピックIssue about the MC9S12XEP100 CLOCK Circuit

Issue about the MC9S12XEP100 CLOCK Circuit

xiaochenqiang — Mon, 27 Nov 2017 03:26:58 GMT

Hi all:

My issue is about the MCU clock circuit,,below is the description of the issue:

===========================Issue Description=============================

My product 's MCU is MC9S12XEP100,MCU circuit and the product can work nomaly,but some products(About 2% of the total, so far) occur the phenomenon of abnormal work after about two years used,The specific phenomenon is：the MCU's PLL can not stable,or occur crashes after run some time.

The MCU's clock circuit design as below:

the picture above is the Crystal circuit,

and the XCLKS pin not connect the pull-down，it is suspended;

the MODA and MODB connect the pull-down;

===========================Troubleshoot=============================

In troubleshooting the process described above, we found that there are two obvious deviations in the design:

1.The Crystal output the 5V square wave,Obviously exceed EXTAL maximum 2.16V input requirement;

2.XCLKS pin not pulled down;

Then make the following changes to the clock circuit of the MCU on the circuit board of the failed product (the MCU's PLL can not be locked occasionally):

At the same time the XCLKS foot drop pull, MODA, MODB and before, pick up the pull-down;

After the above circuit is modified, the MCU still can not lock the PLL; when the crystal oscillator outputs 5V square wave, the PLL can lock occasionally. After the above modification, basically no PLL lock occurs.

=============================Issue advisory===================================

The above is the specific phenomenon of the problem, then my question is:

1. I suspect the crystal oscillator directly outputs 5V square wave to the EXTAL pin, causing the internal clock module of the MCU to malfunction. As a result, the PLL can not be locked, but I just wonder how to determine that the 5V square wave has caused the MCU internal Clock module failure?

2. Why do the above MCUs have a normal working time of nearly 2 years under the condition that there are 2 parts that do not meet the requirements?

Thanks!

Re: Issue about the MC9S12XEP100 CLOCK Circuit

lama — Mon, 27 Nov 2017 17:46:26 GMT

Using oscillator and any of MCU’s pins out of nominal values is really ticking time bomb. Maximum oscillator voltage can be Vddpll+0.3V. Voltage above this will cause higher current flow through internal circuits and because of electromigration it can cause oscillator circuitry damage anytime. Beyond absolute maximum ratings device might be damaged.

We solved similar issue where the customer had been using the device with 5V voltage level on Extal pin without problems for 2 years and then the device was damaged.

There is electromigration which means the conductive material is carried from place to place which is caused by overcurrent and two things can happen:

1) Wire will be interrupted and MCU loses clocks. The thinner wire the higher local temperature => local overheating can cause degradation of ambient integrated circuits.

2) Accumulated conductive material on another place will cause short cut with another line => anything can happen.

Maybe the failure is detectable but nobody tested it because such conditions are out of specifications required by data sheet. I am sorry but nothing is printed or available or qualified for the public on this topic.

I have also contacted design team to get more info and I received the same answer as I have provided to you; citation:

”FSL cannot give any guarantee for operation when 5V is applied to the EXTAL pin, since this is far out of the specification”

You can find in the attachment EVB design as well as you can find some notes to design in:

https://community.nxp.com/docs/DOC-93813

Best Regards, Ladislav

Re: Issue about the MC9S12XEP100 CLOCK Circuit

xiaochenqiang — Tue, 28 Nov 2017 01:34:05 GMT

Hi lama:

Thanks,I want to know how to verify the MCU's oscillator circuitry has damaged,if the design team tell you more info about it ,please tell me in time,because this is very urgent.

thans again!

Re: Issue about the MC9S12XEP100 CLOCK Circuit

xiaochenqiang — Tue, 05 Dec 2017 01:13:07 GMT

Hi,lama:

Have you get the reply of the design team? My problem is very urgent，Please contact the design team again,thank you very much!!

Re: Issue about the MC9S12XEP100 CLOCK Circuit

lama — Tue, 05 Dec 2017 12:52:58 GMT

Hi,

it looks like you incorrectly understood my message:

I have also contacted design team to get more info and I received the same answer as I have provided to you; citation:

”FSL cannot give any guarantee for operation when 5V is applied to the EXTAL pin, since this is far out of the specification”

Their answer was in the brackets. I have no more info from them.

Best regards,

Ladislav

Re: Issue about the MC9S12XEP100 CLOCK Circuit

xiaochenqiang — Wed, 06 Dec 2017 05:08:37 GMT

Hi,lama:

I understand the wrong used,I want to know how to verify the problem,Can the design team provide any way to verify the MCU'S internal clock module have Damaged？

Thank you!

Re: Issue about the MC9S12XEP100 CLOCK Circuit

lama — Thu, 07 Dec 2017 12:31:53 GMT

Hi, the easiest test is to start the MCU in the special single chip mode (pins MODC=MODB=MODA=0) and test the BUSCLK at the pin. In this mode NECLK bit is 0 which provides BUSCLK at pin PE4 which should be OSCCLK/2. If there is frequency between 500kHz~2MHz then the MCU is in the self clock mode and you have wrong input clock or oscillator circuitry is damaged. If there is nothing at the pin then MCU is damaged.

(For more info about self clock mode please read the data sheet)

Best regrads,

Ladislav

Re: Issue about the MC9S12XEP100 CLOCK Circuit

xiaochenqiang — Fri, 08 Dec 2017 02:18:19 GMT

Hi，based on first picture circuit,I connect the MODC to the GDN,and then test the pin PE4,it is the OSCCLK/2,so the oscillator circuitry is not damaged?but some times, the MCU can not work normaly,because the PLL can not stable.why?

Below is the test scope:

the input clock is 16M,and the test scope is 8M,

Re: Issue about the MC9S12XEP100 CLOCK Circuit

xiaochenqiang — Wed, 20 Dec 2017 08:46:18 GMT

Hi lama:

Maybe the MCU's internal CLOCK circuit realy damaged;following is the scope of the test on the PE4 (the yellow scope），

the test phenomenon is: After the power on,the MCU can work normaly for some time,when it can work normaly,the yellow scope is the 5V High level,

and after some time(the time is random),the MCU can not work normaly,the scope changed from 5V high level to 5V square，and the square's frequency not same every time;

Can you explain why it occur this phenomenon ?

when the PE4 become the square，it enter the internal clock mode?

Thank you!