CC1310: Questions about external crystal oscillator

Hi Zack，

I'm very sorry. Maybe I didn't ask the question clearly.
The customer's device wakes up from sleep once every 8 seconds and flashes the LED once. When the external crystal oscillator stops, the LED will no longer flash. 
After the customer discovered the problem of the crystal oscillator stopping and then conducted some tests, he had the following four questions. I hope you can help and give some advice:

1. After the board crashes, the external crystal oscillator stops working. Will the chip automatically switch to the internal crystal oscillator?
2. After the external crystal oscillator stopped, we tried to start the external crystal oscillator from the power supply. Then the external crystal oscillator was working, but the chip did not work normally. Question: How to make the stopped external crystal oscillator start oscillating again and make the chip work normally?
3. Can the watchdog be set to a low-speed crystal oscillator clock source? How to set it up?
4. When the chip is not sleeping, short-circuiting the external crystal oscillator will directly reset the chip. What is the logic behind this?

Thanks & Best regards,

Yolande

Hi Yolande,

Just so I have a clearer picture of what is being investigated: please could you clarify if when you say that "the crystal oscillator stops" is because the customer is intentionally short-circuiting the load capacitor pad(s) and, if so, what is their intention behind doing this?

Regards,

Zack

Hi Zack，

The customer program uses an external crystal oscillator. On a normal board, the external crystal oscillator starts to oscillate when sleeping, but for this crashed board, the external crystal oscillator does not vibrate. What are the possibilities that could lead to this situation? Because I went to investigate the cause of the board crash and found that the external crystal oscillator was not working, and then I conducted subsequent tests, which are the questions in the previous reply. I hope you can help answer them. Thank you again.

Thanks & Best regards,

Yolande

The 32 kHz xtal oscillator is very power and is therefore sensitive for noise. I know about cases where a line containing a noisy signal is placed close to either the X32K_Q1 and X32K_Q2 lines and where the xtal stops when a noise pulse is present on the line close to the xtal.

Either post a zoomed in portion of your layout around the xtal and/ or submit the layout for review. Also look at the timing when the xtal stops and see if you do something on your board that could produce a noise spike. 

Hi TheChostof,

The customers you mentioned have already given feedback to the hardware engineers.

But can you answer the four questions I asked before? This is very important to customers. They are currently looking for solutions to restore the product even if the cause is not found.

Yolande Wang said:

After the customer discovered the problem of the crystal oscillator stopping and then conducted some tests, he had the following four questions. I hope you can help and give some advice:

1. After the board crashes, the external crystal oscillator stops working. Will the chip automatically switch to the internal crystal oscillator?
2. After the external crystal oscillator stopped, we tried to start the external crystal oscillator from the power supply. Then the external crystal oscillator was working, but the chip did not work normally. Question: How to make the stopped external crystal oscillator start oscillating again and make the chip work normally?
3. Can the watchdog be set to a low-speed crystal oscillator clock source? How to set it up?
4. When the chip is not sleeping, short-circuiting the external crystal oscillator will directly reset the chip. What is the logic behind this?

Thanks & Best regards,

Yolande

Hi Yolande,

Is this thread related to this other thread you have started? https://e2e.ti.com/support/wireless-connectivity/sub-1-ghz-group/sub-1-ghz/f/sub-1-ghz-forum/1313391/cc1310-cc1310-is-set-to-long-receiving-mode-brepeatok-and-brepeatnok-are-both-1-after-the-rf-receives-the-data-the-output-through-the-serial-port-is-normal-however-after-the-serial-port-receives-the-data-the-rf-switches-and-s 

Otherwise I am unsure where the board information for review is. The customer can submit their design for review using: Sub-1 GHz Design Review Submission: https://www.ti.com/tool/SIMPLELINK-SUB1GHZ-DESIGN-REVIEWS

I will follow-up on the 4 questions asked as well (I need to check with our SW Apps team for some of them).

Regarding question 4:

• I am still unclear why the customer is intentionally short-circuiting the external low frequency crystal.
• Section 7.7.1.1 of SWCU185 (CC13x2, CC26x2 SimpleLink Wireless MCU Technical Reference Manual): https://www.ti.com/lit/swcu185 indicates that the Clock Loss Detection feature could be enabled (a detected loss of SCLK_LF results in a system reset if enabled).
• Can the customer review Sections 7.5 and 7.7 of SWCU185 and check this?

Regards,

Zack

Yolande Wang said:

• After the board crashes, the external crystal oscillator stops working. Will the chip automatically switch to the internal crystal oscillator?
• After the external crystal oscillator stopped, we tried to start the external crystal oscillator from the power supply. Then the external crystal oscillator was working, but the chip did not work normally. Question: How to make the stopped external crystal oscillator start oscillating again and make the chip work normally?

1) I believe not since you the 32k oscillator you want to use is set in CCFG

2) My experience is that if the 32 kHz oscillator stops due to noise it restarts again after the noise source is gone. 

But: In my opinion question 1,2 and 4 are not relevant. What I mean by that is that you have to try to find out why you have an issue in HW and fix this. Trying to solve this in FW is most likely a waste of time. 

Hi ZC，

ZC said:

Is this thread related to this other thread you have started? 

No, these two posts are not related. They were asked by two different customers.

ZC said:

I am still unclear why the customer is intentionally short-circuiting the external low frequency crystal.

The customer wants to simulate the phenomenon of the external crystal oscillator not working, and then check the original problem: the crystal oscillator of the crashed device cannot vibrate.

Customers have a built-in battery, so they cannot avoid touching the chip with their hands during production.

ZC said:

Section 7.7.1.1 of SWCU185 (CC13x2, CC26x2 SimpleLink Wireless MCU Technical Reference Manual): https://www.ti.com/lit/swcu185 indicates that the Clock Loss Detection feature could be enabled (a detected loss of SCLK_LF results in a system reset if enabled).

The customer has checked these two options and they are both enabled. However, it has no effect after the crystal oscillator is stopped during sleep.

"In addition, the customer found that the normal prototype has VDDS input of 3.3V and VDDR output of 1.6V, but the abnormal prototype has no VDDR output; DCDC_SW has 3V normally but no voltage abnormally; DCOUPL normally outputs 1.27V but has no voltage abnormally."

Is this phenomenon a protection mechanism of the 1310 chip itself? And when the customer actively short-circuits the external crystal oscillator, the same phenomenon will occur.

Thanks &  Best regards,

Yolande

Hi team,

May i know there is any updates?

And i want to know what crystal oscillator is used for watchdog.

Thanks & Best regards,

Yolande

Hi Sid,

Thank you for your answer.

May I ask where the performance parameters of the LDO inside the cc13130 ​​are? I didn’t find them in the manual.

Thanks & Best regards,

Yolande