CYW20719B2 32kHz Oscillator Issues

Tip / Sign in to post questions, reply, level up, and achieve exciting badges. Know more

cross mob
ChCh_3584941
Level 4
Level 4
50 sign-ins 25 replies posted 25 sign-ins

We have an internally designed Development board that includes the CYW20719 chip.

The boards were originally built with the -B1 version of the chip.  And all the boards were verified to be fully functional and were working reliably.

Based on direction from our Cypress Rep, we migrated our design to the -B2 version of the chip and have had the original boards re-worked to swap the

B1 chip with the B2 chip.  ( This rework has been performed by a commercial, quick-turn assembly / re-work vendor, ... so the rework is being done

professionally).

However, we have had persistent issues with getting the 32kHz to operate on the reworked boards.  (Exact same crystals and load capacitors in-place

as were used with the B1 chips.)  I can load firmware into the chips when they are in 'Recovery Mode', and I can see that the loaded firmware will begin to boot

(by monitoring some logging messages on PUART).  But the firmware seems to freeze after a few instructions are executed, ... the 32kHz Oscillator is NOT running

but, if I probe around the external crystal + load Capacitors I can occasionally get the firmware to "un-freeze", but still the 32kHz is not running.

Do you have any suggestions as to what may be going on, or for what we should look at to debug this?  we have reflowed and replaced the CYW20179 chips several

times, a few of the boards began to operate properly after some re-re-working, but the majority continue to show the problem.

0 Likes
1 Solution

Update: Internal case is going on to fix this issue.

Will be providing API to increase the drive strength of 32K Oscillator.

View solution in original post

0 Likes
19 Replies
AnjanaM_61
Moderator
Moderator
Moderator
5 comments on KBA First comment on KBA 5 questions asked

Hi ChCh_3584941 ,

Can you please share the schematics & 32K crystal details ?

Hope you are using latest Modus Tool Box to develop & download application to 20719B2 design. Can you please confirm ?

Regards,

Anjana

0 Likes

Yes We are using ModusToolbox as the development environment for the B2 version of the chip.

As mentioned in the original post,  the circuits worked with no issue when the boards were populated with the B1 chip, so it is either something different about B2, or there is some issue we are overlooking with regard to the re-work.

The 32 kHz Crystal is TXC 9HT11-32.768KDZB-T, 6pF 

Schematic is as follows:

0 Likes

pastedImage_0.png

0 Likes

Wondering if you have any further input or thoughts on the issue we are seeing?

0 Likes

Hi ChCh_3584941 ,

Sorry for missing to update the thread. I am checking with our HW team with 32K crystal part. Will update you asap.

Meanwhile can you please confirm which application are you testing ? Can you just run hello_sensor app in 20719B2 MTB and let us know if the issue persists.

This will help us to confirm whether issue is on HW side or SW side of the application.

Thanks for your patience.

Regards,

Anjana

0 Likes

Hi ChCh_3584941 ,

Can you please confirm the above queries?

Since there is slight FW changes as well as SDK changes from CYW20719B1 to CYW20719B2, we need to check the application (SW) part as well to confirm your issue is with SW or HW. So Please confirm if you are facing issue running hello_Sensor demo app on your custom board.

Or please test your MTB application code on 20719B2 EVB.

Also if you still suspect the HW side, please share the complete schematics.

Thanks,

Anjana

0 Likes
lock attach
Attachments are accessible only for community members.

I am testing using the firmware example "pwm_GPIO"  from:  GitHub - cypresssemiconductorco/mtb-examples-CYW920719B2Q40EVB-01-btsdk-hal: HAL Examples  ​The code does not seem to be the problem.

0 Likes

Hi ChCh_3584941 ,

Thanks for sharing the schematics. Will have a look.

What exact problem you are seeing when you run pwm_GPIO ? Is the pwm output is stopping after some time? Please confirm.

I understand code may not be the problem. However to confirm , can you program hello_sensor code and test ?

I will try pwm_gpio on my EVAL to check if that code is working fine or not.

Hope you are using latest BTSDK.

Thanks,

Anjana

0 Likes

Right now, the defective boards are back at the re-work house, so I am not able to try different code on them.  Second, since this is our own design, directly running your example code does not show much information since our pin assignments are different from your EVB and even the PUART output is re-assigned.  (So in actuality, the pwm-example I reference above is modified somewhat to match our hardware.)

We modified the PWM base-line by adding a timer, and use that to blink another LED on our Design, we also added some trace messages at the start-up phase and a Trace message whenever the timer callback is invoked.  On the defective boards we can see the initial trace messages are output on PUART, so we can see that the code has initialized and has begun to run.  However then LED does not blink and the Trace messages from the Timer callback do not appear on PUART ( I guess this is expected if the 32kHz oscillator is not running).  On a functioning board we can see the LED blink and the call-back trace messages as expected.

0 Likes

Hi ChCh_3584941 ,

One doubt. Since you have mentioned some of the boards are working fine after the HW rework, can you mention what's exactly the rework ?

So was there any chance if something broken on the PCB  when you replace 20719B2 chip from B1 ?

Regards,

Anjana

0 Likes

Hi ChCh_3584941 ,

Adding few more suggestions from our HW team.

Please check all the power supplies to make sure the voltages are the same compare to their 20719B1 board. The 24MHz crystal should oscillate. Make sure RST_N is not in reset (high state). You may contact your local FAE and create an SFDC case & send us a board for debugging. We have a feeling this is a manufacturing issue since rework the board makes it work.

You can also check if UART_RTS_L is Low. 20719 should pull up RTS when the chip is ready for UART communication. It would be good to also double check the power inductor spec to make sure it meets the requirement in the 20719 datasheet. Also check UART_CTS_L should be low when the host is ready for UART communication.

C53 and C54 value seems high. Did you get a chance to run ICFT test to check for +/-75kHz BQB requirement? I think we use 7pF but the value changes depends on the crystal used and board layout.

Regards,

Anjana

0 Likes
lock attach
Attachments are accessible only for community members.

I have performed some additional measurements and observations.

I have access to a few samples of the Cypress-Produced Evaluation board for the 20719 Chip.  Specifically I have probed the 32kHz Oscillator on two samples on the original B1 version of the Cypress Evaluation Board (CYW920719Q40EVB-01)  these are attached as "B1-CYP-EVBx-Pin3y.PNG".  I also have probed two samples of the updated Cypress Evaluation Board with the B2 Chip (CYW920719B2Q40EVB-01),  attached as "B2-CYP-EVBx-Pin3y.PNG".

My observation from these is that the the two B1 Boards behave consistently and show similar 32kHz waveforms on Pins 31 and 32 and the oscillator seems to have a strong and stable output.

The two B2 boards however appear markedly different from what is observed on the B1 boards with lower amplitude  oscillations.  Actually the second sample that I probed seems to only be marginally oscillating and the action of my probing could stop the oscillation.

To me it seems clear that there is some difference between the 32kHz oscillator on the B1 and B2 chips, and that even your reference platform for the B2 chip is performing marginally.

Our own circuit mirrors this behavior, the original build with the B1 chip showed similar waveforms (amplitude and shape) what I see on the CYW920719Q40EVB-01.  Boards with the B2 chip either don't oscillate or if they do, they show similar low-amplitude oscillation waveforms like I observed on the CYW920719B2Q40EVB-01.

I had previously verified the power supplies, etc. on our own boards and I'm now confident that, in general, there is no hardware problem with our boards, but the issue is with the 32kHz oscillator behavior of the B2 chip.  Please provide updated specifications for the Crystal and load capacitance for the B2 chip to insure reliable operation of the oscillator.

0 Likes

Wondering if there is any further feedback on this issue.  We are continuing to experience the problem and cannot seem to find a explanation / resolution.

0 Likes
lock attach
Attachments are accessible only for community members.

Hi ChCh_3584941 ,

As per our HW team, We didn’t make any change to the 32K xtal nor loading cap values from B1 to B2.

The difference from B1 to B2 is majorly on the FW side. So that make sure for B1 , you use WICED SDK and for B2 MTB.

Attached is the sample probed by HW team on B2 EVAL board with latest BTSDK FW and able to see 300mV with stable oscillations.

Can you please confirm you used MTB 2.1 or higher for this test mentioned last response? Note : Latest IDE is MTB 2.2 , BTSDK 2.8 . If not , can you please try with latest version ?

Regards,

Anjana

0 Likes

I can confirm that for the B2 Chips we have been using MTB2.1, including for all test code referenced above.

For the B1 chip, WICED Studio was used.

0 Likes

Hi ChCh_3584941 ,

As we don't see any distortions on 32K pins of EVB at our side , can you please recheck once again?

If you are using MTB 2.1, kindly make sure you are using BTSDK 2.7.1 (which is latest foor MTB 2.1). or please try with MTB 2.2 BTSDK 2.8

Please use a demo code for example , hello_sensor.

Regards,

Anjana

0 Likes

What are the steps to verify the SDK version? I am sure that MTB version

is 2.1.

On Thu, Oct 29, 2020 at 10:33 PM AnjanaM_61 <community-manager@cypress.com>

0 Likes
lock attach
Attachments are accessible only for community members.

Hi ChCh_3584941 ,

To check the BTSDK version, it will be written in wiced_btsdk of workspace as in screenshot.

For example, in case of MTB 2.1 attached screenshot.

BTSDK 2.7.1 is the latest version in MTB 2.1.

If you want to update from older version to 2.7.1 , then delete the old wiced_btsdk folder -> then click on New application -> click on any BT chip -> next -> select wiced_btsdk -> create

Regards,
Anjana

0 Likes

Update: Internal case is going on to fix this issue.

Will be providing API to increase the drive strength of 32K Oscillator.

0 Likes