nRF5340 QSPI flash

RE: nRF5340 QSPI flash

AHaug — Thu, 11 Apr 2024 07:37:55 GMT

Thank you both for reporting the issue and verifying that it works!

Please feel free to raise new questions on Devzone in the future for queries and bug reports

Kind regards,
Andreas

RE: nRF5340 QSPI flash

roshangeoroy — Thu, 11 Apr 2024 06:20:07 GMT

Glad to be of help!

RE: nRF5340 QSPI flash

Jason — Thu, 11 Apr 2024 06:17:15 GMT

Hi AHaug

this patch works. and it has been merged in NCS 2.6.0

RE: nRF5340 QSPI flash

Jason — Thu, 11 Apr 2024 06:15:52 GMT

hi roshangeoroy,

Thank you! This fix was applied in NCS 2.6.0, and this issue was fixed after updating to NCS 2.6.0

RE: nRF5340 QSPI flash

AHaug — Wed, 03 Apr 2024 07:39:04 GMT

Hi Jason,

Apologies for the long response time. It's been Easter holidays here in Norway for the past week.

The description in https://github.com/zephyrproject-rtos/zephyr/pull/64782 seems to me to be exactly what you're describing, and the PR has been merged into https://github.com/zephyrproject-rtos/zephyr/tree/main

Could you let me know if this patch resolves your issue?

Kind regards,
Andreas

RE: nRF5340 QSPI flash

roshangeoroy — Wed, 27 Mar 2024 05:56:14 GMT

Let me chime into this thread. I was having the same issue, and a fix/hack was to pull down the MISO line (IO1) before nrfx_qspi_init. I think this would mimic data from the external flash, making the Nordic chip think there's connectivity. After that, I reconfigure it. I believe this is the fix detailed in https://github.com/zephyrproject-rtos/zephyr/pull/64782/files but for a higher version of NCS/Zephyr RTOS.

RE: nRF5340 QSPI flash

Jason — Mon, 25 Mar 2024 01:47:22 GMT

[quote userid="107683" url="~/f/nordic-q-a/107195/nrf5340-qspi-flash/475409"]Yes, but the command to release DpD (instruction AB) is sent nonetheless0in https://github.com/nrfconnect/sdk-zephyr/blob/2e2523efe52a7ac89f0567b8798fd857b1e71ae3/drivers/flash/nrf_qspi_nor.c#L1339, so a generic solution should be implemented in the nrfx_qspi driver[/quote]

the DpD (instruction AB) will not be sent if nrfx_qspi_init returns an error.

[quote userid="107683" url="~/f/nordic-q-a/107195/nrf5340-qspi-flash/475409"]I think that your device might be stuck in qspi_wait_while_writing https://github.com/nrfconnect/sdk-zephyr/blob/v3.4.99-ncs1/drivers/flash/nrf_qspi_nor.c#L535 since it keeps polling the status register and not releasing.[/quote]

yes, this is the point, because the nrf5340 sends an RDSR(05H) in nrfx_qspi_init before the cmd release DpD (instruction AB) is sent. but for the gd25q256e , once the device has entered the Deep Power-Down Mode, all commands are ignored except the Release from Deep Power-Down.

[quote userid="107683" url="~/f/nordic-q-a/107195/nrf5340-qspi-flash/475409"]Could you send me all of your overlays and dts files such as the zephyr.dts (generated dts located in build/zephyr folder)[/quote]

devzone.nordicsemi.com/.../4863.zephyr.dts

RE: nRF5340 QSPI flash

AHaug — Fri, 22 Mar 2024 13:46:45 GMT

[quote user="greyorbit"]But I do not think all of the QSPI Flash has such a feature, and this should not be a standard.[/quote]

Yes, but the command to release DpD (instruction AB) is sent nonetheless0in https://github.com/nrfconnect/sdk-zephyr/blob/2e2523efe52a7ac89f0567b8798fd857b1e71ae3/drivers/flash/nrf_qspi_nor.c#L1339, so a generic solution should be implemented in the nrfx_qspi driver

[quote user="greyorbit"]

I found the 05H cmd is sent by function nrfx_qspi_init.

but I can not see any code related to 05H cmd inside the nrfx_qspi_init.

so where does the 05H cmd come from? how can I disable this 05H cmd and make nrfx_qspi_init work correctly?

[/quote]

I think that your device might be stuck in qspi_wait_while_writing https://github.com/nrfconnect/sdk-zephyr/blob/v3.4.99-ncs1/drivers/flash/nrf_qspi_nor.c#L535 since it keeps polling the status register and not releasing.

Could you send me all of your overlays and dts files such as the zephyr.dts (generated dts located in build/zephyr folder)

Heads up, I will be out of office for the next week and a half on vacation and will be back in office the 2nd of April. Unfortunately response time will be slow in this period

Kind regards,
Andreas

RE: nRF5340 QSPI flash

Jason — Fri, 22 Mar 2024 04:14:12 GMT

Hi AHaug，

I checked the mx25 data sheet , it can exit Deep Power-down mode if CS pulses. But I do not think all of the QSPI Flash has such a feature, and this should not be a standard.

[quote userid="107683" url="~/f/nordic-q-a/107195/nrf5340-qspi-flash/475197"]Are you able to wake the external flash using one of the instructions that the gd25 datasheet states that will not be ignored?[/quote]

I think it should be able to wake the external flash. but I can not confirm it with the NCS, because that's the problem this post describes

I found the 05H cmd is sent by function nrfx_qspi_init.

but I can not see any code related to 05H cmd inside the nrfx_qspi_init.

so where does the 05H cmd come from? how can I disable this 05H cmd and make nrfx_qspi_init work correctly?

RE: nRF5340 QSPI flash

AHaug — Thu, 21 Mar 2024 15:20:20 GMT

Hi,

Thank you for the trace. It looks like there is a difference between the mx25 that is on the nRF5340DK and the gd25q that you're using. As you state there are only a set of instructions (see table 118 in https://infocenter.nordicsemi.com/pdf/nRF5340_PS_v1.3.1.pdf) that wakes the gd25, but any instruction from the list wakes the mx25. The wakeup procedure for qspi_nor can be seen in github.com/.../nrf_qspi_nor.c

Are you able to wake the external flash using one of the instructions that the gd25 datasheet states that will not be ignored?

Kind regards,
Andreas

RE: nRF5340 QSPI flash

Jason — Fri, 15 Mar 2024 11:17:57 GMT

Hi AHaug，

sorry for the late relay, because I was handling another urgent task.

[quote userid="107683" url="~/f/nordic-q-a/107195/nrf5340-qspi-flash/466434"]However, we see that the external flash only supports 2.7V-3.6V, so this might be related to one of these[/quote]

I changed the VDD voltage and made no difference.

[quote userid="107683" url="~/f/nordic-q-a/107195/nrf5340-qspi-flash/466434"]Also, the logic analyzer trace Simon requested in the other ticket would likely be very helpful here.[/quote]

Release from Deep Power-Down on nrf5340 DK and no problem

Release from Deep Power-Down on our PCBA with gd25q256e fash, return error, and always outputs the below waveform, no stop.

According to the logic analyzer, it seems to be due to the flash doesn't respond to the 05H cmd.

my question is why did the nrf5340 send the 05H cmd before sending the ABH cmd, why not just send the ABH cmd to Release from Deep-Power-Down.

According to the FLASH datasheet, "Once the device has entered the Deep Power-Down Mode, all commands are ignored except the Release from
Deep Power-Down and Read Device ID (RDI) (ABH) or Enable Reset (66H) and Reset (99H) commands."

RE: nRF5340 QSPI flash

Jason — Mon, 11 Mar 2024 06:24:52 GMT

Hi roshangeoroy，

I do not quite understand how to implement this ‘’hack' , could you provide an example?

[quote userid="130181" url="~/f/nordic-q-a/107195/nrf5340-qspi-flash/472882"]After calling nrf_qspi_task_trigger(NRF_QSPI, NRF_QSPI_TASK_ACTIVATE) the QSPI peripheral will try to read the status register of the external flash[/quote]

do you mean this is done by the QSPI peripheral automatically? I don't see any code to read the status.

RE: nRF5340 QSPI flash

roshangeoroy — Fri, 08 Mar 2024 09:54:08 GMT

Hi Jason,

Have a look at RE: QSPI Polling of Status Register by QSPI peripheral at ACTIVATE and WRITE?

After calling nrf_qspi_task_trigger(NRF_QSPI, NRF_QSPI_TASK_ACTIVATE) the QSPI peripheral will try to read the status register of the external flash.

In the above mentioned ticket there's a hack provided. Let me know if it works for you.

Thanks,

Rosh

RE: nRF5340 QSPI flash

Jason — Thu, 07 Mar 2024 09:33:51 GMT

[quote userid="85932" url="~/f/nordic-q-a/107195/nrf5340-qspi-flash/463448"]

update a finding.

Some PCBAs can wake up from deep-power-down mode, some can not, and some have a chance to wake up from deep-power-down mode.

[/quote]

yes, as I mentioned, it seems to be related to the HW, but here's what I don't understand is that the error occurs when initiating the QSPI peripheral of Nordic MCU instead of sending the wakeup CMD to the flash. so we can not say this relate to the HW from this view point.

RE: nRF5340 QSPI flash

roshangeoroy — Thu, 07 Mar 2024 09:25:26 GMT

Oh okay. Could the function pm_device_action_run work without that configuration enabled?

RE: nRF5340 QSPI flash

Jason — Thu, 07 Mar 2024 09:21:27 GMT

[quote userid="130181" url="~/f/nordic-q-a/107195/nrf5340-qspi-flash/472400"]I thinkNRFX QSPI driver is also talking about this issue.[/quote]

I already disabled the CONFIG_PM_DEVICE_RUNTIME.

RE: nRF5340 QSPI flash

roshangeoroy — Wed, 06 Mar 2024 09:48:58 GMT

Hi,
I was doing some tests and noticed that this error occurs only when there is a system sleep in between suspend and resume of the peripheral. System going into the idle thread might be causing this. I thinkNRFX QSPI driver is also talking about this issue.

Regards,

Rosh

RE: nRF5340 QSPI flash

Jason — Wed, 06 Mar 2024 01:49:35 GMT

unfortunately, no progress so far. if you have any progress, please let me know, Thanks!

RE: nRF5340 QSPI flash

roshangeoroy — Tue, 05 Mar 2024 12:57:14 GMT

Hi,

I'm facing the same issue of getting a timeout error while resuming the external flash from suspended state. Please let me know if any solution was found. Additionally, I will create a private ticket myself sharing more information about my setup.

RE: nRF5340 QSPI flash

AHaug — Mon, 29 Jan 2024 10:08:51 GMT

Hi,

I see that my colleagues have had a look at the case.

Could you follow up on the questions Simon asked as well?

However, we see that the external flash only supports 2.7V-3.6V, so this might be related to one of these erratas: https://infocenter.nordicsemi.com/topic/errata_nRF5340_Rev1/ERR/nRF5340/Rev1/latest/anomaly_340_133.html or https://infocenter.nordicsemi.com/topic/errata_nRF5340_Rev1/ERR/nRF5340/Rev1/latest/anomaly_340_99.html.
1. Do you know what speed you're running the QSPI operations at in your project, as there are some issues running the QSPI at "high" voltages and high transfer speeds.
Also, the logic analyzer trace Simon requested in the other ticket would likely be very helpful here.

Kind regards,
Andreas

RE: nRF5340 QSPI flash

Jason — Fri, 26 Jan 2024 09:53:03 GMT

Hi,

We got the HW review feedback, they can't find any obvious flaws in our design.

RE: nRF5340 QSPI flash

AHaug — Thu, 25 Jan 2024 08:40:13 GMT

Thank you for letting me know

Kind regards,
Andreas

RE: nRF5340 QSPI flash

Jason — Thu, 25 Jan 2024 04:03:02 GMT

Hi,

my colleague has created a private case, and waiting for the feedback.

RE: nRF5340 QSPI flash

AHaug — Tue, 23 Jan 2024 10:00:02 GMT

Happy to help.

You can request a HW review through Devzone instead of through email. If you have anything confidential you can create a private case and we'll assign it to an available hardware engineer from there.

Kind regards,
Andreas

RE: nRF5340 QSPI flash

Jason — Tue, 23 Jan 2024 07:32:34 GMT

Thanks, I will let our HW team contact you by email.