AM623: Questions for UBS during in deepsleep

Koichiro Tashiro

Part Number: AM623
Other Parts Discussed in Thread: SK-AM62, SK-AM62B, SK-AM62-LP, SK-AM62B-P1

Tool/software:

Hi,

My customer is considering below scenario using AM6232 and an external WLAN module.

- AM6232 and WLAN module is connected by USB. AM6232 is USB Host.
- When WLAN is waiting for data receive from WiFi, AM6232 is put in deep sleep for power saving.
- During above period, USB VBUS needs to be supplied to WLAN module.
- As soon as WLAN module receives data, WLAN module wakes up AM6232 by WoWLAN (Wake on Wireless LAN).

Questions:
Please answer if both AM623x hardware and TI Linux SDK support below functions.
Q1) Is 5V VBUS is kept provided while AM623x is in Deep Sleep?
Q2) Does AM623x USB support Type-C?
Q3) There are two USB channels. Each channel can work either Host and Device, correct?

Thanks and regards,
Koichiro Tashiro

over 2 years ago

0 Bin Liu over 2 years ago

TI__Guru**** 181546 points

Hi Tashiro-san,

Koichiro Tashiro said:
Q1) Is 5V VBUS is kept provided while AM623x is in Deep Sleep?

I don't think so, since the USB module will be powered off during Deep Sleep, but I will validate this on the EVM tomorrow and get back to you.

Koichiro Tashiro said:
Q2) Does AM623x USB support Type-C?

Yes, please refer to the SK-AM62 EVM USB0 port design. Technically USB Type-C defines the USB connectors, it has nothing to do with USB controllers.

Koichiro Tashiro said:
Q3) There are two USB channels. Each channel can work either Host and Device, correct?

Correct.

0 Koichiro Tashiro over 2 years ago in reply to Bin Liu

TI__Mastermind 22945 points

Hi Bin,

Bin Liu said:
I don't think so, since the USB module will be powered off during Deep Sleep, but I will validate this on the EVM tomorrow and get back to you.

Did you have a chance to try above?
If not, it will be done early next month?

Thanks and regards,
Koichiro Tashiro

0 Bin Liu over 2 years ago in reply to Koichiro Tashiro

TI__Guru**** 181546 points

Hi Tashiro-san,

Sorry, I have been in a full-day training, and haven't got time on this yet. But yes it will be done next week.

+1 Bin Liu over 2 years ago in reply to Koichiro Tashiro

TI__Guru**** 181546 points

Hi Tashiro-san,

I tested the USB1 port on SK-AM62B EVM, the 5V VBUS still presents during deepsleep.

0 Koichiro Tashiro over 2 years ago in reply to Bin Liu

TI__Mastermind 22945 points

Hi Bin,

Thanks for confirming.

Bin Liu said:
I tested the USB1 port on SK-AM62B EVM, the 5V VBUS still presents during deepsleep.

Just in case, this is different result compared to your first guess.
Is this correct behavior, right?

Thanks and regards,
Koichiro Tashiro

+1 Bin Liu over 2 years ago in reply to Koichiro Tashiro

TI__Guru**** 181546 points

Hi Tashiro-san,

Yes the behavior is correct. Though the processor goes into low power mode, USB VBUS should keep 5V, otherwise the USB device would got reset when VBUS 5v absent.

My initial guess was thinking the SoC MAIN domain, which the USB controller sits, would be powered off during deepsleep, then the USB controller would lost control of its pins. But it seems there are some SoC logic, which I don't know, that will keep the USB pin status. That is good.

0 Koichiro Tashiro over 2 years ago in reply to Bin Liu

TI__Mastermind 22945 points

Hi Bin,

Thanks for confirming.

Thanks and regards,
Koichiro Tashiro

0 Bin Liu over 2 years ago in reply to Koichiro Tashiro

TI__Guru**** 181546 points

Hi Tashiro-san,

I tested both SDK9.2.1.9 and SDK9.1.0.8, USB1 VBUS is on in Deepsleep in both SDK Linux.

I am using SK-AM62B (PROC114A).

You mentioned the customer uses SK-AM62-LP, what is the PROCxxxx silk print on the board? I want to check what exact the board which the customer has.

0 Koichiro Tashiro over 2 years ago in reply to Bin Liu

TI__Mastermind 22945 points

Hi Bin,

Thanks for checking.
The customer uses SK-AM62-LP (PROC124E2A).

Thanks and regards,
Koichiro Tashiro

0 Bin Liu over 2 years ago in reply to Koichiro Tashiro

TI__Guru**** 181546 points

Hi Tashiro-san,

I will try to find a PROC124E2 board next week to check its VBUS state. I am out of office today.

0 Bin Liu over 2 years ago in reply to Koichiro Tashiro

TI__Guru**** 181546 points

Hi Tashiro-san,

I am able to reproduce the behavior on SK-AM62-LP (PROC124E1), in which USB1 VBUS is 0v while AM62x is in DeepSleep.

I think the reason is that SK-AM62-LP has a pulldown on USB1_DRVVBUS, while SK-AM62B has a pullup on USB1_DRVVBUS. This also confirms my initial thought that during DeepSleep the main domain is powered off and the USB module doesn't drive its DRVVBUS pin. Then during DeepSleep SK-AM62B has 5v on VBUS because USB1_DRVVBUS has external pullup, but SK-AM62-LP has 0v on VBUS because its USB1_DRVVBUS has external pulldown.

0 Koichiro Tashiro over 2 years ago in reply to Bin Liu

TI__Mastermind 22945 points

Hi Bin,

I see. So the external pull-up is needed to use WoWLAN in the system.

Thanks and regards,
Koichiro Tashiro

0 Bin Liu over 2 years ago in reply to Koichiro Tashiro

TI__Guru**** 181546 points

Hi Tashiro-san,

I am not familiar with WoWLAN, but we know that USB wakeup from DeepSleep works on SK-AM62B or SK-AM62B-P1, but not on SK-AM62-LP, due to the DRVVBUS pullup/pulldown difference on the two EVMs.

Processors

Processors forum

AM623: Questions for UBS during in deepsleep