Wake up event of Target Wake Time (TWT) on nRF7002-DK

RE: Wake up event of Target Wake Time (TWT) on nRF7002-DK

Håkon Alseth — Tue, 17 Oct 2023 14:41:20 GMT

Hi,

Thanks for posting the setup. This explains the difference between our results.

The problem is that the nRF5340 is not power optimized in this sample (logging is on, qspi etc), it is only the nRF7002, which explains why you're seeing a base current of ~2 mA.

Kind regards,

Håkon

RE: Wake up event of Target Wake Time (TWT) on nRF7002-DK

d_7 — Tue, 17 Oct 2023 14:10:11 GMT

Hi,

I thought that turning off VCOM0 VCOM1 would be enough to measure the current consumed in this way, but I have seen that there are other components consuming current. By connecting the multimeter to the P22 port, I was able to verify a consumption of approximately 27uA during sleep time.

Thank you

RE: Wake up event of Target Wake Time (TWT) on nRF7002-DK

Håkon Alseth — Mon, 16 Oct 2023 15:07:19 GMT

Hi,

I am running on the exact same tag, firmware unchanged, except for SSID/PWD used.

Here's the current profile:

And the log output:

[00:00:00.465,728] <inf> net_config: Waiting interface 1 (0x20001460) to be up...
[00:00:00.465,850] <inf> net_config: IPv4 address: 192.168.1.99
[00:00:00.465,911] <inf> net_config: Running dhcpv4 client...
[00:00:00.466,156] <inf> twt: Starting nrf7002dk_nrf5340_cpuapp with CPU frequency: 64 MHz
[00:00:01.466,308] <inf> twt: Static IP address (overridable): 192.168.1.99/255.255.255.0 -> 192.168.1.1
[00:00:03.387,786] <inf> twt: Connection requested
[00:00:03.387,847] <inf> twt: ==================
[00:00:03.387,847] <inf> twt: State: SCANNING
[00:00:03.687,988] <inf> twt: ==================
[00:00:03.688,018] <inf> twt: State: SCANNING
[00:00:03.988,159] <inf> twt: ==================
[00:00:03.988,189] <inf> twt: State: SCANNING
[00:00:04.288,299] <inf> twt: ==================
[00:00:04.288,330] <inf> twt: State: SCANNING
[00:00:04.588,439] <inf> twt: ==================
[00:00:04.588,470] <inf> twt: State: SCANNING
[00:00:04.888,580] <inf> twt: ==================
[00:00:04.888,610] <inf> twt: State: SCANNING
[00:00:05.188,751] <inf> twt: ==================
[00:00:05.188,781] <inf> twt: State: SCANNING
[00:00:05.488,922] <inf> twt: ==================
[00:00:05.488,952] <inf> twt: State: SCANNING
[00:00:05.789,062] <inf> twt: ==================
[00:00:05.789,093] <inf> twt: State: SCANNING
[00:00:06.089,202] <inf> twt: ==================
[00:00:06.089,233] <inf> twt: State: SCANNING
[00:00:06.389,373] <inf> twt: ==================
[00:00:06.389,404] <inf> twt: State: SCANNING
[00:00:06.689,514] <inf> twt: ==================
[00:00:06.689,544] <inf> twt: State: SCANNING
[00:00:06.989,654] <inf> twt: ==================
[00:00:06.989,685] <inf> twt: State: SCANNING
[00:00:07.289,825] <inf> twt: ==================
[00:00:07.289,855] <inf> twt: State: SCANNING
[00:00:07.589,996] <inf> twt: ==================
[00:00:07.590,026] <inf> twt: State: SCANNING
[00:00:07.890,136] <inf> twt: ==================
[00:00:07.890,167] <inf> twt: State: SCANNING
[00:00:08.190,277] <inf> twt: ==================
[00:00:08.190,307] <inf> twt: State: AUTHENTICATING
[00:00:08.437,957] <inf> twt: Connected
[00:00:08.495,849] <inf> net_dhcpv4: Received: 192.168.32.128
[00:00:08.495,971] <inf> net_config: IPv4 address: 192.168.32.128
[00:00:08.496,002] <inf> net_config: Lease time: 43200 seconds
[00:00:08.496,032] <inf> net_config: Subnet: 255.255.255.0
[00:00:08.496,063] <inf> net_config: Router: 192.168.32.1
[00:00:08.496,154] <inf> twt: DHCP IP address: 192.168.32.128
[00:00:08.498,382] <inf> twt: ==================
[00:00:08.498,382] <inf> twt: State: COMPLETED
[00:00:08.498,413] <inf> twt: Interface Mode: STATION
[00:00:08.498,443] <inf> twt: Link Mode: WIFI 6 (802.11ax/HE)
[00:00:08.498,474] <inf> twt: SSID: OpenWrt                         
[00:00:08.498,504] <inf> twt: BSSID: FC:34:97:0B:F6:DC
[00:00:08.498,535] <inf> twt: Band: 5GHz
[00:00:08.498,535] <inf> twt: Channel: 36
[00:00:08.498,565] <inf> twt: Security: WPA2-PSK
[00:00:08.498,565] <inf> twt: MFP: Optional
[00:00:08.498,596] <inf> twt: RSSI: -60
[00:00:08.498,596] <inf> twt: TWT: Supported
[00:00:10.498,687] <inf> twt: AP is TWT capable, establishing TWT
[00:00:10.511,627] <inf> twt: TWT setup requested
[00:00:10.519,897] <inf> twt: TWT response: TWT accept
[00:00:10.519,927] <inf> twt: == TWT negotiated parameters ==
[00:00:10.519,927] <inf> twt: TWT Dialog token: 1
[00:00:10.519,927] <inf> twt: TWT flow ID: 0
[00:00:10.519,958] <inf> twt: TWT negotiation type: TWT individual negotiation
[00:00:10.519,989] <inf> twt: TWT responder: true
[00:00:10.520,019] <inf> twt: TWT implicit: true
[00:00:10.520,019] <inf> twt: TWT announce: false
[00:00:10.520,050] <inf> twt: TWT trigger: false
[00:00:10.520,080] <inf> twt: TWT wake interval: 65024 us
[00:00:10.520,080] <inf> twt: TWT interval: 5005000 us
[00:00:10.520,080] <inf> twt: ========================
[00:00:10.520,812] <wrn> twt: TWT SLEEP
[00:00:15.507,629] <wrn> twt: TWT AWAKE
[00:00:15.573,364] <wrn> twt: TWT SLEEP
[00:00:20.513,671] <wrn> twt: TWT AWAKE
[00:00:20.580,688] <wrn> twt: TWT SLEEP
[00:00:20.594,268] <wrn> twt: TWT AWAKE
[00:00:20.594,390] <wrn> twt: TWT SLEEP
...

I am not sure what the difference is in our setups. I'm supplying the VBAT from the PPK2, with 3.6V and measuring that current.

Can you share a picture or similar with your measuring setup?

Kind regards,

Håkon

RE: Wake up event of Target Wake Time (TWT) on nRF7002-DK

d_7 — Mon, 16 Oct 2023 09:07:55 GMT

Hi,

Yes, I believe so. At least, I'm obtaining the following output:

Now I'm also using SDK v2.4.99-dev2 and a Asus RT-AX53U router.

This is my sample code:

devzone.nordicsemi.com/.../twtNordicSample.zip

Thank you

RE: Wake up event of Target Wake Time (TWT) on nRF7002-DK

Håkon Alseth — Mon, 16 Oct 2023 07:22:48 GMT

Hi,

I set the TWT to 5.24 seconds, and wake interval to 65 ms, and this is my observed current during TWT:

Are you certain that you actually enter TWT? TWT shall have a floor current of less than 20 uA.

Kind regards,

Håkon

RE: Wake up event of Target Wake Time (TWT) on nRF7002-DK

d_7 — Fri, 13 Oct 2023 14:23:59 GMT

Hello again,

Now I can get the TWT events, but I have a question. At sleep time I'm supposed to have a uA consumption, right?

I've been looking at the power consumption with the Joulescoupe and the result is as follows:

From the image above, I can see that the average current consumption is 2,697mA (approx).

If I only pay attention to the sleep time, we have an average current consumption of 1.96mA (approx), as we can see in the following image:

Isn't this too much? Shouldn't it be in the order of the uA?

I'm testing with CONFIG_TWT_INTERVAL=5 sec and CONFIG_TWT_WAKE_INTERVAL=65 ms. I've tested with the default values (524ms and 65ms) and I have the same results.

Thank you

RE: Wake up event of Target Wake Time (TWT) on nRF7002-DK

Håkon Alseth — Tue, 03 Oct 2023 12:25:29 GMT

Hi,

I'm using a Asus RT-AX53U, where I have enabled TWT on both 2.4 GHz and 5 GHz bands.

Their firmware requires to enable it explicitly on both bands, where the default is "disabled":

[quote user="diogo_"]Since TWT is a wifi 6 feature, if the AP has wifi 6, shouldn't it support TWT?[/quote]

Yes, but this is a feature that most access points need to explicitly enable.

[quote user="diogo_"]Do you have any AP list that allows TWT?[/quote]

Most wifi-6 capable APs support it. I have both the Google nest wifi pro and the above mentioned Asus (most Asus wifi 6 capable routers have this feature)

Kind regards,

Håkon

RE: Wake up event of Target Wake Time (TWT) on nRF7002-DK

d_7 — Tue, 03 Oct 2023 12:01:46 GMT

HI,

I didn't understand what I was supposed to do because using the Toolchain Manager, I could only install up to SDK v2.4.2.

I managed to install v2.4.99-dev2 through vscode, but now I have more questions.

While through v2.4.2 I got an output saying that the AP supported TWT, the output I have now is the following:

Therefore, my questions are the following:

Since TWT is a wifi 6 feature, if the AP has wifi 6, shouldn't it support TWT?
What AP are you using to test this?
Do you have any AP list that allows TWT?

RE: Wake up event of Target Wake Time (TWT) on nRF7002-DK

Håkon Alseth — Tue, 03 Oct 2023 11:48:05 GMT

Hi,

you will need to use a terminal or a command prompt, which has "west" and "git" in the path. This can be done by opening the toolchain manager and selecting command line or git bash, or "terminal" on mac/linux:

You should be placed in the "root" of ncs, where you have to change to the "nrf" directory:

cd nrf
git fetch
git checkout v2.4.99-dev2
west update

The above commands can take some time, depending on your internet connection and so forth.

Note: your repositories must be clean, ie. no changed files or uncommited changes.

Kind regards,

Håkon

RE: Wake up event of Target Wake Time (TWT) on nRF7002-DK

d_7 — Mon, 02 Oct 2023 17:08:00 GMT

Hi,

Yes, I'm testing on NCS v2.4.2.

What you are telling me is to replace the nrf folder (inside v2.4.2) with sdk-nrf-2.4.99-dev2? If yes, I get multiple errors when I create new build configuration in vscode.

RE: Wake up event of Target Wake Time (TWT) on nRF7002-DK

Håkon Alseth — Mon, 02 Oct 2023 12:55:34 GMT

Hi,

My apologies, but I've been testing on the devtag "v2.4.99-dev1" (dev2 is also out now), which includes this specific fix for the TWT sleep:

https://github.com/nrfconnect/sdk-nrf/pull/11488/files

Are you testing on NCS v2.4.2 or older? If yes, could you try to test on tag v2.4.99-dev2 and see if you are able to get the sleep events?

Kind regards,

Håkon

RE: Wake up event of Target Wake Time (TWT) on nRF7002-DK

d_7 — Fri, 29 Sep 2023 14:48:11 GMT

Hi,

Yes, I can send the algorithm.

devzone.nordicsemi.com/.../twtUpdated.zip

The AP that I'm using is the UniFi 6+.

RE: Wake up event of Target Wake Time (TWT) on nRF7002-DK

Håkon Alseth — Fri, 29 Sep 2023 14:17:27 GMT

Hi,

Yes, as far as I see, you should now have gotten TWT events.

As your log output shows "wrn", it should be configured with the correct. Are you certain you set CONFIG_SCHEDULED_TX for this build? Could you try removing that ifdef and see if it starts printing? Here's the output when I tried adding the same events and added a LOG_INF print in the event:

...
[00:00:08.700,500] <inf> twt: DHCP IP address: 192.168.32.128
[00:00:08.703,063] <inf> twt: ==================
[00:00:08.703,094] <inf> twt: State: COMPLETED
[00:00:08.703,094] <inf> twt: Interface Mode: STATION
[00:00:08.703,125] <inf> twt: Link Mode: WIFI 6 (802.11ax/HE)
[00:00:08.703,155] <inf> twt: SSID: OpenWrt                         
[00:00:08.703,186] <inf> twt: BSSID: FC:34:97:0B:F6:DC
[00:00:08.703,216] <inf> twt: Band: 5GHz
[00:00:08.703,216] <inf> twt: Channel: 44
[00:00:08.703,247] <inf> twt: Security: WPA2-PSK
[00:00:08.703,247] <inf> twt: MFP: Optional
[00:00:08.703,277] <inf> twt: RSSI: -55
[00:00:08.703,277] <inf> twt: TWT: Supported
[00:00:10.703,369] <inf> twt: AP is TWT capable, establishing TWT
[00:00:10.715,850] <inf> twt: TWT setup requested
[00:00:10.817,718] <inf> twt: TWT response: TWT accept
[00:00:10.817,749] <inf> twt: == TWT negotiated parameters ==
[00:00:10.817,749] <inf> twt: TWT Dialog token: 1
[00:00:10.817,749] <inf> twt: TWT flow ID: 0
[00:00:10.817,779] <inf> twt: TWT negotiation type: TWT individual negotiation
[00:00:10.817,810] <inf> twt: TWT responder: false
[00:00:10.817,840] <inf> twt: TWT implicit: true
[00:00:10.817,840] <inf> twt: TWT announce: false
[00:00:10.817,871] <inf> twt: TWT trigger: false
[00:00:10.817,901] <inf> twt: TWT wake interval: 65024 us
[00:00:10.817,901] <inf> twt: TWT interval: 524000 us
[00:00:10.817,901] <inf> twt: ========================
[00:00:10.818,664] <inf> twt: NET_EVENT_WIFI_TWT_SLEEP_STATE
[00:00:10.818,725] <inf> twt: NET_EVENT_WIFI_TWT_SLEEP_STATE
[00:00:11.015,960] <inf> twt: TWT Setup success
[00:00:11.195,495] <inf> twt: NET_EVENT_WIFI_TWT_SLEEP_STATE
[00:00:11.654,846] <inf> twt: NET_EVENT_WIFI_TWT_SLEEP_STATE
...

[quote user="diogo_"]Q2: When I obtain the result shown in Figure 5, the current consumption always remains around 173mA and never decreases again (even after those 60 seconds).[/quote]

173 mA? That is alot. I have seen RX current levels for 60 seconds if a operation fails (on-air) on previous ncs versions, but this seems like a larger scale problem. If you have an algorithm, or a sample, where this can be seen, I would be very interested in replicating and reporting it internally.

Kind regards,

Håkon

RE: Wake up event of Target Wake Time (TWT) on nRF7002-DK

d_7 — Thu, 28 Sep 2023 15:00:44 GMT

Hi,

I decreased the TWT interval to 200ms but the problem remained (I did not detect the wake up and sleep events).

Regarding your questions:

Q1: What is seen in Image 5 does not occur every time. Although the code is always the same, sometimes it presents the result shown in Image 3 and other times it presents the result in Image 5.
Q2: When I obtain the result shown in Figure 5, the current consumption always remains around 173mA and never decreases again (even after those 60 seconds).
Q3: I'm testing with a Ubiquiti AP. I'm using the UniFi 6+.

I was also testing the recently updated twt sample in main and I still can't detect the events. In fact, it seems that it doesn't even enter the NET_EVENT_WIFI_TWT_SLEEP_STATE case.

In this sample I made the following changes:

main.c :
1. Added the NET_EVENT_WIFI_TWT_SLEEP_STATE to the WIFI_SHELL_MGMT_EVENTS:
2. Added LOG_WRN to the switch...case to check if these events occurred:

Kconfig (samples/wifi/twt) :
1. As I want to use TWT, I changed the default value for SCHEDULED_TX

prj.conf :

Added the following configugurations:

CONFIG_TRAFFIC_GEN_CLIENT=y
CONFIG_TRAFFIC_GEN_SERVER=n
CONFIG_TRAFFIC_GEN_DURATION=10
CONFIG_TRAFFIC_GEN_TYPE_TCP=y
CONFIG_TRAFFIC_GEN_REMOTE_IPV4_ADDR="10.10.10.76"
#
CONFIG_TWT_TRIGGER_ENABLE=y
CONFIG_TWT_ANNOUNCED_MODE=y
CONFIG_SCHEDULED_TX=y

The result is as follows:

So my questions are:

With the changes I've made, if TWT was working correctly, shouldn't I have gotten the LOG_WRN("NET_EVENT_WIFI_TWT_SLEEP_STATE event")?
And the LOG_INF("SLEEP state") and LOG_INF("AWAKE state")?

As far as I can see, I'm still not able to detect these events. Right?

RE: Wake up event of Target Wake Time (TWT) on nRF7002-DK

Håkon Alseth — Wed, 27 Sep 2023 14:47:57 GMT

Hi,

[quote user=""]What am I doing wrong in the wake up and sleep events detection (in order to send data via UDP during the wake time)?[/quote]

I would recommend that you first try with a low TWT_INTERVAL, and then increase this when you have a working setup.

We are continuously improving our handling in this case. Have you looked at the recently updated twt sample in main? It includes handling for this scenario:

https://github.com/nrfconnect/sdk-nrf/commit/e05d9c964c8d7a5f77418347deda6f5795cf9b7e

[quote user=""]Sometimes looks like the TWT is not set (Image 5):[/quote]

Q1: Does the firmware run as expected here?

Q2: Does the current consumption drop after 60 seconds?

Q3: Which AP are you testing against?

Kind regards,

Håkon