XIAO nRF54LM20A

[msfujino]

I received an email from Seeed announcing that the XIAO nRF54LM20A is scheduled for release. I don’t have the full details yet, but judging by the photos, the ceramic antenna (and probably the RF_SW as well) has been removed and replaced with a smaller, more fragile reset/boot switch.
@lolren, if possible, I’d like you to consider developing an Arduino Core for this.

[lolren]

They told me a sample will be sent to me as soon as one is available :D
I will change this core to make it generic NRF54 instead of NRF54L15 specific, so I can support each individual MCU :D

[msfujino]

I plan to order it as soon as it’s released.

[msfujino]

I received a notification from DHL that two samples provided by Seeed, complete with FPC antennas, will arrive next week.
First, I’ll try running them on the Zephyr.

[lolren]

Josie from Seeed contacted me yesterday as well to send me as well. Looking forward to receiving them :D

[lyusupov]

USB of nRF54LM20A is the feature that I was missing in nRF54L15 to compare with nRF52840

[lolren]

I loved the take on NRF54L15 with a different chip.
Made development so much easier due to not requiring a hard reset every time😅
Hope porting the core for this would be easy as they should not be much different.

[msfujino]

I received two samples. I can’t get them to work yet because the device tree and schematics aren’t available. I’ve been told they’ll be released sometime this week.

[lolren]

Im curious of they regulator they used more than everything. You can test using a led on the 3.3v line and power it via vbat and power it with 3.4v and lower voltage 😅

[msfujino]

I hear they're using the NPM1300 power supply.

[lolren]

NPM1300 just seen that in the picture you posted above! must have missed it 🗡️

[lyusupov]

Correction

XIAO nRF54LM20A ( A ) likely has no access to native USB yet.
Native USB is expected to be a part of nRF54LM20B ( B )

According to this Seeed-Studio/OSHW-XIAO-Series#1 (reply in thread) the 20B is scheduled onto

DVT Samples: Mid-to-late August

Official Launch: Late September or early October

[lolren]

Xiao nRF54LM20A should be supported. Serial is not yet supported, as I don't have the pinout yet. I've requested the schematic . I've also patched pyocd to support the target but I have yet to test it on another OS or another pc than mine. Will try later.

[lolren]

Update: Finally, managed to fix the upload on Linux. I suspect Windows upload is broken for this board. Will fix it in the next days. Serial is broken because I don't have the pin-out yet. Most of the stuff should work on the board. Haven't tested much.

[msfujino]

Using v0.9.19, I was able to make the LED (RGB) blink.

I thought the only way to evaluate the 54LM20 was to compile it with Zephyr for now. If ArduinoCore has started supporting the 54LM20, I’ll proceed with the evaluation using ArduinoCore.

It appears that SAMD11 is being used again instead of USB. Porting from the 54L15 might be straightforward.

[lyusupov]

It appears that SAMD11 is being used again instead of USB

It is valid for 20A ( this topic name ) only.

20B will use native USB
Seeed-Studio/OSHW-XIAO-Series#1 (comment)

[msfujino]

I took a quick look at the schematic, and the power supply circuit seems a bit overdesigned. This is just my personal opinion, but it gives the impression that it was designed with the nPM1300 in mind. I plan to evaluate whether the “XIAO” actually needed the nPM1300.

[lolren]

There is also a RGB blink sketch in the examples in board menu😅

[lolren]

BLE advertising worked yesterday but it's the only Ble feature I've tested ( because the kids were very noisy around me, I could not do much work)😅.
Windows upload is broken, serial is broken. A lot of addresses are very different than Nrf54L15.
Looks like the upper 64k is reserved by another core. I might be wrong, need to check more.
I'm more curious of the power consumption with the new PMIC IC🙂.

[msfujino]

It feels slightly warm to the touch. That might be because it contains three regulators.

[lolren]

WoW . hope it's only a issue from USB. I'm hopping that it has very low power consumption via batt pads.

[lolren]

I've written some PMIC examples. And a PMIC driver. Check it out in examples, PMIC folder.

[lolren]

Most of the stuff are working. Some optimisations needs to be written for the PMIC driver as power goes to about 2.5mA when talking to it .
BLE works, IMU should work, mic doesn't work yet.
Somehow, I think its worse than a LDO on the power.

[msfujino]

It was a misunderstanding on my part.
According to Table 19, VOUT2 was set to 3.3 V with R10 = 470 kΩ.

[lolren]

PMIC driver is still work in progress, but I doubt we will get better optimisation than auto. It looks a bit worse while sleeping, but better overall.
But I'm waiting for the professional results ;)

[msfujino]

Please explain what this means.
Better optimization than auto
Professional results

[lsw031105guge-bit]

Wow, you are really amazing! We recently updated the mic section. Welcome to discuss and raise questions together!

[lolren]

Wow, you are really amazing! We recently updated the mic section. Welcome to discuss and raise questions together!

Thank you for the comment! This was what we needed to get the mic working!

[lolren]

The PMIC used has 3 modes:
Auto, Forced PWM, Hysteretic
Currently , I think it's in auto mode( but i need to investigate further).
Also. Professional results I've meant your results from PPK2 testing. For me, you are the professional tester here. I never heard of a PPK before you mentioned it😅

[msfujino]

I mistakenly thought that Zephyr was still developing the PMIC driver and that we were waiting for the results.

[lolren]

Nope. We are on our own :D We can compare with Zephyr when its supported.
For now, just test like that. I will develop a better driver in the following days. 💯
By professional, I've meant you. Looking at your provided results for power consumption and bug reports, they really look professional. My measurements are very sloppy and basic :)

[msfujino]

Here is a comparison of the sleep currents for the LM20A and L15.
When using delaySystemOffNoRetention(), both draw approximately 3 μA, which closely matches a rough estimate based on the datasheet values.
When using delay(), the current drawn by the LM20A is nearly double that of the L15, suggesting that there is some difference between the two.

[lolren]

Looking at delay now.

[lolren]

I've seen this here.. after last core fix, both take roughtly the same current, delay and delaySystemOffNoRetention() so there is a bug there.

[lewuq]

Hi, I'm curious about your test conditions. I'm conducting measurements with a PPK2, powered by a 4.15V lithium battery with a nominal capacity of 1080mAh. The project runs in System Off mode with GPIO interrupt wakeup using the on-board Boot button, yet the current draw hovers around 5.2 µA.

[lolren]

All tests are done by connecting the PPK2 set on 3900mv on the battery pads.

[msfujino]

Since there was a version update, I measured the sleep current. The current values have increased compared to v0.9.59, when I posted the graph. Please take another look at this when you have time.

　　　　    v0.9.59　　v0.9.148　　v0.9.194
Flash ON　   22uA       28uA        28uA
delay         7uA       13uA        13uA
SystemOff     3uA       13uA        13uA

EDIT:
In every version, even when I set NRF_REGULATORS->SYSTEMOFF = 1;, the sleep current is still as high as 7uA.

[lolren]

Although I've said I will take a break, I could not help myself from doing something today.
I've created this replace pyocd as the default upload method.
This is to get rid of all python deps. Still not perfect, but its getting there

[lewuq]

You're absolutely amazing!

[lolren]

Thank you. It's a community effort actually.

[msfujino]

Have you checked out the link below? It's an excellent document.
https://academy.nordicsemi.com/courses/designing-low-power-bluetooth-le-products/

[msfujino]

I've posted a few sample sketches on the SeeedForum.
https://forum.seeedstudio.com/t/the-xiao-nrf54lm20a-runs-on-nrf54-arduino-core/295519

[lolren]

Looks nice. I can see you are very passionate about this. Kudos for all your work.
Things are starting to look good for this board as well
I was merging Thread and Matter work yesterday (hence the errors). I hope Zigbee, Thread and Matter will work in the near future as expected.

[msfujino]

I posted “Battery Charge/Discharge Monitor Using the Built-in nPM1300” on the SeeedForum. If you're interested, please take a look.

https://forum.seeedstudio.com/t/the-xiao-nrf54lm20a-runs-on-nrf54-arduino-core/295519/5

[msfujino]

I posted “0.5 μA Sleep Current in Hibernate Mode Using the Built-in nPM1300” on the Seeed Forum. If you're interested, please take a look.
I think it would be nice to have a helper function that sets the registers required for hibernate mode.

https://forum.seeedstudio.com/t/the-xiao-nrf54lm20a-runs-on-nrf54-arduino-core/295519/11

[lolren]

Wow! that is very low indeed. I will add a function to set it like that! nice find!

[lolren]

Added this in v0.9.213.

There is now a PMIC helper for the LM20A hibernate timer path:

npm1300_enter_timed_hibernate_ms(2000);

I also added npm1300_enter_hibernate_after_ms() as an alias, plus an Arduino example under File > Examples > Nrf54L15-Clean-Implementation > PMIC > nPM1300_TimedHibernate.

It uses the same nPM1300 timer setup from your test: 16 ms prescale, timer target load, timer start, then TASKENTERHIBERNATE. Please test from the battery/VBAT pads, because USB/debug wiring can still dominate the current reading.

[msfujino]

I verified this in v0.9.213.
The code npm1300_enter_timed_hibernate_ms(2000); worked as expected, and the sleep current dropped to 0.5 uA.
Thank you very much.

[lolren]

No worries. Is your code under the hood. You are welcome.