nRF9160 Feather - Suggestions for next version

zpm1066 · 2023-05-17T02:51:26+00:00

Hi Jared, Any chance you can share what new changes you have in mind for the next iteration of nRF9160 Feather? Although I do understand that board space i...

AchimKraus

Get really more than 12 bits from a ADC isn’t easy in my experience.
And sometimes it requires an additional amplification. So I’m not sure, how many will then be happy with it.

jaredwolff

Not sure if you guys have had this issue but modem traces don’t work well with non-Nordic devices since a while back. (The nRF9161 Feather will only have one USB-UART port)

Working on a PR for the Cellular Monitor so it can use any available port on your machine.

Great little tool especially to find out that these devices are no longer being pre-programmed with IMEIs! I will be pre-programming them here so no worries about that.

Nice thing is both the onboard LNA works for GPS. A single antenna is all that’s needed for the nRF9161 Feather!

Tracking:  5 Using:  5 Unhealthy: 0
-----------------------------------
Latitude:       41.XXX
Longitude:      -72.XXX
Altitude:       89.0 m
Accuracy:       10.0 m
Speed:          0.3 m/s
Speed accuracy: 1.2 m/s
Heading:        0.0 deg
Date:           2024-07-23
Time (UTC):     20:19:55.091
PDOP:           5.1
HDOP:           2.6
VDOP:           4.4
TDOP:           3.7

jaredwolff

Another very cool thing that the nPM1300 includes is an extremely low power mode which replicates what I did in the early designs of the nRF9160 Feather: deep powerdown/hibernate.

Here’s the measurement of the nRF9161 Feather in hibernate. About 600nA!

jaredwolff

On the Joulscope it’s actually a little less on avg (~300nA) :

AchimKraus

deep powerdown/hibernate.

Does this mean, the nRF9161 is power off?
If so, how does it power on again?

jaredwolff

Great question @AchimKraus!

There’s an internal timer on the nPM1300 which allows to shut down all regulators and then powers them back on after a certain time interval. I was testing every 10 seconds but I’m sure it can go longer than that. 😃

jaredwolff

Also if you insert USB or send a pulse on the EN pin it will also turn on.

zpm1066

jaredwolff
Do you have pull-up resistors on the Qwiic I2C? If not, it would be good to have them. Thanks.

jaredwolff

zpm1066 yup! 20k

zpm1066

jaredwolff Great! Thanks.

jaredwolff

Hey all

I’ll be assembling Proto 3 boards this week and I’d like you to give them a go. They should work nicely on the 2.7.x branch of nfed and the latest Zephyr Tools plugin. If you have time to try one in the immediate future, let me know and we can coordinate outside the thread.

zpm1066

jaredwolff Hi Jared.
I should be able to assist in testing the new nRF9161 Feather. I did install NCS 2.7.x a few weeks ago and have moved some of my CLI projects to Zephyr 3.7, although I did have to fix a couple of SSD1306 driver bugs. Also, ended up updating ST7789 displays because of the new MIPI Display Bus Interface (DBI) support in Zephyr 3.7. Anyway, it’s worth updating projects to Zephyr 3.7.x LTS.

Btw - What debug probe firmware did you end up using on the RP2040 interface MCU?

I’ll drop you an email later today. Thanks.

AchimKraus

Hi Jared,

I guess, it’s easier to ship the new boards within the U.S. than to ship them to Europe ;-).
Let’s see, which feedback your call gets.

My own app is still on NCS 2.6.1 and starts supporting the nRF9161-DK. Not sure, if that would also be a setup for testing the new feather. Update my app to 2.7.0 is planed, but for now without timeline.

jaredwolff

AchimKraus I guess, it’s easier to ship the new boards within the U.S. than to ship them to Europe ;-).

It is but not impossible. I can send it as a sample so it can result in lower/no VAT for you. 😃

AchimKraus My own app is still on NCS 2.6.1 and starts supporting the nRF9161-DK. Not sure, if that would also be a setup for testing the new feather. Update my app to 2.7.0 is planed, but for now without timeline.

I did build out the board definition for 2.6.1 as well. I may have to sync it with the 2.7.x branch.

AchimKraus

OK. To make it easier:
I will order/buy 3 feathers as usual in your shop, you send me 2x feather v6 and 1x new feather nRF9161.
No need to handle it special as sample or so from my side, the costs are not that high.
Would that be OK for you?

jaredwolff

That’s OK with me!

Use the code NRF9161 for some $ off as a thank you. 😃 Everyone else in the thread is also welcome to use the same. I’ll likely have them fully assembled and tested by tomorrow. I’ll also throw in the antenna I’ve been testing with that is specifically designed to work with GPS + LTE.

Jared

AchimKraus

Ordered ;-).

jaredwolff

zpm1066 Btw - What debug probe firmware did you end up using on the RP2040 interface MCU?

For now the standard one that the RPi foundation developed. The Probe-rs version, though excitingly developed in Rust, doesn’t have all the features yet to make it a viable option.

jaredwolff

Update!

So I ran into some interesting issues over the past week. I wanted to share some of them with you guys as I’m going.

Board Quiescent Current

When I first went over the quiescent current measurements, I was pleased with the results. I went to revisit on some of the newer boards I have here and something was wrong. I was getting anywhere between 10-120uA which is not great!

It’s difficult to isolate the exact cause with a fully assembled board so I built a one-side board for nPM1300 validation.

While their power supply chip does seem fairly robust, it seems there are some bugs around the usage of I2C. I was able to play with it more last night and got about 1uA which includes the Iq for the nPM1300 with one buck running along with the LIS2DH12 in deep sleep (typical 0.5uA)

I’m going to build another one today to confirm I can get the same results. Then attempt to assemble the other side. Assembling the module is tricky especially when hand applying solder paste. I threw together some jigs to make that process easier.

I do need to adjust this guy a bit to raise the PCB higher. Also will be adding some mounting holes so I can put one on the bed of the pick and place.

En pin operation

I was able to validate the EN pin which is hooked up to the SHPHLD pin on the nPM1300. This allows you to completely sleep the device and wake it up by pulling EN to ground for 96ms or plugging in USB. It was < 400nA which was great.

Other open items

Get the board as low current as possible. More on this today.
I have seen one issue so far when “recovering” the device when switching off buck2 for power saving purposes. This keeps the RP2040 powered off so it can be used to program the nRF9161. Usually you can hold the RST button and program (default power on is with both buck converters). This prevents the nRF9161, if programmed to do so, from turning off buck 2 which powers the RP2040.
Confirm the new QWIIC works as expected

zpm1066

jaredwolff
The solder paste application by hand is certainly tricky. You should do a short video at some stage to share the process of assembling the board.

« Previous Page Next Page »