Oh, sorry. UART resets using driver_delete, SPI just calls "free" which I2C doesn't even have an equivalent for
#circuitpython-dev
ionic elk
tulip sleet
so the IDF API is not very consistent
is that true?
but UART does not have this problem, is that correct?
I would look at the IDF impl to see what's different about driver_delete for UART vs I2C, if one works but the other doesn't
ionic elk
Not sure what you mean by "works", works as in it doesn't shut down wifi?
They work fine for actually resetting the bus
serene harness
That information from before is for a magtag project
tulip sleet
Doesn't produce the same bugs we are seeing with I2C and wifi interactions
ionic elk
Right, ok, sorry wanted to make sure we were synced
TBH UART could cause the same issues and that simply was never tested
tulip sleet
it would be interesting to test that, then
ionic elk
I'll give it a shot and see
tulip sleet
@serene harness You have some kind of issue with line 63, also could you ask in #help-with-circuitpython. There are a few more helpers over there. This is more for circuitpython dev issues. Thanks.
serene harness
Oh, I apologize. Thought that I was in that forum.
tulip sleet
np! don't upload as a .odt. It's hard to figure out the line numbers. Just upload the code.py
simple pulsar
@idle owl For data logging to CIRCUITPY, are there any recommendations for how to log? I've been doing this over weekend and I realised yesterday that if I keep doing this I might wear out my flash a bit prematurely: https://forums.adafruit.com/viewtopic.php?f=65&t=175527
idle owl
@simple pulsar Oooh that's a good question. I'm not sure.
onyx hinge
@slender iron apologies for repeating myself, but better to mention it in public. You were looking for a lost thread on double-tap-for-bootloader in rp2040. is this it? I wonder if this is the discussion of double-tap bootloader for rp2040 that scott was looking for: https://github.com/raspberrypi/pico-sdk/issues/87#issuecomment-773929609 -> https://github.com/raspberrypi/pico-examples/tree/master/system/double_tap_usb_boot
tulip sleet
<@!330227457296957440> For data logging to CIRCUITPY, are there any recommendati...
answered you in the forum. short answer is that 10 second cycles is pretty fast.
simple pulsar
@tulip sleet thanks, useful to know
idle owl
@tulip sleet So on the Pico, only the three pins can do analogue? 26-28?
tulip sleet
looking ...
@idle owl, yes, the pins are fixed. GPIO29 is actually a real chip pin, but on the Pico, it is connected to a voltage divider to measure the 3.3V supply, and is not broken out to the edge.
tulip sleet
ADC0, 1, 2,3 are GPIO26-29.
idle owl
Right, fourth channel is temp.
slender iron
@onyx hinge I was trying to remember where I thought I saw a schematic for doing the double tap in circuitry
idle owl
Thanks Dan!
idle owl
um...
tulip sleet
channel #4 is temp 🙂
idle owl
The book seems to indicate there are only 4 channels
3 channels to GPIO, and a fourth connected to temp
slender iron
it's wrong then
idle owl
Hah ok
tulip sleet
they are not counting the GPIO29 voltage monitor?
book is talkign about the board, not the chip
idle owl
Oh right
So am I though
ok
It's super not clear though if that's the case. They talk about the chip in the same paragraph. Bleh. Whatever.
slender iron
they were probably confused themselves
ionic elk
@tulip sleet on a debug optimized build, I still don't see anything weird going on during the Wifi/I2C conflict hang. It's just looping in run_code_py, and should be running background tasks, but it isn't maintaining the USB or keyboard detection. Makes me think that there's something going haywire in the I2C interrupt system, messing with USB interrupt priority
tulip sleet
hmm! maybe time for some debug uart printouts
scenic echo
I just received a FeatherS2. I’m trying to play with toggling the LED via a webpage. Has anyone gotten it to serve a page in AP mode? (I don’t want to connect to an existing network, but create its own point)
slender iron
@scenic echo I recommend asking on the unexpectedmaker discord first
scenic echo
@slender iron they sent me here 🙂
slender iron
haha, ok. I don't know of any examples for that
ionic elk
@tulip sleet I remembered another weird thing I wanted to ask about. When testing the original issue, where the whole chip resets when you tried to use I2C after a soft reboot, it would only happen if I used board.I2C(), not a busio.I2C instance. Poking around, I figured the only difference between these two is that a board.I2C might have the object preserved for longer in a soft reboot? Or something?
Honestly I'm running out of ideas of things to try.
oh, and UART doesn't cause a wifi conflict. The underlying reset code is totally inconsistent between them though, so it'll be hard to narrow down what actual change might be causing this.
tulip sleet
check where the board.I2c instance is stored, whether it's in the heap or separate. If it's separate, maybe it is being garbage-collected by accident. If it's not in root pointers or not handled in a _gc routine, then it's not being walked during gc and any objects it points to will be gc'd and possibly smashed
The underlying reset code is totally inconsistent between them though
that is telling, maybe
ionic elk
I think the board.I2C instance is just STATIC mp_obj_t i2c_singleton = NULL; in shared-module
tulip sleet
when it's set, is it set to be a heap object? See how it's implemented on some other port, like atmel-samd
idle owl
If anyone has a moment to review: https://github.com/adafruit/Adafruit_Learning_System_Guides/pull/1457
ionic elk
@tulip sleet no, it's statically allocated
// Statically allocate the I2C object so it can live past the end of the heap and into the next VM.
// That way it can be used by built-in I2CDisplay displays and be accessible through board.I2C().
STATIC busio_i2c_obj_t i2c_obj;
but if its struct has internal pointers that might be heap allocated, then that's an issue
ionic elk
Oh, hmm, I see your point
tulip sleet
so the second time around it will use the existing one. but maybe it got messed with during vm restart
i would again check if the port-specific code differs from, say atmel-samd
@idle owl where is guide page for this?
ionic elk
None of the above is port specific, it's all shared-module
idle owl
@tulip sleet Not published yet. I can link you in Slack if you want.
tulip sleet
yah
idle owl
It's in the Pico guide
tulip sleet
tnx
idle owl
Linked
ionic elk
But I'm wondering here, if the IDF tries to allocate something in the heap, does that get cleared? Or does it have its own heap?
idle owl
@tulip sleet If you have suggestions, we can chat here, I try to avoid putting the internal links in public channels.
tulip sleet
aprvd
idle owl
Ah thanks!
ionic elk
I'm going to try what Bruce suggested and update the IDF to something more recent, see if that helps
manic glacierBOT
modern wing
Good afternoon all you wonderful folks -- happily lurking/text only. Updated the notes doc with those details. Will be updating myself with a hot cup of tea. 🍵
idle owl
<@&356864093652516868> CircuitPython Weekly is in ~15 minutes. Please add your notes to the notes doc. If you're listening in or lurking, please add your name to the lists under Hug Reports and Status Updates alphabetically with (lurking) after it so we know to skip over you. Thanks much!
tulip sleet
wow, forgot all about that
fossil gorge
Mostly lurking, but should be able to hop in for my "in the weeds" topic if need be. However, I'm fairly certain I'm not a CircuitPythonista. Probably more CircuitPythonistish 🙂
idle owl
@fossil gorge Welcome to the role!
slender iron
I just added you 🙂
idle owl
We both did 😄
slender iron
great minds think alike 🙂
fossil gorge
lol, thanks. It's all downhill from here folks!
Mostly coz it's snowy and iced over in Dallas
modern wing
Thankfully it wasn't bad in NJ....early projections were 1/3rd inch of ice. We barely got any.
solar whale
What happend to the pinned messages?
idle owl
@solar whale They're showing up for me.
supple gale
I'm going to try what Bruce suggested and update the IDF to something more recen...
I have other issues with the 4.3 beta idf. When I get home later I will try to see if it’s i2c related or some other incompatibility. @analog bridge did a branch of the IDF that compiles but has issues. I fixed some but it still crashes often. Unrelated to i2c probably.
idle owl
I think it's a Discord glitch.
solar whale
they're back 😉
idle owl
Magical!
ionic elk
@supple gale thanks for the update - I've spent more time digging at the I2C bug but I don't think I've turned up anything useful, unfortunately.
solar whale
When we lose power -- it's squirrels....
ionic elk
I think I'm going to put it down until we get the 4.3 update changes together, and start afresh from there. No point in solving a super-involved bug if the next version update breaks it again.
tidal kiln
lurking
blissful pollen
Had squirrels take out my internet before
fossil gorge
Had squirrels take out my internet before
Had that happen at work before. Fun times..., after they sent us home for the afternoon
modern wing
I've had squirrels in my home....put cashews....in the shoes under my bed. That was a great thanksgiving vacation, coming back to that.
solar whale
What -- no squirrel emoji!
modern wing
🐿️
solar whale
That's not a squirrel!
fossil gorge
I've had squirrels in my home....put cashews....in the shoes under my bed. That ...
"Where's Eddie? He usually eats these things"
- C.W. Griswold
errant grail
🔥
onyx hinge
We're soon changing how lurkers work a little bit, stay tuned. (not today)
solar whale
anyone else having trouble accessing the Google Doc?
onyx hinge
@solar whale it's fine for me, was just editing it
idle owl
@solar whale No... it's all issues for you today!
solar whale
just came back to life -- must have been a squirrel 😉
modern wing
[they think I'm a real human!]
stuck elbow
I'm not a cat!
still zephyr
Lurking
errant grail
Not my favorite designation. I'm just listening.
onyx hinge
Meeting times are also available in iCal format using the following link: https://raw.githubusercontent.com/adafruit/adafruit-circuitpython-weekly-meeting/master/meeting.ical for use with standard calendar applications or in your browser: https://open-web-calendar.herokuapp.com/calendar.html?url=https://raw.githubusercontent.com/adafruit/adafruit-circuitpython-weekly-meeting/master/meeting.ical
onyx hinge
https://github.com/adafruit/circuitpython-weekly-newsletter/tree/gh-pages/_drafts -- submit a pull request. You may also tag @anne_engineer on Twitter, or email anneb AT adafruit.com.
idle owl
Haha I didn't read it off because the name was so long 😄
old smelt
Lurking
tulip sleet
it took me days to memorize lywsd03mmc well enough to type it without copy/paste
tidal kiln
looks like a cheat code
supple gale
Tuned in just in time for a listen.
fossil gorge
TIL that even with PTT, you need to unmute Discord. The more you know...
analog bridge
just stopping by to say hi 👋 to all <@&356864093652516868>
bleak spear
I had no problem getting the gamepad examples setup but those examples have the same problem I was talking about. No way to implement analog trigger buttons. Only the analog sticks and 16 buttons. It appears to just not be implemented in circuitpython at all and I can't parse the micropython well enough to see how to add it. @steep finch mentioned they have a fork exposing those buttons, but I couldn't find it. @steep finch do you have a link? I should asked earlier.
cobalt grail
Sorry. Lurking
onyx hinge
yeah you did, I didn't get it in the doc
slender iron
@low sentinel are you lurking?
low sentinel
Yeah! Just realized this was happening with the role ping.
slender iron
😄
tidal kiln
onyx hinge
in case anyone else wonders about those pieces of software: http://astyle.sourceforge.net/ https://clang.llvm.org/docs/ClangFormat.html
oh there's another one -- http://uncrustify.sourceforge.net/
fossil gorge
Issues and PRs - a good place to contribute and enhance, and scratch some of that "code it from scratch" itch
onyx hinge
belated hug report to dan for doing that sdcard work
low sentinel
2nd usb channel is amazing
tulip sleet
^^ uncrustify is what MicroPython finally settled on
low sentinel
I use that with a car dashboard with particle photon. I can log and drive the display with a serial test harness over usb. Just luxury 🙂
steep finch
@bleak spear https://github.com/alexbartlow/circuitpython/tree/hotas This is my fork that gives the default gamepad 6 axes, 6 hat switches, and 32 buttons
@bleak spear I'm going to rebase it after the pico gets UART support
fossil gorge
I seem to recall that the sphinx generator also creates man pages, PDF, and some other file formats. Are those not published or available anywhere?
Alternatively (but bandwidth and time consuming), clone the repos, and build the docs?
onyx hinge
@fossil gorge at a page like https://circuitpython.readthedocs.io/en/6.1.x/README.html click the triangle in the lower left, there's "Downloads - PDF"
so now I know that there are 248 pages in the circuitpython (6.1) manual. wow.
fossil gorge
<@!361158808501747735> at a page like https://circuitpython.readthedocs.io/en/6....
Thanks @onyx hinge - heads up for @thorny jay about it!
onyx hinge
My suspicion is that there's some rtos thing like a mutex or a task that is getting confused. just a theory, I couldn't prove it.
steep finch
@bleak spear https://github.com/alexbartlow/pico-hotas I have a build for the pico here that has some sample gamepad code just to kick the tires
thorny jay
Thanks <@!426200547024961539> - heads up for <@!681528999520108564> about it!
My plan was more to backup my internet provider that is "cable/TV" with a second internet provider by a ADSL/Telephone. Maybe this is overkill for one outage a year, but I don't want to live an internet cut of 3 days again... 4G/LTE is a good enough backup but maybe not for voice/video and a full family at home.
solar whale
Never let the boss test...
marble hornet
Great to hear about everyone's work and projects!
Have a great week all 👋 , have to duck out.
tidal kiln
🐈
onyx hinge
that sounds like a lot of work
@idle owl will there potentially be a new twitter handle to tag?
thorny jay
@idle owl cpnews@adafruit.com and for Twitter, you put @CircuitPython in the tweet?
onyx hinge
Widgets: Button, Progress bar, Sliding switch, Needle dial gauge, Scrolling value selector, Widget label, Widget indicator (label plus arrow)
gloomy shuttle
I second it being in pre-commit
lone axle
+1 from me for moving it into pre-commit as well. I still sometimes forget to run one or the other.
still zephyr
As new user I will love a pre-commit
low sentinel
Have a great week y’all!
onyx hinge
the rp2040 datasheet states that I2C peripheral mode is supported
still zephyr
Thanks I will see the M4 implementation
onyx hinge
Is the concern about the time to sort, or the dirty rectangle that results from it?
thorny jay
It's GPLv3 so you may not want to add it...
Compatible with MIT but will be "contaminated".
So "macro" kind of feature?
onyx hinge
was just going to mention .wrap_target
thorny jay
Ok, great. I'll see how I can use it for testing.
More about regression checking.
Thanks.
I will start with accumulating example and expected result.
@narrow dirge was testing with the comma with or without a space on the left and the right...
onyx hinge
too big to fail test
idle owl
<@!330227457296957440> cpnews@adafruit.com and for Twitter, you put @CircuitPyth...
For David and for everyone else, yes I think tag (at)CircuitPython or (#)circuitpython on Twitter with your newsletter projects and ideas. I haven't gotten confirmation from Anne on that, but I'm pretty sure that will work.
thorny jay
Anna was picking up project like that.
At a time I was doing "one project a week", but I failed a bit, or not interesting enough to be spread in newsletter.
errant grail
Thank you!
mental nexus
Thanks everybody!
still zephyr
Thanks!
modern wing
Thanks y'all!
gilded cradle
Thanks
thorny jay
Bye
modern wing
Another successful meeting in the bag 🙂
You know it's cold when C and F are the same...
thorny jay
I was considering checking all the meeting notes and pick what I did over a year... and what I had plan and did not.
fossil gorge
Cookiecutter issue opened. Feel free to assign the issue to me
modern wing
+45F right now in NJ
fossil gorge
That is correct, yes
fossil gorge
Thanks! 🙂
thorny jay
@idle owl in the "how to contribute" do you have separate instruction (and cookiecutter things) for Adafruit library and Community library?
idle owl
@thorny jay I haven't actually checked. Foamyguy updated the guide, I'm not sure exactly how he did it.
lone axle
It's the same basic instructions. The first prompt asks you which you want to do.
thorny jay
Yeah, there was that list of change to do to use it for non Adafruit library.
It's automated, great.
lone axle
The defaults are all good now I believe. If you choose community bundle it should have sane defaults for many of the fields. If you end up running it and notice there are specific prompts that could use better or additional wording in the guide feel free to ping me.
idle owl
Dropping off. Later!
idle owl
The defaults are all good now I believe. If you choose community bundle it shou...
Thank you again!
@tulip sleet Disconnected you from the chat so nobody jumped in and started talking.
tulip sleet
tnx - i was loading the dishwasher
idle owl
@tulip sleet Any thoughts on this one? https://forums.adafruit.com/viewtopic.php?f=60&t=175459
tulip sleet
@idle owl sleep(0.00001) is 10 microseconds. time.sleep() is not good for that. It might not sleep at all or be highly inaccurate. Try microcontroller.delay_us(10) instead, but that might not be great either.
something like pulsein would be a lot better
manic glacierBOT
onyx hinge
hmmm here's another puzzler -- PioPixl8 needs consecutive pins. There's not actually a way to check that is there?
from python
tulip sleet
you could get the microcontroller.pin equivalents and deconstruct the pin names ...
but no easy way
slender iron
Here is the notes document for Monday’s CircuitPython Weekly meeting. It is at the normal time of 11am Pacific / 2pm Eastern here on Discord. Everyone is encouraged to attend! Please add your hug reports and status updates even if you’ll be attending the meeting - it’s super helpful! If you are unable to attend but would still like to include updates, feel free to include them in the notes and we’ll read them off during the meeting. Hope to see you there! <@&356864093652516868> https://docs.google.com/document/d/10c1pc8x_d5PM5QPUSj9Ex0OfgyFvTiLZV9Cuno_UQDI/edit?usp=sharing
idle owl
@onyx hinge @slender iron FYI I updated the CP Weekly template with the new suggestions for submitting newsletter ideas. It now reads You may also tag a tweet with #circuitpython on Twitter, or email cpnews@adafruit.com. Letting you know so you're aware for when you next run the meeting, in case you have that part memorised and don't read it from the doc.
onyx hinge
wheee! "piopxel8" now works with the LED animations library. I'm running a different animation on each of the 8 independent strands right now.. (only 2 are wired up but I imagine the rest in their glory)
slender iron
nice!
onyx hinge
it seems reasonably fast with 240 pixels but I don't necessarily have a point of comparison. @slender iron do you think that (A) I should implement piopixl8 as a core module or (B) expose this extremely specific bit_transpose routine to Python? I don't want to go down the wrong route with a pull request.
idle owl
wheee! "piopxel8" now works with the LED animations library. I'm running a dif...
Brilliant!
slender iron
(can't you do 16 that way too?)
onyx hinge
@slender iron the transposition routine is specific to working in 8-bit (8-strip) units. as for working on other ports, I guess there was a long ago failed attempt to do something similar to this on samd with spi dma but it didn't pan out (maybe not even in CP; limor mentioned paintyourdragon worked on it)
(so not 2, 3, 4, ..., 7 strips either -- only 8)
idle owl
@onyx hinge Unrelated, check the #books-and-tutorials channel - potential feedback on the canio guide. I asked them to submit feedback through Learn, but I know it's difficult to get through all of it.
slender iron
only 8 if you are outputting via shift register
you could do an odd number with a single PIO state machine doing similar
onyx hinge
the "bit transposition" code is very specific to working with a unit of 8 strips, in order to be efficient.
it has to take the first bit for each strip and pack it into 1 byte, etc
slender iron
right, I get what you are doing
I'd probably lean towards B, a bit manipulation library
time for me to eat lunch
onyx hinge
To do 16, aside from needing a different transposition code, you'd need to double the speed of the pio peripheral to ....83.2MHz?
idle owl
Do all Adafruit CircuitPython compatible microcontrollers have super specific PWM "slices" like the Pico does? As in are there usually conflicting pins where if you used a certain two pins it could be conflicting PWM? I'm not sure how to ask this question.
onyx hinge
@idle owl short answer is often they do.
onyx hinge
and, like unhappy families, they are each different in their arbitrary-seeming requirements
idle owl
We have that script that tells you which pins have PWM, but not if they're conflicting.
onyx hinge
you can go read a datasheet? which we like to do for fun, yes we do
idle owl
Oh.
Fair enough.
And is it true that if you have two things on conflicting PWM pins, and you send an instruction to one of them, it affects the other one too?
onyx hinge
But to be less flip, no, I don't think we've tried to docuement in a "regular people"-compatible way, the specifics of this; because it can be complicated.
idle owl
But to be less flip, no, I don't think we've tried to docuement in a "regular pe...
Also fair enough.
onyx hinge
Generally speaking, the core should raise an exception if you ask for something the hardware can't do. There could always be bugs, of course.
idle owl
This book seems to indicate that it won't error, it'll simply send the same instruction to both pins.
onyx hinge
is that describing mp or cp?
idle owl
Presumably meaning an LED each on two conflicting pins, and you PWM one pin, it will affect both.
MP.
presumably
I thought it might be a hardware thing though
I wasn't attributing that concept to MP.
onyx hinge
If they have an example of that circumstance, you should adapt it to CP and try it. I think we would want it to throw an exception.
idle owl
I don't think they have an example. It was an explanation in the copy.
onyx hinge
for instance, if pins 99 and 53 both use PWM unit "sqrt(2)", then they have to agree on the frequency when they are created
idle owl
Says basically you'll know if you use conflicting PWM because it will affect both pins, and if that happens, check the pinout and try again.
I could try it, the pinout is in the book. I know two conflicting pins.
onyx hinge
0/1/16/17 are all tangled together for instance
idle owl
Right.
Quick hardware question, the ground rail on a breadboard is simply connected pins right? So I could use it to route an LED via a resistor? Or is there something special about the power and ground rails on breadboards other than connectivity?
onyx hinge
>>> q = pwmio.PWMOut(board.GP1, frequency=800)
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
ValueError: All timers for this pin are in use
>>> q = pwmio.PWMOut(board.GP1, frequency=400)
>>>```Kind of dumb question I guess. But still, question stands.
idle owl
```>>> p = pwmio.PWMOut(board.GP0, frequency=400)
>>> q = pwmio.PWMOut(board.GP1...
Ah ok. Timer in use.
onyx hinge
so in CP if you ask for different frequencies, but the pins have to SHARE a frequency, it says "no"
idle owl
Good news! Alright
jaunty juniper
yes that's what it's there for, all pins connected on a rail
idle owl
That is way easier to explain than to say "read the datasheet!"
idle owl
yes that's what it's there for, all pins connected on a rail
Cheers
onyx hinge
Because GP0 and GP1 are on the same "PWM slice" 0 but different "PWM channels" 0A and 0B. If you use GP0 as PWM, you can't use GP16 as PWM even if they share frequency=, because they are the same "PWM channel" 0A. But yes, short story, CP tells you about this kind of limitation with an exception
idle owl
Because GP0 and GP1 are on the same "PWM slice" 0 but different "PWM channels" 0...
Ok. I think that's good enough for the purposes of this guide. The book gets into that level of detail, but I would have to explain a lot of other things in that level of detail as well... I'm not prepared to do that I think.
onyx hinge
@idle owl thanks for the other mention .. I fixed the typo
idle owl
<@!330227457296957440> thanks for the other mention .. I fixed the typo
Right on - can you ping the poster to let them know?
onyx hinge
oh sure
idle owl
Thanks!
@onyx hinge But I'm reading this right that every pin can do PWM? but not at the same time obvs.
Wait. Put simply for the Pico, you can do PWM on one side of the board or the other, but not both at the same time if you want to be guaranteed to avoid conflicts. Not sure that's worth including in the guide though.
onyx hinge
@idle owl yes "All 30 GPIO pins on RP2040 can be used for PWM" though with CP you won't be able to do more than 16 simultaneously
idle owl
Right on. Thank you.
onyx hinge
(with MP or C++, you could sort of go past 16 but it's not clear whether that's useful)
(because some pairs of them would be exactly identical)
idle owl
(hmm yah)
onyx hinge
(actually ... no, wait, I don't want to go further into the weeds than we have)
idle owl
😄 Appreciated.
onyx hinge
(using a feature of the RP2040 chip that CP doesn't make available, I think you can cause the PWM output on pins 0 and 16 to be exact opposites which is useful to someone, somewhere. don't write that in the guide)
onyx hinge
OK drop everything & update idf now
simple pulsar
For RP2040/Pico, it sounds like it can do 16pwm on GP0-GP15? I can't get this to work: https://github.com/adafruit/circuitpython/pull/4167#issuecomment-779491927 Should i log that as an issue?
tidal kiln
@idle owl hey. i'm about to start something that'll essentially be a data logger. been seeing your recent traffic on that, but haven't really followed it. just curious if anything has changed?
idle owl
<@!330227457296957440> hey. i'm about to start something that'll essentially be ...
I wrote up a clearer boot.py file, and I added a comment to the basic data logging code.py that explains that errorno: 25 is filesystem full because it was dreadfully unclear. Other than that, not really. Want links?
tidal kiln
yes please. i'll take a look.
idle owl
yes please. i'll take a look.
https://github.com/adafruit/Adafruit_Learning_System_Guides/blob/master/Getting_Started_With_Raspberry_Pi_Pico/data_logger_boot.py and https://github.com/adafruit/Adafruit_Learning_System_Guides/blob/master/Getting_Started_With_Raspberry_Pi_Pico/data_logger.py
idle owl
@tidal kiln As far as I know?
stuck elbow
@slender iron why do you keep both native_layer and original_layer in the group children?
tidal kiln
okie doke. cool. thanks for the links. yah....looks pretty much like i remember.
idle owl
@tidal kiln But you can do it like I linked where it requires a pin to be grounded or a button to be pressed (depending on what board you're using) at startup for it to data log. Or not. Up to you.
Yeah, it's pretty much the same.
idle owl
Yeah. Otherwise, you can get out of a writeable state using the REPL and os.rename or os.remove on boot.py.
But a button makes it a little easier if it's a reasonable addition to your project.
jaunty juniper
by the way I'm using that in boot.py (and a button on a breadboard) to have safe mode on the pico, I needed it to test a bug that makes the board hang
https://gist.github.com/Neradoc/8056725be1c209475fd09ffc37c9fad4
idle owl
Oh interesting. I guess I should have known you could reboot into safe mode using software, but I've never seen it done.
tidal kiln
wondering if i could use nvm to maybe do this from user space....set a flag there, reset, read that in boot.py.....hmmmm
tidal kiln
you can!
here's boot.py:
import storage
from microcontroller import nvm
storage.remount("/", nvm[0])
and then in user code, can do something like:
>>> nvm[0] = False # or True
>>> microcontroller.reset()
Fancy!
jaunty juniper
note that currently on the S2 nvm[0] will raise an exception if it was not set before
tidal kiln
like if you just did print nvm[0] without do anything else first?
jaunty juniper
tidal kiln
oof. thanks. good to know. luckily i'm using nrf52 for this particular project.
jaunty juniper
it's implemented with a key-value storage in the background but there is no way to probe the key it seems, so the only way is to try/except, which is less than ideal
tidal kiln
@idle owl anyway to check current fs state? other than trying to open and catching exception?
idle owl
<@!330227457296957440> anyway to check current fs state? other than trying to op...
Not that I'm aware of, but that doesn't mean that's not the case.
tidal kiln
@tulip sleet @solar whale just fwiw/fyi - still getting USB hard locks after upgrading to ubuntu 20.04 laser eyeball tiger cat
solar whale
I get them occasionally. Totally unpredictable.
tidal kiln
it seems to always be related to a soft reset
but as to when and why...???? yah totally random
idle owl
<@!329766224093249548> <@!343570430042308609> just fwiw/fyi - still getting USB ...
laser eyeball tiger cat 😆
slender iron
<@!252717193496756235> why do you keep both native_layer and original_layer in t...
If a child is a subclass we internally want to use the native superclass but need the original (subclass) to return when folks access it.
simple pulsar
@tidal kiln I've found holding down the button on the CLUE works well for enabled write access in boot.py
It's more useful than I thought because it means the data logging code doesn't start writing when I plug it into desktop to retrieve data. I append so it would be safe (in term of not losing data) but it's still more convenient for my case to have the writing be very explicitly enabled.
tidal kiln
thanks. yah, i might end up going that route also.
blissful pollen
@slender iron You were still planning to change PIO pin initialization correct? I (think) I figured out how to get the 8 pins for ParallelBus going, though to run an init program more then 1 instructions have to move the pio_sm_set_enabled to after the init instructions get ran. But if it's all going to change i'll just sit on this until then.
tidal kiln
this is something i'm wanting to be sort of plug-n-go and auto-magic. at least for my initial thinking.
slender iron
@blissful pollen ya, I'm planning on it
gonna just add a ton of initial_*_values as kwargs
with sensible defaults
blissful pollen
okay i'll sit on ParallelBus until after that, if that makes sense?
blissful pollen
no problem, just wanted to make sure (and had meetings all CP meeting earlier today)
slender iron
np, always feel free to ping me to coordinate
I'm hoping to go heads down on it tomorrow
finishing up email now and then I'm making dinner tonight
idle wharf
This sounds very cool Jeff... can't wait to see the demo !
onyx hinge
It doesn't "look" like much yet. ```code.py output:
?0100010-001000010-000100000-
-00100011-001000001-000000100-010100010-001000010-000100000-
(21, 45, 21, 13, 0, -200, 0)
-00100100-001000001-000000100-010100010-001000010-000100000-
(21, 45, 21, 14, 0, -200, 0)
The symbols 0, 1, - and ? appear, about 1 per second. When 60 of them in a row "make sense" it prints the time: 2-digit year, day-of-year, hour and minute GMT, and some other information (about leap seconds and daylight saving time). The time is the minute that _just ended_.the receiver is very noise-sensitive so I have to test & debug it from a battery-powered laptop. I think 60Hz noise from the mains may be particularly able to interfere with the 60kHz radio signal.
blissful pollen
What are you using to receive it?
onyx hinge
Looking for an interesting display concept, but I also need to figure out the power issues
https://www.sparkfun.com/products/retired/10060 [no longer available]
blissful pollen
Oh neat. I've received it on an RTLSDR. I wonder if ferrite beads on the cable would help but the antenna wire is pretty short
stuck elbow
@slender iron but calculating the native type doesn't seem to be that expensive?
onyx hinge
My friend who has built several WWVB clocks just says to add filtering on the power input to the reciver until it works
blissful pollen
Makes sense. Radio always hurts my head when all the interference can come into play.
onyx hinge
seems like a modern alternative is https://www.universal-solder.ca/product/canaduino-60khz-atomic-clock-receiver-module-wwvb-msf-jjy60/
onyx hinge
@slender iron any name suggestions for this new function & its module for re-ordering the bits for piopixl8? I called the function bit_transpose and put it in _pixelbuf for testing but I don't think we want it in pixelbuf permanently.
slender iron
<@!252717193496756235> but calculating the native type doesn't seem to be that e...
probably pre-mature optimization 🙂
onyx hinge
.. it's tempting to see if we can get to 16. can the PIO divisor go all the way to "1"?
onyx hinge
(16 strands)
onyx hinge
see ya
stuck elbow
@slender iron ok, thanks, I will try with storing only the original type then and see how that works
slender iron
@onyx hinge you should support less than 8 inputs and just pad the byte with 0s
some folks will want to use it with fewer than 8
onyx hinge
@slender iron I can benchmark it but I think it will make the speed worse
slender iron
people will still want to do it
onyx hinge
I agree, but it'll hurt performance, or burn flash space
slender iron
we've got a ton of flash space on the rp2040
onyx hinge
adding 6 conditional branches to each byte handled
or adding 6 copies of the critical function
~200 bytes apiece, so +1 to 1.25kB?
slender iron
or making 6 extra buffers to get bytes from
I allocated 1MB for the CP flash space
for a reason 🙂
onyx hinge
I'll consider your feature request 🙂
slender iron
gotta give folks a path from 1 to 8
ornate breach
If you’ve got the space split it in two 😉
onyx hinge
I could do one optimized for 8, and one branchy for 2..7.
slender iron
sure
onyx hinge
it'll satisfy my peculiarities for optimization, but give flexibility, and only cost ~200 bytes
onyx hinge
awwww but they were comfortable there
simple pulsar
@onyx hinge Did you do the PWMAudioOut for RP2040? I've noticed it's a bit quiet on the Cytron Make Pi Pico's speaker - similar volume to CLUE which has a much smaller speaker and is driven in similar way.
onyx hinge
nope, it was scott
manic glacierBOT
simple pulsar
I'll try compressing it (the sound sample) a lot to see if that improves things. There's a huge difference between PWMOut square wave and square wave played over PWMAudioOut.
tidal kiln
@simple pulsar i'm trying the hold button approach on CLUE for enabling write access. but it seems to go into bootloader mode when a button is held during reset. have you seen that behavior?
jaunty juniper
I believe that holding both buttons puts it in dfu mode, but that's different
tidal kiln
i'm doing just one. and either one does it.
idle owl
Does CircuitPython do something special with duty_cycle using pwmio on Pico? I wrote some simple code that uses a potentiometer to fade an LED, but it is not behaving like this MicroPython book seems to indicate it should be. Turning the pot fades the LED from off to bright, over the full rotation of the potentiometer. The book seems to make it sound like it should reach max brightness at 1024/65353 in the potentiometer rotation range unless I do something else to the code.
@slender iron Is this the expected result of using a pot with PWM to fade an LED? Or is there something weird going on...
Here is my code ```python
import board
import analogio
import pwmio
potentiometer = analogio.AnalogIn(board.GP26)
led = pwmio.PWMOut(board.GP14, frequency=5000)
while True:
led.duty_cycle = potentiometer.value```
slender iron
it should be over the full rotation
idle owl
Ok, so CP is doing the duty_u16 MicroPython thing in the background?
slender iron
I don't know what you are referring to
slender iron
both duty_cycle and value are defined as 16 bit values in circuitpython
idle owl
Ok then yes is the answer to my question. Thank you.
idle owl
How do I determine a good frequency for PWM? There's an example using 5000 for fading an LED, but 1000 seems to work as well.
There must be a reason to it.
idle owl
Ah ok
tidal kiln
there can be secondary things as well, like electrical interference
or even auditory
i forget what the technical terms is for "the freq response of human eye", you'll see it come up in POV discussions. as i remember, it's pretty low
idle owl
Fair enough
tidal kiln
tv/movies for example are O(24-30)Hz
and if other sensors, like digital cameras, need to be considered, then that can up the threshold
jaunty juniper
you might want to avoid being too close or a multiple of a camera's frequency like 60fps, lest the picture flickers when filming
idle owl
Thanks @tidal kiln
tidal kiln
"why not just set it crazy high and forget about it?" <-- also a good question 🙂
idle owl
Well? 😄
slender iron
too high and you'll lose resolution
idle owl
I mean, I went with 1k to match the book example. But still worth knowing. 🙂
slender iron
ya, that'll be fine I think
idle owl
The code and hardware works anyway.
slender iron
its 125000000 / (2 ** 16) for the fast full resolution
idle owl
hmm. Fair enough.
tidal kiln
>>> 125000000 / (2 ** 16)
1907.3486328125
it's the 6 in the decimal that really matters 🙂
but, yah, 1k better than 5k to get you the full res.
ionic elk
@tulip sleet I'm working with the Monsoon today and I just wanted to double check - I need to actually have two USB B cables in order to measure USB power draw? IE the monsoon itself can't provide power over USB. Is that right?
It seems like what the manual is implying but I'm not totally sure
tulip sleet
<@!329766224093249548> I'm working with the Monsoon today and I just wanted to d...
I think that's right. I actually didn't use it much or at all to measure USB current draw, because I wanted to see sleep current, and if I powered via USB, I would be measuring the regulator dropping 5V to 3.3V, etc.
so I used the binding posts
ionic elk
Oh, that's a good point. I didn't do that and was surprised it wasn't lower. I'll have to go back and re-measure my control.
tulip sleet
there is a dialog box to set what voltage is coming out of the binding posts
tulip sleet
no, I sent it into the battery connector on the MagTag, using 3.7V
since the idea is to measure consumption with a battery
ionic elk
Sure, makes sense.
I'll see if I can dig up a JST I can use instead of trying to cram duponts or something in there
tulip sleet
be careful not to leave it connected if you power via USB temporarily, since you'll be backpowering the Monsoon from the Feather battery charging circuitry
Hopefully there's a protection diode, but I didn't check that.
The MagTag has no charging circuit, so it's not an issue
But Saola has no charger, is that right?
ionic elk
What does that mean, no charging circuit? On the magtag?
Saola doesn't, at least that I'm aware of. Nothing marked Vbat or that looks like an MCP
tulip sleet
The Feathers all have a LiPo battery charger on board that charges an attached battery if they are powered via USB. When there's no USB power then the charger is off.
The MagTag does not have that circuitry
That is a basic feature of Feathers
ionic elk
It doesn't? So it won't charge a Lipo when plugged in?
tulip sleet
I don't think so, but I'll check
ionic elk
I'm just surprised because I'm pretty sure I've charged my project devices on a Magtag
tulip sleet
oh, you're right!
tulip sleet
I have a switched JST cable I was using between the Monsoon (or the PPK) and the battery connector
but I would hope both the PPK and Monsoon have protection
ionic elk
handy, did you make that yourself or do you happen to have a purchase link?
tulip sleet
The switch is weird, it is on when the button is up
that was confusing
I would suggest ordering several other battery-ish cables at the same time, like 261 , 3814, etc. https://www.adafruit.com/?q=jst+ph+2-pin&sort=BestMatch
I put some Dupont style connectors on the ends of some
manic glacierBOT
tidal kiln
for anyone wanting a one liner to convert image files to CP bitmaps:
convert foo.png +dither -colors N foo.bmp
seems to work nicely.
set N as low as possible to cover desired number of colors
lone axle
neat! I've never seen this convert command. I made a python script that uses PIL that does something similar, but there are so many other nice options built into this
jaunty juniper
it's imagemagick I believe
tidal kiln
yep. although i think i'm still getting slightly better results going through gimp.
vague latch
Hi all, just started with the raspberry pi pico, i am trying to do a series of asyncronous data collects and was trying to use asyncio in its micropython version. However i have since found libraries for the sensors im using in the circuit python github. Is there asyncio in circuitpython.
idle owl
@vague latch There is not. There may, however, be ways to do what you're trying to do without specifically using asyncio.
vague latch
thank you for your response. Do you have any examples on how this may be done instead.
Or is it in the pipelines to port asyncio to circuitpython
idle owl
I would need to first know what you're trying to do.
It's on our radar, but it's not a priority for the core devs at this time.
vague latch
I have a gps, bme280 (pressure, humidity..) and a bno055(IMU) i want to collect the data from each of these at a different sampling rate and parse them to the SD card
i was thinking asyncio would allow me to do a OS like event handler
idle owl
Ah ok. I might not be the best person to help you as my knowledge gets a little fuzzy in this area. I feel like you should be able to set each one up, and have a loop within a loop that uses something like time.monotonic() to poll at different rates... But I'm not certain.
It might do that, yeah.
vague latch
ok, ill have a look into that
idle owl
I am fairly certain it's possible to do what you're suggesting in CircuitPython. But I don't think I can make any helpful suggestions.
vague latch
ok, can you mix micropython and circuitpython at all in these systems
idle owl
No. They're not really compatible.
The code will always look really similar, but it won't work together.
vague latch
thanks for your help. ill look into time.monotonic or ill port the adafruit libraries to micropython
idle owl
You're welcome. Sorry I couldn't help more!
lone axle
@vague latch we do have the async and await keywords in CircuitPython from this PR: https://github.com/adafruit/circuitpython/pull/3540 there is a library called tasko written by the person that made that contribution: https://github.com/WarriorOfWire/tasko I'm not certain but I think it's goal is to be similar to asyncio is for CPython.
vague latch
that looks interesting. thankyou @lone axle
idle owl
<@!706647997798416455> we do have the async and await keywords in CircuitPython ...
Thank you!
vague latch
Another quick question. I searched for the better part of yesterday and couldn't find the answer and would suggest someone write a blog post on it if they are so inclined. I need to add libraries onto the circuitpython based raspberry pi. My mu editor isnt working properly so i began copying across the code using the file storage but would like a way where i can use pip to install the libraries straight onto the board. Further i would like to get someones github library and add it to the device like i would with python. ie pip3 install git+https://github.com/....
clarification raspberry pi pico
lone axle
@vague latch check out circup if you haven't seen it yet: https://github.com/adafruit/circup It's a CLI tool for installing circuitpython libraries on your devices. If you have experience with pip then the workflow of this will likely be familiar.
vague latch
That looks good. however i cant see an indication as to whether anything other then adafruit based libraries can be added, this will help with the other libraries that i was manually copying across. Im just thinking in regards to the tasko library is there a simple way to get this onto the board directly from git. I don't see any real way of doing. It is not a huge hassle to git clone then copy to lib folder.
idle wharf
@vague latch Community Bundle support is on the to do list https://github.com/adafruit/circup/issues/82
vague latch
hmm i just had a thought, i can cd into the circuitpython location and git clone directly into the lib folder
slender iron
that will likely fill up your filesystem
idle wharf
You can pip install to a folder... let me lookup how that is done
vague latch
yeah, i could then remove the .git files and other stuff but it is slower
@slender iron what is the current best solution to put these libraries that are sourced from github onto the pico
slender iron
generally we expect you to download the bundle zip and copy libraries over
as you need them
circup is replacing that over time though
vague latch
yeah ok that sounds good
idle wharf
assuming the package is published (not all are)
pip install <PACKAGE_NAME> --target /Volumes/CIRCUITPY/lib
or your OS specific version of /Volumes/CIRCUITPY/lib
I'm on a Mac.
idle wharf
Yes, I think you're right. If you're making your own library, start with the cookiecutter for library projects
https://github.com/adafruit/cookiecutter-adafruit-circuitpython/
But honestly 99.9% of the libraries are a single .py file... you can download onesie twosies
Ultimately, circup will do the magic
vague latch
yeah, i am looking at using this for multiple identical picos so was looking at automating some of the installation
but i definitely see that the majority of cases a single copy works
Do the .py files need to be at the top level of the lib folder or can they be in their own folders.
idle wharf
in /lib
lone axle
I would set up one master "image" one and then copy from that to a PC and then from there to all of the others. instead of direct install each lib on each device.
idle wharf
folder under /lib are for modules (libraries with many files)
vague latch
Thanks for your help all
manic glacierBOT
Since the datasheet cast some doubt on the strength of the "rosc_hw->randombit", I use the SHA256 hash function to create a high quality random seed from random values of uncertain entropy, as well as to generate a sequence of random values from that seed using SHA256 as a cryptographically-secure random number generator.
In practice, it produces over 100kB/s of random data whi...
vague latch
<@!706647997798416455> check out circup if you haven't seen it yet: https://gith...
circup is perfect. It installs dependencies as well. Can i recommend adding this into the getting started tutorial on adafruit if it is not already there
idle wharf
I just added that, glad you like it. 🙂
tidal kiln
anyway to recover when both CP and PC see filesystem as readonly?
jaunty juniper
that sounds like it's corrupted
tidal kiln
could be. i've been going back and forth a lot. just hoping for something before i storage.erase
jaunty juniper
I would say try safe mode but you probably should erase anyway for safety
idle owl
circup is perfect. It installs dependencies as well. Can i recommend adding this...
I'll take this recommendation into consideration. Circup is still in active development so I want to make sure it's ready for us to document when we do.
prime flower
anyway to recover when **both** CP and PC see filesystem as readonly?
Is the boot file set to readonly?
tidal kiln
there currently is no boot file
vague latch
I can confirm that if you create a simple setup.py file which i checked through forking tasko. You can then do pip install git+https://github.com/qwertimer1/tasko.git --target /media/<user>/CIRCUITPY/lib and the library is saved into the lib folder with setuptools install working without an issue
vague latch
I can confirm that if you create a simple setup.py file which i checked through ...
Also confirmed it works by running import tasko in the repl
tidal kiln
stroage.erase fixed it
idle wharf
@tidal kiln I was surprised you ended up there, do you think it was from disconnecting from USB without ejecting or some other action?
tidal kiln
unknown. could be. i'm working on a data logging app. so there's lots of back and forth.
need to set write access to allow PC to push code update
then change to allow code to write
rinse, repeat
jaunty juniper
you could ampy put it should still work
gloomy shuttle
I'll take this recommendation into consideration. Circup is still in active deve...
I have been using it for a couple of weeks not and it is a life saver!
manic glacierBOT
vague latch
anyone know how to completely wipe the storage for the pico. I tried the bootsel and re add circuitpython but all my files are still there. I have an 'OSError: [Errno 5] Input/output error' occuring, I am hoping i havent brcked the pico
idle wharf
anyone know how to completely wipe the storage for the pico. I tried the bootsel...
storage.erase_filesystem()
https://circuitpython.readthedocs.io/en/6.1.x/shared-bindings/storage/index.html#storage.erase_filesystem
vague latch
storage.erase_filesystem()
https://circuitpython.readthedocs.io/en/6.1.x/shared...
It looks like this isn't working at the moment. https://github.com/adafruit/circuitpython/issues/4066
slender iron
what I/O error are you getting?
vague latch
OSError: [Errno 5] Input/output error
i cant open the code.py file
it wont let me write to the circuitpython folder
vague latch
The repl still works
slender iron
hrm. I don't see a uf2 of nuke
(it's very weird to me that erase_filesystem doesn't work)
but it didn't work for me today either
vague latch
can we add erase filesystem and not reload circuit python
jaunty juniper
ah I was gonna ask: would a "erase flash UF2" be hard to make ?
vague latch
i was trying to find the library in the git repo but am still looking for it
slender iron
can we add erase filesystem and not reload circuit python
not sure what you mean by this
vague latch
not sure what you mean by this
https://circuitpython.readthedocs.io/en/6.1.x/shared-bindings/storage/index.html#storage.erase_filesystem says erase and recreate the CIRCUITPY filesystem
slender iron
right
vague latch
i was thinking you could just erase but not recreate the filesystem
slender iron
that's what it actually does. it just rewrites the FS data structures
vague latch
ok
slender iron
it's the same code that's run the first time when a filesystem isn't found
so it's weird that it doesn't work
manic glacierBOT
vague latch
found the nuke uf2 https://www.raspberrypi.org/forums/viewtopic.php?t=300903&p=1806928
this works
jaunty juniper
I'm gonna bookmark that nuke uf2
vague latch
yeah, very useful
glossy pecan
? Using Segger J-Link Edu Mini adapter to debug ESP32-S2 Saola 1R dev kit?
Kind folks, I have followed the instructions for setting up to build the CircuitPython binary for the ESP32-S2 Rover module on the Saola 1R dev kit, using my MacBook Pro "Big Sur". Great. Got a live REPL, etc., Thanks. Now I'd like to use my Segger J-Link Edu Mini to debug changes I want to make to the code. Has anyone succeeded in doing this? How? Using OpenOCD or J-Link Commander to control the adapter?
Note: This article describes how to do it for the J-Link "Edu" adapter, but says it doesn't work for the "mini" -- https://dzone.com/articles/eclipse-jtag-debugging-the-esp32-with-a-segger-j-l
manic glacierBOT
glossy pecan
? esp-idf version in CircuitPython repo?
While I'm asking questions, I note that the Espressif esp-idf code repo is included in the CircuitPython source repo. I presume this is to ensure some version control on which release of esp-idf is used. Anything else I'm missing here? Why not just check out a fixed tag from the general esp-idf repo, e..g., "Release 4.2"?
analog bridge
? esp-idf version in CircuitPython repo?
While I'm asking questions, I note that...
Hi @glossy pecan the current IDF is based on slightly patched v4.2 release... this will soon be updated to v4.3 directly from espressif.
calm oar
I am trying to find a good use of my MagTag and thought of making a news headline viewer. Are there any rss feed parsing libraries that will work with circuitpython? This is the only one that I found -> https://github.com/kurtmckee/feedparser
solar whale
Does anyone have a Fingerprint sensor ( not the R503)? It would be nice if someone could test this PR https://github.com/adafruit/Adafruit_CircuitPython_Fingerprint/pull/28 with a different sensor - I have tested it with the R503. The new example can run on a Raspberry Pi, not on an MCU.
tulip sleet
I am trying to find a good use of my MagTag and thought of making a news headlin...
I don't know of any XML parsing examples. We have plenty of JSON examples. Could you get a feed in JSON?
also maybe re-ask in #help-with-circuitpython
this is more the dev channel (fine to ask here, but there are more user eyes in the other channel)
onyx hinge
@tulip sleet could it be a dict with meaningful keys, 'repl' and '???' to avoid the problem with numbers vs com port numbers?
manic glacierBOT
tulip sleet
<@329766224093249548> could it be a dict with meaningful keys, 'repl' and '???' ...
it could be - right now it's just an index into the CDC descriptors, basically. I was using integer indices because I could imagine adding a third one, etc. The key could be the identifier included in the USB descriptor. Feel free to add a comment to the PR.
could be .serials with integers now, and add .serials_by_name now or later
onyx hinge
so nice to see a number of these bars nearly full!
we just got possibly the first weblate contributions in Italian, nice!
calm oar
I don't know of any XML parsing examples. We have plenty of JSON examples. Could...
Thanks for the reply. I already put my question in there as well. I'll check to see if there are any news feeds that provide json.
onyx hinge
@slender iron if time.monotonic_ns() doesn't seem to be advancing while executing a pio program, is that its own bug or is it https://github.com/adafruit/circuitpython/issues/4067 ?
.. but maybe monotonic_ns is? or maybe my test was wrong?
....................................................................................................
Neopixel show() 2.93066s, 34.122FPS, 32757.084370pixels/second
....................................................................................................
PioPixl8 show() 0.429688s, 232.727FPS, 223418.116570pixels/second
``` it seems to be working now, huh.also speedup ahoy
slender iron
@onyx hinge it should work because it's based on the single 64 bit timer I think
tulip sleet
@slender iron I am working on the short I2C write issue on RP2040 by using bitbangio. Just about to test.
slender iron
thanks!
idle wharf
so nice to see a number of these bars nearly full!
Is there a link to that graph? It would be a good link to tweet ...
onyx hinge
@idle wharf good point -- I dunno if we've tweeted about weblate lately. There are share links right on the weblate page... dunno if it shows that image in particular though
idle wharf
Look like a good landing page
https://circuitpython.org/contributing/weblate
jaunty juniper
@tulip sleet I know pyserial can get this information on mac, is pyserial platform-independent enough ?
import serial.tools.list_ports
for port in serial.tools.list_ports.comports():
print(port.device,port.interface)
/dev/cu.Bluetooth-Incoming-Port None
/dev/cu.usbmodem7CDFA100E6201 CircuitPython CDC data
jaunty juniper
I've been using a tool so I never have to look up serial ports for my boards, though it's mac/linux only because I had to use platform specific tools to identify which drive matches which board/serial port. https://github.com/Neradoc/discotool
(I need to add differentiating between serial ports now)
I whish to make it multiplatform and hopefully the methods used could be used to identify boards in Mu or other similar environments more reliably
tulip sleet
@slender iron ok, rathole commencing: bitbangio I2C not working well on RP2040, so that needs to be fixed first. Are there known time-keeping issues?
slender iron
ya, it's something I stubbed out but didn't exhaustively do
generally sleeps appear to be right to me
I wanted to get WFI working but had to punt on it
tulip sleet
so common_hal_mcu_delay_us() is probably not doing great: that's the basic timekeeping in bitbangio I2C.
i will do some testing and open issue(s) as appropriate
tulip sleet
haven't gotten there yet ; that's next. Just was trying a couple of simple sensors to make sure it worked, after TCS34725 still did not work,
slender iron
if you have some time to fix an issue this would be good to fix: https://github.com/adafruit/Adafruit_CircuitPython_BLE_BroadcastNet/pull/16
tulip sleet
looks like it's worth fixing but I have a higher prio BLE bug to fix as well. I thought I could knock off this I2C thing but it's not a one-hour fix. I will save what I have and come back to it.
simple pulsar
<@404073847885922323> i'm trying the hold button approach on CLUE for enabling w...
No, I have classic CLUEs. I think I saw a note about ESP32S2 CLUE's using both buttons to go into bootloader mode.
simple pulsar
Hello. What are the units for the frequency arg for rp2pio.StateMachine ?
simple pulsar
@tidal kiln Have you got ESP32S2 CLUE? https://www.youtube.com/watch?v=h_kmnOEpAYk&t=25m48s
tidal kiln
@simple pulsar not yet. i'm using the current nRF52840 one.
ionic elk
How would people feel about a new "port_" prefix for functions that are implemented in the common-hal layer and used elsewhere in Circuitpython (like main, shared-bindings, and supervisor) but are not represented in the python API?
I've been feeling like it'd be nice to group them - Alarm has a lot of them, Microcontroller/Pin has a few, RTC probably has candidates too.
slender iron
I wouldn't do port_ because that implies to me that it's unique to a port
but maybe that's wrong. port_init and port_reset don't do that
ionic elk
I'm thinking stuff like alarm_enter_deep_sleep, reset_all_pins, and claim_pin, which have unique port level implementations, but don't have python equivalents
Maybe hal_? Since it's not the part that's "common" with python?
But I think the unique to a port part is what I'm going for? Maybe I'm misunderstanding you.
tulip sleet
@ionic elk something like internal_hal, maybe? I think the point of common_hal is that the same functions are provided by all ports: they're common across ports. I think common_hal routines would not HAVE to be called in service of implementing a python api. There is the old mp_hal name too
slender iron
@ionic elk is this the i2c issue you are tracking on the stm32? https://forums.adafruit.com/viewtopic.php?f=60&t=175632
manic glacierBOT
ionic elk
@tulip sleet if we define common_hal as items that should be accessible across all ports, then that would apply to many of the contents of alarm, like alarm_enter_deep_sleep and alarm_woken_from_sleep. Would you be open to changing those to a common_hal prefix?
@slender iron yes, that issue should be resolved by my I2C PR.
back on the prefix thing, I originally brought it up because I was noticing a lot of alarm functions in main that weren't common-hal prefixed, and it reminded me of a similar issue I worked on last year with the pin claiming functions spread across displayio that didn't have a common prefix or API
ionic elk
@slender iron just merged it, I'd forgotten to check on the CI passing.
slender iron
ok, great!
analog bridge
@slender iron can can you create a new branch for idf update so folks can easily collaborate
slender iron
is it not working to use patrick or bruce's?
analog bridge
patrick did some initial work but I had already gotten further in my branch... bruce is working on i2c off of my branch...
spi has a lot of changes in v4.3 which need to be tested... I am currently focusing on getting things running which were stuck due to idf@v4.2
https://github.com/adafruit/circuitpython/compare/main...microDev1:update-idf
tidal kiln
if one is using the pyportal in BLE mode, the pp lib can't really be used, correct? that seems sort of wifi specific.
tidal kiln
@gilded cradle hey. you'd probably know. pyportal question ^^ got lost in scroll.
gilded cradle
@tidal kiln, I believe that's correct. I haven't used bluetooth mode, but likely functionality just needs to be added and a firmware version check would need to be added.
tidal kiln
thanks!
gilded cradle
👍
ionic elk
I have a git fu question: I've made changes to a bunch of files but haven't committed anything yet. I think I need to break out some of those file changes into a separate branch so it can go in before the others. What's the best way to do this? I'm thinking maybe git stash push -p all the stuff I don't want, then make new branch with the unstaged material, commit it, return to the original, merge the branch I just, then pop the stash so I can keep working. Is there a better way?
jaunty juniper
alternatively there's git add -i, leaving the unstaged changes in the files
simple pulsar
Does anyone have an example of PIO code for RP2040 doing PULL operations? I'm struggling to get this to work and write to the rp2pio.StateMachine seems to lock up the code - cntrl-c can't interrupt it.
idle wharf
I have a question about https://github.com/adafruit/actions-ci-circuitpython-libs
I see this being pulled in when the build action runs for libraries and circup. But when I. look through it, it looks like (and I might be wrong) that all this is now covered by the cookiecutter and github actions in general.
If that's true, can we remove it from the cookiecutter (and thus libraries)? It would speed up the builds a bit.
onyx hinge
@slender iron if you have an idea what to call the module for driving the neopixels with shift register & pio, let me know -- I know you didn't dig the piopxl8 name for reasons...
.. I'll read back later, just remembered we didn't get to it on the stream (like that was gonna happen)
manic glacierBOT
slender iron
Does anyone have an example of PIO code for RP2040 doing PULL operations? I'm st...
I just tried pull with pdm but need to debug it. write will wait for a tx stall currently
onyx hinge
I also saw un-ctrl-c'able during PIO operations when I had bad code -- didn't debug it. It had to do with trying to shift in or out in the program when there wasn't an in/out buffer provided @simple pulsar @slender iron
the code's supposed to permit ctrl-c
I'd concentrate on debugging (by print statement if necessary) the two different places in the core that StateMachine write (and soon readinto and write_readinto) try to wait for keyboard interrupts
RUN_BACKGROUND_TASKS;
if (mp_hal_is_interrupted()) {...```The pull instruction does work, generally speaking. I have one in this shift-register neopixel program https://gist.github.com/jepler/a6968d731f78cf0f72e99f7b08e09956#file-adafruit_piopixl8-py-L41
tulip sleet
@onyx hinge I'm trying to track down a ulab issue. Are you still up?
tulip sleet
^^ it is this: https://github.com/v923z/micropython-ulab/issues/328
glossy pecan
@tulip sleet Pinging you because of an earlier msg of yours. Have you (or anyone you know) gotten Segger J-Link EDU Mini adapter working with ESP32 or ESP32-S2?
wise wigeon
Hello and thank you for maintaining this awesome library! I am currently working on a project where extreme power-efficiency is important. I bought the nRF52840 to see if it was fit for my project, and as I've always been a fan of Micropython, I was excited to see that you support this chip. My question is: how far along is the nRF52840 port of the alarm library, is it even being worked on? I see that it currently only supports ESP32-S2, but I could not find anyone working on the port to nRF52840.
The nRF52840 currently goes in to soft sleep on timeouts, but so far there is no interpretation of sd_power_system_off() - is that correctly understood?
onyx hinge
^^ it is this: https://github.com/v923z/micropython-ulab/issues/328
@tulip sleet no but I can take a look today if you like. (also have a guide to work on)
tulip sleet
thanks - I pinged you before I looked more closely. It's really a question for Zoltan. I just got on and haven't seen if he's replied yet.
fossil gorge
<@!252717193496756235> if you have an idea what to call the module for driving t...
NeoPIO?
tulip sleet
Hello and thank you for maintaining this awesome library! I am currently working...
@wise wigeon there has been no work on alarms for nRF yet, though it's on our radar. It does appear that you get very reduced power consumption on time.sleep() and similar. Your last sentence is correct, we don't call sd_power_system_off)_ explicitly yet.
tulip sleet
<@329766224093249548> Pinging you because of an earlier msg of yours. Have you ...
I haven't tried any J-Link with ESP32 yet. @onyx hinge might have gotten it working with either a regular J-Link or a Mini. It is not directly supported by the Segger software: it is only a conduit for commands from openocd or similar.
onyx hinge
@glossy pecan When I have needed to use a debugger with ESP32-S2 I've used the Kaluga development board which has a built-in debugger that works with openocd and is well documented by Espressif how to start & use the debugger. In principle the j-link hardware works but you still have to use (esp-idf's version of) openocd. I just updated this guide page https://learn.adafruit.com/adafruit-metro-esp32-s2/debugging-with-openocd because it gave the wrong "gdb" program name to use (it didn't call for the "esp32s2" version). When you run the wrong gdb, it can sort-of work sometimes, and I suspect that was related to my problems...
I had been meaning to go back and check whether I'd made that error so thanks for causing me to actually do thta
onyx hinge
@tulip sleet I suggested a (partial) fix for the buffer bug, it should allow any 1D arrays to be converted to buffers but not 2D arrays.
tulip sleet
@onyx hinge thanks. Since there is no ndarray_get_buffer() in the 2.x code, does it work completely differently?
ornate breach
Hey @tulip sleet what do you use for your Linux dev environment? I’m going to be setting up a nice little Linux dev server after I move 🙂
tulip sleet
Hi - are you asking about hw or sw? I use Ubuntu 20.04. I use the cinnamon desktop.
tulip sleet
desktop is a personal choice. I like the thin taskbar on cinnamon. I'm not that fond of unity. Never really tried kde for any length of time
taskbar. i have two monitors
ornate breach
Makes sense. I was basically going to set up a server and use a cheaper mini desktop to virtualize on in my office
tulip sleet
what kind of server?
tulip sleet
I am just using an old Intel i5 4-core processor and 16gb, with ssd's. But i have a 6-core i7-9500t arriving in a day or two. Both of these are used Dell Optiplex desktops.
scott has a hefty zillion-core AMD, but i am not going to spend that much
the latest Dell was $420 from dellrefurbished.com
there is always a coupon, or just wait a week or two
ornate breach
Haha, yeah I might go AMD but I’m also considering a r240 from Dell
They have one for $500 right now
tulip sleet
do you have other uses for a server? do you need a home media server or something like that?
ornate breach
It will manage NAS storage and whatnot
I’ll be upgrading my Netgear readyNAS 2 bay NAS to a rack mounted NAS
tulip sleet
i have been going for single used machines instead: tv computer, development desktop, another desktop for Windows testing, another leftover tiny one for Windows Insider Preview testing, etc. I was tired of managing VM's
onyx hinge
For building CircuitPython—that is, repeatedly building one particular firmware.elf—having a zillion cores doesn't help that much.
(I use Debian & xfce but I don't necessarily recommend it)
manic glacierBOT
tulip sleet
this "new" machine I am getting is 6 cores 12 threads, tshould maybe be twice as fast.
it is micro-form-factor, but the speed penalty is only about 10%
tulip sleet
i used debian for a long time but I like the up-to-date-ness of ubuntu
ornate breach
I’m also going to probably host a website or two. I’m not sure yet. It’s more to play around with, get better setting up and developing in Linux and creating an isolated development environment away from my regular computer
I’m probably going to load a docker container on it to to manage my 3D printers
Plus circuitpython development will happen there
onyx hinge
Throwing 16 threads at the build process only speeds it up by less than 4x...
tulip sleet
make sure the remote desktop setup you proposes works well for remote USB. I am sticking to native so I can see USB problems for what they are, instead of being confused by VM USB issues
does the remote desktop sw you are using do remote usb?
ornate breach
I’ve not chosen Remote Desktop software yet, I’ve used VMware Horizon Client
But I’m not sure if it supports native USB
tulip sleet
so plugging/unplugging boards, using J-Link etc. might be problematic
ornate breach
Potentially
Maybe I’ll set up CI on the server that I can fetch on my desktop
Like a local git server or something
tulip sleet
i haven't really had the need for that kind of thing. Usually I just push to github at least every day for backup reasons, and when I'm ready for a PR, I rely on GitHub to do the big CI. But if you are doing other dev tasks too, that could be handy.
onyx hinge
It's a good idea to have some kind of backup plan for any data you're working with. Anything that backs up your user files will work to back up your working files & local (non-pushed) git changes. but pushing regularly to a remote system like github is also not a bad idea, because it means you have different independent kinds of back-ups.
ornate breach
Yeah, that’s the thought
I’m just mostly looking to brush up understanding and learn more stuff related to developing software and devops things.
ebon ingot
Hi! I was directed here by @slender iron, as I'm in the process of adding definitions for Pimoroni RP2040 boards. I've got them written and compiling, and am looking to create a draft PR for review. The only issue is, as we already have a fork of Micropython, Github won't let us create fork of Circuitpython too. I could fork to my personal account, but that will become problematic when several of the team wish to contribute.
Scott mentioned something to me about "adding pimoroni/micropython as a remote to push to" but I'm a bit of a beginner when it comes to Git so have no clue what that actually means 😆 . Any help with this would be greatly appreciated
tulip sleet
Hi! I was directed here by <@!252717193496756235>, as I'm in the process of addi...
I think you should be able to add adafruit/circuitpython as a remote to your local pimoroni/micropython clone. Then do git fetch adafruit/circuitpython. You can then checkout adafruit/main, branch it, and create a PR. But you might see "detached head" warnings - I'm not sure.
analog bridge
@ebon ingot you can create a new branch from CircuitPython & MicroPython break-point (i.e. before Jan 1 2017) and then pull in the changes from adafruit/circuitpython...
fossil gorge
I’m just mostly looking to brush up understanding and learn more stuff related t...
For this, I would strongly recommend NOT using that environment as your sole/primary backup or storage plan. If you want to try stuff out, but you risk losing data, or want to reset
ebon ingot
Thanks for the suggestions @tulip sleet & @analog bridge . No idea how to do either of them, so any guides you could point me too that would step me through the process would be welcome
tulip sleet
@ebon ingot I think this is going to be painful for anyone else trying to do PR's to circuitpython in your organization. You might consider creating a new organization to hold a new fork. I know that's not great. You could segregate all circuitpython dev there (e.g. also CPy libraries), and then it begins to make more sense
I wish github handled this better
i will do a little research
analog bridge
git checkout -b branchname 0ae344841
git pull https://github.com/adafruit/circuitpython.git
@ebon ingot 👆 and you are good to go
manic glacierBOT
ebon ingot
And that's to be applied to a clone of our micropython repo, yes?
analog bridge
I'd suggest doing another clone as the change is too big and switching branches would be cumbersome
git clone https://github.com/micropython/micropython.git circuitpython
Okay, thanks. I'll give that a try
tulip sleet
@analog bridge so @ebon ingot and colleagues will need to repeat git checkout branchname; git pull https://github.com/adafruit/circuitpython.git in each new clone, right?
where branchname is circuitpython-dev or similar
whatever it was called the first time
(and that branch should be pushed at some point)
@ebon ingot Also be careful when you submit a PR that it's against adafruit/circuitpython and not micropython/micropython, which can be embarassing (I've done that by accident a couple of times).
analog bridge
<@!123644956421259264> Also be careful when you submit a PR that it's against ad...
ya... the gui will default to micropython
onyx hinge
to do circuitpython work, why don't you suggest making the initial clone be of adafruit/circuitpython, @analog bridge ?
analog bridge
I'd suggest doing another clone as the change is too big and switching branches ...
oh shooot I suggested the wrong repo here... I meant the clone of your forked micropython repo
onyx hinge
I see.
analog bridge
to do circuitpython work, why don't you suggest making the initial clone be of a...
this can also be done... I think this requires a remote change to the forked repo
onyx hinge
Git accommodates a lot of different workflows; I typically start by cloning where I eventually will pull-request to, so that origin/main or origin/master refers to the place I want to start branches. I add my own fork with git remote add jepler .... Starting new work is git fetch origin; git checkout -b project-name origin/main (or origin/master). Pushing it (first time) is git push --set-upstream jepler HEAD and subsequently is git push. Creating a PR is visiting the intended repository and going through the steps manually to create a PR from my repo to the intended repo. Take that all for what it's worth.
tulip sleet
the base problem is that they already forked micropython in their org, so they can't fork circuitpython also, maybe your workflow accomodates that. I always clone my own dhalbert/circuitpython, because it contains my un-PR'd branches, etc.
analog bridge
the base problem is that they already forked micropython in their org, so they c...
this is also a problem at the user level too and not just for org
manic glacierBOT
ebon ingot
```bash
git checkout -b branchname 0ae344841
git pull https://github.com/adafrui...
Just to confirm, which repo was this meant to be to. I saw your comment about having said the wrong one, so I took that to mean I should clone pimoroni/micropython instead of micropython/micropython?
analog bridge
yep... pimoroni/micropython it is
ebon ingot
Hmm okay. Well it shouts at me when trying the checkout line
fatal: '0ae344841' is not a commit and a branch 'circuitpython-dev' cannot be created from it
hmm... let me give you an older commit.... how about a081b49d
ebon ingot
That worked
ornate breach
For this, I would strongly recommend _NOT_ using that environment as your sole/p...
Everything will have redundancy in terms of storage
ebon ingot
hmm... let me give you an older commit.... how about `a081b49d`
So when I come to push, what should I add on to the command? Git suggests ```git push --set-upstream origin circuitpython-dev
@tulip sleet would it become more difficult for us to merge from Micropython if Circuitpython became a non-fork repo?
analog bridge
So when I come to push, what should I add on to the command? Git suggests ```git...
yes... its like any other branch
tulip sleet
<@!329766224093249548> would it become more difficult for us to merge from Micro...
Yes, that's the reason we haven't separated.
ebon ingot
yes... its like any other branch
After all that, it seems I cannot do the push because of some stupid permissions thing, despite having successfully created many branches for this repo before
fatal: unable to access 'https://github.com/pimoroni/micropython.git/': The requested URL returned error: 403
analog bridge
this is for the new clone?
ebon ingot
Yes
analog bridge
you can make branch upstream and then modify it locally and force push
analog bridge
should we switch to #help-with-git as this is getting more git focused
ebon ingot
Ah yes, good idea
wise wigeon
<@208541388307038208> there has been no work on alarms for nRF yet, though it's ...
I've looked into the nRF52 SDK, but I have not yet established a complete overview of working with the sleep bindings. If I were to initiate the work of connecting your alarm bindings to the SDK, how and where would you suggest that I begin?
tulip sleet
@wise wigeon You'll need to reimplement ports/esp32/common-hal/alarm for nrf. I am not sure if TouchAlarm makes sense for nrf, so that might be omitted. Note the discussion of light sleep and deep sleep in the documentation and the Sleep Learn Guide. We initially thought that light sleep on the esp32-s2 would preserve enough state that we could use the ESP light sleep, but it turns out even its light sleep shuts down a lot of things.
The nrf52 chips have excellent power-down capabilities, and you may be able to keep more things on at a low power draw. It may not be necessary to switch to system OFF to achieve very low draw.
wise wigeon
<@208541388307038208> You'll need to reimplement `ports/esp32/common-hal/alarm` ...
Thank you for the swift responses!I will begin running power consumption tests on monday, and as you mention, it might not even be necessary to invest time into optimizing more that it is.
As can be seen on https://infocenter.nordicsemi.com/pdf/nRF52840_PS_v1.1.pdf page 56, the difference between system OFF and ON is insignificant in most cases.
tulip sleet
yes, I was surprised by that, but they've done an excellent job in power reduction.
wise wigeon
yes, I was surprised by that, but they've done an excellent job in power reducti...
Indeed! Thanks again 👋
tulip sleet
and thank you for working on this!
simple pulsar
I also saw un-ctrl-c'able during PIO operations when I had bad code -- didn't de...
I was doing 4 bytes which I believe means I'm not in the DMA code path
ionic elk
@tulip sleet quick couple of questions for you:
- is
common-hal/alarm/pin/__init__.can obsolete leftover? - it references
alarmio, is there a particular reason why we didn't go with that name? I kind of prefer it toalarm, since there's a lot of included IO with the pin interrupts and adjustments to other parts of the processor.
@wise wigeon I am currently implementing alarm for STM32 so I'd be able to help/provide reference for NRF once that's done.
tulip sleet
- yes, that is junk; the routines there are not used elsewhere.
- I thought for a while about alarm vs alarmio, and liked alarm better, because some of its functionality is not io, like sleep alarms
orchid basinBOT
ionic elk
Also, was esp_sleep_pd_config not useful in narrowing the domains that ESP32S2 shuts down in light sleep? The docs make reference to it as a way to change the behavior of the RTC
tulip sleet
i don't think we are going to change it now, since it's in stable releases. Originally it was sleepio, but I thought we might have alarms not realted to sleep later
esp_sleep_pd_config I think might be called one or two places. the ESP-IDF already calls it properly in several places to do the right thing depending on what's enabled
there is discussion with igrr about that in some of the PR's or issues
ionic elk
i don't think we are going to change it now, since it's in stable releases. Orig...
Ok, I didn't feel especially strongly about it, it's fine to have it in the same category as rtc and os and stuff.
tulip sleet
primarily, light sleep turns off wifi, and that's not just pd_config.
ionic elk
Turning off wifi is a big issue?
tulip sleet
it changes wifi state, it doesn't just suspend things
ionic elk
hmm. Is that irrecoverable?
ionic elk
That's not such a big deal, though, is it? As long as it's clear upfront. If you have a program that sleeps for 1 hour, your connections are probably borked anyway.
tulip sleet
yeah, but light sleep otherwise doesn't save that much power, so you may as well just time.sleep()
ionic elk
Ah, yeah if it doesn't save anything that makes sense. I'm surprised it doesn't though, if it shuts all this stuff down.
simple pulsar
@ionic elk It would be polite/efficient to close the connections if you know you're going to sleep for that long unless they are to a private resource.
ionic elk
I'm trying to figure out parallels to the STM32, since ST chips have three states, not two: Sleep, Stop and Standby. Standby is the closest to deep sleep, Sleep is the ideal version of Light Sleep (some savings with no state loss) and Stop is kinda in between
@simple pulsar I'm simply wondering if light sleep should actually handle a clean Wifi shutdown automatically, and warn users of that in the docs.
I'm just speculating though, it doesn't sound like it's priority since the light sleep doesn't save much anyway
tulip sleet
we wanted the semantics of alarm.sleep_until_alarms() not to turn things off. It also resets the pins
my light-sleep measurements at the time (no longer in the guide), showed it drew about 1.5mA
so ... maybe losing the pin state was also a big lose
if you are going to lose all that, you may as well just deep sleep and restart
ionic elk
But right now light sleep doesn't actually do anything at all, right?
tulip sleet
it just does WFI, it's like time.sleep()
on the nrf you can do a lot better
with a light sleep; you can preserve all the pin state and the RAM and still just draw microamps
the ESP32-S2 is just not a good model for light sleep that preserves state
simple pulsar
I just tried pull with pdm but need to debug it. write will wait for a tx stall ...
The PWM PIO example is a good one to try and replicate to prove this code. The init vs exec looks problematic though: https://forums.adafruit.com/viewtopic.php?f=60&t=175662
marble hornet
@tulip sleet can I bounce a quick question off you regarding the nrf52840 board config? it looks like there are two conflicting BOARD_HAS_ definitions.
tulip sleet
sure
orchid basinBOT
marble hornet
what's the difference between BOARD_HAS_32KHZ_XTAL and BOARD_HAS_CRYSTAL.
EX: I dead bugged an nrf module and found that adding a 32.768 khz crystal makes the _bleio module importable.
orchid basinBOT
tulip sleet
nrf likes to have both a 32kHz crystal and a something-MHz crystal (32, I think, but I'd have to look)
orchid basinBOT
tulip sleet
without one of those, you don't have accurate enough timing for BLE.
The CPB is unusual and has only one crystal, I think. Let me look.
yeah, it only has the 32MHz crystal. Most have both, for more accurate timekeeping, I think, when the CPU clock is off
Long ago I wanted to rename BOARD_HAS_CRYSTAL to something more specific for nrf, but didn't get around to it yet.
in CPB's mpconfigboard.h:
// Board does not have a 32kHz crystal. It does have a 32MHz crystal.
#define BOARD_HAS_32KHZ_XTAL (0)
turbid radish
?serverinfo
digital shoreBOT
adafruit
marble hornet
thank you @tulip sleet ! Ah, I thought (assumed) crystal referred to the 32khz. the watch now has BT! it's the last known bug! switch BOARD_HAS_CRYSTAL on and BOARD_HAS_32KHZ_XTAL off.
tulip sleet
great! yes, it's confusing, I spent a while with schematics. The nrf modules often don't mention they have an internal crystal, so I was pretty confused for a while
marble hornet
would it be acceptable to propose a change from BOARD_HAS_CRYSTAL to BOARD_HAS_MAIN_CRYSTAL or BOARD_HAS_32MHZ_CRYSTAL in a pr?
tulip sleet
would it be acceptable to propose a change from `BOARD_HAS_CRYSTAL` to `BOARD_HA...
I think that would be great! You can leave BOARD_HAS_CRYSTAL alone in ports/atmel-samd, but change all uses in ports/nrf. The original use for it was to store USB clock calibration data on atmel-samd. It is used only in those two ports.
marble hornet
gladly! i'm leaning towards MHZ over MAIN solely because it draws attention to the units, thoughts or input?