no matter how many times i "proofread" my comments, i always miss something. ๐ฐ
#circuitpython-dev
raven canopy
manic glacierBOT
manic glacierBOT
slender iron
@solar whale have you used bluepy? I want to make sure we target its API and not ubluepy
solar whale
@slender iron no - so far all the examples I have tried have been using ubluepy.
slender iron
ok, I think the largest thing will be using attributes where bluepy does instead of methods like ubluepy
solar whale
i'm very new to this but will be happy to take a look at it.
@slender iron only ubluepy is in the moules list at this time - with the current build
slender iron
yeah, it'd be good to get it reorganized
solar whale
@slender iron I found the ubluepy code in nrf/modules --- is bluepy somewher else? Not finding it.
solar whale
ah - so you would like to see that replace ubluepy ??
slender iron
yeah
the api should be the same
they are close but got adapted for mciropython-isms like not using properties
solar whale
@slender iron ah - saw your note in the PR - sounds good. This has been fun to play with.
slender iron
great! thanks as always for the testing
solar whale
Kevin has made it pretty easy!
slender iron
he's a rock star
solar whale
Now I just need to find time to study his book!
slender iron
yeah, I thought about picking it up
I realized on fs issue coudl be that we don't flush the flash cache when repl is active unless a different block is written
solar whale
could taht be related to the import problems I reported?
slender iron
I honestly haven't read in depth about the issue. I saw it go by
solar whale
I can give a report on it tomorrow - in the issue - I tracked it to down to a specific build but I don't see the cause.
slender iron
I did see that. its super helpful
solar whale
easy workaround when it happens
the potential FS bug is good to find. Htere have been some "funny" things taht are not reproducible. They are so hard to find..
slender iron
yeah, really hard
any memory issues you've come across with libraries would be helpful
my heap tools are very neat
solar whale
I saw your "map" very nice!
slender iron
yeah, I think it'll be useful to find bad libraries
idle owl
I tried to get an nRF52 setup, pretty sure I hosed the board. Although my friend who is a wizard with hardware says he bets he can fix it.
raven canopy
it definitely isn't an easy process. @solar whale is a champ though. helped me get through mine.
solar whale
just follow the steps in ports/nrf/boards/feather52 readme
idle owl
I'm talking about getting the connector on in the first place.
solar whale
teh swd connector?
idle owl
yes
raven canopy
oh. you were going the Segger/JLink route
solar whale
you really don't even need it for the feather52
idle owl
Oh. Thought I did, heh
solar whale
the docs use the dfu port so it is all done via USB
I used j-link initially before this was updated but have been using the dfu recent;y - have not tried j-link in quite awhile.
idle owl
I see
solar whale
but it is nice to have in case of "emergency"
raven canopy
...told ya he was a champ. ๐
solar whale
@raven canopy I just like banging my head against things....
raven canopy
it's one of my favorite past-times as well. ๐คผ
i also like to change thoughts mid-stream. example, 2 minutes ago i ended up with a +=, and was surprised when the output value kept climbing exponentially.
idle owl
@slender iron I need to get Sphinx setup. You helped me with getting the venv setup, but we never got to getting Sphinx up and running again. Do you have a few minutes?
raven canopy
@slender iron now this looks a little closer. upped the the raw_id[] size to 128 since 16 was causing the bootloader to fail. however, i'm pretty sure 128 is too much...
Auto-reload is on. Simply save files over USB to run them or enter REPL to disable.
main.py output:
^...SAMD CPU Information...^
Frequency: 48000000
Unique ID len: 128
Unique ID: 15007828
Adafruit CircuitPython 3.0.0-alpha.1-62-g111faa1-dirty on 2018-01-15; Adafruit Feather M0 Express with samd21g18
>>> import microcontroller
>>> microcontroller.cpu.uid
bytearray(b'T\x00\xe5\x00\xbe\xa9\xcb\x00|\xa9\x07\x00\xf0\x00\x0f\x00R\x00%\x00R\x00\xd5\x00M\x90\x04\x00P\x06\x05 J\xa4\x14\x001\x00\xe3\x00.\x002\x003\x00\x03\x00=\x00C\x00$\x00b\x00\x06\x00\xf0\x00\xff\x00\x0f\x00\x02\x00\x00\x00\x04\x00\x00\x00\x00\x00\x00 \x0f\x00\x00\x00\r\xa8\x01\x00\x91\xa9\x01\x00\x01\x00\x00\x00@\x00\x00\x00\x9f\xa9\x01\x00\x01\x00\x00\x00\xc0\x13\x00 ]i\x00\x00\x00\x00\x00\x00\n\x00\x00\x00\x0b\x00\x00\x00\x00\x00\x00\x00')
>>>
mathing tells me it should be 64...
slender iron
@idle owl are you free in an hour or two? I need lunch soon
@raven canopy I think 16 was correct
idle owl
@slender iron Yah should be
manic glacierBOT
slender iron
its four words of four bytes each so 16 total
raven canopy
then i am adding to the raw_id array incorrectly. back at it
slender iron
yeah, did you see my comment?
raven canopy
i did. working it...in baby-step fashion. ๐
slender iron
I think you need the loop from esp8266 code inside the code for the samd and nrf
raven canopy
well, that one was easier for me since it's just a single nest. just need to get the left side of the second-level assignment figured out. (that's where I got the 64 a minute ago). I'll get there.
idle owl
@solar whale I did manage to terribly get the button on the metrowing and it resets the feather/featherwing on top, so apparently it at least works.
solar whale
@idle owl nice! It does not have to be pretty to work!
same goes for the SWD connector on the nrf52!
@idle owl I'm looking ove the makefile for using the jlink with the nrf52 - I'll let you know how it goes.
flashing the image woks fine - now to try the soft-device
it seems to have survived
manic glacierBOT
idle owl
@solar whale Bonus! Survival is good
solar whale
@slender iron cool! I was just able to run a .py file from an sd card!!! who knew!
somehow I thought that had been discussed and could not be done....
@slender iron I did this on a MEtro_m0 Express with an adalogger featherwing. is there any reason it should not also work with a non-express board like a feather m0 basic?
@idle owl using the j-link is faster and simpler than the DFU method so when you get to that point, it may be an option. May as well put that connector to use after all the work getting it installed. I just want to get confiration on one point before I deem it a complete success.
However, it does wipe out the DFU - boot loader so you have to choose your poison. It is not hard to re-install the DFU loader.
idle owl
good to know on all counts.
I don't think the one I tried to connect will work. It's bad looking. Not on straight or whatever.
solar whale
not hard - if you remember how ๐
idle owl
But I have a rework station now, so.... maybe I can get another one connected?
I don't have solder paste though
Kind of didn't think this through. From looking at the images of it, I thought I could do it. I was wrong.
And I forgot to have Brian do it.. he's the solder wizard. Would have had it done and perfect looking.
solar whale
I think I just got lucky!
cunning crypt
@idle owl What are you attempting to do?
I've missed everything before "rework station"
slender iron
@solar whale I think it should work for the other boards as well
solar whale
That's really nice - sorry I gave the wrong answer on the forum. It will be great for some of the basic boards.
slender iron
np, thats how we learn ๐
solar whale
I was sure it had come up and had been ruled out. Lesson learned to do simple tests before relying on vague memories!
manic glacierBOT
raven canopy
@slender iron i cheated and just setup a "counter" in the nested loop. truth tabling with a recurring zero beat me. :D
>>> microcontroller.cpu.uid
bytearray(b'\x04\x0e\x0c\x00\x02\x02\r\x00\n\x01\x0e\x03\r\x04\x06\x0f')
>>> struct.unpack('<i', microcontroller.cpu.uid)
(790020,)
>>>
uint8_t id_tick = 0;
for (int i=0; i<4; i++) {
for (int k=0; k<4; k++) {
raw_id[id_tick] = (*(id_addresses[i]) >> k * 8) & 0xf;
id_tick++;
}
}
i think id_tick is 4 * i + k
raven canopy
well now it's staring right at me...too many iterations and i lost it (forest/trees). thanks!
idle owl
Excellent, only one library left on the Lint All The Libraries list!
@cunning crypt Trying to attach a connector with very tiny feet to very tiny pads on a Feather. Is what I am trying to do.
cunning crypt
Oof. I'd be worried about the rework station melting the plastic on the connector, but it might be possible with care.
idle owl
Yeah same.
I am fairly certain I already hosed one, so I think I might leave the rest to someone with more skill than I have.
cunning crypt
Which feather, which connector, and which location?
idle owl
nRF52, SWD, center of the board.
cunning crypt
Oooh, yeah. My personal attempt would be with a soldering iron.
idle owl
I did do that. And I did end up removing half of one of the pads on the feather, and the connector isn't straight, so maybe it actually connected? I am highly skeptical.
cunning crypt
How time-pressed are you?
idle owl
Not pressed. It's something I'd like to get going but there's nothing waiting on me. I have a mate who is a solder wizard who might visit again next weekend if he's got time and he could do it. I simply forgot to have him do it when he was here this past weekend. Unfortunately.
cunning crypt
I'm confident I could get one of the feathers and solder a header on there, but probably not before your friend gets to it
idle owl
Hmm. Depends on whether he makes it, I suppose.
cunning crypt
You're putting a new bootloader on it, right? Why not use the NRF52 Pro? https://www.adafruit.com/product/3574
idle owl
Because I wasn't thinking when I ordered it ๐
I looked right at it and thought, oh this doesn't work for what I need. When I knew bloody well it would work perfectly.
idle owl
Exactly.
manic glacierBOT
slender iron
bus device is built into some firmware right?
raven canopy
@slender iron i'm looking at giving a Trellis library a go. ladyada mentioned a couple weeks ago, and I've been studying since last night, the HT16K33 library. Which is preferred: standalone Trellis, or Trellis functionality (button reading) added to the HT16K33 library?
slender iron
i don't know what trellis is
in general, its better to have functionality split so only what you need can be imported
raven canopy
slender iron
I'd suggest building a separate trellis library that uses the ht16k33 one
raven canopy
well, I could just import the current library inside a Trellis library. yeah. same page.
slender iron
yup yup ๐
raven canopy
i doubt i'll be of much help, but are you talking about native bus device? or CP implemented?
slender iron
gah! ```>>> from adafruit_featherwing.motor_featherwing import MotorFeatherWing
wing = MotorFeatherWing()
s = wing.stepper2
s.onestep()
MemoryError: memory allocation failed, allocating 512 bytes```
idle owl
Oi
slender iron
stupid parser
ok, it works
code.py works fine
dear repl,
(โฏยฐโกยฐ๏ผโฏ๏ธต โปโโป
manic glacierBOT
The main tool is tools/analyze_heap_dump.py tools/analyze_mpy.py is still very coarse but the Preamble stuff is used by the dump tool.
The heap dump tool creates graphs of the heap objects like this:
raven canopy
well, i think i 
my Trellis. assumptions...get you everytime. "Of course i put the right color wires on the right connections..."
blazing trail
lol
covert birch
hey peeps. quick question: can i run/install circuitpython on my bluefruit feather? i;m guessing no as it does not appear in the release list.
covert birch
@slender iron does that mean i wouldn't be able to use it with the bluefruit app?
slender iron
@covert birch correct
covert birch
ok thx
slender iron
np, sorry its not there
covert birch
it's cool. just glad to clarify
slender iron
we'll have bluetooth support with the nrf52 first
sick creek
so cool
covert birch
i just want to play around with my dotstar featherwing. i'm just learning circuitpython, my first exposure to any python and was having trouble with the dotstar featherwing sitting on my mo express which does not have female headers and i've already soldered pins to the dotstar.
slender iron
@covert birch it should be usable then, you just won't be able to use bluetooth
the drive will be much smaller on the bluefruit board though since its not an express
covert birch
oh, well i trying to work with the mo now. may just breadboard each and try that way, if i can get the jumpers in the right spots
slender iron
that'll work too!
manic glacierBOT
upbeat plover
will there be a library for RFM69 to use with circuitpython?
stuck elbow
I'm sure that sooner or later someone writes one
slender iron
@upbeat plover @timber lion has started working on it here: https://github.com/adafruit/Adafruit_CircuitPython_RFM69
timber lion
yeah the library works great, see simpletest.py for example of sending an receiving
we'll have a guide update later this week
but one thing, you might want to familiarize yourself with the radiohead library for arduino
these radios are pretty complex and support a ton of different modulations
this lib tries to emulate radiohead's GFSK, 250kbit/s, 250khz frequency deviation by default
so it can talk back and forth to itself and to radiohead's default GFSK_250Rb_250Fd config
if you want to do something different you'll need to configure the radio yourself
and also be careful these radios send 60 byte packets at most.. radiohead hides this and adds extra layers on top like guaranteed delivery, multipart messages, etc
we don't have that, it's just like the 'raw' radiohead usage
think of it more like UDP packets
i do use the same 4 byte packet header as radiohead though so its compatible with radiohead receiving our packets
but you can't control node address (it's just always broadcast to everyone) or do other advanced things that radiohead might add
so yeah check out the simpletest.py, it shows the basic usage of send and receive and kind of the intended use
if you try to receive kilobytes and kilobytes of data at a time it will be troublesome.. you'll lose packets because we can't read them fast enough
so stick with simple 60 byte control messages
solar whale
@timber lion thanks for this addition! Looking forward to working with it.
cunning crypt
@timber lion Is a LoRa RFM95 planned? I'm assuming you're going to get this library complete first, of course
timber lion
yeah lora's not on the list yet, haven't look at it but it might have a bit more complexity
there's a handful of sensors to do next
eventually will get to lora i bet though
this RFM69 should be good to go for folks now, just working on a guide update to show the basic usage that simpletest.py demos
idle owl
@slender iron I tried updating api.rst and it didn't do what I think it was supposed to. I'm also getting an import error on when Sphinx is trying to build anyway, so I'm not sure if that's the issue even though it builds. Doesn't seem to include the new lib, so the issue may not be with api.rst.
slender iron
whats the import error? you may need to automock something
idle owl
That sounds like it would be right. reading sources... [100%] api /Users/kattni/repos/Adafruit_CircuitPython_FeatherWing/api.rst:6: WARNING: autodoc: failed to import module 'adafruit_featherwing.ina219_featherwing'; the following exception was raised: Traceback (most recent call last): File "/Users/kattni/repos/venv3/lib/python3.6/site-packages/sphinx/ext/autodoc.py", line 658, in import_object __import__(self.modname) File "/Users/kattni/repos/Adafruit_CircuitPython_FeatherWing/adafruit_featherwing/ina219_featherwing.py", line 34, in <module> import adafruit_ina219 ModuleNotFoundError: No module named 'adafruit_ina219'
that's from the middle of the Sphinx build
slender iron
yup yup
idle owl
No idea how to do that though
idle owl
ok
slender iron
I think I left an example in there commented out
idle owl
autodoc_mock_imports = ["adafruit_motor", "adafruit_pca9685", "board", "busio"] this?
slender iron
yeah but delete the contents and list the things for your library that you need
idle owl
I'm building that lib inside adafruit_featherwing though. So don't I need to leave those too?
idle owl
ah
slender iron
just add to the list
idle owl
Hah! Sorted a bunch of other errors that popped up once it dealt with that. And Sphinx is now happy
slender iron
great!
slender iron
@opaque patrol @umbral dagger @stuck elbow gonna make the meeting?
stuck elbow
yes
slender iron
sweet!
umbral dagger
On my way
slender iron
kk
prime flower
using txt ๐
tidal kiln
hear me?
raven canopy
Ditto txt
prime flower
nope, sorry for the lack of generosity this week
raven canopy
HUGS: @slender iron @solar whale for putting up with my myriad of windows + toolchain = headache issues. And @tonyd for his awesome videos on setting up the VMs.
prime flower
status update(s): metro express explorers guide code is now linted (thx kattni/mrmcwethy) and conforms to 2.2.x instead of what it ran over the summer and runtime errors are gonee, physical booklet for the metrox classic kit is almost done too (it's really awesome - just wait for the cirpy edition..SOON!), and I started a badge design for the BSides security conference in boston (its gonna use circuit python whooo)...oh and I applied to grad school so fingers crossed (my thesis proposal incorporates circuitpython)
timber lion
wow good luck with grad school app! ๐
prime flower
thanks tony!
idle owl
@prime flower That's great! Good luck again!
umbral dagger
@prime flower Woot! Grad school is a blast.
raven canopy
uid as you said earlier is done (minus pending PR). Started working on a Trellis library...+/- wires crossed... came to town. So up next: start working on VEML6070 UV Sensor library, and practicing SMD repair once a new HT16K33 chip shows up. Other than that, cruising the Issues pages for ideas.
tidal kiln
โฌ #sparky
sick creek
raven canopy
UICR is also available on SAMD in NVM. User Info Control Register
But, that's Express only ATM.
raven canopy
Thanks everyone! 
umbral dagger
This was awesome... glad I finially made it.
idle owl
Yeah glad to have you!
prime flower
thx everyone (was deep in illustrator listening to the call ๐ )
slender iron
@fading solstice thanks for doing the is31.... I think I'll skip it on my end then
raven canopy
someone is busy tugging on the issue thread... ๐
hot surge
can someone help me with a general python question?
tulip sleet
@hot surge go ahead
hot surge
im trying to convert an opencv/numpy array into a format that can be used to send a request to the google cloud vision api
basically, i cant figure out how to go from a numpy array to a json serializable utf8 base64encoded something
this works
with open('image2.png', 'rb') as image_file: image_data = base64.b64encode(image_file.read()).decode('utf-8')
but i can't figure out how to take a numpy array and get to the same type
tulip sleet
do you know how to make the json? Are you using a json library?
what's the code that creates the numpy array? The base64 thing looks likes it's assuming that it's a base64 encoded png, which is not going to be like a raw image in an array
.png files have headers and others stuff in them: they are not just raw pixels
could you give a link to the REST cloud API call you are making?
looks like you could use opencv cv2.imencode('png', imgdata) to convert a raw buffer of pixels to png format, then you can pass that to base64.b64encode()
that's basically the in-memory version of imwrite, which you could use to write a png file that would be read by your code above
timber mango
I modified Adafruit_CircuitPython_seesaw/adafruit_seesaw.py
and copied it to /lib on the Metro M0 Express. It is not an .mpy file.
I do not know how to make an .mpy from a .py. ;)
I did remove the existing /lib/adafruit_seesaw.mpy that was already present.
soft reboot
Auto-reload is on. Simply save files over USB to run them or enter REPL to disable.
main.py output:
Traceback (most recent call last):
File "main.py", line 3, in <module>
MemoryError: memory allocation failed, allocating 130 bytes
Press any key to enter the REPL. Use CTRL-D to reload.
hailp ;)
solar whale
@timber mango get the appropriate mpy-cross from https://github.com/adafruit/circuitpython/releases then just execute mpy-cross your.py
timber mango
Oh! wow. Neat. Thanks, Jerry.
solar whale
good luck. That should help witht he memory allocation at startup.
timber mango
Thanks. Didn't see any errors. Seems to work just fine! @solar whale
timber mango
https://github.com/wa1tnr/Adafruit_CircuitPython_seesaw.git
I now have the ability to setup the PWM Freq for the Seesaw Breakout, in CircuitPython.
Checked with the Extech EX330 in 'Hz' mode (duty cycle also checks).
Also repaired analog_write() which was only passing 8-bit values (unless
I've misunderstood something basic, which is possible). It works now. ;)
Oddly enough, the Seesaw Breakout seems to have a power-on default of 44.59 Hz (but cannot be relied upon to stay there during complex operation).
upbeat plover
For CPX what are the MOSI and MISO, is it TX/A7 and RX/A6?
tulip sleet
@upbeat plover board.MOSI and .MISO refer to the SPI flash chip, but there are other pins you can use for SPI. Looking and will return shortly...
covert birch
ok, i screwed something up and now my computer is showing two volumes of circuitpy, the second one is called circuitpy 1 and it is the only one that shows up in atom - i can only see the two volumes in terminal - it's not in my finder window. how can i repair this situation?
tulip sleet
@upbeat plover these work:
busio.SPI(board.A3, board.A2, board.A1)
busio.SPI(board.A1, board.A3, board.A2)
and maybe more. The order is clock, mosi, miso. The TX and RX are UART designations
@covert birch do you have two boards plugged in? Also, what OS?
upbeat plover
thank you
covert birch
no, mac os 10.11.16
tulip sleet
did you just try rebooting?
covert birch
no, guess i should try that
tulip sleet
we've seen MacOS get confused or refuse to mount CIRCUITPY. A reboot has fixed it.
let us know if it doesn't
covert birch
ok, that makes sense. i've had a lot of problems trying to use my dotstar wing because i don't have any pins soldered to my mo express
i will let you know
tulip sleet
you mean you were just trying to make them touch? yeah, that will be quite unreliable. I know "reboot" seems like Windows advice, but in this case it seems to apply to Mac as well.
covert birch
yes, it does. i will solder in some stacking headers too before i plug in after rebooting
i think that must be cause of all the problems
thanks @tulip sleet
timber mango
raven canopy
are we able to handle Exceptions from submodules? I'm getting a ValueError that I can't seem to catch:
with I2CDevice(i2c_bus, _VEML6070_ADDR_ARA) as ara: # address will only exist if there is an Alert Response Address (aka INTerrupt)
try:
ara.readinto(ara_buf)
print('ARA: ', ara_buf)
except ValueError: # also tried except Exception & OSError
print('err0r')
@raven canopy are you getting the ValueError on the ara.readinto() or on the with? Maybe you need to wrap the with in try?
raven canopy
@tulip sleet sorry, could've made that clearer. getting the error on ara.readinto(). "No I2C device at address..."
tulip sleet
@raven canopy I think you will get that error if there's no device at that I2C address earlier, in the I2CDevice constructor. Checking...
raven canopy
moving the try: up a level worked. thx!
idle owl
@slender iron My example code for the INA219 doesn't pyllint, but looking, neither does yours for the motor_featherwing. Are examples exempt from pylint or something?
slender iron
the travis.yml needs to be adjusted. its not set up for subfolders
idle owl
so makes more sense to make sure the examples pass pylint now for less updating later?
robust coral
QQ: #circuitpython-dev is great. Is there a channel in a similar vein for discussing Arduino/C++ interfaces?
tidal kiln
@robust coral nothing as specific, best is #help-with-projects
robust coral
Ok. That was my guess as well. Thanks.
prime grove
@robust coral we just added a Arduino channel on the Sudomod server and the discussion includes C/C++ stuff
do you have a specific question?
robust coral
I was just looking to bounce some ideas about the Adafruit_SSD1306 driver.
robust coral
I've made a lot of private improvements and I want to see what it would take to upstream some of it.
prime grove
I ported it to Arduino, but haven't shared that code yet. What kind of improvements did you make
?
I
tidal kiln
@robust coral submit a PR to the repo
prime grove
I'm also working on a remote API to use the Arduino as an I/O slave
today, my itsy bitsy (Atmega32u4) stopped being recognized by my computer. Another one I have works fine, so it's not a driver issue
robust coral
A PR would be after the work is done. I wouldn't want to waste reviewer time if the direction isn't where Adafruit wants to go with it, so I was hoping to get a sense of what direction they're looking for. I'll just submit a PR and see what they say.
Summary: Eliminate the #define shenanigans to support different sizes. Add a small Python script to generate Adafruit logo bitmaps properly scaled to various sizes. Support for multiple displays. Rename variables to avoid macro conflicts (e.g. CS -> cs_pin). Concentrate #ifdefs with output pin abstractions. Fix dim() to emit the right size-dependent electrical current values. displayOn/displayOff. Like I said, lots of stuff. I'd have to break it down in a few PRs I think.
tidal kiln
@robust coral let's continue in #general-tech
robust coral
ok
tidal kiln
which is more pythonic? or just proper i guess
if value != 0 and value != 1:
print("must be 0 or 1")
or
if not (value == 0 or value == 1):
print("must be 0 or 1")
IMO it is more about what's important to communicate.
regardless of language
To me, personally, the first is usually easier to understand
But the details of the code and context effect that.
robust coral
Yeah, looks 50/50 to me. I like the second in this case only because what you print() matches the language of the conditional. That is, the phrasing with the word "or".
tulip sleet
if not value in (0, 1)
that's pretty pythonic
if value not in (0,1) is even better
robust coral
I wonder if there's a slight runtime cost to making the (0,1) tuple and iterating over it vs just doing the == directly.
Or if Python is smart enough to optimize it out.
tidal kiln
i like that second one @tulip sleet
tulip sleet
halbert@salmonx:~/bin$ python3 -m timeit 'value = 1; value not in (0,1)'
10000000 loops, best of 3: 0.0363 usec per loop
halbert@salmonx:~/bin$ python3 -m timeit 'value = 1; not (value == 0 or value == 1)'
10000000 loops, best of 3: 0.052 usec per loop
that's not in CircuitPython, so I don't know how it would do, but one could do a similar test
robust coral
Cool. "-m timeit" just changed my life.
slender iron
welp, I hosed my usb bus ๐
fringe trench
yikes!
slender iron
lost my stream audio
tulip sleet
did you fry it, or is this a software problem?
slender iron
lol
idle owl
Ok I actually laughed out loud.
raven canopy
@opal elk careful, you might set off a pun-battle with Andon. ๐
tulip sleet
took me a while to get that joke
opal elk
I mean I was a little offended he had the title but I really love my purple color so I won't complain
I only contribute to open source for the discord flair 
I think I have to write an example for a library to make up for the sins of that pun though
raven canopy
lol
upbeat plover
how do I wire this to be controlled by a trinket? https://www.adafruit.com/product/3191
tulip sleet
@upbeat plover just pick a pin on the trinket and wire that pin to a pin on the relay wing. Solder the on-board jumper for the pin on the relay wing. Don't forget a ground to both as well
upbeat plover
Before I purchase can a trinket handle the relay, Si7021 ,RFM69HCW, and a Magnetic contact switch ?
tulip sleet
@upbeat plover you mean trinket m0?
upbeat plover
I have a project in mind, and want to do it all in circuitpython.... Control the trinket with my CPX using 2 RFM69 breakouts.... Want the relay to turn on and off Humidifier to keep plant turgor pressure good....
yes trinket M0
tulip sleet
we have no libraries yet for RFM69, but they are being worked on
solar whale
@upbeat plover I think you will run out of pins for an I2C sensor and the SPI rfm69 and a switch on a trinket.
tulip sleet
@upbeat plover easier to use feather m0 and stack all that stuff
or use a feather doubler or tripler
solar whale
I would go with a feather M0 express so you have the UF2 Bootloader as well.
upbeat plover
does Feather M0 RFM69HCW work with the sample ?https://github.com/adafruit/Adafruit_CircuitPython_RFM69
solar whale
It should. But you wonโt have the UF2 Bootloader.
That library is very new and under development. It only supports some of the rfm69 modes at this time but should be useable. I have not tried it yet. Hope to soon.
tulip sleet
you may also need the extra code space available in the Express version (2MB instead of 64kB)
upbeat plover
Okay, thanks for advise @solar whale & @tulip sleet
solar whale
Youโre welcome. Good luck with your project.
raven canopy
taking the lazy route: do we have frozenset available? might need something dictionary-like...but ordered.
required less effort than i expected py/objset says we do in fact have frozenset! yay!
raven canopy
VEML6070 driver done (I think). Now to tackle the tough part: {cookiecutter} and linting on a windows machine...oh, and non-WIP documentation.
stuck elbow
@slender iron I see you are working on the i2c/spi registers stuff โ some time ago I started work on an alternate approach to the whole thing: https://github.com/deshipu/circuitpython-bus_registers
@slender iron maybe it will give you some ideas
it uses a shared buffer for all the registers
manic glacierBOT
tidal kiln
@idle owl you've been working with joy featherwing?
timber lion
@upbeat plover probably not yet, the M0 feather needs a special circuitpython build to expose the internal CS and RST pins
later this week when the guide is published it will probably be available
and yeah a trinket + SPI + I2C is going to be really hard to do pin-wise
the RFM69 library isn't too bad on memory usage but be prepared just in case you can't fit both the Si sensor and RFM in memory
you might need to rethink the project or switch some parts to arduino
and just FYI @tulip sleet the RFM69 library is done and ready for folks to use in case people ask
solar whale
@timber lion looking at the pinouts for the M0 RFM69 aren;t the CS and RST availble on 11 and 13? https://cdn-learn.adafruit.com/assets/assets/000/046/255/original/feather_Feather_M0_RFM95_v1.2-1.png?1504886606
timber lion
i don't see that here: https://cdn-learn.adafruit.com/assets/assets/000/032/914/original/feather_schem.png?1465421956
it's 3, 4, 8
not sure what those black boxes are
solar whale
sorry - my confusion - you are correct 3,4,8 - the black boxes are Physical Pin names for the processor chips
timber lion
ah no worries, yet i'm still kinda confused by the fancy graphic diagrams.. a lot of info there ๐
er yeah i'm still...
i always go straight for the schematic these days
solar whale
good advice! Sorry for the misleading advice @upbeat plover
upbeat plover
could you use a wire from the radio over to the feather?
solar whale
@upbeat plover - not sure what you mean. It is not a problem for the breakout board, only for the integrated unit where the radio is on the board with the processor. Building a new version of Circuitpython that allows access is not a big deal. Just has to be done to use that particualar board.
If you go withe the M0 express and the breakout board, you are all set.
upbeat plover
I won't be getting the parts for another week. I want to order a bunch of stuff at once.
solar whale
But as @timber lion pointed out, you still may run into problems with memory availabilty.
I hope to run some tests with the RFM69 and a temp/humidity sensor this weekend. I am using the sht31. I'll let you know how it goes.
@upbeat plover oops - just realized there is no CP library for the sht31 - gues I'll try the si7021. ๐
slender iron
@timber mango do you have examples about how the drivers would change?
tidal kiln
@slender iron when you get a chance:
https://github.com/adafruit/circuitpython/issues/535
some first pass code to take a look at.
slender iron
kk, will look
solar whale
Is anyone working on a CP driver for the SHT31-D temp/humidity sensor? https://www.adafruit.com/product/2857
slender iron
@solar whale I feel like I saw one go by
but maybe not
@tidal kiln looks really good overall. the only thing you could change is lazily creating the pin objects when people read the property
solar whale
guess I have a challenge ๐
It was dicsuused a long time ago and not recommended as a "first try". I'll do some digging.
tidal kiln
@solar whale not seeing an issue for it, nor an entry in the spreadsheet
solar whale
@tidal kiln thanks - guess I'll add one after I look into it a bit.
timber lion
awesome yeah let me know if you run into any RFM69 trouble.. and if you make a fun script we could throw it in the examples to show people more usage too
i have a feeling sensor reading and sending is something a lot of folks would find useful there
oh interesting, FYI for pycharm users: https://blog.jetbrains.com/pycharm/2018/01/micropython-plugin-for-pycharm/
could be nice to use its repl view and code editing at least
the upload i bet uses the pyboard stuff
solar whale
@tidal kiln where is the "speadsheet"?
tidal kiln
@slender iron would that be like setting them to None in __init__ and then having a conditional in the @property function to create them if they are None
@solar whale linked from this issue:
https://github.com/adafruit/circuitpython/issues/269
solar whale
thanks!
upbeat plover
The other day I said "plant turgor pressure", but I really mean is "vapour pressure deficit". How do you write the equation in python?
tidal kiln
i assume it's still being used?
slender iron
@tidal kiln yeah, exactly
tidal kiln
@upbeat plover what do you mean by that? 'write the equation'?
tidal kiln
try exp() from math
@upbeat plover https://docs.python.org/3/library/math.html#math.exp
upbeat plover
thank you
tidal kiln
math.log is natural log (ln):
https://docs.python.org/3/library/math.html#math.log
solar whale
sorry - my mistake -- having a bad day .... yes math.log
solar whale
@slender iron so you can access any pin like this from microcontroller.pin import PA08 pin=digitalio.DigitalInOut(PA08)
solar whale
cool!
idle owl
@tidal kiln Yes, re joy featherwing
stuck elbow
@slender iron sorry, I don't have an example, I will add it
slender iron
np, just trying to understand the differences
<@&356864093652516868> Here is the recording of the Weekly meeting from this week: https://youtu.be/a34DH1FSqJc
stuck elbow
stuck elbow
the main difference is that the registers are handled all together, at the level of the whole driver, not at the level of individual registers, so they can use a common shared buffer and so on
I also merged the busdevice and register together, to remove overhead
I tried to also come up with a way to not have to have a separate object for every register, but couldn't think about anything that didn't seem artificial
@slender iron I also wonder if it would make sense for busdevice to have a buffer, so that all drivers on the bus could share it
and methods like i2c.writeto_struct(device, data_format, data)
then the registers don't need a buffer at all
slender iron
yeah, thats a possibility
the register data descriptors can always add a share buffer onto the object
have you measured a memory or speed increase with this approach? it seems to me like its a lot of the same code organized in a different way
stuck elbow
I didn't, I just started rearranging things to see if I can come up with something more efficient
and yes, it's the same thing โ it's supposed to do the same thing, after all
slender iron
๐ experimentation is good. did you see my latest heap stuff?
stuck elbow
I only saw that you started to come up with ideas, but didn't look at any code
stuck elbow
so the top row is root pointers?
stuck elbow
are you using dot?
stuck elbow
try javascript and d3
slender iron
yeah, I thought about that
I'm trying to not to incur a bunch more work
the layout stuff is because I'm experimenting with allocating "long lived" objects at the top of the heap
stuck elbow
I love graphviz, but it has its limitations
slender iron
yeah, having fixed positions is being finnicky
timber lion
is something broken with 2.x @slender iron @tulip sleet ? I'm doing a 'git submodule update' in 2.x and getting: "fatal: no submodule mapping found in .gitmodules for path 'frozen/Adafruit_CircuitPython_BusDevice'"
something is very strange, i did a git pull origin 2.x and it's really mad
i'm just going to have to delete and start fresh
for some reason it's changing a ton of stuff that i haven't modified
golly git stash even fails
i've never seen that
atmel-samd/boards/circuitplayground_express/mpconfigboard.h: unmerged (f9ff05084a26653dc97a6d6300453572a81de331)
...
fatal: git-write-tree: error building trees
Cannot save the current index state
that seems like quite a nasty bug in git
i wonder if all of this is completely lost now
ah esoteric git commands to the rescue ๐ https://stackoverflow.com/questions/6006737/git-merge-errors
stuck elbow
perhaps it's time for a fresh clone
timber lion
hrm i can't get a fresh clone to build either
i've done a gitsubmodule init
but fail to build any board
tony-imac:atmel-samd tony$ make BOARD=feather_m0_basic
Use make V=1, make V=2 or set BUILD_VERBOSE similarly in your environment to increase build verbosity.
make: *** No rule to make target freetouch/adafruit_ptc.c', needed by build-feather_m0_basic/genhdr/qstr.i.last'. Stop.
are we missing submodules somewhere?
@slender iron anything up with 2.x branch? fresh clone and git submodule init after is failing to build the existing boards
tony-imac:circuitpython tony$ git submodule init
Submodule 'freetouch2' (https://github.com/adafruit/Adafruit_FreeTouch.git) registered for path 'atmel-samd/freetouch'
Submodule 'frozen/Adafruit_CircuitPython_BusDevice' (https://github.com/adafruit/Adafruit_CircuitPython_BusDevice.git) registered for path 'frozen/Adafruit_CircuitPython_BusDevice'
Submodule 'frozen/Adafruit_CircuitPython_LIS3DH' (https://github.com/adafruit/Adafruit_CircuitPython_LIS3DH.git) registered for path 'frozen/Adafruit_CircuitPython_LIS3DH'
Submodule 'atmel-samd/frozen/Adafruit_CircuitPython_NeoPixel' (https://github.com/adafruit/Adafruit_CircuitPython_NeoPixel) registered for path 'frozen/Adafruit_CircuitPython_NeoPixel'
Submodule 'frozen/Adafruit_CircuitPython_Thermistor' (https://github.com/adafruit/Adafruit_CircuitPython_Thermistor.git) registered for path 'frozen/Adafruit_CircuitPython_Thermistor'
Submodule 'lib/axtls' (https://github.com/pfalcon/axtls) registered for path 'lib/axtls'
Submodule 'lib/berkeley-db-1.xx' (https://github.com/pfalcon/berkeley-db-1.xx) registered for path 'lib/berkeley-db-1.xx'
Submodule 'lib/libffi' (https://github.com/atgreen/libffi) registered for path 'lib/libffi'
Submodule 'lib/lwip' (http://git.savannah.gnu.org/r/lwip.git) registered for path 'lib/lwip'
Submodule 'tools/python-semver' (https://github.com/k-bx/python-semver.git) registered for path 'tools/python-semver'
Submodule 'tools/uf2' (https://github.com/Microsoft/uf2.git) registered for path 'tools/uf2'
tony-imac:circuitpython tony$ cd atmel-samd
tony-imac:atmel-samd tony$ make BOARD=feather_m0_basic
Use make V=1, make V=2 or set BUILD_VERBOSE similarly in your environment to increase build verbosity.
make: *** No rule to make target freetouch/adafruit_ptc.c', needed by build-feather_m0_basic/genhdr/qstr.i.last'. Stop.
tony-imac:atmel-samd tony$
slender iron
@timber lion sorry, I don't know the answer, thats why I didn't reply
its green on travis: https://travis-ci.org/adafruit/circuitpython/builds/327788515
timber lion
yeah i don't see a makefile in this submodule: https://github.com/adafruit/Adafruit_FreeTouch
unsure how is it supposed to build?
is there some state folks have that isn't checked in?
slender iron
travis wouldn't have it
timber lion
ah ok, is there a step i'm missing?
just following the standard build steps we've had
timber lion
ah so i think the problem is you need to git submodule init --recursive
looking at what travis does differently
that pulls in a ton more files
slender iron
ah, I see
timber lion
yeah that seems to be the issue
slender iron
great!
timber lion
would be awesome to keep the readme up to date with that, we don't mention it here: https://github.com/adafruit/circuitpython/tree/master/ports/atmel-samd
slender iron
please update it ๐
timber lion
i'd suggest the creator of the travis config update it
since that seems to be the canonical source
i can guess but will get it wrong ๐
slender iron
travis does the submodule stuff automatically
timber lion
something is telling it to do this:
2.26s$ git clone --depth=50 --branch=2.x https://github.com/adafruit/circuitpython.git adafruit/circuitpython
Cloning into 'adafruit/circuitpython'...
$ cd adafruit/circuitpython
$ git checkout -qf 8251abba8a39991c4b3475161c1565f4fa8d105b
git.submodule
10.21s$ git submodule update --init --recursive
slender iron
ยฏ_(ใ)_/ยฏ
timber lion
wow i cannot for the life of me get github to push this change up
tony-imac:circuitpython tony$ git push -u origin missing_feathers
error: src refspec missing_feathers does not match any.
i'm in a local 'missing_feathers' branch, it's all committed and ready
it refuses to let me push to a remote missing_feathers branch
everything looks fine branch-wise
tony-imac:circuitpython tony$ git branch -vv
2.x 8251abb [origin/2.x] Merge pull request #520 from pewpew-game/2.x
- missed_feathers 23cad67 Add missing Feather M0 boards: RFM69, RFM9x, nRF51, WINC1500.
what's up?
it's tracking origin/2,x
slender iron
your local branch is "missed_feathers"
timber lion
yes
slender iron
and your push is "missing"
slender iron
๐
timber lion
oh boy now we enter the world of pain figuring out how to send the pull against 2.x and not master ๐
slender iron
world of pain = drop down menu
timber lion
yeah looks like they made it a lot easier than last i did it
slender iron
๐
idle owl
@tidal kiln Finishing up a lib for the joy featherwing right now
tidal kiln
@idle owl cool. just started using it this morning. saw ur previous comments. so was wondering if one was in works.
idle owl
yeah. The bulk of it is done. I want to add a calibration option, so I'm trying to figure out how to do that. Otherwise needs docs and it'll be set.
Mine reads -4, 0 when it's centered.
tidal kiln
also, saw that issue in the main seesaw lib, was wondering if there was any relation
idle owl
I don't know about that issue
tidal kiln
to make the api match cp
idle owl
Ah. In this case we're building a featherwing library. Where you can import the featherwing and don't have to deal with setup.
tidal kiln
yep. nvm then. basically just looking for a next thing to work on.
tidal kiln
and sounds like joywing is covered ๐
idle owl
At least from this angle anyway! I'll send you a link once I post it so you can test it. That would be super helpful
tidal kiln
๐
upbeat plover
# temp in C
def vpd(temp, rh):
# Saturation Vapor Pressure
es = 0.6108 * math.exp(17.27 * temp / (temp + 237.3))
# Actual Vapor Pressure
ea = rh / 100 * es
# Vapor Pressure Deficit
vpd = ea - es
return vpd
I'm having troubles with VPD calculations, I don't know how to do log stuff. Equations from
http://bioma.jrc.ec.europa.eu/components/componentstools/evapotranspiration/help/Actual_vapor_pressure.html
Can someone else do it for me? ๐คง
simple quarry
so, im trying to create something that can access the internet from the computer, and read something back to the device.... How would i do this?
thanks in advance!
slender iron
@simple quarry to use it in conjunction with your computer you'll need code running on it as well. It might be easier to use a Feather Huzzah ESP8266 to connect directly to the internet
simple quarry
thats fine to be able to have code running on the computer, how would i do that?
slender iron
to use python you'll want to look into pyserial as a way to communicate with circuitpython
unfortunately, I don't know of any guides that do this
simple quarry
huh
solar whale
@timber lion good start - I have an RFM69 featherwing on a feather m0 Express using your rfm69 driver and the example - and it is communicating with my "base" station running an Arduino sketch.
@upbeat plover not sure what you are looking for use log just like exp ```
import math
math.pow(2.,4.)
math.exp(2.)
math.log(2.)
tidal kiln
@upbeat plover your code looks OK to me
es = 0.6108 * math.exp(17.27 * temp / (temp + 237.3))
ea = rh / 100 * es
and i don't see any log functions
upbeat plover
how do you do these?
raven canopy
In today's episode of "Windows Strikes Again..." ๐
Python 3.6.2 (v3.6.2:5fd33b5, Jul 8 2017, 04:57:36) [MSC v.1900 64 bit (AMD64)] on win32
Type "help", "copyright", "credits" or "license" for more information.
>>> cookiecutter gh:adafruit/cookiecutter-adafruit-circuitpython
File "<stdin>", line 1
cookiecutter gh:adafruit/cookiecutter-adafruit-circuitpython
^
SyntaxError: invalid syntax
@raven canopy lol ๐
raven canopy
the VM doesn't like cookiecutter either...hrm
oh man. i gotta make dinner! later taters
idle owl
@raven canopy have a good one!
slender iron
@tulip sleet I think I found a speed up: https://github.com/adafruit/circuitpython/blob/master/py/persistentcode.c#L127
str can be allocated on the stack
upbeat plover
best VPD I can get
# temp in C
def vpd(temp, rh):
# Saturation Vapor Pressure method 1
es1 = 0.6108 * math.exp(17.27 * temp / (temp + 237.3))
# Saturation Vapor Pressure method 2
es2 = 6.11 * 10**((7.5*temp)/(237.3+temp)) / 10
# Saturation Vapor Pressure mean
es = (es1 + es2) / 2
# Actual Vapor Pressure
ea = rh / 100 * es
# Vapour Pressure Deficit
vpd = es - ea
return vpd
@slender iron ! is the qstr obj ever possibly returned from the routine that calls load_qstr()?
slender iron
tulip sleet
@slender iron that's great! was that a module load? Looks like you already did the temp/perm partition. What was all the leftover stuff on the left at the end?
slender iron
@tulip sleet its loading the motor featherwing code
and doing one step
code.py is four lines or something ๐
this video only has qstr stuff being "long lived"
tulip sleet
@slender iron should I approve this https://github.com/adafruit/circuitpython/pull/529 or is it still very much in process? (obviously you are still working on the display stuff).
slender iron
@tulip sleet yeah sure
solar whale
@timber lion Success! I now have my rfm69 featherwing transmitting temperature and humidity from and si7021 either on receipt of a packet from the "base" station or at a fixed time interval. This is using the new adafruit_rfm69 driver with CP2.2 on a feather m0 express. The base station (arduino) is an rfm69 featherwing on a feather m0 WINC1500 that then relays the data to my adafuit.io page. I can even request a reading from the adafruit.io Dashboard! It needs some clean up, but it is working great! Thank you for getting the dirver out there.
solar whale
@simple quarry try restating your question - it's not quite clear what you are trying to do.
What type of computer aare you using and what have you tried so far. How are you communicating with the CP Express now?
raven canopy
cookiecutter is going to fight me just as hard as everything else has. little does it know, i'm an aires... head down, ramming speed!
simple quarry
how do you print something in serial, and then read it back in python on the pc
raven canopy
@simple quarry by "print in serial", i'm assuming you mean having a microcontroller send data through the serial port, and then reading that data on the PC, yes? someone mentioned pySerial iirc. which if you're locked into using Python on the PC, would be what you're looking for.
http://pyserial.readthedocs.io/en/latest/pyserial.html#
timber mango
(they didn't have a spot-on emoji, so I chose 'on' here)
raven canopy
take that, cookiecutter! I win! i may be too excited about that
timber lion
oh rad super cool @solar whale !
stuck elbow
what happened to that blinka logo in that tutorial?
half her stomach scales are missing
upbeat plover
okay, so I found another Saturation Vapor Pressure equation, I think I like this one the best. I have no way of checking other then using online calculators and charts. Use plant canopy temperature for best results. The 3 methods of getting SVP are all nearly synonymous.
# temp in C
def vpd(temp, rh):
# Saturation Vapor Pressure
es = 6.112 * math.exp(17.62 * temp / (temp + 243.12)) / 10
# Actual Vapor Pressure
ea = rh / 100 * es
# Vapour Pressure Deficit
vpd = es - ea
return vpd
while True:
if vpd(10, 50) < 0.6:
#turn on humidifier
if vpd(10, 50) > 1.6:
#turn off humidifier
bronze geyser
I am attempting to try out the RFM60 CircuitPython library @tdcola. The challenge is the library is 33K, too big for the Feather M0.... I assume I need to get it to a .mpy (frozen library). The only directions say to use vagrant on my mac...I can do this but is it possible to post the .mpy on GitHub? Or am I to do something else? I can do the vagrant step...but wanted to ask before investing time. THanks.
OOps. I meant RFM69.
stuck elbow
the .mpy file doesn't have to be smaller
you can try just removing all the comments
Here is the current RFM69 library compiled for CircuitPython 2.2.0
bronze geyser
Thank you @beshpu
@beshipu and thank you about reminding I could just remove the comments! I forgot about that.
Ugh @beshipu still complaining there is not enough room. GetInfo shows 10Kb...when I try to copy the 10Kb rfm69 lib you kindly provided, it says it needs 6Kb more... I'm going to Yoga...will figure something out when I get back.
stuck elbow
make sure your system didn't put any hidden files in there
fading solstice
@stuck elbow who is responsible for adding a release to a bundle library? The review (merging) or the coder?
stuck elbow
@fading solstice no idea, really
fading solstice
the ds1307 lib last got release 11/2016 but has important changes made in the last month. no release
i.e. do not work
split ocean
hey @slender iron or anyone else w IR experience, got any time to help me figure this thing out?
Attempting to use adafruit_irremote to decode and then transmit, but getting some odd results.
slender iron
@fading solstice feel free to release as long as you've tested it
@split ocean what are you trying to do?
split ocean
decode a button press from the little Adafruit IR remote on a Feather M0 Express, and then code a Gemma M0 to be able to send that same command.
so far I'm using this to decode:
slender iron
did you start with the circuitplayground code?
split ocean
>>> import array
>>> array.array('B', [223, 32, 239, 16]).tostring()
or whatever the decode array is for the given button
slender iron
can you post your full code somewhere? how is it not working?
split ocean
will do. the way it is not working is that the serial output of the reader isn't the same between the actual remote button press and the Gemma transmit.
slender iron
ah, so its transmitting different things?
split ocean
yes.
slender iron
kk
split ocean
this is the reader code: https://gist.github.com/anonymous/7e64e42545030d9cf10095c0f75767e8
this is the sender code:
timber lion
oh check your byte string
ONE=bytearray(b'xff\x02\xcf0')
did you mean: b'\xff\x02\xcf\x00' ?
i.e. 0xFF, 0x02, 0xCF, 0x00
because without that leading slash it's actually: ascii ordinal value for x, ascii ordinal value for f, ascii ordinal value for f, 0x02, 0xCF, ascii ordinal value for 0
if it's more clear you can do:
ONE=bytearray([0xFF, 0x02, 0xCF, 0x00]) etc
split ocean
oops, I definitly messed that up
timber lion
yeah byte strings are tricky, i'm starting to realize they're a bad idea to use in practice
if you miss that leading b it silently fails too
generates a completely different string value
burned me once
and easy to miss the \x slash.. i think best to be explicit with arrays and numbers
split ocean
ok, i wasn't signed into git when I made those gists, so here they are again, but ediatable this time: https://gist.github.com/jedgarpark/7920289fd1d19399daa908426d11f258
I've fixed the bytestring in the transmitter code, but still don't get correct read from the gemma transmit
failed to decode: ('Pulses do not differ',)
slender iron
@tulip sleet @timber lion should I switch my brain to merge only for github?
I like being able to squash things but its constantly confusing people
split ocean
this is my output from the remote vs. that of the gemma:
Heard 3 Pulses: [9122, 2241, 569]
NEC repeat!
Heard 67 Pulses: [9144, 4504, 573, 568, 567, 580, 565, 599, 546, 568, 567, 602, 543, 598, 547, 595, 542, 599, 544, 1679, 621, 1630, 570, 1679, 571, 1680, 570, 1680, 570, 1685, 565, 600, 545, 1678, 573, 569, 566, 602, 543, 572, 572, 596, 540, 1683, 567, 574, 571, 571, 574, 570, 565, 1682, 568, 1683, 567, 1683, 567, 1686, 565, 574, 570, 1680, 576, 1675, 570, 1680, 570]
Decoded: [255, 2, 247, 8]
Heard 67 Pulses: [9179, 4477, 563, 568, 597, 535, 590, 568, 567, 564, 591, 541, 594, 537, 587, 572, 583, 547, 588, 574, 561, 566, 588, 570, 595, 562, 563, 569, 595, 562, 593, 539, 596, 561, 594, 538, 587, 544, 591, 567, 587, 544, 591, 567, 568, 563, 592, 540, 595, 563, 562, 569, 595, 563, 592, 570, 565, 563, 592, 565, 590, 543, 592, 565, 589, 568, 567]
Failed to decode: ('Pulses do not differ',)```@split ocean where did you come up with the values for the generictransmit init?
split ocean
@candid sun had some notes on that here: https://www.hackster.io/BlitzCityDIY/circuit-python-ir-remote-for-apple-tv-e97ea0
split ocean
she said she trial and errored it, so I'd love to know more about how that should be determined.
is there a protocol we can reference for different remote types?
slender iron
ok, I think you can get values for it from the decoder
try passing debug=True into decode_bits
there is a print for the bins that should be what you want
upbeat plover
@split ocean if you get it working can I see an example? I ended up just sending the whole array... and what your doing is so much better
slender iron
@split ocean these numbers are too close (570, 570), (570, 580)
split ocean
w debug on for my real remote i get this output:
slender iron
my guess is that ones are (570, 1600) or so
split ocean
length: 3
NEC repeat!
Heard 67 Pulses: [9123, 4510, 568, 572, 573, 568, 567, 574, 571, 570, 575, 566, 570, 572, 572, 569, 566, 576, 572, 1678, 572, 1678, 569, 1682, 568, 1682, 569, 1681, 569, 1681, 572, 596, 545, 1681, 570, 595, 540, 574, 571, 569, 566, 575, 570, 1680, 570, 599, 546, 569, 566, 575, 570, 1680, 570, 1681, 570, 1683, 567, 1681, 569, 600, 550, 1673, 572, 1679, 571, 1680, 570]
length: 67
new length: 66
evenbins: [[4510, 1], [588, 17], [1679, 15]] oddbins: [[569, 33]]
evenbins: [[588, 17], [1679, 15]] oddbins: [[569, 33]] outliers: [4510]
Pulses: [4510, 572, 568, 574, 570, 566, 572, 569, 576, 1678, 1678, 1682, 1682, 1681, 1681, 596, 1681, 595, 574, 569, 575, 1680, 599, 569, 575, 1680, 1681, 1683, 1681, 600, 1673, 1679, 1680] & Bins: [[588, 17], [1679, 15]]
Space: 1679 Mark: 588
32 [True, True, True, True, True, True, True, True, False, False, False, False, False, False, True, False, True, True, True, True, False, True, True, True, False, False, False, False, True, False, False, False]
Decoded: [255, 2, 247, 8]```i don't know what in that is meant to go into my GenericTransmit line
slender iron
so looking at this: evenbins: [[588, 17], [1679, 15]] oddbins: [[569, 33]] outliers: [4510]
upbeat plover
one of the even and odd are almost the same
slender iron
right, so bits are sent as a high pulse and a low pulse
and Generic transmit take a tuple of the high time and the low time
- note my highs and lows may be backwards
the low time is always the same, hence oddbins are centered around 569
(33 is the count of pulses in that "bin")
so that means you're params would be (one, 569), (zero, 569)
where one and zero are different
looking at the even bins there are two values 588 and 1679
my guess is one is longer
so it'd be (1679, 569), (588, 569)
timber lion
@split ocean this might be a good guide to check out for reference: https://learn.adafruit.com/ir-sensor/circuitpython i go through some basics of capturing a remote press
from what your gemma is sending it looks like it's a bunch of zeros
all those 500 ms changes
at least IIRC those are zero
not super familiar with the generic IR transmit code, i haven't used that one but let me take a quick look at what it does
i wonder if there's a diff in what it expects to be told vs. what you have captured
split ocean
@slender iron i'm trying to understand this, but coming up short so far. is the adafruit_irremote.GenericTransmit() looking for two even bins pairs, and odd pair and an outlier? I don't get how to derrive those values from the debug, despite your efforts to help me :)
timber lion
oh yeah
so you're initializing it like this:
remote = adafruit_irremote.GenericTransmit((9135, 4505), (570, 570), (570, 580), 565)
that's telling it the header is (9135, 4505)
split ocean
is this it?
remote = adafruit_irremote.GenericTransmit((1679, 15), (588, 17), (569, 33), 4510)
timber lion
that's the start of a standard NEC pulse
9ms high, 4.5 low
that's all good
but then the transmit code is told a one and a zero are the same
its initializer expects the next two tuples to be the time for a zero and one
so you're saying
one bit = (588, 17)
zero bit = (569, 33)
they differ a teeny bit with that 17 vs 33
but that's super quick and i think too fast
most remotes are a diff of like 500 and 1000 IIRC
see in your capture
Heard 3 Pulses: [9128, 2236, 572]
length: 3
NEC repeat!
Heard 67 Pulses: [9123, 4510, 568, 572, 573, 568, 567, 574, 571, 570, 575, 566, 570, 572, 572, 569, 566, 576, 572, 1678, 572, 1678, 569, 1682, 568, 1682, 569, 1681, 569, 1681, 572, 596, 545, 1681, 570, 595, 540, 574, 571, 569, 566, 575, 570, 1680, 570, 599, 546, 569, 566, 575, 570, 1680, 570, 1681, 570, 1683, 567, 1681, 569, 600, 550, 1673, 572, 1679, 571, 1680, 570]
you can see pairs like 1680, 570
that 1600 500 is different from a 500 500
and the diff between a one and zero
so basically your transmittier is being told one bits and zero bits should send out the effectively the same pulse lengths
and it all looks the same
i think you want something a little different there, like
remote = adafruit_irremote.GenericTransmit((1679, 15), (588, 17), (1600, 33), 4510)
but i can't remember exactly what the remote codes want
try changing that and see what you capture on the board from gemma sending it
it shouldn't be all 500s now
i might have it swapped too, a one might actually be 1600 first
so
remote = adafruit_irremote.GenericTransmit((1679, 15), (1600, 17), (569, 33), 4510)
oh wait be careful
i just copied your line you pasted
but you want that first parameter to be what you originally had
that's the header
it needs to be 9000, 4500 for most devices
remote = adafruit_irremote.GenericTransmit((9135, 4505), (1600, 570), (570, 580), 565)
split ocean
?
split ocean
k
timber lion
but the issue is the same if you look at the transmit code in the gist, a one and zero are basically the same right now
and explains why you only see 500's coming from the transmit.. it blinks the LED the same rate for both a one and zero because of this in the transmit gist code: remote = adafruit_irremote.GenericTransmit((9135, 4505), (570, 570), (570, 580), 565)
split ocean
that's closer for sure
timber lion
(570, 570) and (570, 580).. they're both effectively the same
split ocean
the gemma transmit is now understood as being sorta properly formed, but not decoded correctly just yet
timber lion
this is good background too i think mention in the guide: http://techdocs.altium.com/display/FPGA/NEC+Infrared+Transmission+Protocol
Logical '0' โ a 562.5ยตs pulse burst followed by a 562.5ยตs space, with a total transmit time of 1.125ms
Logical '1' โ a 562.5ยตs pulse burst followed by a 1.6875ms space, with a total transmit time of 2.25ms
so that's what those 500 and 1600 numbers mean
split ocean
so this is what the gemma output looks like now:
timber lion
so this code in the gist is saying a logical one is a 570, 570 pulse and a logical zero is a 570, 580 pulse (the order of the initializer is one bit first, then zero bit it seems): https://gist.github.com/jedgarpark/7920289fd1d19399daa908426d11f258
split ocean
length: 67
new length: 66
evenbins: [[4490, 1], [571, 32]] oddbins: [[1612, 17], [579, 16]]
evenbins: [[571, 32]] oddbins: [[1612, 17], [579, 16]] outliers: [4490]
Pulses: [1613, 1646, 1617, 1646, 1617, 1621, 1615, 1595, 582, 586, 594, 591, 594, 595, 1614, 593, 1614, 1618, 1615, 1618, 588, 1599, 1612, 1616, 588, 565, 589, 585, 1612, 586, 564, 586, 579] & Bins: [[1612, 17], [579, 16]]
Space: 1612 Mark: 579
33 [False, False, False, False, False, False, False, False, True, True, True, True, True, True, False, True, False, False, False, False, True, False, False, False, True, True, True, True, False, True, True, True, True]
Decoded: [0, 253, 8, 247, 1]```but if you check the protocol you want a logical one as 570, 1600 and a logical zero as 570, 570
cool yeah do you see how you want to change the tuple values to change those timings to be like the NEC protocol?
tidal kiln
@slender iron lazy'd, examples, and some bug fixes:
https://github.com/caternuson/Adafruit_CircuitPython_DS2413
slender iron
@split ocean the second number in the bin (like 17 and 33) is a count and you don't want to use it
upbeat plover
GenericTransmit((a, b), (e, f), (c, d), z)
#where the values that appeared in the serial monitor were:
[a, b, c, d, e, f...z]
from that apple remote thing
hmm i dont see where you would put the 1600 though >.<
timber lion
if it helps write out the values in the parenthesis too
a zero should be about 570 ms of pulses followed by 570 ms of nothing
a one should be 570 ms of pulses followed by 1600 ms of nothing
the tuples tell it (time in ms for pulses, time in ms for nothing)
split ocean
so, what is each space in the set of values meant to represent in GenericTransmit, then? Sorry, I'm trying to piece this all together :)
upbeat plover
GenericTransmit((a, b), (e, f), (c, d), z)
(a, b) = timing
timber lion
ah yeah so that's not super clear unfortunately unless you look at thecode
but these are the tuples
so header, one, zero, tail
the tail/trail is just one value, how long to hold low i assume
split ocean
I'm going to try to correlate the remote blast to what I need in my GenericTransmit line: so header here is 9146 because that's the first thing read when I hit the real remote button
timber lion
yep check out the NEC protocol doc
pretty much all remotes start with a header of 9ms on 4.5ms off
split ocean
4490 is the highest value of the even bins, so that's the next one
timber lion
so everything will be about (9000, 4500) for the header
split ocean
(rounded)
timber lion
yeah, i wouldn't get too wrapped up in decoding exact values
there's jitter and timing diffs
split ocean
yep
timber lion
as long as its near 9000, 4500 you're seeing the start of a remote code for sure
split ocean
then the value of a one
which is the higher number in the odd bin pairs?
so 1610?
and a zero which is 570?
upbeat plover
1615
timber lion
yeah check the protocol http://techdocs.altium.com/display/FPGA/NEC+Infrared+Transmission+Protocol
Logical '0' โ a 562.5ยตs pulse burst followed by a 562.5ยตs space, with a total transmit time of 1.125ms
Logical '1' โ a 562.5ยตs pulse burst followed by a 1.6875ms space, with a total transmit time of 2.25ms
oh they have a typo
it says 1.687ms but should b us
oh wait no
they're right ๐
ignore me, convert ms to us though
split ocean
then the zero which is 580
timber lion
1.6875 turns into, 1687.5
split ocean
and then the trail which is 580?
so adafruit_irremote.GenericTransmit((9145, 4490), (1615, 570), (590, 570), 580)
roughly?
timber lion
pull it apart to double check, it might help to write out each value
head = (9145, 4490)
head on = 9145
head off = 4490
one = (1615, 570)
head on = 1615
etc
then compare against the protocol
specifically:
Logical '0' โ a 562.5ยตs pulse burst followed by a 562.5ยตs space, with a total transmit time of 1.125ms
Logical '1' โ a 562.5ยตs pulse burst followed by a 1.6875ms space, with a total transmit time of 2.25ms
pulse burst = on in their nomenclature, and space = off
one
ack
aw geeze discord is really bad at editing
so i meant
split ocean
:)
timber lion
head = (9145, 4490)
head on = 9145
head off = 4490
one = (1615, 570)
one on = 1615
etc
i have no idea what discord just did above
split ocean
inception
head = (9145, 4490)
head on = 9145
head off = 4490
one = (1615, 570)
one on = 1615
one off = 570
zero = (580, 570)
zero on = 580
zero off = 570
trail = 580
does that look right based on this:
length: 67
new length: 66
evenbins: [[4510, 1], [572, 17], [1677, 15]] oddbins: [[572, 33]]
evenbins: [[572, 17], [1677, 15]] oddbins: [[572, 33]] outliers: [4510]
Pulses: [4510, 571, 571, 568, 574, 576, 576, 571, 596, 1676, 1677, 1679, 1681, 1686, 1684, 604, 1681, 600, 567, 576, 573, 1677, 596, 574, 570, 1678, 1678, 1679, 1680, 569, 1676, 1678, 1678] & Bins: [[572, 17], [1677, 15]]
Space: 1677 Mark: 572
32 [True, True, True, True, True, True, True, True, False, False, False, False, False, False, True, False, True, True, True, True, False, True, True, True, False, False, False, False, True, False, False, False]
Decoded: [255, 2, 247, 8]```(note, some values changed due to fuzziness after I wrote the list and then output from the ir remote)
timber lion
check against the protocol first
so how what you wrote out compares to
Logical '0' โ a 562.5ยตs pulse burst followed by a 562.5ยตs space, with a total transmit time of 1.125ms
Logical '1' โ a 562.5ยตs pulse burst followed by a 1.6875ms space, with a total transmit time of 2.25ms
ignore the numbers being measured for now, just get the configuration the same as the protocol
manic glacierBOT
upbeat plover
[True, True, True, True, True, True, True, True, False, False, False, False, False, False, True, False, True, True, True, True, False, True, True, True, False, False, False, False, True, False, False, False]
this part tells you what is a one or zero?
timber lion
i wouldn't get too into decoding the debug output
check the protocol config first, there's a small difference right now
i can point it out but am just trying to kinda direct towards it ๐
but closely compare the protocol timing for one and zero to what the config written out above has
look at times on and off
ignore all the debug output, etc. this is purely with this line: adafruit_irremote.GenericTransmit((9145, 4490), (1615, 570), (590, 570), 580) and the NEC protocol timing
lucid arch
hi all,
Can the circuitpython board do soft serial?
sorry, this is the wrong forum ..
split ocean
@timber lion I'm not sure I'm following. the protocol wants double the time I've listed for the space after the one?
timber lion
ahh, yeah that does look like a difference
you might try changing it to match
but it's not really just double, 570 vs 1600.. more like 3x
split ocean
huh, my attempts at that fail
split ocean
oh, wait, now i have it working. my zero and one were reversed.
yes, @upbeat plover we hit the solution at the same time :)
upbeat plover
i still dont see where you get the tail lol
split ocean
@slender iron you were right, the one and zero is reversed in the GenericTransmit
I think
it seems to be (header, zero, one, trail)
split ocean
yoinks!
timber lion
think of it this way, on and off is just a concept in software
but it can change depending on the hardware
if you hook up the output of a digital line to the cathod or anode of a LED it swaps how it works
split ocean
of course!
timber lion
if you hook it up to the negative side of the LED when the digital line is high the LED turns off
but if you change it so its hooked up to positive, now when the digital line is still high the LED is on
so you might actually be reversed in your hardware setup
i.e. how you hooked up the LED
and then to throw another wrench into it IIRC the IR pulse output also have a concept of inversion
so it can separately change its concept of on vs. off to swap them
so potentially two places where polarity might change
i haven't dug through all the code and setup but you might check and see
because that can drastically change the behavior and timing from what you see
timber mango
@upbeat plover I like that someone's working on plant life / agrarian stuff here.
re: > [10:55 UTC] hexthat: okay, so I found another Saturation Vapor Pressure equation
bronze geyser
Trying to debug RFM69 library. Used a .mpy ..ran simpletest.py RuntimeError raised (version != 0x24). Thought I'd try some debugging using the .py file...but get a MemoryError ... I tried deleting most (many?) of the comments, but still having a challenge. Admittedly, I'm this side of clueless. What is best way to debug RFM69 on Feather M0 RFM69HCW? (thank you).
manic glacierBOT
slender iron
@bronze geyser I think you are on the right track with trying to slim down the py file. Maybe @timber lion can provide insight on the error raised
tidal kiln
@bronze geyser also, you can try and mpy your modified file, grab a mpy-cross binary from here:
https://github.com/adafruit/circuitpython/releases
timber lion
@bronze geyser like the error mentions check your wiring
it's trying to validate it can read the ID register from the chip
it's the very first thing the library does to talk to the chip
so if that fails it's very likely power or wiring to it
if it's getting a different value from 0x24 then something is up, make sure you're using a RFM95HCW from adafruit
er RFM69HCW
simple quarry
im trying to get my circuit playground express to read something that i put into code on the computer, and i cant figure it out...
slender iron
@simple quarry you need the computer to tell the circuitplayground express something over serial
upbeat plover
Si1145 or VEML6070 library?
raven canopy
@upbeat plover i'm actually putting the finishing touches on a VEML6070 library. hope to have it done over the weekend.
upbeat plover
Im just looking for parts to put in my wishlist, no hurry. Any plans for the si1145?
slender iron
thanks for the good review @stuck elbow ! I'll get to it next week. brain is in the heap now ๐
stuck elbow
TGIF
slender iron
๐
stuck elbow
I on the other hand finally have some time to do something...
raven canopy
@upbeat plover i've haven't heard of anyone working on the SI1145, but that doesn't mean it isn't happening.
raven canopy
@slender iron is Sphinx/RTD a requirement for libraries? not trying to dodge it, just seeing some disparity among libraries...
slender iron
its been something we should maintain but have been bad about
raven canopy
ok. i'll jott down what i come across...
slender iron
we'll add checking to adabot when we want them to be consistent
simple quarry
yes, @slender iron
golden tartan
hi
slender iron
@simple quarry take a look at pyserial as a way to connect to the circuitpython serial connection
hi @golden tartan
slender iron
I've never done it myself either
simple quarry
oh, huh
slender iron
thats why I'm not being more helpful ๐
slender iron
@simple quarry let me know if you figure it out or have other circuitpython questions!
simple quarry
i will
last drum
oh hello!
slender iron
hi @last drum
solar whale
@bronze geyser what pins are you using for RST and CS for the rfm69 on your M0?
raven canopy
got a 7.26/10 on my first pylint... this close to passing the test! ๐