hmm, i can't force a build, not my project, or something
#circuitpython-dev
tulip sleet
idle owl
hmm
tidal kiln
@idle owl @raven canopy please take a look at:
https://github.com/adafruit/Adafruit_CircuitPython_HT16K33/pull/15
will fix things for now at least...
idle owl
@tidal kiln Done, I'm leaving it for sommersoft to take a look at as well.
tulip sleet
@idle owl I think rtd is churning the themes, not something we can fix for now
idle owl
@tulip sleet Ok good to know.
tidal kiln
@idle owl thanks!
haughty bobcat
Just got my Gemma M0 up and running
so much better than arduino so far
but Arduino does have more libraries
tidal kiln
print("yay!")
haughty bobcat
so is there a place where all the pinouts are listed or am I missing something?
so far I only know pin 13 is red led
I found a guide
tulip sleet
Are you looking at the Learn Guide? [we crossed]
haughty bobcat
well I went to the github listed above so I think I'm good on guides
fingers crossed that adabox 008 has circuitpython and a crickit
tidal kiln
@haughty bobcat this? https://learn.adafruit.com/adafruit-gemma-m0/pinouts
haughty bobcat
yes!
tidal kiln
the other little trick is to do this in REPL:
import board
dir(board)
Ok thanks!
So is Arduino on its way out?
Personally I haven't heard any new things about Arduino for a while
stuck elbow
plus, all the boards you can program with circuitpython can also be programmed with arduino
haughty bobcat
yeah it's a win win
tidal kiln
think of it as more programming options, not one thing replacing another
haughty bobcat
so is adafruit just going to start making everything circuitpython since the chipsets support both?
stuck elbow
@haughty bobcat you still can do more with Arduino — it gives you more resources on the same hardware
so there will be both
scarlet fjord
this is probably more of a general Python question, but is there any way for me to have an array of variables where some of those variables are a single thing and some are themselves an array of variables?
cunning crypt
@scarlet fjord Yes, absolutely!
scarlet fjord
or a better way would be to have everything as an array of arrays, but then have the single-thing ones be just an array of one?
cunning crypt
Say you have variable [X, Y, Z]
With python, variables can be freely changed.
So you can do something like X = 10
X is now an int.
Well, more accurately
Var[X, Y, Z]
Var[0] = 10 would make X an int
scarlet fjord
oh ok, so something like this then, should work?
cunning crypt
Var[0] = [0, 10, 5] would make X a list
scarlet fjord
kbd_p1 = [Keycode.A, Keycode.S, Keycode.F]
kbd_p2 = Keycode.B
kbd_p3 = Keycode.C
kbd_p4 = Keycode.D
kbdkeys = [kbd_p1, kbd_p2, kbd_p3, kbd_p4]```Yep, that should work.
scarlet fjord
ok cool
cunning crypt
But, you'd have to re-update kbdkeys
Which could be useful in some cases, IE to see if it has changed.
scarlet fjord
what do you mean update it?
cunning crypt
Oh, right. Sorry.
If you changed kbd_p1, it wouldn't automatically update inside kbdkeys
tulip sleet
Keyboard.press() will take multiple arguments. So if you make them all lists or tuples, you can do this:
kbdkeys = ( (Keycode.SHIFT, Keycode.A), (Keycode.B,), (Keycode.CTRL, Keycode.D))
...
kbd.press(*kbdkeys[i])
im assuming then;
Traceback (most recent call last):
File "main.py", line 48, in <module>
File "adafruit_hid/keyboard.py", line 116, in release
File "adafruit_hid/keyboard.py", line 156, in _remove_keycode_from_report
File "adafruit_hid/keycode.py", line 305, in modifier_bit
TypeError: unsupported types for __le__: 'int', 'list'```
I'm assuming im not doing what Dan just said i should do 😛two new concepts there: * in arg list, and (1,) for a tuple with one element
scarlet fjord
(whats a tuple) 😅
stuck elbow
it's a math term generalizing the idea of pair, triple, quadruple, etc.
quadruple = 4-tuple
scarlet fjord
i see
cunning crypt
Deshipu and Dan know a lot more than I, and it's dinner time for me anyawy.
scarlet fjord
thanks for your help! 👍
stuck elbow
most of the time you can use tuples and lists exchangeably
the exception is for dict keys, where you can't use mutable objects, so you can only use tuples
scarlet fjord
hmm ok
so
kbd.press(*kbdkeys[i])
is looking for whatever '*' is? in an array called kbdkeys, for the array item at index [i]?
stuck elbow
the asterisk in this place is a little tricky to explain
it lets you pass a list (or tuple) as parameters to a function as if you separated them with commas
for instance, instead of f(1, 2) you can have x = (1, 2) and then f(*x)
scarlet fjord
i see, ok
stuck elbow
you can also use two asterisks to pass a dict as keyword arguments, for example x = {'a': 1, 'b': 2} and then f(**x) gives you f(a=1, b=2)
don't worry if you don't understand now
you will get there
scarlet fjord
haha yeah i wasnt expecting to understand any of this right away 😅
theres just so much that i didnt expect to need to understand to get this to work 😛
the issue im having is this:
so
kbd_p1 = [Keycode.A, Keycode.S, Keycode.F]
kbd_p2 = Keycode.B
kbd_p3 = Keycode.C
kbd_p4 = Keycode.D
kbdkeys = [kbd_p1, kbd_p2, kbd_p3, kbd_p4]
keypress = kbdkeys[keyindex]
kbd.press(*keypress)```gives an error:
TypeError: can't convert list to int
stuck elbow
which line?
scarlet fjord
File "main.py", line 49, in <module>
stuck elbow
which of those lines is the 49?
scarlet fjord
which in my code is a print str...
waiduminnit
yup, lol... changed the do thing but didnt change the print thing 😅
stuck elbow
it's always good to look at the line to which the error points, Python is a bit better than C in guessing where the error is
though of course it's not always perfect
scarlet fjord
print("Pressed #%d" % *keypress)```
`SyntaxError: invalid syntax`?
im assuming im not allowed to just do what i just did 😛yeah, you can only use * this way in a function call
scarlet fjord
fair
kbd.release(*keypress)```
` File "main.py", line 48, in <module>
TypeError: 'int' object is not iterable`I suppose you could change the %d to %r
scarlet fjord
oooh is that like a type thing, so %d would be int, so i can change that to str or something?
stuck elbow
if you look carefully at Dan's example, he used 1-element tuples
yes, %s is string, %f is float, and %r gives you the "representation" of the object
scarlet fjord
ah
scarlet fjord
oh, of course...
woohoo! thanks a bunch 😁
@stuck elbow and @tulip sleet thank you both for your infinite wisdom and patience 😛
tulip sleet
np
fading solstice
@tulip sleet when building circuitpython 3, should i be using branch master or 3.x
tulip sleet
depends on what you're testing. master is now for nRF board development, but we are merging from 3.x to master. either is probably fine. if you're trying to pretend to be a regular user, use 3.x.
tidal kiln
@slender iron TLDR: something's up with busio.OneWire
deets follow...
Adafruit CircuitPython 2.3.1 on 2018-05-07; Adafruit Itsy Bitsy M0 Express with samd21g18
>>> import board, busio
>>> ow = busio.OneWire(board.D7)
>>> ow.write_bit(1)
Adafruit CircuitPython 3.0.0-rc.0 on 2018-06-18; Adafruit ItsyBitsy M0 Express with samd21g18
>>> import board, busio
>>> ow = busio.OneWire(board.D7)
>>> ow.write_bit(1)
and there can be additional confusion due to the fact that D5 is a special pin on the Itsy (the example should be updated to use something else)
@tulip sleet ⬆ FYI
stuck elbow
that is a looooong bit
tidal kiln
sure thing
manic glacierBOT
REF: https://forums.adafruit.com/viewtopic.php?f=60&t=137405
D5 is a special pin on the Itsy, but even using D7 for testing indicates something is not working:
CP 2.3.1
Adafruit CircuitPython 2.3.1 on 2018-05-07; Adafruit Itsy Bitsy M0 Express with samd21g18
>>> import board, busio
>>> ow = busio.OneWire(board.D7)
>>> ow.write_bit(1)
CP 3...
strange pumice
@idle owl is it possible to fork an empty repository? I'm not able to get going with the one you setup for me today for the MLX90614 driver. https://github.com/adafruit/Adafruit_CircuitPython_MLX90614/
idle owl
@strange pumice I would have thought so, but maybe not. I'm at a meeting. I'll take a look when I get home. Worst case, I'll upload the first version.
raven canopy
just scanned over some of Adafruit_Blinka. very nice!! "Blinka, for all the things!" 😄
slender iron
@strange pumice I don't think so. I usually have github put in a bogus README temporarily. need me to?
raven canopy
haha. i just looked at that too, and was about to suggest the same thing (a bogus file). fork is definitely disabled since it is empty.
strange pumice
@slender iron if you don't mind placing a bogus file in there for me that would be great.
slender iron
kk
done @strange pumice ! https://github.com/adafruit/Adafruit_CircuitPython_MLX90614
strange pumice
thanks...tryign now.
slender iron
😃
idle owl
Back.
manic glacierBOT
raven canopy
Front. Side-to-side. (yes, i did a desk dance while i said that out loud) 🕺
idle owl
As well you should have!
raven canopy
@tulip sleet so drive strength is probably a red herring? i trust that you and Scott are light years ahead of me on pin handling... 😄
and no errata or other datasheet goodness that points to DRVSTR.
tulip sleet
@raven canopy we could easily be wrong, it's just that the DRVSTR change was made a long time ago, and a similar change was also made in 2.x.
i am going to try a onewire device in old and newer 3.0's
tulip sleet
@raven canopy I don't want to discourage you at all from trying to debug this. I was just trying to understand why you did what you did. I think an interesting test would be to duplicate the OPEN_DRAIN and PULL_NONE settings, and re-test with an external pullup, since that's how the devices work. There may well be a port setting issue.
raven canopy
I was thinking of using Atmel Studio to verify that the pin config registers were changing accordingly. but, that was under my assumption that my test was "valid". i can keep poking at it. abstract testing is largely an experience base; my base is shallow. 😄
tulip sleet
I am testing with a DS18B20 temp sensor. I went all the way back to 3.0.0-alpha.2, and it doesn't work there either. Also verified 2.3.1 works. I am just scanning the OneWire bus for the device.
solar whale
watchhing tihis discussion made me think that the DHT is "like a one-wire" so I tried a DHT22 and it does work with master. Just FYI - probabbly of no help.
tulip sleet
@solar whale I looked, because that was intriguing. The DHT library uses pulseio, not bitbang, so it's not exercising the same paths.
@raven canopy I did comment out the DRVSTR setting, and it still doesn't work. But the way setting the pin value works is kinda different in 2.3.1 vs 3.0.0. Diff screenshot attached. The gpio_... routines in ASF4 actually do more than the register operations in 2.3.1. Worth looking at carefully. I also wonder if there's something wrong with the microsecond timer for some reason.
ruby atlas
@sharp rain were you using CP2.x or 3.x?
(I'm getting memory errors on my branch of CP3, trying master)
(Feather M0 express)
Huh, build error. make: *** No rule to make target `peripherals/samd/clocks.c', needed by `build-feather_m0_express/genhdr/qstr.i.last'. Stop. on newest master
Trying a fresh checkout
ah oops, git submodule update --init --recursive : forgot the --init
also fails on master 3.x
tulip sleet
I'll continue on this OneWire thing in the morning.
sweet wedge
Hello ... I'm new around here. Just began using CircuitPython last weekend. Would anyone know why the method described to install Mu with the plotter doesn't seem to work for me? Tried on both of my Macs following the steps described here: https://learn.adafruit.com/welcome-to-circuitpython?view=all#installing-mu-editor I've been using Mu just fine, wanted to try the plotter so I installed Homebrew, entered the next commands described. Each appeared to run successfully, but when I type Mu in Terminal ... "Command not found"
tulip sleet
@sweet wedge Did you type "mu" or "Mu" to run?
In any case, there is now a real installer for the Mac, available here: https://codewith.mu/ -- click the Download button. We'll be revising the Mu installation instructions imminently to reflect this rather than the complicated procedues described in the guide.
solar whale
@sweet wedge when imu is installed via the new MacOs installer, it isntalls to "Applications" as mu-editor and can be started by double clicking on the mu-editor icon -- it does not install a command line script to start it.
you can start if rom the command line by entering /Applications/mu-editor.app/Contents/MacOS/mu-editor
or even simple open -a mu-editor
tulip sleet
@solar whale thanks for elaborating
solar whale
I thought I'd finanlly give it a try myself!
sharp rain
@ruby atlas Its CP2.x but I'm using fairly basic python so I would be surprised if the code is not compatible.
ruby atlas
Compatible. Just out of memory.
timber mango
Any chance we can do something about very long usernames?
It's extremely annoying to me.
Not everybody is 25 years old.
stuck elbow
what do you mean?
timber mango
well your name is 7 letters and in lowercase.
right? Mine's 3 letters. 'adafuit' is 8 letters but one of them is an 'i' which takes up less room here (we use proportional fonts)
stuck elbow
I think I follow so far
timber mango
MYNAMEISJOEKAVEMAN is a bit larger than those examples.
stuck elbow
that is true, what of that?
timber mango
Well it's distracting and it serves no purpose here. It doesn't facilitate communication; it obstructs it.
stuck elbow
you lost me there
timber mango
MYNAMEISJOEKAVEMAN suppose I say
MYNAMEISJOEKAVEMAN lots of things that have no real value
MYNAMEISJOEKAVEMAN but in each utterance
stuck elbow
there is tab completion, so it doesn't really matter
timber mango
MYNAMEISJOEKAVEMAN I state that my name is joe kaveman
MYNAMEISJOEKAVEMAN in all caps.
I believe I've just illustrated the problem.
timber mango
That's your real name though.
It looks like a real name and the eye forgives the length because it has a natural aesthetic quality about it.
(I think it's excessive practice for every single utterance, but it's your name and it's nice to see it once in a while).
Dido Florian Cloud de Bounevialle O'Malley Armstrong
is a real person's name, too.
But she uses 'Dido'. ;)
I'm back quiet (and afk).
sweet wedge
@tulip sleet @solar whale Thank you both. I ran the new installer you linked to Dan, and I'm in-business including the plotter. With the old method, I had tried typing mu and when it didn't work, tried Mu ... no luck either way. Anyway thanks again!
tulip sleet
@slender iron I'm out for a couple of hours. Working on fixing OneWire. It's really mostly a fix to DigitalInOut.c, but the timing is still flaky sometimes because calling tick_delay(small numbers of microseconds) is not always accurate. Using some NOP loops but need to tune it.
slender iron
@tulip sleet k. it could be related to the monotonic issues notro reported
tulip sleet
i don't know if those issues are cause or effect. I think the time it takes to go through tick_delay varies depending on whether it's on a ms boundary or not and maybe other stuff.
onewire turns off interrupts but needs to do delays of like 6us, which is not always accurate, it appears. I get CRC errors from the DS18B20 on and off. Didn't see these on 2.3.1.
i think we can't use pulseio because it has to be an open-drain output, which is simulated by switching back and forth to an input
stuck elbow
I hate time-based protocols
ruby lake
time-division multiplex forever 😉
teal bear
"OUT OF STOCK" 😮 any guess to how soon they will be shipping again?
raven canopy
Usually that is answered in terms of "few weeks". Depends on what "they" you are referring to. At any rate, best suggestion is to use the "Notify" button; it's usually the best and quickest indicator. @teal bear
timber mango
If it's a good selling new item you have to watch it manually. ;)
teal bear
thanks - isn't this the CPX channel? (that's the they I was hoping to buy
timber mango
I've been burned several times by thinking 'Notify' was going to give me a good shot at a newly introduced item.
CarlFK there are no channels that are board specific. ;)
teal bear
what else does circuitpyuthon run on?
timber mango
Every SAMD21 board (M0 is usually in the name).
Feather M0 Express. Metro M0 Express. Gemma M0. Trinket M0. ..
tidal kiln
timber mango
The ones without 'Express' in the name usually don't have a 2MB SPI flashROM on them, and so even when they run CircuitPython, there are restrictions on how large a library can be stored on the target board at any one time.
The 'Express' ones all have 2MB SPI flashROM and can hold the entire populated /lib folder.
teal bear
neat - thanks
i the cpx was more expensive I might look at the smaller ones .. but for $25 it's hard to justify going cheaper
timber mango
The form factor is an important aspect of the CPX as well as the wide variety of sensors.
teal bear
I show them to people that want to learn Python and they keep trying to buy mine
timber mango
Haha yeah have a few spares to sell!
teal bear
so today I decided to to finely do that... and Bzzz
timber mango
What happened?
raven canopy
@teal bear DigiKey has 59 in stock. https://www.digikey.com/product-detail/en/adafruit-industries-llc/3333/1528-2280-ND/7310913
idle owl
@tulip sleet Do you want me to wait for Scott to review this or merge it?
tulip sleet
@idle owl thanks, it was fine you did
solar whale
@tulip sleet did you run into any issues with servos getting very hot with the extended min/max pulse limits?
tulip sleet
@solar whale, i didn't really test that, I was just looking at the angle. I think the 550-2400 is a compromise, some are wider
have you seen that?
solar whale
OK - I did some playing when I first got my CRIKIT and one servo go vry hot when I pushed the limits out. Not an exhaustive test. I did not repeat the test 😉
not sure if it was the limit or something else I did. It kept working.
I'll try to reproduce it int the next few days. more systematically.
tulip sleet
I went to the Tower Pro website and they don't actually seem to give the specs for pulse width! I am giong to ask CGrover
he added one extra to some info the library
solar whale
I had the same issue about needing more range.
tulip sleet
I think if you go too far and hit the stops it wil get hot
so just back off a little
1000-2000 is clearly too low
tulip sleet
I think it is just shy of the stops. I just tested the micro servo and the continuous servo
solar whale
I don't recall how far I went. sorry.
tulip sleet
this website https://servodatabase.com/ lists some limits, but they're missing for a lot of the servos.
that's ok, I think it's trial and error.
solar whale
it seems to be a bit empirical!
tulip sleet
there may be sample variation as well
I did a bunch of googling and came up with very little, which I don't really understand. I thought the RC world would be all over this issue
solar whale
just something to watch out for
tulip sleet
ya, I'll ping @errant grail now - he's offline. @errant grail, how did you come up with the pulse limits for the servo you put in a PR to the motor library, and do you know in general where this stuff is documented and explained?
solar whale
thanks - I was surprised how hot it got, but it recovered nicely when I stopped!
tulip sleet
Limor just told me in a mtg earlier this evening that the default limits should be changed. Brought up in several ways including this forum post: https://forums.adafruit.com/viewtopic.php?f=8&t=137545
Python 3.7 released: See https://python.org and https://pythoninsider.blogspot.com/2018/06/python-3.html
solar whale
I'm not objecting to the change. Just curious if others have seen issues. Again, I need to do more systematic testing.
what will 3.7 break 😉
Its going to take me a day to clear enough space on the workbench to test servos again! I'm buried in radios....
errant grail
@tulip sleet The values I provided were for just the Tower SG92R servo. Those values were determined experimentally using a batch of three servos recently purchased from the Adafruit store. I subtracted 5% from the empirically-measured range to determine new recommended values. I searched far and wide for specifications or guidance on those specific servos, but found nothing. A suggestion would be to provide an obvious note in the examples that the values can vary from batch to batch and that driving the servo to the min or max physical positions can cause permanent damage.
tulip sleet
Sounds like a good idea. We can add that. A guide is underway as well. Thanks! Yes, the whole thing seems to be poorly documented. Maybe in RC applications one rarely goes to the limits?
errant grail
My dad was an avid RC enthusiast and rarely pushed the limits of servos. Having trim and gain always handy on the controller helps, too. Wish the hobby servos were smarter about that -- mechanically and electrically.
( Sorry about the delay in my response. I missed S&T and AskEng, too. 😦 )
tulip sleet
no problem - I was just making a new motor library release and noticed you gave some servo data and thought maybe you had insider knowledge.
solar whale
Thanks @tulip sleet and @errant grail - glad I'm not only one finding it hard to get clear guidance for the servo limits
errant grail
I'm wondering if the servo universe will change now that Tower is out of the picture. I wonder who (in China) really makes the Tower servos...
errant grail
Tower/Hobbico selective product lines were sold to Horizon Hobbies, I believe. Many of the parts I use are out of stock on the current Tower site.
solar whale
Bed time here - good night all! -- I will try to set up some servo testing tomorrow.
tulip sleet
I did some looking. I think that may be name overlap. https://www.towerhobbies.com/ does not seem to be the same as http://www.towerpro.com.tw/.
errant grail
@tulip sleet That's good news. Have been ordering Tower servos from Tower Hobbies for so long I thought they were the same. 😜
umbral dagger
I'm seeing some odd behaviour that I haven't found an explaination for. ItsyBitsyM0 Express. Double reset goes into bootloader mode fine. Drop the uf2 onto it (CP3rc0). ITSYBOOT unmounts, but CIRCUITPY never mounts. Reset does get it to mount either. The dotstar indicates that it's in the CP runtime (solid green). Thoughts?
I can load an arduino sketch with no problem.
errant grail
Just did the same with the ItsyM0x this morning and didn't encounter any issues. FYI
umbral dagger
Yeah, I've been using them since they came out.
tulip sleet
@umbral dagger which OS and version? Try changing ports and rebooting, especially for Linux and Mac. If that doesnt' work, try beta.1 and see if it's different, and 2.3.1 if necessary? Post an issue if it seems consistent. Thanks.
umbral dagger
Latest Ubuntu. I lost bootloader mode at one point, but a reboot fixed that. Changing from a hub to a frontpanel USB port doesn’t bring back CIRCUITPY.
stuck elbow
@umbral dagger can you do os.listdir() from the repl?
@umbral dagger the circuitpy disk won't mount if the filesystem is corrupt
or otherwise inaccessible
tulip sleet
did you try on an another computer? did you reload the .uf2?
umbral dagger
It’s a cable I’ve been using successfully before this.
tulip sleet
make sure it's the right .uf2 for the board
solar whale
@umbral dagger - just my experience -- on my Ubuntu system -- the CIRCUITPY drive often shows up in the file manager but does not auto mount. Clicking on the drive icon mounts it and I can copy file to it -although drag/drop often does not work in the file manager.
stuck elbow
for me it often mounts, but the directory is read-only, I have to do chmod +w .
fading solstice
@umbral dagger i built circuitpython on my ubuntu machine from source and had the same problem as you. i was thinking i did not have my machine tools installed correctly. are you also building circutpython from source?
solar whale
nevermind - if you are not getting REPL, its a different issue. I can always get to REPL
tulip sleet
solid green is unusual, usually it's pulsing green
manic glacierBOT
umbral dagger
@fading solstice No, using the release uf2
tulip sleet
what was in the main.py? anything unusual?
"adafruit-circuitpython-itsybitsy_m0_express-3.0.0-rc.0.uf2", right?
solar whale
@umbral dagger when you connect it does /dev/ttyACM0 or any /dev/ttyACMx appear ? just wondering if when you connect to REPL, soes it connect but not respnd (possibly running code) or does it fail to connect at all.
umbral dagger
@tulip sleet yes
The board was fresh from the factory, so whatever demo code was preloaded.
@solar whale ttyACM0 shows up, screen connects but there’s nothing responding. Unplug and screen exits. (ttyACM0 goes away)
tulip sleet
so ctrl-C more than once does nothing
solar whale
sounds like it is booting and running.
tulip sleet
it was probably loaded with 2.3.1, so the lib directory is 2.x .mpy's. Can you load 2.3.1?
solar whale
is there an itsybity_m0 eraser
tulip sleet
no
if 2.3.1 can be loaded, then try getting to repl there, and easy to erase fs from repl
i could build one, just haven't
solar whale
let me check - I may have built one...
solar whale
I do have a .uf2 that will create a new FS - built in April --- I can post it if useful.
umbral dagger
@tulip sleet Will do
scarlet fjord
hello it me again with another question 😅
following this tutorial for more GPIO on my Feather M0 Express:
https://learn.adafruit.com/using-mcp23008-mcp23017-with-circuitpython/software
it makes mention of the I2C addresses:
Either class initializer can also be passed an optional address keyword to override the I2C address
but it doesnt say how to use those address keywords.
i have two MCP23017 chips, one at 0x20 and one at 0x24, but i dont see how to tell CircuitPython that?
tulip sleet
mcp = adafruit_mcp230xx.MCP23017(i2c, address=0x24)
scarlet fjord
ah, yep thanks, i just saw that on the github page 👍 (this one: https://github.com/adafruit/Adafruit_CircuitPython_MCP230xx/blob/master/examples/simpletest.py)
does the library have any way of inverting the pin value?
cause the MCP23017 can be set up as pullup inputs, which would give False if the switch is pressed, and that is confusing to me 😅
unless im dumb and it can also be configured as pull-down?
tulip sleet
looks like there are only pull-ups. I know it's confusing, but kinda typical to use pull-ups instead of pull-downs, so I'm afraid you'll have to live with it. you could write a helper function in Python to invert the value if you don't want to do if not blah.value
I added an example of using address= to the guide.
ruby lake
"active low"
umbral dagger
@tulip sleet 2.3.1 works fine
3.0rc0 works fine on a different ItsyBitsyM0Express board
solar whale
in 2.3.1 you can erase the FS import storage storage.erase_filesystem() then reload 3.0 -- but @tulip sleet may have had other things to check first.
tulip sleet
@umbral dagger if there is a main.py/code.py you could rename it in 2.3.1 and see if 3.0.0rc0 starts, or just erase the filesystem in 2.3.1 and see how 3.0.0rc0 works
umbral dagger
@solar whale That did it. It now seems fine in 3.0
I'm so bleeding edge.... never occurred to me to go "back" to 2.3.1
interesting...
solar whale
was the othe IitsyBityM0 you tried also "fresh" that is did it have the same 2.3.1 based main.py and libs on it?
umbral dagger
It would have initially, but I've been working with it for a while. This "new" one was from the saem order. It's been built into a project (i.e. I wanted a headerless one) so I grabbed a fresh one from the drawer.
solar whale
As Dan said, probaly 3.0 did not like trying to run the .mpys from the 2.3.1 lib. Still a bit surpriing there was no way to "break" out in the REPL.
scarlet fjord
is there any way to ask the adafruit feather what sort of devices are connected to it over i2c, and specifically what addresses those devices have?
I ask because I can't seem to find these mcp23017 devices
nvm, taking the picture helped, that darned resistor there was plugged in to the wrong thing
stuck elbow
@scarlet fjord you can do i2c.scan()
@scarlet fjord you have to do i2c.try_lock() before that, though
fading solstice
@tulip sleet I am trying to build metro m4 circuitpython from github source. it build (with a warning), but emits a uf2. when i copy the uf2 to my metro m4, i get the exact same behavior as dastels. maybe my machine does not have the coreect setup. is there a doc i should be following to setup my machine. UBUNTU 16.04
scarlet fjord
oh cool, thanks for letting me know 😃
ive figured that out now though, it was just that resistor not doing its job that caused them to not be detected 😅
hmm...
File "main.py", line 27, in <module> MemoryError: memory allocation failed, allocating 584 bytes
all im doing is this:
line: 17 pin00 = mcp1.get_pin(0)
line: 18 pin01 = mcp1.get_pin(1)
line: 19 pin02 = mcp1.get_pin(2)
line: 20 pin03 = mcp1.get_pin(3)
line: 21 pin04 = mcp1.get_pin(4)
line: 22 pin05 = mcp1.get_pin(5)
line: 23 pin06 = mcp1.get_pin(6)
line: 24 pin07 = mcp1.get_pin(7)
line: 25 pin08 = mcp1.get_pin(8)
line: 26 pin09 = mcp1.get_pin(9)
line: 27 pin10 = mcp1.get_pin(10)
line: 28 pin11 = mcp1.get_pin(11)
line: 29 pin12 = mcp1.get_pin(12)
line: 30 pin13 = mcp1.get_pin(13)
line: 31 pin14 = mcp1.get_pin(14)
line: 32 pin15 = mcp1.get_pin(15)
have i hit some memory limit im not aware of or something??
solar whale
@fading solstice if you reset the Metro M4 - what does the LED show ? if you type `ls /dev/ttyACM*' do you see anything?
stuck elbow
@scarlet fjord you ran out of memory, but that's strange if it's the only thing your program is doing
scarlet fjord
its not all im doing, but... im not doing much? i dont think...
fading solstice
the serial port does not connect on Ubuntu and does not connect to Window 10. the rgb led is solid green
the metroboot shows after double click
scarlet fjord
i did a thing and now it does this:
File "adafruit_hid/keycode.py", line 31, in <module> File "adafruit_hid/keycode.py", line 165, in Keycode MemoryError: memory allocation failed, allocating 584 bytes
fading solstice
circuitpy does not show
solar whale
does screen /dev/ttyACM0 connect
fading solstice
@solar whale no
solar whale
does ls /dev/ttyACM* show anything -- may conenct as ACM1...
fading solstice
@tulip sleet @solar whale nothing connected. BTW last 3.0 release uf2 works.
I did the erase_filesystem trick. did not help
same on windows 10
fading solstice
no
solar whale
one thing to try is boot a working 3.0 version and erase the FS then reload RC0
fading solstice
extern mp_obj_tuple_t common_hal_usb_hid_devices;
^
common-hal/usb_hid/__init__.c:127:16: note: 'common_hal_usb_hid_devices' was previously declared here
mp_obj_tuple_t common_hal_usb_hid_devices = {
^
that is normal
fading solstice
ok thanks
scarlet fjord
@stuck elbow is there anything i can do? im at an impasse here, i have no idea how to progress
solar whale
its a known issue with -lto and the frozen modules
fading solstice
oh, anything to get around it?
solar whale
the solid green LED indicates the Metro is running somthing --- odd
It is just a warning -- not a problem
no need to do anything about it
fading solstice
if i was able to build a uf2, then either my machine is not setup correctly or the source i am build is incorrect
solar whale
are you building current master?
fading solstice
i tried master then i tried 3.x. same result from both
solar whale
git submodule foreach --recursive 'git fetch --tags'
i cut 39 lines of code (all variable assignments identical to above) and i got the number to go up 🙃
File "adafruit_hid/keycode.py", line 195, in Keycode
MemoryError: memory allocation failed, allocating 776 bytes```try running thos commands after you do git pull in master
fading solstice
i used git fetch adafruit; git pull adafruit. i am thinking i should try to clone the source instead
@solar whale i see your message now. i did the first, but not the second
solar whale
try the second -- won't hur t 😉
stuck elbow
@scarlet fjord I suppose you can save memory by using the mcp1.gpio() method directly, instead of creating all those pin objects
solar whale
@fading solstice see if it helps -- it may depend on the version of git running. It has been a recent change needing it for me. @tulip sleet is more familiar with the reason for it.
scarlet fjord
oh, how does that work? @stuck elbow
stuck elbow
are you reading or writing?
solar whale
@fading solstice jsut to be sure, you are using a "new" metro M4, not one of the older "revbs"
scarlet fjord
how much memory should i have access to? im really not doing all that much
im reading, i think
listening for switches
stuck elbow
oh, sorry, gpio is a property, not a method, so you simply read it, and it gives you a value where each bit corresponds to a state of pin
strange pumice
@idle owl https://github.com/adafruit/Adafruit_CircuitPython_MLX90614/pulls is ready for QA.
stuck elbow
so for example to check if the first pin is high, you would do if mcp1.gpio & 0x01:
for third it would be & 0x04 etc.
using powers of two
solar whale
@fading solstice if you load a working 3.0 into the board and at the repl tyrp import os os.listdir() what do you see?
stuck elbow
0x01, 0x02, 0x04, 0x08, 0x0f, 0x10, 0x20, 0x40, 0x80, 0xf0
fading solstice
@solar whale
>>> import os
>>> os.listdir()
['.fseventsd', '.metadata_never_index', '.Trashes', 'boot_out.txt']
what if i want to read the state of all 16 pins from 3 MCP23017 chips? surely id need to make an array of all the pins and read them in a for loop, instead of writing out a lot of if mcp1.gpio & 0x01?
whats different between mcp1.gpio() and mcp1.get_pin()?
stuck elbow
@scarlet fjord what do you plan to do with the values once you read them?
fading solstice
@solar whale the board is the Metro M4 Express BETA
solar whale
ok - that shpuld be fine.
scarlet fjord
im trying to make a keyboard, so i have a list of keys and a list of what those keys do, then a for loop that runs and checks each key to see if its pressed then passes on the keycodes that that key presses over USB as a HID if that key was pressed
fading solstice
@solar whale i rebuit with the extra git command. same result
stuck elbow
@scarlet fjord ok, so perhaps something like this will work:
solar whale
@fading solstice did you do make BOARD=metro_m4_express clean make BOARD=metro_m4_express that is make clean first
fading solstice
that is what i do all the time
stuck elbow
for mcp_id, pressed in enumerate([mcp1.gpio, mcp2.gpio, mcp3.gpio]):
for bit in range(8):
if pressed & 1 << bit:
print("key %d of chip %d has been pressed" % (bit, mcp_id))
running out of ideas -- hopefully Dan will have magic answer when he gets back.
stuck elbow
(you might need to negate it)
solar whale
does dmesg shown anything about the USB connection when if you run it after a RESET of the board
stuck elbow
1 << bit calculates the power of 2 corresponding to the bit
scarlet fjord
woah that looks fancy
idk what it does, but ima try it out? 😅
stuck elbow
@scarlet fjord you might need to enable the pullups with mcp1.gppu = 0xff first
aaaand I found a bug
this won't work
scarlet fjord
im assuming i have to set them as inputs too?
stuck elbow
they are inputs by default iirc
fading solstice
@solar whale the METROBOOT has a connection, but after the uf2 is coopied and after a single click reset, no connections are logged
solar whale
it is METROM4BOOT, correct?
fading solstice
sorry yes
solar whale
whew 😉
fading solstice
@solar whale just to make sure i am doing the right thing. cd circuitpython; cd ports; cd atmel-samd; make clean BUILD=metro_m4_express; make BUILD=metro_m4_express
solar whale
looks correct (as long as you spell BUILD correct on the second make 😉 -- then copy the build_metro_m4_express/firmware.uf2 to the METROM4BOOT
solar whale
just for a test, I can post a .uf2 I have built and used.
fading solstice
sure
solar whale
metro4m
fading solstice
@solar whale that worked.
solar whale
that is very odd -- it was from 2 days ago
I can do a ne pull and buld - but I can't test it if you want to try it.
fading solstice
yes i will
solar whale
building - just a sec
fading solstice
thx
solar whale
new metrom4
scarlet fjord
well its 2:30 am and i need sleep, but if anyone wants to have a dig through and figure out why im running in to memory issues when using the adafruit_hid.keycode library, that would be really cool...
Just @me and ill see it in the morning 😬
(for context, the keycodes im defining in macros are arbitrary and random at the moment, i just wanted something there)
fading solstice
@solar whale yeah, that worked as well.
has to be something wrong with my environment
solar whale
weird - waht arm-none-eabi-gcc are you running -- I have 7.2.1
fading solstice
how to i tell
solar whale
arm-none-eabi-gcc -v
fading solstice
gcc version 7.2.1 20170904 (release) [ARM/embedded-7-branch revision 255204] (GNU Tools for Arm Embedded Processors 7-2017-q4-major)
solar whale
same -- one other suggestion -- after you cloned did you build mpy -cross -- make -C mpy-cross
solar whale
not sure -- worth a try. If for some reason it is building 2.x versions ....
at top level make -C mpy-cross
idle owl
@tulip sleet @slender iron I'll be out for a couple of hours this morning/afternoon. Should be back before 2pm.
solar whale
then rebuild the .uf2
also jsut to be sure, git branch shows you on master, correct?
stuck elbow
can someone please review this? https://github.com/adafruit/Adafruit_CircuitPython_MCP230xx/pull/2/files
fading solstice
@solar whale master branch, rebuilt after building mpy-cross; still fails 😣
solar whale
😦 sigh
fading solstice
do you have the same number of bytes as me?
4068 bytes free in ram for stack out of 32768 bytes ( 32.0 kb ).
Create metro_m4_express/firmware.bin
Create metro_m4_express/firmware.uf2
../../tools/uf2/utils/uf2conv.py -b 0x2000 -c -o metro_m4_express/firmware.uf2 metro_m4_express/firmware.bin
Converting to uf2, output size: 457216, start address: 0x2000
Wrote 457216 bytes to metro_m4_express/firmware.uf2.
Use make V=1, make V=2 or set BUILD_VERBOSE similarly in your environment to increase build verbosity.
install -d build-metro_m4_express/genhdr
python3 tools/gen_usb_descriptor.py \
--manufacturer "Adafruit Industries LLC"\
--product "Metro M4 Express"\
--vid 0x239A\
--pid 0x8021\
--output_c_file build-metro_m4_express/autogen_usb_descriptor.c\
--output_h_file build-metro_m4_express/genhdr/autogen_usb_descriptor.h
Generating build-metro_m4_express/genhdr/mpversion.h
GEN build-metro_m4_express/genhdr/qstr.i.last
QSTR updated
GEN build-metro_m4_express/genhdr/qstrdefs.generated.h
../../shared-bindings/usb_hid/__init__.h:33:23: warning: size of 'common_hal_usb_hid_devices' differ from the size of original declaration [-Wlto-type-mismatch]
extern mp_obj_tuple_t common_hal_usb_hid_devices;
^
common-hal/usb_hid/__init__.c:127:16: note: 'common_hal_usb_hid_devices' was previously declared here
mp_obj_tuple_t common_hal_usb_hid_devices = {
^
276476 bytes free in flash out of 499712 bytes ( 488.0 kb ).
49104 bytes free in ram for stack out of 196608 bytes ( 192.0 kb ).
Create build-metro_m4_express/firmware.bin
Create build-metro_m4_express/firmware.uf2
../../tools/uf2/utils/uf2conv.py -b 0x4000 -c -o build-metro_m4_express/firmware.uf2 build-metro_m4_express/firmware.bin
Converting to uf2, output size: 446976, start address: 0x4000
Wrote 446976 bytes to build-metro_m4_express/firmware.uf2.
not entirely sure
fading solstice
Use make V=1, make V=2 or set BUILD_VERBOSE similarly in your environment to increase build verbosity.
install -d metro_m4_express/genhdr
python3 tools/gen_usb_descriptor.py \
--manufacturer "Adafruit Industries LLC"\
--product "Metro M0 Express"\
--vid 0x239A\
--pid 0x8014\
--output_c_file metro_m4_express/autogen_usb_descriptor.c\
--output_h_file metro_m4_express/genhdr/autogen_usb_descriptor.h
Generating metro_m4_express/genhdr/mpversion.h
GEN metro_m4_express/genhdr/qstr.i.last
QSTR updated
GEN metro_m4_express/genhdr/qstrdefs.generated.h
../../shared-bindings/usb_hid/__init__.h:33:23: warning: size of 'common_hal_usb_hid_devices' differ from the size of original declaration [-Wlto-type-mismatch]
extern mp_obj_tuple_t common_hal_usb_hid_devices;
^
common-hal/usb_hid/__init__.c:127:16: note: 'common_hal_usb_hid_devices' was previously declared here
mp_obj_tuple_t common_hal_usb_hid_devices = {
^
25168 bytes free in flash out of 253696 bytes ( 247.75 kb ).
4068 bytes free in ram for stack out of 32768 bytes ( 32.0 kb ).
Create metro_m4_express/firmware.bin
Create metro_m4_express/firmware.uf2
../../tools/uf2/utils/uf2conv.py -b 0x2000 -c -o metro_m4_express/firmware.uf2 metro_m4_express/firmware.bin
Converting to uf2, output size: 457216, start address: 0x2000
Wrote 457216 bytes to metro_m4_express/firmware.uf2.
@solar whale even the first part is different for example
--output_h_file build-metro_m4_express/genhdr/autogen_usb_descriptor.h```is different
solar whale
I "think" it identifies the board configuration. -- may be mistaken.
if you type git diff master do you get any output
fading solstice
no output
the python3 command is generating the usb descriptor
i think we are getting somewhere byt i don;r know where
solar whale
do you see a folder name metro_m4_express as well as a build_metro_m4_express?
fading solstice
in atmel-samd? no
solar whale
good
fading solstice
just metro_m4_express
solar whale
oh - I have build_metro_m4_express -- that is odd
ahhh --- igot it
use make BOARD=metro_m4_express not BUILD
solar whale
the defaiultfor BOARD is metro_mo_express -- see Makefile
fading solstice
ok, that has to be it
stuck elbow
why is there a default?
fading solstice
thank s thanks.
solar whale
@stuck elbow good question
@fading solstice the build will go in build_metro_m4_express
you can (and were) renaming that via the BUILD parameter
@stuck elbow was taht just a typo you are fixing in the MCP230xx -- looks good to me - do you want me to reviewi it
I can't test it
@stuck elbow never-mind - it's done!
@fading solstice is it working OK now?
fading solstice
@solar whale you are the best. i was blind and you have allowed to see. it works!
solar whale
Excellent! I was really puzzled -- I was going to suggest standing on your left foot next 😉
fading solstice
i might suggest a change to the makefile that requires a BOARD
solar whale
seems reasonable -- with so many boards, it's not clear a default is helpful
tulip sleet
@fading solstice PR's gratefully accepted
tidal kiln
@tulip sleet @slender iron interesting question:
https://forums.adafruit.com/viewtopic.php?f=57&t=137582
what is on the non-express feather M0's? SAM-BA?
tulip sleet
@tidal kiln I think it may just work fine, since uf2 bootloaders work on trinket and gemma. only issue is if the pin mappings are different on Feather M0 Basic vs Feather M0 Express. I'll look.
meager fog
@tulip sleet replied
tulip sleet
@tidal kiln pin mappings are almost the same. no NeoPixel on Feather basic, so that wouldn't work and the USB PID would be wrong, but not a big deal to change.
meager fog
we're putting in some feather m4's today!
i know nobody here wants em
just throws all the feather m4's in the trash
scatters them to the wind
j/k! lots for everyone!
they are being QA'd now 😄
errant grail
Time to spend some $$!
... and just in time. I'm giving a CPy talk at our local makerspace tonight. They'll be excited to hear that it's real!
slender iron
just got my odroid-go 😃
errant grail
Had a late-night idea a few days ago for a CPy-based musical instrument. They just published the learning guide for it today. Woo hoo! https://learn.adafruit.com/blues-playground 
🥁 🎵
meager fog
@errant grail nice work, this was a cool guide!
errant grail
Thanks! I had a lot of fun making it and playing it!
solar whale
@errant grail Amazing. Nice guide.
idle owl
@slender iron Do you want to look over the new driver? The changes I thought needed to be made are in there, it's set to go in my opinion. There's a PR in for it. You don't have to look at it if you're busy, I'll take care of it, but I wanted to offer. https://github.com/adafruit/Adafruit_CircuitPython_MLX90614/pull/6
slender iron
@idle owl I'll take a peek
main meteor
I've been looking forward to a Feather M4. I managed to snatch an ItsyBitsy M4 before they were all gone again (I'm aiming to build a CRT art project out of it, using the 2 DACs for X and Y signals). I can't help wondering if there will be a Trinket M4, or is that just silly.
tidal kiln
@main meteor seen this?
https://learn.adafruit.com/circuit-playground-express-dac-hacks/composite-video
main meteor
Oooo, I hadn't! That's cool! I wonder what the mighty M4 can do with that. I should grab some clip leads and play with it (I don't have the ItsyBitsy M4 in my hot little hand yet, but I have a Metro M4).
meager fog
we've got some more itsy m4s in stock
trinket m4 is a bit unlikely for now since the chip is bga
and i think itsy does a good job
of 'small powerul'
in stock now https://www.adafruit.com/product/3857
@ruby lake ping pong
we'll get more in stock shortly but this is the first bath
ch
errant grail
I used the trinket on quite a few projects where the number of GPIO pins was perfect. Kept things really small. Designed some Trinket-compatible doublers and other PCB interfaces based on the form/factor/pinout. Sad to hear that it may not make it to the next generation, but I'm slowly overcoming that and moving on... 😉
meager fog
well ya never know - its just not a high priority is all 😃
main meteor
I just realized the AdaFruit Discord icons don't include a feather!
errant grail
... the Trinket-compatible Rover family:
errant grail
Trinket OCD
ruby lake
@meager fog Just ordered a couple 😉
meager fog
@ruby lake i should have labeled the D+ D- pads "FOR OLDCROW USAGE ONLY" 😄
ruby lake
"crow's feathers"
meager fog
but i think this design should satisfy your synthneeds - we put as many i2s pins as we could fit 😃
slender iron
@river quest we need a feather emoji
meager fog
we have a feather logo
idle owl
Oh yeah, the drifting feather.
ruby lake
I will work out the next board for it, will be a bit wider
errant grail
@stuck elbow Very cool. With a pressure sensor?
errant grail
I'm thinking about an array of these for a MIDI harmonica: https://www.digikey.com/products/en?keywords=z3639-nd
... with a touch pad array for note bends.
Your ocarina project sounds interesting. I'll be watching for details...
stuck elbow
I was thinking about BMP280
in fact, I should probably start designing the PCB if I want it to arrive next month
do we have usb midi support for circuitpython?
errant grail
That sensor should work nicely. The harmonica design needs 10 independent sensors, though.
meager fog
we're making more feather m4s - sign up if you missed out
errant grail
I've played around with MIDI on CPy a lot. There's nothing native, but CPy is pretty capable. The versions on M4 will be fantastic.
opaque patrol
How long did it take, I was 15 minutes after the email and already gone
errant grail
bummer. I put two in my cart when they were available, was adding other items and the M4s were dropped.
meager fog
n000
errant grail
I'll be patient.
opaque patrol
That happened to me yesterday with the Metro M4
ruby lake
gotta move fast
errant grail
High demand is a good thing.
main meteor
Get your enamal pins fast, too.
opaque patrol
(And the arcade feather 😃 )
stuck elbow
@errant grail I was thinking about just doing pwm, or maybe a bit nicer DAC output, but with midi, I could actually skip the speaker and make it a usb midi device
saves me a pin
ruby lake
I see a ping from ladyada and I don't even look here, I just go see if they are in stock on the site 😉
meager fog
lollll
errant grail
@stuck elbow You can get some nicely natural sounds out of dedicated MIDI voice modules or add-ins that include the ability to respond to expression as well.
meager fog
thaz rite!
stuck elbow
@errant grail but I have no idea about midi, so unless there is something ready, I think I will go for the built-in speaker
errant grail
That's simpler and more portable, too.
ruby lake
n-voice key assigner in cp runs on the m0, m4 will be quick
stuck elbow
yeah, with a coin cell
errant grail
@stuck elbow Have you consider using the wind noise from a microphone? Do you need variable pressure sensing for volume or expression?
errant grail
Indeed.
stuck elbow
to be honest I have no idea how it will all work in the end
it's an experiment
if I wanted something working, I would go and buy https://www.tindie.com/products/yoe/twi-a-teensy-based-usb-midi-woodwind-controller/?pt=ac_prod_search
I actually even have that sensor in my drawer
errant grail
Haven't played with a breath controller (yet). @ruby lake have you?
ruby lake
@meager fog my only remaining experiment will be how to do time division multiplex without a direct hook into a hardware timer. I might be able to use time.monotonic to do a polled loop test, if I can get it to 1ms or better, we win
@errant grail I have, I have yamaha breath controllers from the olden days
they are not EWI, but rather the mouthpiece used with various Yamaha digital synthesizers
idle owl
@slender iron Great catches, thanks for looking at the driver.
stuck elbow
I have never tried any electronic woodwind instruments
I can't even imagine how it feels like
slender iron
@idle owl np
stuck elbow
but I would very much like a woodwind instrument that I can practice with headphones, because my gf hates me when I practice
slender iron
@ruby lake what do you need to do?
ruby lake
@slender iron Ideally, I'd like to start a timer and bind it to a callback that sets my channel output states. That is not a cp thing, at least not yet, so I will have to try grabbing a time.monotonic value and waiting for a certain interval of ticks before running through an iteration of the output channel update.
It should work fine, it is just that I have to be better than 1ms of total time taken or I risk missing a midi byte
slender iron
can you explain what you want to happen without saying how?
I don't know what the output channel is supposed to do
ruby lake
ok, say you have 2 lists, the input list and the output list
every interval (say 500 microseconds for now) a bit of code pulls an item from the input list and moves it to the output list.
the input list is populated based on how the midi keys are assigned. The output list is always written in a linear order. The lists have a certain number of entries (voices), say 8.
slender iron
hrm
ruby lake
The time matters because the output hardware is a series of sample/hold circuits and their output voltages can droop if left unrefreshed for too long
feather-> dual -> mux -> S/Hs
slender iron
what protocol is the output? midi (aka uart)?
ruby lake
er dac not dual
analog voltages. Multiplexed from a single dac channel into eight channels
I will experiment
slender iron
k, sounds like something we could do in C land but still uncertain exactly what it is
ruby lake
well, when I get to my home office I will link my C version
slender iron
k, perfect
plucky flint
@ruby lake o/ I sent you an email earlier today... 😃 Thanks for submitting something.
ruby lake
@plucky flint yep, got it. Will write a proper response when I get to the home office
rigid path
CircuitPython upgrade question: I am currently running CP 2.3.1 on my Feathers and I am looking at testing the version 3 of CP. Do I need to upgrade the uf2/bin files on the individual boards or just the folders/files in the lib folder, or everything? Thanks.
plucky flint
@ruby lake +1
idle owl
@rigid path Everything. You'll need to install a new uf2 for each board, and you'll need the 3.x version of the library files you're using in your /lib folder.
@plucky flint Hey!
idle owl
Pretty good! How about you?
I'm finishing up a guide to using Git and GitHub to contribute to CircuitPython or any open source project. It's massive. I'm super excited about it.
plucky flint
Busy: getting Mu 1.0 ready (beta.17 next week), also some behind the scenes work on CodeGrades which you'll hear about very soon. 😉
idle owl
Excellent!
plucky flint
Oooh... that's a really useful resource. Using GitHub can be v.intimidating.
Happy to proof read / test etc... if you need extra eyes... just assign me on GH (@plucky flint)
idle owl
It came out of us adding Travis to the Learn guide code repo, and forcing people to do PRs instead of pushing directly, but we get this question a lot "what's a good workflow?" And we usually say "whatever works best for you... here's a few ideas" which isn't all that helpful.
That would be great, but none of it's on GitHub. It's all in the Learn system. I can email you a preview link though.
plucky flint
oooh yes please!
plucky flint
I know millions (* may be a slight exaggeration) of teachers in the UK who'd find a guide like yours really useful.
Happy to help! 😃
idle owl
Sent. With caveats included 😄
plucky flint
gotit!
idle owl
Did the link work? I feel like they don't always work
Oh you know what, hold on. I don't think it will work yet
fading solstice
@idle owl also would like to review it, if you want a reviewer that is
idle owl
Ok refresh
@fading solstice That would be amazing. It's a huge guide. I haven't read back through it myself yet, I'm about to start. If you have the time, I'd love to have you review it. https://learn.adafruit.com/git-and-github-workflow/overview?preview_token=w7AAAYjAwqDDoemj6gGAoA
plucky flint
That worked! 😃 Thanks...
fading solstice
@slender iron do I need to do anything else to get my FTC fix included in next build of circuitpython 3.0
solar whale
@tulip sleet my servos don't like 550 -2400 heard snap on a sg-5010 and it does a funny travel at the extreme -- the sg92r vibrates at 2400 seems to work. still testing
slender iron
@fading solstice the submodule commit reference needs to be updated
in your circuitpython copy you can change into the directory and git pull
then back out and you should see a diff you can commit and PR
rigid path
@idle owl Thanks for the info.
solar whale
@tulip sleet its the 2400 that is problematic
tulip sleet
what's a safer value?
tulip sleet
@meager fog @solar whale testing servo limits here
tulip sleet
do you have more than one sample of each servo?
solar whale
I have 2 of each
tulip sleet
I'll try them later. I did try the extremes manually in the REPL and they seemed to be slightly under the extremes of travel (I tried 500 and 2450, which went a little further), but that was mine. Tried mostly micro servo
solar whale
I need to take some time to set this up for systematic testing... but 600 - 2400 caused bad behavior and heating of the sg-5010 -- the sg52 just made some noise - so far.
I think the microservos are more forgiving
stuck elbow
are those analog or digital?
in analog servos, the frequency of the pwm signal determines the power (and thus heating)
tulip sleet
they are just the regular servos from the adafruit store. we're using 50hz.
stuck elbow
so if you feed more than the standard 50Hz, you can overheat them
(I actually use that to power them with lower voltage than they are made for)
tulip sleet
I think @solar whale is seeing overheating at the extremes of travel, when the pulse width is too short or too long and they hit the stop. Does that make sense?
stuck elbow
ah, yes, they really are not made to go beyond 90° travel normally
so around 1000-2000
they have large margins, but they are inconsitent
you have to tune them per servo
solar whale
still not making +/- 90
stuck elbow
no, +/-45°
solar whale
trying for full 180 degree swing
stuck elbow
90° total
solar whale
ah - they do that at 1000-2000
tulip sleet
1000-2000 was the original default, but that is a lot less than 180deg, more like 140. The Tower Pro ones seem more like 550 or 600 to 2300 or 2400 now. Limor said to change the defaults to 550-2400. They seem poorly spec'd: the datasheet (such as it is) doesn't say
stuck elbow
@tulip sleet it really varies per individual servo
solar whale
I wonder if the limits should stay safe at 1000 - 2000 and let the user play beyond if they want to.
slender iron
There is enough clock instability in the M0 that the actual output value can be significantly different
I added a calibration example that allows you to compensate for it
tulip sleet
The motor library assumed 1000-2000 would get you 180 degrees of rotation, but you're saying we should set that to 90 instead
hmm... I should look at the arduino libraries too
@slender iron I fixed OneWire by adding some delay routines that do not use tick_delay(). They are just busy-wait NOP loops. The OneWire read/write routines turn off all interrupts, and then call common_hal_mcu_delay_us(). But that's a bad idea, I believe, because then the SysTick handler doesn't get called. If tick_delay() is called for a delay that will cross a SysTick reset, the timing is thrown off. Couple of ideas:
- Add and call these busy-wait loops (they are calibrated based on M0 48MHz and M4 120 MHz, with cache on) instead.
- Make some interrupt enable/disable routines that don't disable SysTick interrupt. RIght now it is prio 1, same as USB. I did make it prio 0, to see if USB was messing things up. we could disable interrupts below prio 0 for critical sections that need to call delay routines.
I wonder if this is causing other timing issues, like notro's clock drift issues.
slender iron
2 sounds better to me
tulip sleet
yeah. I like it better also. Time must march on no matter what.
I did check calibration of tick_delay() for small us values, checking systick values before and after calling it (compensating for systick reset if necessary). It is actually quite good. 6us does really seem to be 6us most of the time. Occasionally 6, 10,15, or 100us was 1us more. Checked on both M0 and M4
@solar whale @stuck elbow Interesting reading re servos: http://forum.arduino.cc/index.php?topic=48773.msg348806#msg348806
solar whale
I wan to update every thing to latest crikit build - I'm a few days old... occasionally getting an I/0 error (#5)
tulip sleet
and following posts
solar whale
also need to retest SD card on nrf52...
tulip sleet
try this. my PR is not merged. I buitl this test one for Mike B
it has the 550-2400 default range and everything else is very new
tulip sleet
in later post: http://forum.arduino.cc/index.php?topic=48773.msg348809#msg348809 : "Keep in mind that R/C servos were developed for the model industry. They typically use adjustable mechanical linkages from the servo horns (or wheels) to the thing they are actuating, here is a link explaining: http://www.powertorque.co.za/epf/pdf/servoset.pdf . That way they can adapt the servo a given servo rotation to the range of mechanical motion they need. There was never a need for all servos to move the exact same number of degrees per usec. "
solar whale
is the limit the only change from master for the cpx-crikit build?
tulip sleet
master has new nrf-related code
solar whale
feather m4s are back in stock!
tulip sleet
allowing '/' mounted on '/' is the last change to 3.x
solar whale
right -- got all that - I was just wondering what was differnt in your build
stuck elbow
@tulip sleet yeah, I know those, the thing is, even individual servos will have different ranges -- I suspect that's because of variations of components used
tulip sleet
if R/C timing circuits, then component variation would account for a lot. I don't even know what' inside: I haven't seen a schematic for the insides. I thought there are some active components
stuck elbow
@tulip sleet there is a generator of the base frequency, changed with the pot, and a comparator with the input, and the difference drives the motor
tulip sleet
Doing some self-education. http://www.aishack.in/tutorials/servo-motors/ and https://www.voti.nl/docs/M51660L.pdf
was NE544 before M51660
stuck elbow
that's why giving higher frequency signal gives more power
tulip sleet
thanks for all this explanation - it's almost folkloric. people just use them without really knowing
stuck elbow
yeah
to be honest, I have never seen a schematic myself, and even if I saw one now, I wouldn't understand it
of course digital servos just have a microcontroller in them
usually a PIC
tulip sleet
but I am thinking that there were servos way before IC's, so there was a simpler analog circuit. There were RC airplanes way before IC's.
indigo wedge
@slender iron Any suggestions who I should also add to review to offload you from the mountain of my PRs? 😉
manic glacierBOT
slender iron
@indigo wedge I mean to look today. just got distracted
tulip sleet
I am supposed to be doing nRF stuff as well but am finishing up some other stuff.
so i haven't gotten very far in the self-education needed
indigo wedge
No, of course you have every right to get distracted, I just don't want to put too many patches on just you.
slender iron
I can take a look now. ladyada has wandered off 😃
indigo wedge
I tried to make the PRs smaller this time around so they are easier to review one by one
slender iron
np, I'm not the most thorough reviewer. we can always fix bugs after @solar whale finds them 😃
indigo wedge
😄
solar whale
@slender iron I'll be happy to do tome testing of the Blnka stuff on Raspberry pi. any caveats or special docs I should look for?
slender iron
@solar whale @meager fog would know better than I. I'm just packaging it
solar whale
ah -- I just saw a lot of stuff going into it. looks intriguing!
indigo wedge
I actually do wonder a but about Adafruit's process, at work we have a Jenkins slave that has real HW connected to it so we run tests for each PR. Don't think this can be done with Travis. So I what I'm wondering is, for example with the coming 3.0 release, what is the testing process, someone sits with all the boards and runs python scripts with different modules, or is there some automated process, or no process? 😄
stuck elbow
@tulip sleet found the website of Rue who did teardown of some servos: http://ruemohr.org/~ircjunk/tutorials/elex/hobbyservo/servo101.html
slender iron
@solar whale Yup! She wants to get rpi crickit going
solar whale
cool!
stuck elbow
but that looks like the same block diagrams :/
slender iron
@indigo wedge I've spent time on Rosie which gets webhooks from Travis and runs basic stuff on the boards
but its super flaky
we mostly rely on our testing and our community's testing
solar whale
move fast and break things
slender iron
and then fix them 😃
tulip sleet
@smoky swallow https://github.com/adafruit/rosie-ci
indigo wedge
ah you mean https://github.com/adafruit/rosie-ci ? I can have a look, I did a lot of work on the Jenkins config at work 😉
thanks Dan
slender iron
@indigo wedge yup thats it. not related to jenkins at all though. it only knows how to load code through uf2 atm
indigo wedge
yeah, I'm just using Jenkins as a example of CI system, anything that can execute scripts and take commands remotely would do
solar whale
@indigo wedge Does your build have all the new stuff - being merged in now in it -- could that impact SDCard. I can retry after the dust settles.
indigo wedge
No, I tested on what was master as I wrote the comment, which I think was same commit as you used.
solar whale
oh well
stuck elbow
@tulip sleet ok, looks like pre-IC the servos just looked at the energy of the PWM signal, not at the pulse width
indigo wedge
But I'll test on the feather52832 tomorrow, closing shop for today 😄
solar whale
No rush - thanks and good night!
indigo wedge
@slender iron 🚓 Do you know how fast you've been merging, sir?
slender iron
😃
idle owl
Scott's a merging ninja
stuck elbow
@indigo wedge is that a Heisenberg joke?
indigo wedge
Nope, just a police pull over joke
stuck elbow
right and he answers "no, but I know where I am"
slender iron
gave you a comment on the latest 😃
idle owl
@fading solstice Unless you're actively going through the guide, I'm going to submit it. If you're looking at it, I'll wait for your review.
indigo wedge
Oh no I realized I wrote a "Well, actually" response 🤦
slender iron
heh
indigo wedge
I think we will have to come back to the BLE API discussion again, but not today, have a nice TIME_OF_DAY everyone!
slender iron
thanks @indigo wedge !
idle owl
later @indigo wedge!
solar whale
@tulip sleet is there a new release for the seesaw on the crikit that I should load? I'm still getting I/O errors.
tulip sleet
@solar whale no new seesaw release. how are you producing io errors?
I never have seen any
solar whale
moving the servo --
tulip sleet
intermittent or you have a consistent example?
solar whale
the green light on the CRICKIT goes out and I get an Error 5
tulip sleet
what kind of 5v power are you supplying to crickit
tulip sleet
so maybe you are drawing too much power with your testing and activating the electronic fuse
fading solstice
@idle owl no reviewing at this time.
idle owl
@fading solstice Ok np.
tulip sleet
or otherwise causign the power to sag and crickit board to reset
solar whale
even happend at 1000 - 2000 limits
tulip sleet
so you are just running the servo for minutes at 1000-2000?
solar whale
I can switch to batteries.
tulip sleet
many servos or just one
tulip sleet
only if you were hitting the limits maybe
solar whale
I'll keep trying an put in an issue if I can make it reproducible. with 1000 2000 I would not have expected it. I also have a 4A supply - not sure I want to give it too much!
tulip sleet
yes, make an issue, but in https://github.com/adafruit/seesaw, maybe. Dean made some kind of PWM fix a few days ago, not sure what it's about
solar whale
ok -- I'll poke at it a bit more first. Thanks
oh yeah - there was a new release 3 days ago. I'll check the commit
main meteor
The analog out page says "Express boards like the Circuit Playground Express, Metro M0 Express, ItsyBitsy M0 Express, Metro 4 Express, or Feather M0 Express have more code space and can perform audio playback capabilities via the DAC. Gemma M0 and Trinket M0 cannot!" I know the Gemma and Trinket use the E18 version of the chip and the Metro uses the G18, but I thought that was a package difference, not RAM. Then again, the Metro also has the SPI flash chip, maybe that's the reason?
tulip sleet
@main meteor it's the flash chip. we don't have room on the small boards for the code, because the 256kB of internal flash on those boards is also used for CIRCUITPY (a 64kB filesystem). On the express boards, we can use the entire 256kB of flash for code.
only 192kB for code on the non-Express boards
main meteor
I'm still intrigued by the SAMD51's built-in microSD card support. There's room on the Metro for a microSD socket. That would give Circuitpython a ton of room!
I'd be tempted to make my own board and try it (and write the microSD support for Circuitpython), but I'm not set up to solder BGA chips!
tulip sleet
we don't use BGA chips ourselves
main meteor
Whoops, when I asked about an M4 Trinket, LadyAda said it would be tricky because it's BGA.
tulip sleet
you can just use regular SPI to talk to the uSD card
that's what we do now
because only BGA will fit on the narrow width of the trinket. no 32-pin SAMD51
main meteor
Ah, that's the thinking! I had it, um, not backwards, but sort of sideways. Now I understand!
tulip sleet
np, yeah we don't do BGA for anything because it's tricky to manufacture and get good yields
main meteor
I was integrating a resource-hungry peripheral and an ordinary ATmega328 wasn't even close to managing it. The SAMD51 didn't even break a sweat.
solar whale
@tulip sleet I think I am losing it -- whne I command the servo on pin 17 - I am movin servo 4 -- pin 14 povew servo 1 - isn't that backwards?
main meteor
I wondered about that: BGA requires a lot of process support, and while I know there's a bunch of very cool gear in use at AdaFruit, I didn't know if BGA was practical there yet.
tulip sleet
i'm not sure of the manufacturing issues, but also rework is hard. We have a high yield, though, right now
@solar whale that's the numbering I have in the Crickit library. They were numbered in reverse for a while. Are you looking at a guide?
solar whale
ah - yes
tulip sleet
SERVO1 14, SERVO2 15 ,etc
tulip sleet
so SERVO1 corresponds to Servo 1 on the board?
solar whale
guide says OK that was fun but you want MORE servos right? You can control up to four! The servos are on the seesaw pins 17, 16, 15, 14 in that order after demo using 17 for servo 1
tulip sleet
so what worked for you? is 17 terminals 4 or 1 for you
solar whale
now it is terminals 4 -- was 1 before.
tulip sleet
so the guide is wrong, right?
solar whale
is SERVO1 defined somewhare?
tulip sleet
from adafruit_crickit.crickit import crickit
from adafruit_crickit.terminals import *
servo1 = crickit.servo(SERVO1)
servo1.angle = 30
that's how the new library works
solar whale
ah - I am using Seesaw
tulip sleet
the library hides all that boilerplate. It is doing the same thing, but hiding the I2C and seesaw devices (though you need seesaw if you are using the digital I/O pins)
solar whale
tulip sleet
crickit.seesaw is the seesaw object
yes, the library is brand new and not in the guide yet
solar whale
right but teh guide does not use crikit version --- that clarifies things!
tulip sleet
solar whale
thanks!
solar whale
ok - I'll clean things up before generating any issues....hard to keep up!
tulip sleet
hardly anyone has used the library yet - Mike B added as an alternative example in one guide so far
solar whale
I also need to check my path to make sure which version of the libs is getting used but on this system I deleted seesaw from the /lib folder
on CPX's I usually remove all the libs that are "built-in"
to avoid confusion
tulip sleet
@solar whale when you said the servo pin mapping reversed, do you mean you were using 17 for SERVO1 (as labeled on the board) before my new CPy .uf2?
solar whale
yes
tulip sleet
but not today, ok
solar whale
Am I not the only one confused? was there a consious effort to revers the ordering?
tulip sleet
i am really confused. I just tried with the seesaw library (not my library) and 17 is servo4, so how did that work for you before, with 17 as servo1
can we get in voice chat for a minute
raven canopy
just listening for possible future support
meager fog
more feather m4s in the coming days
we sold out fast!
oh wait we ahve 22 left
GET EM WHILE WE GOT EM
idle owl
😄
ruby lake
omg featherzz
tulip sleet
@solar whale ok, turns it that the seesaw library really does do pin mapping, in an obscure way. I am fixing it: 17,16,15,14 is correct. I accidentally changed the mapping and "fixed" it. WIll change it back.
ruby lake
sad we lost Harlan Ellison today. I have a vivid memory of him guest lecturing in my spring 1985 university English course
solar whale
@tulip sleet are you sure it should be changed back. 14-17 makes a lot of sense. It does not matter to me.
tulip sleet
@solar whale the correct numbering is 17-14. I made an accidental change in the seesaw library which made me think it was reversed, so I reversed it in the crickit library as well. Just went over this with Limor
I'll have a new .uf2 for you in a few minutes after I test it
tulip sleet
@solar whale
Adafruit CircuitPython 3.0.0-rc.0-7-g6908eed7e-dirty on 2018-06-28; Adafruit CircuitPlayground Express with Crickit libraries with samd21g18
>>> from adafruit_crickit.crickit import crickit
>>> from adafruit_crickit.terminals import *
>>> SERVO1
17
>>> servo1 = crickit.servo(SERVO1)
>>> servo1.angle = 70
>>> servo1.angle = 0
>>> # works!
@slender iron disabling interrupts below a certain prio level is implemented only on M4, not M0+ 😠 , so looks like I need busy-wait loops at least for M0. Also since no true open-drain outputs, can't use pulseio for OneWire.
slender iron
how did it work before?
tulip sleet
because in 2.3.1 i think the common_hal_mcu_delay_us() impl didn't use systick interrupts, so it busy-waited itself and was not subject to interruption. Disabling interrupts and then doing a delay is not done anywhere except in onewire.
slender iron
ah, because it uses asf?
tulip sleet
no, it's the tick_delay impl; just checking that
common_hal_mcu_delay_us() eventually calls delay_cycles(), which actually just uses SysTick as if no one else is using it (which I think is true on 2.x?)
did you start using SysTick for timekeeping in 3.0?
no, I see we had systick interrupts even in 2.x. But delay_cycles just overrode that and used systick for its own devious purposes.
it would just load it and count down (the delay was broken up into 10us chunks)
i know the busy-wait loops work - i tested that already, so i'll go back to that
main issue is that calibration assumed a fixed cpu freq, I can probably calculate that anyway without losing too much time. tick_delay does such a calculation and it still has ok accuracy even for small times like 6us
solar whale
@tulip sleet Sorry, I'm not going to be able to do any more servo testing tonight. I'll set up some systematic testing tomorrow to check the range and see if the I/O errors are reproducible. Thanks for the help today.
tulip sleet
Thanks!! np, we can always revise the ranges. I really appreciate your looking at this more thoroughly than I did.
slender iron
@tulip sleet yup, systick switch was with 3.x
tulip sleet
how did time.monotonic keep time in 2.x?
ok, I see you were using TC5. So you gained a timer by using systick.
slender iron
and saved a timer peripheral
mental marsh
Hi everyone, can an itsybitsy M4 Express with CircuitPython drive a servo and sound fx board and neopixels or leds?
ruby atlas
@tulip sleet you mentioned doing a cpu frequency calculation. would it be better to store that in board details, or compute something akin to the linux bogomips at cpy startup?
tulip sleet
I don't actually calculate the frequency. I just call common_hal_mcu_processor_get_frequency(), which right now returns a constant (it could look it up in the clock details, but not really necessary now).
ruby atlas
ah okay, using an abstraction like that is ideal anyhow 😃
i missed some context by not looking at code
once my next order arrives i'll have 3 m4 boards 😃
tulip sleet
more to test with 😃
scarlet fjord
what does this mean?
File "main.py", line 22, in <module>
File "adafruit_hid/keyboard.py", line 74, in __init__
File "adafruit_hid/keyboard.py", line 123, in release_all
OSError: USB Busy
line 22 in main.py just does this:
kbd = Keyboard()
and i wasnt having this issue yesterday
i unplugged it and plugged it back in and idk if this is better or worse:
File "main.py", line 35, in <module>
File "adafruit_hid/keyboard.py", line 100, in press
File "adafruit_hid/keyboard.py", line 135, in _add_keycode_to_report
File "adafruit_hid/keycode.py", line 305, in modifier_bit
TypeError: unsupported types for __le__: 'int', 'list'
line 35 in main.py is:
kbd.press(Keycode.A)
whelp, its fixed now, idk how tho
i think im just running out of ram tbh
i need me a feather m4 express, but theyre already out of stock 😭
my M0 has 32kb of ram and the M4 has 192kb!
i needs it
indigo wedge
Nordic has a new dev kit coming in the form of a usb dongle, there is no jlink chip, the usb is connected to the nrf usb peripheral, so it's perfect for running CP with Adafruit bootloader:
https://infocenter.nordicsemi.com/index.jsp?topic=%2Fcom.nordic.infocenter.nrf52%2Fdita%2Fnrf52%2Fdevelopment%2Fnrf52840_dongle%2Fintro.html&cp=2_0_4
Ah also on their website, https://www.nordicsemi.com/eng/Products/nRF52840-Dongle
stuck elbow
@scarlet fjord I was looking at your code
@scarlet fjord I think it will use less ram if you don't use so many global variables, but instead put the values in a list directly
so instead of doing a1 = 1 a2=2 a3=3 and as = [a1, a2, a3] simply do as = [1, 2, 3] if that makes sense
scarlet fjord
that would make a lot more sense yes o:
can i still do that if a1 = [1, 2, 3] and a2 = [1, 2, 3]?
like, is there a way to nest arrays inside arrays in that way?
stuck elbow
like a = [1, [2, 3]]?
scarlet fjord
yeah, does that work..?
stuck elbow
yes
scarlet fjord
cool 😮
stuck elbow
I edited your code, but I didn't try it
scarlet fjord
oh wow, thanks!
ill give it a shot
wow, i see what you mean about too many variables 😅
you made it look so clean
it errors out though:
File "main.py", line 1, in <module>
File "mcpsetup.py", line 7, in <module>
File "adafruit_mcp230xx.py", line 266, in __init__
File "adafruit_bus_device/i2c_device.py", line 62, in __init__
AttributeError: 'tuple' object has no attribute 'try_lock'
you are passing a tuple instead of a i2c object
ah, there is an extra comma in mcpsetup.py
sorry, I added the commas with search-and-replace
three extra commas, in fact
scarlet fjord
ah i see them
ooh ooh there were a few more issues, but the compiler told me about them and i fixed them and guess what ❗
it works 😄
stuck elbow
awesome
scarlet fjord
thanks a bunch 💖
stuck elbow
I'm glad I could be of assistance
tidal kiln
@mental marsh it can. doing them all at once may take a little effort though.
mental marsh
@tidal kiln thanks
tidal kiln
@mental marsh servos: https://learn.adafruit.com/circuitpython-essentials/circuitpython-servo
neopixels: https://learn.adafruit.com/circuitpython-essentials/circuitpython-neopixel
and you'd control the sound fx board using digital io: https://learn.adafruit.com/circuitpython-essentials/circuitpython-digital-in-out
mental marsh
Thank you so much!!
tidal kiln
np. good luck. don't try to do it all at once. work up to it piece by piece.