Conversation
|
The tests failed because of the numpy dependency. I guess I could move the check out of |
|
such a pity numpy is so bloaty and time consuming to just add it to requirements |
I'm not sure that this has a real world impact. I guess most, or at least a significant portion also use input shaping and have it installed anyway. |
|
i think its the sheer time to compile numpy is the issue if the host is sub par, it can take hours on some |
|
I would love to get on my "just require numpy" soap box (Binaries are installed automatically for 32bit, you don't need to compiling it. Its shockingly fast vs plain python. "Your job as a Python programmer is to write C") but... I want to ship this and be 'done' with the project. So what do I have to do? I can push the numpy require check back to the point where its clear that you are configuring a probe. Most of what I'm using it for in this PR ( For probing, I need |
|
Is there any way to provide / use a prebuilt v1.26 wheel? |
|
|
||
| This document describes Klipper's support for load cells and load cell based | ||
| probes. | ||
|
|
There was a problem hiding this comment.
not knowing what a load cell does/means for a 3d printer, could you add a sentence or two describing it's usage/benefits for 3d printers/klipper?
There was a problem hiding this comment.
I made some updates to the documentation including a little bit more exposition at the top and a section about how to read the load cell gram force
garethky
force-pushed
the
pr-load-cell-gram-scale
branch
2 times, most recently
from
a3f6aa5 to
dde606a
Compare
For users: if you use Python 3.0 on a 32 bit RPI OS then PiWheels will just install a binary. That's the default behavior. You have to kinda work hard to not get that to happen. i.e. willfully use Python 2 or pick a 64 bit OS. For the git build system: I'm not sure how that works. |
PyPi has some arm64 wheels as well
There are examples of project who built the necessary wheels as CI actions. E.g. https://github.com/matrix-org/synapse/pull/14212/files Maybe it would be an option to built the needed wheels (Python 2 and Python3?) as Klipper CI and the install scripts (or KIAUH) could automatically draw upon them. |
|
This was updated to remove the numpy dependency |
| # Load Cells | ||
|
|
||
| This document describes Klipper's support for load cells. Basic load cell | ||
| functionality can be used to read force data and to weigh things like filament. |
There was a problem hiding this comment.
Consider adding a note that weighing things is only possible on printers with the load cells in the bed, and not those with the cell in the extruder.
There was a problem hiding this comment.
I don't want to get into applications because there are many. You don't even have to put it in the printer! Several people want to use this functionality to weigh filament spools mounted outside the printer. E.g. https://klipper.discourse.group/t/filament-spool-scale-hx711/19800/
- You could check before a print if there is enough filament on the spool based on the print weight
- Track filament used by weight.
- Track drying if the filament by weight lost
In the future this may be in an application that is just a dry box and not even a printer.
One community member is even working on runout detection in a resin printer.
KevinOConnor
left a comment
KevinOConnor
left a comment
There was a problem hiding this comment.
Thanks. Sorry for the delay in responding.
In general it seems fine to me. I have a few comments and questions - see below.
Cheers,
-Kevin
| ### load_cell/dump_force | ||
|
|
||
| This endpoint is used to subscribe to force data produced by a load_cell. | ||
| Using this endpoint may increase Klipper's system load. |
There was a problem hiding this comment.
It seems odd that there is an additional endpoint added for this information. I would have thought the information could be added to the existing hx71x/dump_hx71x endpoint (or that endpoint replaced with this one).
There was a problem hiding this comment.
Well there are 2 ADC endpoints hx71x/dump_hx71x and ads1220/dump_ads1220.
My goal with this was to try to keep everything "load cell" related out of the ADC code. The load_cell is acting as a decorator to the raw ADC data stream. The raw stream is still useful as a diagnostic tool (or at least it was during development).
This goes with your other question about extending/updating the socket code. What we decide here is going to determine what to do about that comment.
At a high level, what are you thinking about here?
- Is the performance penalty of the duplicate streams a concern?
- Is the duplication of the raw counts data in 2 streams confusing?
- Everything is one stream is desired but then the ADC stream is redundant/confusing?
There was a problem hiding this comment.
A bit of all three. Mostly concerned with maintaining two different interfaces that provide redundant capabilities, and the resulting developer confusion on why there are two redundant interfaces.
Perhaps take a look at how ldc1612.py takes a calibration object and calls self.calibration.apply_calibration(samples), which then allows probe_eddy_current.py to populate a "Z" field in the samples.
There was a problem hiding this comment.
I look at the ldc1612 code and I think the other option is a better fit for this use case. The "that endpoint replaced with this one" idea.
First, I think its more developer friendly to have the endpoint be named according to the load cell in the config and not the sensor implementation. That's a leaky abstraction. That way a front end doesn't need to look up the sensor type in the config and try to build the endpoint URI. The URI can be built from the load cell name alone and its always consistent. Force data comes from a thing that is about force. That feels consistent and logical. (You could argue that LDC1212 is only about Z so that's not necessarily inconsistent)
Second but less important: the 'B' channel on the HX717 is used to measure an RTD and a hall effect filament sensor in Prusa's new machines. I think it would make supporting that kind of thing less awkward later on if we deleted the endpoint from the sensors. Objects that consume the data (LoadCell, Temperature etc.) can transform and publish it (or not) as required.
I'll work on this first.
There was a problem hiding this comment.
less important: the 'B' channel on the HX717 is used to measure an RTD and a hall effect filament sensor in Prusa's new machines
Hey, that is important! ;)
I've currently got a very rough proof of concept implementation for the filament sensor, based on your load-cell-probe-community-testing branch -- ie, using the WebhooksHelper/WebhooksTransformer in this PR. Probably the main feature that would be useful for my code is being able to subscribe to a particular channel only. At the moment my code has 3 or 4 places all receiving the whole data stream and doing their own filtering on it.
In day-to-day usage the raw counts data is not that useful, especially to the user -- dropping the raw sensor API endpoint seems fine. However, if a load cell ADC were to be reused by existing modules that currently just rely on regular ADC pins (gcode_button, hall_filament_width_sensor), the user will need access to the raw counts or the normalized 0..1 value somehow during the setup process in order to calibrate their new gadgets, since they won't necessarily have specific commands like in this PR. For the filament sensors it doesn't have to be the existing endpoint, though. The values change dramatically enough that getting even just one data point each with filament inserted and not inserted would allow for a reliable configuration.
There was a problem hiding this comment.
Updated to not have the sensors register websocket endpoints.
There is still a class for handling both internal and websocket clients: ApiClientHelper. I need it in subsequent PR to emit events for the probe. If that looks like a class that should be in webhooks.py let me know.
For now I've kept things in separate commits so the changes are easier to see. But my intention is to squash them before its merged.
| `{"id": 123,"result":{"header":["time", "force (g)", "counts", "tare_counts"]}}` | ||
| and might later produce asynchronous messages such as: | ||
| `{"params":{"data":[[3292.432935, 40.65, 562534, -234467]]}}` |
There was a problem hiding this comment.
Just out of curiosity, does "tare_counts" change with each sample?
There was a problem hiding this comment.
No, it only changes when the tare command is executed.
There was a problem hiding this comment.
Okay - seems a little odd to repeat this value for every sample. Maybe just export the information via get_status()?
There was a problem hiding this comment.
The load cell probe implementation currently re-tares before homing, so API consumers may find this valuable over time?
There was a problem hiding this comment.
Because a Tare can happen in the middle of a batch of results you wouldn't know which samples used what tare value. A UI would not correctly reflect whats really going on. And as pointed out, tare can happen frequently while probing.
There was a problem hiding this comment.
Correction: it cant happen in the middle of a batch. The tare value would always apply to a whole batch.
docs/Config_Reference.md
Outdated
| #reverse: | ||
| # Reverses the polarity of the load cell. This is a boolean value, the | ||
| # default is False. |
There was a problem hiding this comment.
Out of curiosity, is "reverse" effectively the same as a negative "counts_per_gram"?
There was a problem hiding this comment.
Basically. This lets you do what you wanted, where if the force graph is naturally negative you can reverse its polarity.
| ### LOAD_CELL_DIAGNOSTIC | ||
| `LOAD_CELL_DIAGNOSTIC [LOAD_CELL=<config_name>]`: This command collects 10 | ||
| seconds of load cell data and reports statistics that can help you verify proper | ||
| operation of the load cell. This command can be run on both calibrated and | ||
| uncalibrated load cells. | ||
|
|
||
| ### CALIBRATE_LOAD_CELL | ||
| `CALIBRATE_LOAD_CELL [LOAD_CELL=<config_name>]`: Start the guided calibration | ||
| utility. Calibration is a 3 step process: | ||
| 1. First you remove all load from the load cell and run the `TARE` command | ||
| 1. Next you apply a known load to the load cell and run the | ||
| `CALIBRATE GRAMS=nnn` command | ||
| 1. Finally use the `ACCEPT` command to save the results | ||
|
|
||
| You can cancel the calibration process at any time with `ABORT`. | ||
|
|
||
| ### TARE_LOAD_CELL | ||
| `TARE_LOAD_CELL [LOAD_CELL=<config_name>]`: This works just like the tare button | ||
| on digital scale. It sets the current raw reading of the load cell to be the | ||
| zero point reference value. The response is the percentage of the sensors range | ||
| that was read and the raw value in counts. | ||
|
|
||
| ### READ_LOAD_CELL load_cell="name" | ||
| `READ_LOAD_CELL [LOAD_CELL=<config_name>]`: | ||
| This command takes a reading from the load cell. The response is the percentage | ||
| of the sensors range that was read and the raw value in counts. If the load cell | ||
| is calibrated a force in grams is also reported. |
There was a problem hiding this comment.
As a general comment, it's generally preferable to implement less commands. For example, could this be reduced to just two commands: LOAD_CELL_READ (for reading and diagnostics) and LOAD_CELL_CALIBRATE (for calibrating and changing the tare).
There was a problem hiding this comment.
Ok, I can make those changes.
There was a problem hiding this comment.
Does the gcode parser work if you dont assign a value to things. Like, could this work:
LOAD_CELL READ
LOAD_CELL TARE
LOAD_CELL DIAGNOSTIC
There was a problem hiding this comment.
This does not work, it results in a !! Malformed command error. I don't see anywhere in the existing GCode command that tries to make 1 command perform different actions. The only way this would work would be:
LOAD_CELL READ=TRUE
LOAD_CELL TARE=TRUE
LOAD_CELL DIAGNOSTIC=TRUE
That looks horrible and is hard to use/discover in the front ends. I'd like to stick with separate commands. I have no problem prefixing them all with LOAD_CELL_, that actually improves discoverability. Looking at Bed Mesh, Angle, Accelerometer etc, that's the pattern they all follow.
There was a problem hiding this comment.
Updated commands to use a consistent prefix LOAD_CELL_
klippy/extras/load_cell.py
Outdated
| # Adapter for WebhooksHelper that transforms the response using a function | ||
| # Anything that implements the add_client contract can be a message source | ||
| # Outputs to its own clients | ||
| class WebhooksTransformer(WebhooksHelper): |
There was a problem hiding this comment.
Can you elaborate on what this does (as well as WebhooksHelper)?
At first glance, it seems a little odd to make a slightly different interface for just loadcell. If the BatchWebhooksClient code isn't flexible enough, maybe we should expand the webhooks interface and refactor the other users (like ldc1612 and accelerometers).
There was a problem hiding this comment.
This goes back to your first comment, lets resolve what we want for a data stream structure. If this is still required I can do a refactor.
| self.printer.send_event("load_cell:calibrate", self) | ||
| if self.is_tared(): | ||
| self.printer.send_event("load_cell:tare", self) |
There was a problem hiding this comment.
Can you elaborate on what these events will be used for?
There was a problem hiding this comment.
The probing code uses these events to get notified about the state of the gram scale. When the user interactively calibrates the load_cell for the first time it goes from uncalibrated -> calibrated. That state transition allows probing to work without a restart. Before that it will throw an error that you are not allowed to probe or home with a load cell that isn't calibrated.
I still have the probing code in a separate file because I believe this chunk of functionality is a nice cleanly separate unit. My intention is to pitch you on keeping them split up. These events are one part of the decoupling that makes that work.
|
Thank you for your contribution to Klipper. Unfortunately, a reviewer has not assigned themselves to this GitHub Pull Request. All Pull Requests are reviewed before merging, and a reviewer will need to volunteer. Further information is available at: https://www.klipper3d.org/CONTRIBUTING.html There are some steps that you can take now:
Unfortunately, if a reviewer does not assign themselves to this GitHub Pull Request then it will be automatically closed. If this happens, then it is a good idea to move further discussion to the Klipper Discourse server. Reviewers can reach out on that forum to let you know if they are interested and when they are available. Best regards, PS: I'm just an automated script, not a human being. |
garethky
force-pushed
the
pr-load-cell-gram-scale
branch
5 times, most recently
from
d12b00c to
b0533b9
Compare
|
Okay, thanks. In general it seems fine to me. The only thing I noticed is that docs/API_Server.md should be updated now that the hx71x/ads1220 api is no longer available, and Config_Changes.md should have a note that the previous API endpoints were removed. As for the commands, I was thinking something like FWIW, the "reverse" config option seems a little odd to me, though I'm not sure if a negative counts_per_gram would be better. It's not a big deal either way. Note that the added docs have described "reverse" twice. Cheers, |
|
Thanks for taking a look @KevinOConnor
🍻 |
Thanks.
Okay.
Okay. Maybe a longer name may help - for example: Cheers, |
|
Just checked it. Appears to be working with hx711 — alas the calibration doesn't fix the sign for some reason. |
Can you explain what you mean? Seems like are reporting a bug? |
* Convert sensor counts to grams and make this available via unix socket and object status * Basic GCodes for tearing and reading the load cell * Guided Calibration * Diagnostic gcode to check the health of the load cell Signed-off-by: Gareth Farrington <gareth@waves.ky> Refactor - remove WebhooksHelper and WebhooksTransformer Signed-off-by: Gareth Farrington <gareth@waves.ky> Capture errors and overflows in LoadCellSampleCollector Signed-off-by: Gareth Farrington <gareth@waves.ky> Refactor GCode Commands
garethky
force-pushed
the
pr-load-cell-gram-scale
branch
2 times, most recently
from
759290d to
fd9e9f9
Compare
I think this is ready to merge. |
|
Okay, thanks. I'll give a few days to see if there are comments and otherwise look to commit. I must admit, I got quite a chuckle out of that default. Separately, if Load_Cell.md is a new document, it should also be added to Cheers, |
* Add API server load_cell/dump_force endpoint * Update [load_cell] config with calibration fields * Add G-Code commands for working with load cells * Add status reference for load_cell objects Signed-off-by: Gareth Farrington <gareth@waves.ky>
garethky
force-pushed
the
pr-load-cell-gram-scale
branch
from
fd9e9f9 to
e03740f
Compare
😂 I'm doing too many things today! Its 'normal'. fixed mkdocs.yml |
All of our measurements displayed going into negative values, yet inversion wasn't reflected in the final results, nor was asked to be attributed in config during calibration itself |
In the calibration tool this is the expected output. Everything there that is negative is the raw counts from the sensor. The only measurement that gets inverted is the grams when reading the load cell. |
The problem is that after the calibration it shows negative values, cause the fact that counts per gram is negative is not saved) |
Yes. If the wiring is 'backwards' it will always show negative grams, unless you go into the config and set |
yeah, but if in case of steppers there is no automatic calibration, here code can actually deduce your orientation during calibration process maybe this is actually good, and is an additional protection from assembly mistakes or something like that — but I still thought it was worth asking about |
|
Thanks. -Kevin |
* stm32: Add support for USART6 on STM32F401 STM32F401 has USART6 on PA12/PA11 and PC7/PC6 with alternate function mapping AF08. This can be used, for example, to connect to the Elegoo Neptune 3, where PA12/PA11 are wired to an RJ10 plug going to the stock screen. Signed-off-by: Marius Petcu <marius@petcu.me> * gcode_macro: Expand template syntax errors (Klipper3d#6839) Signed-off-by: Timofey Titovets <nefelim4ag@gmail.com> * axis_twist_compensation: Remove the auto parameter from axis_twist_compensation Signed-off-by: Jorge Apaza Merma <yochiwarez@gmail.com> * docs: Note AXIS_TWIST_COMPENSATION_CALIBRATE AUTO removal in Config_Changes.md Signed-off-by: Kevin O'Connor <kevin@koconnor.net> * input_shaper: Fix for polar kinematics Forward post_cb calls from itersolve to the original kinematics. Signed-off-by: Dmitry Butyugin <dmbutyugin@google.com> * axis_twist_compensation: allow compensating both axis at once Restores the behavior before Klipper3d#6739 since people seemed to rely on it, even if the math is not exact. Signed-off-by: Philippe Daouadi <philippe@ud2.org> * load_cell: Load cell gram scale (Klipper3d#6729) * Add gram scale features to load_cell * Convert sensor counts to grams and make this available via unix socket and object status * Basic GCodes for tearing and reading the load cell * Guided Calibration * Diagnostic gcode to check the health of the load cell * Update load_cell Documentation * Add API server load_cell/dump_force endpoint * Update [load_cell] config with calibration fields * Add G-Code commands for working with load cells * Add status reference for load_cell objects Signed-off-by: Gareth Farrington <gareth@waves.ky> * buttons: Debounce gcode_button and filament_switch_sensor (Klipper3d#6848) Add `debounce_delay` config option which sets the debounce time, defaults to 0 Signed-off-by: Gareth Farrington <gareth@waves.ky> * stepper: Support step on both edges with custom minimum pulse duration Add support for "step on both edges" to the main stepper_event_full() code. This makes that mode of operation available even when the micro-controller is not compiled for "optimized step on both edges". It also enables the custom pulse duration support (step_pulse_ticks) when in "step on both edges" mode. Signed-off-by: Kevin O'Connor <kevin@koconnor.net> * stepper: Support disabling optimized "step on both edges" in "make menuconfig" Add a new "low level option" to allow users to configure if they want to optimize for Trinamic drivers or traditional stepper motor drivers. Signed-off-by: Kevin O'Connor <kevin@koconnor.net> * buttons: fixes incorrect parameters Signed-off-by: Pedro Lamas <pedrolamas@gmail.com> * config: Add LED definitions to generic-bigtreetech-skr-mini-mz.cfg Signed-off-by: Kevin O'Connor <kevin@koconnor.net> * fan_generic: Fix handling of template rendering errors Make sure to assign 'value' on a rendering error to avoid an internal error. Signed-off-by: Kevin O'Connor <kevin@koconnor.net> * output_pin: Fix handling of template rendering errors Make sure to assign 'value' on a rendering error to avoid an internal error. Signed-off-by: Kevin O'Connor <kevin@koconnor.net> * rp2040: define spi bus on pins 12,11,10 Mellow FLY SHT36 Pro toolboard uses those pins Signed-off-by: Timofey Titovets <nefelim4ag@gmail.com> * stm32: Added PH13/14 CAN pin option for stm32h743 (Klipper3d#6857) Added the option to select PH13/PH14 as CAN pins. Signed-off-by: Christoph Frei <fryakatkop@gmail.com> * ldc_1612: Supports configurable external crystal frequency (Klipper3d#6734) You can use the 40Mhz crystal oscillator recommended by TI official manual to get the best performance. refer to: [ldc1612.pdf](https://www.ti.com/cn/lit/ds/symlink/ldc1612.pdf) 7.3.4 Signed-off-by: Xiaokui Zhao <xiaok@zxkxzk.cn> * docs: Fix typos in installation.md Fixed typos, hyphenation, and minor phrasing for better readability. Signed-off-by: Tobias Rumiz <TobiasRumiz@gmail.com> * docs: Fix typo in docs generation documentation Signed-off-by: Russell Cloran <rcloran@gmail.com> * docs: Fix link syntax typo for bed_screws Signed-off-by: Russell Cloran <rcloran@gmail.com> * stm32: Fix RESERVE_PINS_CAN pin ordering in fdcan.c Always report the reserved pins in the same order (rx,tx). Signed-off-by: Kevin O'Connor <kevin@koconnor.net> * stm32: Add support for additional i2c bus Signed-off-by: Russell Cloran <rcloran@gmail.com> * stm32: Turn on can.c error interrupts It seems both ERRIE and LECIE must be enabled to get hardware error interrupts. Without this, the rx_error and tx_error reports are likely to always be zero. Signed-off-by: Kevin O'Connor <kevin@koconnor.net> * output_pin: Make it possible to assign dicts/lists as template parameters The output_pin template code has a cache to speed up duplicate rendering of templates. However, this cache doesn't work if one of the parameters is a Python list or dictionary. Just disable the cache in this case. Signed-off-by: Kevin O'Connor <kevin@koconnor.net> * mkdocs-requirements: Update to Jinja 3.1.6 A security vulnerability was found in Jinja 3.1.5 . The software is not impacted by this vulnerability, but there is no harm in updating to the fixed version. Signed-off-by: Kevin O'Connor <kevin@koconnor.net> * icm20948: Fix sample rate and accels scale selection To set a value in the SET_ACCEL_CONFIG register, you must first go to BANK_2. Signed-off-by: Maksim Bolgov maksim8024@gmail.com * icm20948: Transition from 8g to 16g accels scale During standard resonance measurements, the icm20948 in 8g mode may reach the accels max threshold. Signed-off-by: Maksim Bolgov maksim8024@gmail.com * icm20948: Formatting refactor Signed-off-by: Maksim Bolgov maksim8024@gmail.com * adxl345: Allow read and write 127 register address icm20948 accelerometer has an ACCEL_CONFIG register at address 127 Signed-off-by: Maksim Bolgov maksim8024@gmail.com * docs: Add icm20948 description Signed-off-by: Maksim Bolgov maksim8024@gmail.com * temperature_combined: delay initialization Signed-off-by: Timofey Titovets <nefelim4ag@gmail.com> * temperature_combined: avoid crash with temperature monitors Signed-off-by: Timofey Titovets <nefelim4ag@gmail.com> * axis_twist_compensation: Fix AttributeError on klippy connect state (Klipper3d#6881) Object 'configfile' has no attribute 'error' Signed-off-by: Maksim Bolgov <maksim8024@gmail.com> * docs: Fixup G-Codes.md so that sections are sorted alphabetically Signed-off-by: Kevin O'Connor <kevin@koconnor.net> * i2c_software: Fix i2c_delay() The i2c_delay() function did not properly handle counter rollovers. It also performed an expensive run-time divide. Signed-off-by: Kevin O'Connor <kevin@koconnor.net> * force_move: Support a SET_HOMED parameter to SET_KINEMATIC_POSITION Commit 7083879 added support for clearing the homing state in SET_KINEMATIC_POSITION commands. However, it can be difficult to use that support as the default for SET_KINEMATIC_POSITION is to set all axes as homed. Add a new SET_HOMED parameter to allow one to explicitly request which axes to consider in a homed state. Also introduce a CLEAR_HOMED parameter and prefer that to the existing CLEAR parameter. Signed-off-by: Kevin O'Connor <kevin@koconnor.net> * force_move: Make sure to use lower() on SET_KINEMATIC_POSITION CLEAR_HOMED Signed-off-by: Kevin O'Connor <kevin@koconnor.net> * ci-install: Use prebuilt pru gcc binaries Don't build the pru binaries directly in the build test cases, instead use the upstream binaries provided. Signed-off-by: Kevin O'Connor <kevin@koconnor.net> * led: Fix off-by-one bug in SET_LED_TEMPLATE INDEX parameter Signed-off-by: Kevin O'Connor <kevin@koconnor.net> * docs: Update Features.md to reflect recent work Signed-off-by: Kevin O'Connor <kevin@koconnor.net> * docs: Note the release of v0.13.0 Signed-off-by: Kevin O'Connor <kevin@koconnor.net> * Makefile: Don't disable gcc's use-linker-plugin option This option seems to be confusing ld's region usage checks (builds that could fit in small chips are being reported as not fitting). The option was disabled back in commit 4e8674d because it showed worse results. However, recent versions of gcc seem to produce the same results even if this option is enabled, so change the build to avoid disabling that option on newer versions of gcc (those that have the -ffat-lto-objects option - which is needed to ensure compile_time_requests sections can be extracted with objcopy). The PRU build is dependent on -fuse-linker-plugin, so enable that option explicitly in its build. Signed-off-by: Kevin O'Connor <kevin@koconnor.net> * test: Disable optional features in atmega328 build Signed-off-by: Kevin O'Connor <kevin@koconnor.net> * workflows: Update github build-test.yaml to ubuntu-22.04 Signed-off-by: Kevin O'Connor <kevin@koconnor.net> * Kconfig: Replace WANT_GPIO_BITBANGING with individual options Support setting individual options instead of one global option ( WANT_BUTTONS, WANT_TMCUART, WANT_NEOPIXEL, WANT_PULSE_COUNTER, WANT_HX71X). Signed-off-by: Kevin O'Connor <kevin@koconnor.net> * Kconfig: Replace WANT_DISPLAYS with individual options Support setting WANT_ST7920 and WANT_HD44780 individually. Signed-off-by: Kevin O'Connor <kevin@koconnor.net> * avr: Enable gcc -Os option on CONFIG_HAVE_LIMITED_CODE_SIZE Signed-off-by: Kevin O'Connor <kevin@koconnor.net> * stm32: Enable gcc -Os option on CONFIG_HAVE_LIMITED_CODE_SIZE Signed-off-by: Kevin O'Connor <kevin@koconnor.net> * Kconfig: Add new WANT_HARD_PWM option to reduce code size Make it possible to not compile in support for hardware pwm on chips with small flash sizes. Signed-off-by: Kevin O'Connor <kevin@koconnor.net> * Kconfig: Add new WANT_SPI option to reduce code size Make it possible to not compile in support for SPI on chips with small flash sizes. Signed-off-by: Kevin O'Connor <kevin@koconnor.net> * Kconfig: Add new WANT_I2C option to reduce code size Make it possible to not compile in support for I2C on chips with small flash sizes. Signed-off-by: Kevin O'Connor <kevin@koconnor.net> * Kconfig: Add new WANT_ADC option to reduce code size Make it possible to not compile in support for ADC on chips with small flash sizes. Signed-off-by: Kevin O'Connor <kevin@koconnor.net> * stm32: Simplify Makefile Breakout selection of timer and gpioperiph objects. Signed-off-by: Kevin O'Connor <kevin@koconnor.net> * Kconfig: Add some user visible comments to the optional features menu Signed-off-by: Kevin O'Connor <kevin@koconnor.net> * atsamd: Enable HAVE_LIMITED_CODE_SIZE on small atsamd chips Signed-off-by: Kevin O'Connor <kevin@koconnor.net> * config: Update generic-bigtreetech-skr-2.cfg - SPI Drivers (Klipper3d#6895) Added SPI tmc2130 driver config Signed-off-by: James Hartley <james@hartleyns.com> * test: Disable all additional features in atmega328 build Signed-off-by: Kevin O'Connor <kevin@koconnor.net> * Revert "Makefile: Don't disable gcc's use-linker-plugin option" This reverts commit 8087200. The change can break the build on some versions of gcc. Signed-off-by: Kevin O'Connor <kevin@koconnor.net> * spi_software: respect expected rate On fast MCU software spi may violate maximally supported by TMC driver rate. Add dynamic limits to overcome that. Signed-off-by: Timofey Titovets <nefelim4ag@gmail.com> * software_spi: set rate limiting ticks from the host Signed-off-by: Timofey Titovets <nefelim4ag@gmail.com> * i2c_software: allow freq adjust Signed-off-by: Timofey Titovets <nefelim4ag@gmail.com> * i2c_software: reduce gpio calls count gpio reset calls are heavy. gpio state are persistent between calls. Drop useless calls. Avoid calls if SDA does not changed. Signed-off-by: Timofey Titovets <nefelim4ag@gmail.com> * i2c_software: pass pulse ticks from host Signed-off-by: Timofey Titovets <nefelim4ag@gmail.com> * tmc: add missing freewheel config options Signed-off-by: Timofey Titovets <nefelim4ag@gmail.com> * hall_filament_width_sensor: Add filament switch values to status `hall_filament_width_sensor` contains a runout sensor object internally. This exposes those values in the API status result. ``` SEND: {"id":123,"method":"objects/query","params":{"objects":{"hall_filament_width_sensor":["enabled","filament_detected","is_active","Diameter","Raw"]}}} GOT: b'{"id":123,"result":{"eventtime":199567.823596603,"status":{"hall_filament_width_sensor":{"enabled":true,"filament_detected":true,"is_active":true,"Diameter":1.9499999999999986,"Raw":6113}}}}' ``` The duplication of `is_active` and `enabled` seems confusing, but both of these can be independently manipulated by GCode: ``` SEND: {"id":123,"method":"gcode/script","params":{"script":"DISABLE_FILAMENT_WIDTH_SENSOR"}} GOT: b'{"id":123,"result":{}}' SEND: {"id":123,"method":"objects/query","params":{"objects":{"hall_filament_width_sensor":["enabled","is_active"]}}} GOT: b'{"id":123,"result":{"eventtime":199770.446013297,"status":{"hall_filament_width_sensor":{"enabled":true,"is_active":false}}}}' SEND: {"id":123,"method":"gcode/script","params":{"script":"SET_FILAMENT_SENSOR SENSOR=hall_filament_width_sensor ENABLE=0"}} GOT: b'{"id":123,"result":{}}' SEND: {"id":123,"method":"objects/query","params":{"objects":{"hall_filament_width_sensor":["enabled","is_active"]}}} GOT: b'{"id":123,"result":{"eventtime":199847.927726196,"status":{"hall_filament_width_sensor":{"enabled":false,"is_active":false}}}}' ``` Signed-off-by: Russell Cloran <rcloran@gmail.com> * docs: Update Config_Changes.md to note software spi is now rate limited Signed-off-by: Kevin O'Connor <kevin@koconnor.net> * stm32: Fix prescaler overflow check in hard_pwm.c Signed-off-by: Kevin O'Connor <kevin@koconnor.net> * stm32: Change hard_pwm.c MAX_PWM to 257 Choose a value for MAX_PWM that avoids an expensive run-time division. Signed-off-by: Kevin O'Connor <kevin@koconnor.net> * lcd_hd44780: Make sure nsecs_to_ticks() is always inlined It is a compile-time calculation that needs to be inlined to work. Signed-off-by: Kevin O'Connor <kevin@koconnor.net> * lcd_st7920: Make sure nsecs_to_ticks() is always inlined It is a compile-time calculation that needs to be inlined to work. Signed-off-by: Kevin O'Connor <kevin@koconnor.net> * neopixel: Make sure nsecs_to_ticks() is always inlined It is a compile-time calculation that needs to be inlined to work. Signed-off-by: Kevin O'Connor <kevin@koconnor.net> * Kconfig: Note which chips require software divide operations Add a new HAVE_SOFTWARE_DIVIDE_REQUIRED that indicates which chips require software divide. Signed-off-by: Kevin O'Connor <kevin@koconnor.net> * check-software-div: Add a new build check for software divide Update the build checks to include a check for unexpected software divide operations. Signed-off-by: Kevin O'Connor <kevin@koconnor.net> * neopixel: Add comments on timing Signed-off-by: Kevin O'Connor <kevin@koconnor.net> * toolhead: Avoid LookAheadQueue calling back into toolhead class Avoid lookahead.flush() calling back into toolhead._process_moves(). Instead, rename toolhead._process_moves() to toolhead._process_lookahead(), have it call lookahead.flush(), and consistently use it when flushing the lookahead queue. Signed-off-by: Kevin O'Connor <kevin@koconnor.net> * toolhead: Avoid toolhead.move() and toolhead._process_moves() in drip_move() Implement move checking and trapq loading directly from drip_move(). This simplifies the interactions between these components. Signed-off-by: Kevin O'Connor <kevin@koconnor.net> * manual_stepper: Implement "drip moves" for manual stepper STOP_ON_ENDSTOP Currently, `MANUAL_STEPPER STOP_ON_ENDSTOP=1` type commands will move until hitting the endstop, but it will still always consume the total amount of move time. That is, following moves can't be started until the total possible time of the homing move is completed. Implement "drip moves" so that the code only schedules the movement in small segments. This allows following movements to be scheduled without a significant delay. Signed-off-by: Kevin O'Connor <kevin@koconnor.net> * probe: Add a default probing_move() function to HomingViaProbeHelper Signed-off-by: Kevin O'Connor <kevin@koconnor.net> * probe: Add a new ProbeEndstopSessionHelper class Move the HomingViaProbeHelper() instance from ProbeSessionHelper to a new ProbeEndstopSessionHelper class. Signed-off-by: Kevin O'Connor <kevin@koconnor.net> * probe: Add a new lookup_minimum_z() helper function Signed-off-by: Kevin O'Connor <kevin@koconnor.net> * probe: Add a new ProbeParameterHelper class Split multi-sample config reading from ProbeSessionHelper to a new ProbeParameterHelper class. Signed-off-by: Kevin O'Connor <kevin@koconnor.net> * probe: Change probing_move() to pass a gcmd instead of (pos, speed) Signed-off-by: Kevin O'Connor <kevin@koconnor.net> * probe: Convert probing_move() callback to use regular probe sessions system Use the normal probe_session_start(), run_probe(), pull_probed_results(), and end_probe_session() API from ProbeSessionHelper. This removes the custom probing_move() callback. Signed-off-by: Kevin O'Connor <kevin@koconnor.net> * probe: Add a new LookupZSteppers helper class Split code to lookup the Z stepper from HomingViaProbeHelper to new LookupZSteppers class. Signed-off-by: Kevin O'Connor <kevin@koconnor.net> * probe_eddy_current: Do not support QUERY_PROBE command Report an error if a user issues a QUERY_PROBE command (instead of always returning not-triggered). Signed-off-by: Kevin O'Connor <kevin@koconnor.net> * probe_eddy_current: Separate probe style commands from homing operations Separate homing operations (as called from probe:z_virtual_endstop) from the normal probe command handling. Signed-off-by: Kevin O'Connor <kevin@koconnor.net> * probe: Remove ProbeEndstopSessionHelper Have all callers instantiate the individual helper classes directly. Signed-off-by: Kevin O'Connor <kevin@koconnor.net> --------- Signed-off-by: Marius Petcu <marius@petcu.me> Signed-off-by: Timofey Titovets <nefelim4ag@gmail.com> Signed-off-by: Jorge Apaza Merma <yochiwarez@gmail.com> Signed-off-by: Kevin O'Connor <kevin@koconnor.net> Signed-off-by: Dmitry Butyugin <dmbutyugin@google.com> Signed-off-by: Philippe Daouadi <philippe@ud2.org> Signed-off-by: Gareth Farrington <gareth@waves.ky> Signed-off-by: Pedro Lamas <pedrolamas@gmail.com> Signed-off-by: Christoph Frei <fryakatkop@gmail.com> Signed-off-by: Xiaokui Zhao <xiaok@zxkxzk.cn> Signed-off-by: Tobias Rumiz <TobiasRumiz@gmail.com> Signed-off-by: Russell Cloran <rcloran@gmail.com> Signed-off-by: Maksim Bolgov maksim8024@gmail.com Signed-off-by: Maksim Bolgov <maksim8024@gmail.com> Signed-off-by: James Hartley <james@hartleyns.com> Co-authored-by: Marius Petcu <marius@petcu.me> Co-authored-by: Timofey Titovets <nefelim4ag@gmail.com> Co-authored-by: yochiwarez <yochiwarez@gmail.com> Co-authored-by: Kevin O'Connor <kevin@koconnor.net> Co-authored-by: Dmitry Butyugin <dmbutyugin@google.com> Co-authored-by: Philippe Daouadi <philippe@ud2.org> Co-authored-by: Gareth Farrington <gareth@waves.ky> Co-authored-by: Pedro Lamas <pedrolamas@gmail.com> Co-authored-by: FrY Sennberg <fryakatkop@gmail.com> Co-authored-by: XiaoK <xiaok@zxkxz.cn> Co-authored-by: Tobias Rumiz <63514002+Ruminini@users.noreply.github.com> Co-authored-by: Russell Cloran <rcloran@gmail.com> Co-authored-by: MRX8024 <57844100+MRX8024@users.noreply.github.com> Co-authored-by: JamesH1978 <87171443+JamesH1978@users.noreply.github.com>
* Add gram scale features to load_cell * Convert sensor counts to grams and make this available via unix socket and object status * Basic GCodes for tearing and reading the load cell * Guided Calibration * Diagnostic gcode to check the health of the load cell * Update load_cell Documentation * Add API server load_cell/dump_force endpoint * Update [load_cell] config with calibration fields * Add G-Code commands for working with load cells * Add status reference for load_cell objects Signed-off-by: Gareth Farrington <gareth@waves.ky> Co-authored-by: Gareth Farrington <gareth@waves.ky>
8 participants
This PR adds the features needed to make a load cell actually work as a gram scale. This PR will let you weight filament or measure a force in grams. (no homing or probing yet!)
countsto grams and make this force data available via unix socket and printer object statusThis code has been tested by several people in the community and some last minute bug fixes and changes were landed:
reverseoption allows the polarity of the force readings to be inverted. You can have your probing collision graphs in either orientation now.reference_tare_countsis written to config and read as an integer. linkCALIBRATEwas used beforeTAREA lot of work (months) and bug fixing went into parts of this, particularly
LOAD_CELL_DIAGNOSTICand theLoadCellSampleCollector. These bits have to work when the sensor is buggy and still produce usable output.LoadCellSampleCollectoralso underpins later work for probing.numpyis required for this PR. We could change that. But PR's for probing will absolutely need it. If you want to merge[load_cell]and[load_cell_probe]then it seems there is no point in trying to keep it out of this PR.