Add stm32f1 FAST_PWM feature #19952

victorrar · 2020-10-29T20:48:28Z

Description

Implement stm32f1 FAST_PWM feature

Benefits

Stm32f1 FAST_PWM feature

Configurations

Configuration.h
Any with enabled #define FAST_PWM_FAN

Configuration_adv.h
Tested frequencies from 1 Hz to 60kHz
#define FAST_PWM_FAN_FREQUENCY 31400

sjasonsmith · 2020-10-29T21:01:37Z

Marlin/src/HAL/STM32F1/fast_pwm.cpp

+    //protect used timers
+    if(timer == get_timer_dev(STEP_TIMER_NUM)) return;
+    if(timer == get_timer_dev(PULSE_TIMER_NUM)) return;
+    if(timer == get_timer_dev(TEMP_TIMER_NUM)) return;


PULSE timer doesn't really exist, it is just the STEP_TIMER on this platform.

There are more timers than this which could be interfered with:

Servo Timer

Tone Timer

Software Serial Timer

Timers connected to SPI pins

The SPI bit might seem weird, but I believe on STM32F1 chips pins are not actually disconnected from the timers even if you are using their other functions. This means that running the timer connected to SPI interferes with SPI, even if you are running it for other purposes.

Detecting all the possible timer conflicts is hard, which is why there have been so many mistakes made in this area.
Using your feature won't be any less safe than "normal" PWM on these pins, so perhaps is little to be gained from these partial checks.

Servo Timer

Tone Timer

Software Serial Timer

Timers connected to SPI pins

Am I right, that all timers, allocated by Marlin are listed in timers.h? Or there any implicit usage? (except libraries, but it will be good to know what timers exactly they allocates)

Some are implicit. I think that dealing with timers is a little more mature in HAL/STM32. Maple (HAL/STM32F1) doesn't always give a good way to override the default behavior.

Tone is the hardest to deal with in Maple. It default to 4 or 8 (depending on the chip), but will use the timer connected to a pin if available. This means it is entirely up to the board layout whether enabling SPEAKER will interfere with some other timer.

sjasonsmith · 2020-10-29T21:04:56Z

Could you look at HAL/STM32 to see how hard the same change would be there?

Some HAL/STM32F1 boards are being migrated to HAL/STM32, with the eventual goal of moving everything and deleting HAL/STM32F1. Adding new functionality to STM32F1 which is missing from STM32 makes this transition harder.

Which board are you using for your testing?

victorrar · 2020-10-29T21:07:17Z

MKS Robin Nano V1.2 with TFT35 (no tone) and BLTouch

Marlin/src/HAL/STM32F1/fast_pwm.cpp

rhapsodyv · 2020-10-29T21:12:35Z

MKS Robin Nano V1.2 with TFT35 (no tone) and BLTouch

You can use mks_robin_nano35_stm32 to test.

sjasonsmith · 2020-10-29T21:19:07Z

At a glance, I think this looks ok. Maybe there are bugs, but that doesn't even worry me very much for a brand new feature.

My larger concern is with adding any new features to the STM32F1 HAL. I would love to see this added to the STM32 HAL instead. This could then be a good incentive for people to help migrate boards from STM32F1 to STM32.

victorrar · 2020-10-29T21:28:02Z

Ok, I will try to migrate it to "STM32".

MKS Robin Nano V1.2 with TFT35 (no tone) and BLTouch

You can use mks_robin_nano35_stm32 to test.

Can you, please, explain more? I've can't find the config for that

I'm new here, and is there any guide about the project structure?

thisiskeithb · 2020-10-29T21:29:46Z

Can you, please, explain more? I've can't find the config for that

It's in bugfix-2.0.x:

Marlin/platformio.ini

Lines 1325 to 1333 in 1247f73

    
           # 
        
           # MKS Robin Nano V1.2 and V2 using hal STM32 
        
           # 
        
           [env:mks_robin_nano35_stm32] 
        
           platform             = ${common_stm32.platform} 
        
           extends              = common_stm32 
        
           build_flags          = ${common_stm32.build_flags} -DMCU_STM32F103VE -DSS_TIMER=4 -DENABLE_HWSERIAL3 -DTRANSFER_CLOCK_DIV=8 
        
           board                = genericSTM32F103VE 
        
           board_build.core     = stm32

You'll want to use/check bugfix-2.0.x as it's the most up to date branch.

sjasonsmith · 2020-10-29T21:41:20Z

It's in bugfix-2.0.x:

And it was just added. Maybe this morning?

sjasonsmith · 2020-10-29T21:47:01Z

I'm new here, and is there any guide about the project structure?

For someone who is "new here" I am impressed that you added this feature and got it working!

I am not aware of any good explanatory resources. If you encounter questions while adapting the implementation, please post them here.

This might be even easier in HAL/STM32. HardwareTimer has a setPWM function with frequency and duty cycle arguments.

One caution is that not all STM32 timers use the same base clock, even on the same chip. Make sure you use the framework functions to set rates, and don't divide from the CPU clock yourself.

Once you make it work for your STM32F1 board, we can help you test it on a larger variety of hardware.

zbyrek · 2021-01-02T17:43:26Z

I have problem with enabling it (creality v4.2.2 board), when i disable soft pwm and enable it (it's on PA0) it's just DC output on pin, always enabled, even for 0% fan is spinning at max speed.

sjasonsmith · 2021-01-02T17:47:44Z

@zbyrek many fans don’t respond well to PWM. You can try adjusting the frequency (somewhere in Configuration_adv.h).

For most of my fans they behave best with soft PWM.

zbyrek · 2021-01-02T17:56:29Z

@zbyrek many fans don’t respond well to PWM. You can try adjusting the frequency (somewhere in Configuration_adv.h).

For most of my fans they behave best with soft PWM.

Unfortunately that's not it, i measured frequency and with that option on it's just DC, also i can see it when i connect engraving laser (that's the reason why i need it in the first place), with soft pwm it's just 7.8Hz which is definetly too low for laser engraving (if i want control power and not just hard ON/OFF), however thanks for reply.

sjasonsmith · 2021-01-02T18:17:04Z

I haven’t actually measured it on a board, but it looks like it probably has a conflict on timer 2 between the fan PWM and the timer used for Marlin’s temperature ISR.

zbyrek · 2021-01-02T18:23:19Z

Here, I made some proof measurments, that's definetly not generating any PWM (sorry for crude photo of screen as proof), I'm not here to just complain, I could help fix that, i have some knowledge in programming and in electronics, however never used STM32, I looked at code but without debugger is hard to find what's what.

HW 50% duty with screen of printer:

SW 100% duty:

SW 0% duty:

SW 50% duty:

HW 0% duty:

HW 50% duty:

zbyrek · 2021-01-02T20:57:42Z

Ok, there is definetly problem with timer2 conflict, since PA0 is controled by timer2 which unfortunately cannon't be change:

safe switch just returns from pwm code:

I tried to remap TEMP timer to timer4:

but without success, need to dig in more.

edit:
I found that PA0 could be also controlled by timer5:

sjasonsmith · 2021-01-02T21:01:46Z

It's probably better for you to open an issue and record details there, rather than in this closed PR.

zbyrek · 2021-01-02T21:04:46Z

Right.

thisiskeithb · 2023-08-16T00:18:27Z

@rajhlinux: This Issue Queue is for Marlin bug reports and development-related issues, and we prefer not to handle user-support questions here. (As noted on this page.) For best results getting help with configuration and troubleshooting, please use the following resources:

Marlin Documentation hosted on MarlinFW.org
Marlin Forum hosted on RepRap.org
Marlin Firmware Facebook Group
Marlin Firmware for 3D Printers Facebook Group
Marlin Configuration on YouTube
Marlin Firmware Discord server

Add stm32f1 FAST_PWM feature

d7b57de

sjasonsmith reviewed Oct 29, 2020

View reviewed changes

Marlin/src/HAL/STM32F1/fast_pwm.cpp Outdated Show resolved Hide resolved

Remove dead code

777bfb6

Tidy and adjust

195a0d5

thinkyhead added A: STM32 PR: Improvement T: HAL & APIs Topic related to the HAL and internal APIs. labels Oct 29, 2020

Fini

e6ea835

thinkyhead merged commit de315c9 into MarlinFirmware:bugfix-2.0.x Oct 29, 2020

victorrar mentioned this pull request Nov 4, 2020

Add STM32 HAL FAST_PWM feature #20025

Merged

vgadreau pushed a commit to vgadreau/Marlin that referenced this pull request Dec 9, 2020

FAST_PWM for STM32F1 (MarlinFirmware#19952)

8ae278b

tharts pushed a commit to tharts/Marlin that referenced this pull request Jan 6, 2021

FAST_PWM for STM32F1 (MarlinFirmware#19952)

c663102

kpishere pushed a commit to kpishere/Marlin that referenced this pull request Feb 19, 2021

FAST_PWM for STM32F1 (MarlinFirmware#19952)

a8c0938

zbyrek mentioned this pull request Feb 21, 2021

[BUG] HW PWM not working for stm32f1 PA0 #20638

Closed

W4tel-BiDi pushed a commit to W4tel-BiDi/Marlin that referenced this pull request Apr 5, 2021

FAST_PWM for STM32F1 (MarlinFirmware#19952)

b64beb7

chrisjenda pushed a commit to chrisjenda/Marlin that referenced this pull request Apr 5, 2021

FAST_PWM for STM32F1 (MarlinFirmware#19952)

b323504

thinkyhead pushed a commit to thinkyhead/Marlin that referenced this pull request Apr 28, 2021

FAST_PWM for STM32F1 (MarlinFirmware#19952)

e8bcd11

thinkyhead pushed a commit to thinkyhead/Marlin that referenced this pull request Apr 29, 2021

FAST_PWM for STM32F1 (MarlinFirmware#19952)

93ad703

thinkyhead pushed a commit that referenced this pull request Apr 30, 2021

FAST_PWM for STM32F1 (#19952)

3dbe4ac

This comment was marked as off-topic.

Sign in to view

Lyrth pushed a commit to Lyrth/Aquila_X2_N32 that referenced this pull request Sep 6, 2023

Implement fast PWM from MarlinFirmware/Marlin#19952

2aa73ba

This comment was marked as off-topic.

Sign in to view

Uh oh!

Add stm32f1 FAST_PWM feature #19952

Add stm32f1 FAST_PWM feature #19952

Uh oh!

Conversation

victorrar commented Oct 29, 2020

Description

Benefits

Configurations

Uh oh!

sjasonsmith Oct 29, 2020

Choose a reason for hiding this comment

Uh oh!

victorrar Oct 29, 2020

Choose a reason for hiding this comment

Uh oh!

sjasonsmith Oct 29, 2020

Choose a reason for hiding this comment

Uh oh!

sjasonsmith commented Oct 29, 2020

Uh oh!

victorrar commented Oct 29, 2020

Uh oh!

Uh oh!

rhapsodyv commented Oct 29, 2020

Uh oh!

sjasonsmith commented Oct 29, 2020

Uh oh!

victorrar commented Oct 29, 2020

Uh oh!

thisiskeithb commented Oct 29, 2020

Uh oh!

sjasonsmith commented Oct 29, 2020

Uh oh!

sjasonsmith commented Oct 29, 2020

Uh oh!

zbyrek commented Jan 2, 2021

Uh oh!

sjasonsmith commented Jan 2, 2021

Uh oh!

zbyrek commented Jan 2, 2021

Uh oh!

sjasonsmith commented Jan 2, 2021

Uh oh!

zbyrek commented Jan 2, 2021 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

zbyrek commented Jan 2, 2021 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

sjasonsmith commented Jan 2, 2021

Uh oh!

zbyrek commented Jan 2, 2021

Uh oh!

This comment was marked as off-topic.

thisiskeithb commented Aug 16, 2023

Uh oh!

This comment was marked as off-topic.

This comment was marked as off-topic.

Uh oh!

zbyrek commented Jan 2, 2021 •

edited

Loading

zbyrek commented Jan 2, 2021 •

edited

Loading