Status (part 1)

[luc-github]

Everything need to be done:

• Projects prerequisits
• UI design / flow
• Icon design
• Code project
• Test project
• Add missing feature (e.g: Gcode streamer)

[luc-github]

Project prerequisits :

Step 1: setup environment 👍

• Install IDF 4.4.2 👍
• Compile hello world program 👍
• Install lvgl in idf 👍
• Compile sample tft program 👍

Step 2 : Review Lvgl bricks

• Code lvgl hello world using version 8.3.1
• Review all Lvgl API (orz)
• TBC

Step 3: Review program bricks

• Code basic Wifi AP
• Code basic WiFi Client
• Code LittleFS file access
• Code mDNS usage sample
• Code Basic http server
• Code Basic websocket server
• Code EEPROM sample for ESP3D settings

[luc-github]

I am still learning the idf CMakeList syntax - as I never used it before
e.g.: some basic things like copy binary after compilation is easy in platformIO with python script but in CMakelist.xt in idf I was not able to use cmake add_custom_command POST_BUILD to copy the binary file - seems there are some concept for TARGET I am still not catching
also IDF add some extra commands : https://docs.espressif.com/projects/esp-idf/en/latest/esp32/api-guides/build-system.html

Finally I am currently testing C but I think use CPP would be better - I am not sure if it may have impact in performance but I think it would bring more ease of use in coding

[3DSmitty]

I am good with whatever you decide. For me it will take a little time getting used the espressif IDF extension in vscode. This may or may not effect other developers jumping onboard as most 3D Printer developers are used to platformio, but thats just my opinion.

[3DSmitty]

Yes i did notice the UI make menuconfig or ninja menuconfig. Seems to be alot like building the linux kernel (kconfig).... interesting.

[luc-github]

got the 2.8' today, wow very small...

[3DSmitty]

I got the Panlee 3.5" WT32-SC01 Plus today. Pretty awesome.

[Skorpi08]

[luc-github]

@3DSmitty nice thank you

[3DSmitty]

@luc-github also:
https://github.com/wireless-tag-com/8ms-esp32/tree/release/2.2

[luc-github]

Thank you I have updated the reference doc accordingly - unfortunatly the wireles-tag-com github mostly use already compiled libraries but it should be ok I think I can handle the st7796 in 8080 when necessary as there is an example here https://github.com/espressif/esp-idf/tree/master/examples/peripherals/lcd/i80_controller for st7785 and as it is supported by tft_espi I can also refer to it when I will be on this screen

[3DSmitty]

yeah i did notice that about wireless-tag-com, but they are using lvgl. Does that mean the code for st7796 is closed source?
On this github:
https://github.com/nopnop2002/esp-idf-parallel-tft
it says: ST7796(Same as ILI9486)
Maybe thats a clue?

[luc-github]

haha yes saw this also doing research,
but same issue, I only find spi sample code for ILI9486 not parallel
This will be solved when I do real test soon, I will find a way I am not too worry, I was just hoping to have already working way to focus on what to do instead of how to do.
if was too easy there would be no fun ^_^

[Skorpi08]

I don't understand the IDF stuff, so I use Arduino with TFT_eSPI.

[luc-github]

Ho thank you - Yes I saw the screen is supported in spi mode and parallel mode in tft_espi

[3DSmitty]

I feel like the naming convention of these reference pages should be a little more clearer. they should be named by the screen size and board number. For example:

2.8inch_ESP32-2432S028

and

Panlee_3.5inch_ZX3D50CE02S-USRC-4832

This way they will be more search friendly. That just my opinion.

[Skorpi08]

the first thing I would do, create a design/UI in Photoshop.

[3DSmitty]

@Skorpi08 I was just thinking just hardware reference Wiki pages like Luc already started.
Once the firmware starts to shape then we could create design UI flowcharts etc.

[luc-github]

@luc-github for the moment they are more notes for the devt but if you want to create wiki page feel free - I have unlocked it

@Skorpi08 I am not Photoshop guy - this far away from my limited skills and far too expensive for me- I did/do all my stuff on paint.net and inkscape (paintfully) ^_^
I think before even do the design need to define a flow of what need to be displayed - I guess the design should then come naturaly, but I may be wrong

[3DSmitty]

OK the Wiki Page is looking pretty good for the two boards I have. I have one more datasheet for the ILI9341 to add later. I can add the other two if you like but there may be some info missing as I do not have those boards.

[luc-github]

Ok thank you

[3DSmitty]

@luc-github you mentioned you were watching some vids on the IDF. Can you recommend some good ones to watch?

[luc-github]

@3DSmitty you may be interrested in reading this https://docs.platformio.org/en/stable/frameworks/espidf.html

[3DSmitty]

@luc-github Yes! I like that. I read through that a few weeks ago. I even created a new project (blink) with PlatformIO and the espressif library and played around with it.

[3DSmitty]

Only downside I see is the ULP coprocessor, not sure if we will need that.

[luc-github]

ulp coprocessor is mostly designed for sensors, no ?

[3DSmitty]

Yes, and it's only if the ESP goes into deep sleep mode. I can't think of any reason of using deep sleep mode. That would be more for battery powered ESP projects so I think we are good if you decide to go the PlatformIO route. Totally up to you, I am good either way.

[3DSmitty]

Look what I just found!

https://github.com/wireless-tag-com/lv_port_esp32/tree/master/components/lvgl_esp32_drivers/lvgl_tft

Not sure what the difference is between the ST7796s and ST7796u

[luc-github]

if you do a menuconfig the scu tft is defined as spi not parallel 8080

[3DSmitty]

Ugh yeah i didn't notice:
#include "disp_spi.h"
False alarm...

[Skorpi08]

https://github.com/UsefulElectronics/esp32s3_lilygo_8bit_parallel_display_lvgl

[3DSmitty]

@luc-github I started on the Wiki page. Should I embed the datasheets (drag/drop) in the page or link them to the reference folder?

[3DSmitty]

[3DSmitty]

sorry its been a while since I have used Git. What did I do wrong?

[luc-github]

looks like authentication is not done

I see you are using window, use this application https://desktop.github.com/
it is more user friendly than using command line

[3DSmitty]

ok I think I got it. I had to create a fork first to be able to do a pull request. Its coming back to me, lol.

[3DSmitty]

Just waiting on you to approve the PR

[luc-github]

Got the Panlee today - I only can see the rendering - cannot test the touch reactivity because sample app is only display but refresh rate seems good.

I have now setup a basic project supporting the 4 TFTs - it currently only show started on serial ...

So I think I am done with CMakelist.txt training, I have started to check the sdkconfig options but will see that device by devices, especially the PSRAM config and partitions configuration.
Need to define the storage organisation according features.

• Bluetooth need a lot of space so 4 MB flash TFT may not have it enabled
• OTA/webUpdate need 2 partitions - 4 MB flash TFT may not have it neither TBC
• I would like to limit the number of images because having them means store them, if part of FW it means double space for update which can be a waste, so add them on littlefs partition is may be the way to go, adding a dedicated partition for images could be problematic also if space is not enough but have extra space on partition which serve the webserver, so may be keep simple is better

Because the UI is not defined yet and no idea of final FW size - I keep the partition definition for the end

I will now start to do basic App which should run on 4 tft
Blue background + central button - this should allow to check the lvgl part is well configured and the lcd/touch drivers are properly setup for each TFT.
I will start by ESP32_2432S028R TFT then ESP32 Rotrix because they are pretty similar (all SPI) , then the ESPS3 HMI because we have sample code for 8080 lcd and i2C for it, then port the 8080 for the panlee and touch is same as ESP32S3 HMI

Once done the prerequisites level 1 will be done ^_^

Step by step progress is going on right direction 😺

[3DSmitty]

Interesting. Is there any open source UI Builder solutions out there?

[luc-github]

there was longtime ago - but now there is squareline the official ui builder - free for non commercial use

[Skorpi08]

free for 1 Month.
Smitty, why you need open source?

[luc-github]

Here the size I have :
ESP32-2432S028 : 4194304 = 4MB
ESP32 Rotrix: 8388608 = 8MB
ESP32S3 HMI: 8388608 = 8MB
ESP32S3 Panlee: 8388608 = 8MB

Note: I used the following sketch :

#include <Arduino.h>

//Setup
void setup()
{
	Serial.begin(115200);
}

//main loop
void loop()
{
    Serial.println(ESP.getFlashChipSize());
    log_v("%d",ESP.getFlashChipSize());
    delay(2000);
}

On Panlee the Serial.println seems not working - only the Debug message - I need to see why - it may be explained in doc but I do not remember any mention of change in Serial, and debug is displaying properly but no boot message - so may be something was modified in efuse settings?

ESP-ROM:esp32s3-20210327
[  2335][V][getsize.ino:13] loop(): 8388608
[  4335][V][getsize.ino:13] loop(): 8388608
[  6335][V][getsize.ino:13] loop(): 8388608
[  8335][V][getsize.ino:13] loop(): 8388608
[ 10335][V][getsize.ino:13] loop(): 8388608

[3DSmitty]

@Skorpi08 Haha because i am a open source kinda guy :)

[3DSmitty]

@luc-github is it because on the Panlee it is sending serial output to the serial (rx/tx) on the debug port? I also noticed the Panlee does not have a "boot' button.

[luc-github]

Issues solved

• the lvgl_conf.h related to the target was not in lvgl include path, problem solved with updated CMakeLists.txt
• colors were wrong because in drivers need to not invert color but in lvgvl_conf.h need to invert color because it is spi drivers ...

So first POC is done and updated in repo

Now I will start the rotrix TFT drivers part which should be close to ESP32_2432S028R for the touch driver part , it use ILI9488 vs ILI9341 , it use bigger flash and use PSRAM so need to put these transparent for the main program, resolution is also a new parameter 480 x 320 vs 240 x 320 so it will affect the UI TBD

[3DSmitty]

Awesome!!! Life is insane for me this week, many things going on but I am hoping to be able to test it out and start contributing more next week. Keep up the amazing work!!

[3DSmitty]

@luc-github Question, my 16MB SPI flash chips should finally be here in a few days. Should i go ahead and rework the ESP32_2432S028R board or wait to see how much space it will use up? In your initial testing how much space on the ESP32_2432S028R is used?

[luc-github]

because Ui is not defined yet neither workable fw is yet started , it is hard to say what will be final size
It will depend if UI use native lvgl control only or need extra bitmaps TBD ...

[3DSmitty]

Ok I will hold off on the rework for now and just keep the chips handy :)

[3DSmitty]

Any idea what the female wire connectors are for these boards? The schematic says it is HC-1.25-4PWT for the male connectors on the board.

[Skorpi08]

https://de.aliexpress.com/item/4001235017139.html

[Skorpi08]

[3DSmitty]

Awesome thanks. So their just 1.25mm jst connectors, thats easy.

[luc-github]

I have been busy with webui this week until today, I still need to work on some tasks of webui today, but I will continue on code base preparation of tft tomorrow

once code base is done for all samples it will be time for UI definition and lvgl Api capabilities evaluation

[3DSmitty]

Yahoo!

[luc-github]

The rotrix TFT code base with PSRAM is now added, the PSRAM support definitly make the boot and first display faster vs no psram enabled on same TFT (resolution is almost twice first TFT), that said due to PSRAM initialisation boot is slower of 1s vs first TFT
I needed some time as I do not have any schematics / pinout for the Rotrix TFT and I had to check some missing pins for touch and SD especially the irq pin for the touch, without this pin the CPU go over 50% instead of 3~4% in idle time

next step is the ESP32S3 HMI with i2c touch driver and 8080 parallele display.
Because code for lvgl 8 is already available in reference directory, it should be not too hard to do the porting, tomorrow I am out but I hope to finish this tft this week end.

[3DSmitty]

So awesome! Yeah I figured the PSRAM would be a lot faster loading times. Another crazy week for me, ugh. I am basically working two full time jobs right now. I am going to do some research this weekend and see if I can come up with that schematic. I am pretty good about finding things that are not meant to be easily found ;)

[3DSmitty]

@luc-github Any way you have some high res pics of the Rotrics board you could share?

[luc-github]

No I do not have such pictures, but it is ok, all mandatories pins are now found, only the speaker/ microphone are not identified but we will not use them, so no need to spend time on this IMHO

I have started the S3 HMI let see where it goes :)

[Skorpi08]

https://www.youtube.com/watch?v=KPiDBOOTJ8g&ab_channel=ThatProject

[luc-github]

Spent my sunday to refactoring components to reuse them as unitary elements, looking at ESP32S3 HMI which use some code from https://github.com/espressif/esp-bsp, because ESPS3 use i2C touch driver and parallele bus display dirver , when the esp32 use SPI, the hardware initialisation is different so it definitly needed bsp file (board specific file)
it also simplify main program by moving lvgl to bsp part component

[luc-github]

The ESP32S-HMI code base is finally done for screen and touch

It took more time than I expected because:

• Panel use a pin from IO expander so I had to add support for i2C iO expander also
• Drivers provided in reference code are actually not really used and working code was actually using sample code of lvgl library instead of code provided - I took some time to realize that.. orz..
• My first attempt failed and I had to redo each steps one by one to found my mistakes...orz..orz...

But here we are, even with only 2MB of PSRAM, the ui is super reactive, there is no irq pin for the FT5X06 I2C touch driver, so cpu usage look higher than XTP2046, but consomption increase less when touch is working - so we will see once real UI is implemented...TBD

Next step is to work on Panlee TFT - the chalenge will be the panel as no driver so far on idf but I will see how the arduino library implemented it

[Skorpi08]

It's Arduino Framework, it's a no-go

[Skorpi08]

How it's work's, you can see here
https://youtu.be/tPk6erJif6Y

[3DSmitty]

ugh that sucks... Do you think it adaptable to esp idf framework?

[3DSmitty]

Yeah i couldn't even find the data sheet for the ST7796UI but I did find one for ST7796U2 which has the 8 bit parallel interface.

[luc-github]

I will start with st7789 driver and sample code and see how to adjust commands as the LovyanGFX and TFT_eSPI define.h and TFT_eSPI Init support it

No mention lvgl already support SPI version so commands are also known : C file and H file

[luc-github]

Panlee panel added :

Reactivity is same as HMI - very good

I still need to work on touch driver because Panlee has interupt pin for panel which could save MCU time - it is not yet implemented but should be done today or at latest tomorrow

[3DSmitty]

Ok great. Also from the datasheet:

[luc-github]

looking at https://github.com/sparkfun/SparkFun_TouchInput_Driver_FT5xx6
their interrupt mode does not use this - I have tried the both modes and nothing change the pint does not change its state always 1
I will dig more today - I can flash again the TFT using USB - if no progress I will put in hold and continue on starting UI study

[3DSmitty]

Maybe this driver would work better?
https://github.com/martinberlin/FT6X36-IDF

[luc-github]

not really, the int pin does not change state ever so that is the issue - the code after pins change state is ok - I won't spend more time on this - I will open a ticket so if anyone want to work on it it can share / discuss on ticket

[luc-github]

ticket for FT5X06 is created - I have also opened a new discussion thread for UI #6 and another one for expected features #7

So this thread will be kept to track implemented features and discuss status as planned

[luc-github]

Now can display spinner smoothly when doing background task (Scan SSID List, Scan SD content, TBD)

V_20230802_135123_ES6.mp4

and still at 33fps at 800X480 resolution 😸

[luc-github]

I am reviewing each board / TFT one by one and will end by today (did 3 only 2 left) I will also update the custom esp32/esp32s3 to compile without TFT and same for the freeNov which does not have TFT but has a camera as POC

I spent all sunday checking why the no psram TFT board (ESP32_2432S028R) was always out of memory when I manage to get it 65KB free ram, and I got conclusion that with current UI it is just not possible, it consume too much (extra task / not controled lists) - the no PSRAM need a lighter UI and different display list method , so from now no PSRAM board are put on on not supported, typicaly the ESP32_2432S028R - when I have time (if I have time orz...) I will work on dedicated solution for such target but that is not the priority as now most of latest ESP32 TFT have PSRAM.
Improving performance / memory took me more time that I expected as usual ^_^ , UI is now more fluid

Latest.mp4

I will update esp3d.io accordingly for board support

[luc-github]

@serisman feel free to join the ESP3D-TFT discord : https://discord.gg/PTRpFh7jk2 it may be easier to share and discuss the mod topic

[serisman]

Good idea. I just joined.

Will post another big performance improvement for the 2.8" shortly.

[3DSmitty]

@serisman So I removed the external 4MB flash and it works just fine so I don't think it was really doing anything at all. I also check continuity between internal flash and external flash and they are both the same all pins are connected. So i guess they were connected in parallel. So strange....

[serisman]

Yes, very strange. I still have no explanation why it was working with both populated. They should have conflicted with each other. Unless maybe they happened to respond identically and therefore didn't get into (too much) conflict? If you happen to have a TL866 or similar programmer, I'm curious if there was anything written on the external flash that you pulled off?

Anyway, I'm glad you got it working without the external flash. That should mean you can proceed with upgrading the internal flash to 16MB and then eventually do the PSRAM hack exactly like mine (once you get parts in).

[3DSmitty]

@serisman no I have some old PC Bios chip programmers but not nothing for SOP-8. Anyways yes I will do the 16MB mod and wait until the PSRAM gets here.

[luc-github]

@luc-github Freenove is interesting device. Hey what do you think of James Pearson's PR?

It is the next step planned, I am working on it today, that is why it is now on github we are discussing on discord for this ^_^

[luc-github]

It just do not power on anymore -no idea what the problem is as I tested it few weeks ago - it was a sample board provided by rotrix so it was subject to failure I guess, and the soldering header was damaged so I may destroy it by not handling carfully ...

[luc-github]

I did not found any esp32 TFT + PSRAM on aliexpress , only esp32S3, so It may not be a big issue to not have esp32 + PSRAM reference board
But I know rotrix does not give any support anymore so it would be a good FW alternative for rotrix users

[3DSmitty]

Yes I didn't either. But I did find this other ESP32-S3 camera module that would be cool and is super tiny. It has 8MB PSRAM: https://www.aliexpress.us/item/3256804601970891.html

[luc-github]

yes I know the seeed xiao, the camera / sd reader is actually an addon module for the S3 to it but it is pitty it does not provide the usb host connector
Note: the C3 version is super cheap but does not have PSRAM

[luc-github]

I have updated hardware support to 3 more TFT and one resolution:

[luc-github]

I have added new settings and choice editor, and make loading settings smoother

newsettings.mp4

in parallele I am working on Gcode parser, @jamespearson04 is still working on streaming code service, we had several discussions on how to implement the feature, the gcode streaming service should use the GCode parser to know how to react for each command, when Gcode parser will be used by rendering client to extract data to display to TFT
every steps are on right direction ^_^

[serisman]

@luc-github - I noticed a few minor issues with some of the things that have already been implemented. Do you have a place we should start logging feedback, or would you rather wait until all the features are there?

[luc-github]

@serisman you can edit/pr each part you need by pushing pen icon on targeted page

if not in documentation - please create an issue here and we can review them

[luc-github]

I have implemented the positions screen completly, I am not video guy, I only have my phone and I am not expert in recording, I did not found a software that can do screen recording for 2 windows at once, so I used droidcam to stream my phone on my PC when terminal is visible in visual studio, and to record both, I used licecap which generate a 66MB gif, then use online converter to convert to 5.6MB mp4 so quality is not great, sorry, if you have better solution, suggestions are welcome

tft.mp4

The positions screen (jog) support relative and absolutes commands, poilling engine is also implemented - currently every 3s (M105 and M114)
Default states for Jog type and polling are modifiable in settings

The Screen is also catching position report from serial/serial-usb, when that part is not recorded on video it is actually working

[3DSmitty]

Gotcha :)

[serisman]

@3DSmitty - If you are referring to potential changes to your Marlin configuration, you could probably just remove the LCD (and SD) configuration entirely. At least that's what I'm planning on doing. The entire UI and all SD processing will happen within ESP3D. Marlin will only be responsible for receiving GCode commands (through its serial interface) to control the stepper motors and heaters and then returning status information through GCode responses. This should lighten the load on the Marlin controller and give old 8-bit boards some extra breathing room.

[3DSmitty]

@serisman yes that's exactly what i mean. Thanks for clarifying.

[serisman]

Also, just to clarify. You shouldn't have to make any changes to Marlin if you don't want to. The existing LCD and SD could still be wired up and either ignored or used independently. I believe ESP3D would still be able to receive GCode status reports in this configuration, and could still send GCode commands as well. I imagine you just wouldn't necessarily be able to send commands while a Marlin integrated SD card print is in progress. And, uploads to the Marlin integrated SD card would still be painfully slow.

[3DSmitty]

@serisman Yes I figured that much as I am very familiar with ESP3D and sdcard transfers. Several of my printers are running it including the MKS Tinybee which uses the ESP32 as MCU. I am also very comfortable with Marlin and building my own firmware.

[luc-github]

Next step is doing same for temperatures screen, which should be easier now all API is ready

[3DSmitty]

I got a sd extension for the Panlee that should work good. Now I can start designing a rear cover for it.

[luc-github]

FYI: IDF 5.1.1 is now out, I will test today or tomorrow

[serisman]

I can confirm that IDF 5.1.1 seems ok so far for me as well, and did not require any code changes to compile. I did notice slightly higher IRAM and Flash usage, but I suppose that is to be expected. Hopefully there are useful bug fixes with the extra usage.

[luc-github]

Disable stepper button is now functionnal - as feedback rendering it display a message box and add a status

[3DSmitty]

@luc-github for the fan have you looked into REPORT_FAN_CHANGE (in Configuration_adv.h). There are other TFT's that use this option.

[luc-github]

Yes this is what webui already do : https://github.com/luc-github/ESP3D-WEBUI/blob/3.0/src/targets/Printer3D/Marlin/filters.js#L240-L248
but because it is only a reaction to a change, and only if this setting is enabled it is tricky to only rely on it
but that will be added for sure , as well as the scan of GCODE M106/M107 streamed by sd

[luc-github]

It seems I have already implemented some code for 2 fans so will go for multiple fanes supports, also I will implement a user setting to display or not the fan(s) control instead of rely of M115 report

[luc-github]

Fan screen is now implemented
Last screens to be done are :
1- change filament / filament extraction
2 - leveling : TBD if start by manual leveling only

[luc-github]

need to send G28 before G29 for what I can see , so G29 keep hand until it finish process ? no report at all? just ok ? I do not have BLTouch or any leveling sensor

Also seems need to use A parameter to enable it G29 A and G29 D to disable it

[3DSmitty]

I suppose you could do a "G29 V4" which is the verbose option and output it. Bilinear bed leveling is much more common than UBL or Mesh.
https://marlinfw.org/docs/gcode/G029-abl-bilinear.html

[3DSmitty]

Yes the printer must be homed so a G28 before G29 is appropriate.

[luc-github]

wow depending method enabled parameters.A and D have different action
that is tricky
Better use M420 S1/0 to enable disable bed leveling

[luc-github]

Do you have a G29 V4 output ? I wonder if I could use it to show a kind of progress

[luc-github]

I have implemented the filament change / extruder screen and so far so good.

I have tested my HMI V3 in USB Serial on my ender printer which seems use a custom version of Marlin but TFT and web UI seems OK 😺

The Temperature output looks definitly different but still detected
I tested also Telnet and same
I had a restart when saving sta ssid / password and applied to Sta mode, but no other issue so far, it need definitly more test to know if is reproductible

It seems my ender 5 has bed leveling but it was not working - so I never used it, and even forget I have it actually orz...
the output I got after a G28 then G29 V4:

G28 
G29 V4
G29 Auto Bed Leveling
Bed X: 50.000 Y: 50.000 Z: 0.000
Bed X: 133.000 Y: 50.000 Z: 0.016
Bed X: 216.000 Y: 50.000 Z: -0.013
Bed X: 299.000 Y: 50.000 Z: -0.051
Bed X: 299.000 Y: 133.000 Z: -0.005
Bed X: 216.000 Y: 133.000 Z: -0.041
Bed X: 133.000 Y: 133.000 Z: -0.031
Bed X: 50.000 Y: 133.000 Z: -0.036
Bed X: 50.000 Y: 216.000 Z: -0.050
Bed X: 133.000 Y: 216.000 Z: 0.055
Bed X: 216.000 Y: 216.000 Z: 0.051
Bed X: 299.000 Y: 216.000 Z: 0.026
Bed X: 299.000 Y: 299.000 Z: -0.018
Bed X: 216.000 Y: 299.000 Z: -0.064
Bed X: 133.000 Y: 299.000 Z: -0.036
Bed X: 50.000 Y: 299.000 Z: -0.046
Bilinear Leveling Grid:
       0       1       2       3
 0 +0.0000 +0.0162 -0.0125 -0.0512
 1 -0.0363 -0.0313 -0.0412 -0.0050
 2 -0.0500 +0.0550 +0.0512 +0.0262
 3 -0.0463 -0.0363 -0.0638 -0.0175

So I will see what kind of UI I can do with that

About TFT UI I will up side down up/down command for Z to be consistent with printer movement, for X Y it is not needed because I use only vertical controls instead of crossed control so it cannot be wrong neither confusing
I should say for a first test in real conditions I am quite happy, especially with the USB serial support working flawlessly 😎

[3DSmitty]

@luc-github Awesome! Even if you just show the Leveling Grid in UI that would be very helpful. If not thats still ok as you will still be able to see it from terminal output in Webui.

[luc-github]

So far my ender 3 is also fine in USB-serial - this time the FW is not custom as temperature output is standard

So far so good

[serisman]

Initial support for the two 3.5" Sunlon 480x320 screens (ESP32_3248S035C and ESP32_3248S035R) is in this PR: #28

Most of the code was copy/paste from other hardware targets, with a few tweaks and minor changes. Everything seems to be working correctly except for SD (not enough RAM left). I left the code changes in for the PSRAM hardware mod, but haven't tested it on these boards yet as I am waiting for some more chips to arrive, hopefully in a few days.

There are probably still some optimizations to do, especially with the GT911 capacitive touch sensor. I noticed it uses a fair amount of CPU compared to the XPT2046 touch sensor. It might be due to the lack of IRQ support in the current code.

[luc-github]

Manual bed leveling is now implemented but it rely on GCode Streaming that is not yet implemented for successive commands

I will work on automatic bilinear bed leveling UI and it should be done for UI
I should be able to focus on Gcode streaming implementation part once UI is done for 3DPrinter
Once Streaming Service is implemented I will work on cnc ui implementation

[3DSmitty]

@luc-github I had a thought about the auto bed leveling button. Could the G29 command options be customizable as a option in settings? This way the user could set it up for bilinear, UBL, or Mesh. Kind of like a macro button in WebUI.

[luc-github]

it is already enable disable in settings
Macro support is not for short term - need to have basic features implemented before starting to work on advenced ones

[luc-github]

I have implemented last UI screen for 3d printer so I will now focus on GCode streaming integration
Because this discussion is now really too long I have opened a new one (#32) with new targets which are actually the continuation of current ones
I will also lock discussion to avoid back and forth