MKS DLC32 max support questions

[luc-github]

Hello I have bought a MKS DL32 Max board with its serial TFT
It is a ESP32S3 chip with 8MB flash - no PSRAM detected so far and no IO extender

Using https://github.com/makerbase-mks/MKS-DLC32/blob/main/MKS--DLC32-MAX-main/firmware/hardware.rar
MKS DLC32 MAX V1.0_002 PIN.pdf
and some tests I think am was able to list the global pinout :

GPIO	Function	Description
GPIO0	Flash mode
GPIO1	Air Pump	Header
GPIO2	Laser	Header
GPIO3	??	??
GPIO4	DIR Z
GPIO5	STEP Z
GPIO6	DIR Y1-Y2
GPIO7	STEP Y1-Y2
GPIO8	EN Stepper X Y1 Y2 Z A
GPIO9	??	??
GPIO10	SD Detect
GPIO11	SD D1	SDIO D1
GPIO12	SD D0	SDIO D0
GPIO13	SD CLK
GPIO14	SD CMD
GPIO15	DIR X
GPIO16	STEP X
GPIO17	TX1	UART1 TX
GPIO18	RX1	UART1 RX
GPIO19	DIR A
GPIO20	STEP A
GPIO21	SD D3	SDIO D3
GPIO33	??	??
GPIO34	??	??
GPIO35	Spindle
GPIO36	??	??
GPIO37	??	??
GPIO38	??	??
GPIO39	END STOP X	also probe also A endstop
GPIO40	END STOP Y	also Flame
GPIO41	END STOP Z	also DOOR
GPIO42	Beeper	header
GPIO43	TX0	UART0 TX
GPIO44	RX0	UART0 RX
GPIO45	??	??
GPIO46	??	??
GPIO47	SD D2	SDIO D2
GPIO48	??	??

the serial tft is using RX1/TX1

I have started to see how to build a mks_dlc32_max_1_0_map.h

About SD card how to specify the SD Detect pin ?Or I just ignore it ?
Is SDIO 4 bits Reader supported (CMD, CLK, D0, D1, D2, D3) ?Or should I try to map SPI pins instead ? (Mosi/Miso/SCK/CS)

here what board looks like

the screen is using rj11 connector, but there is another connector on the side if needed

[terjeio]

I have just added SDIO and SD card detect support to the driver:

#if SDCARD_ENABLE

#define SDCARD_SDIO
#define SD_CLK_PIN          GPIO_NUM_13
#define SD_CMD_PIN          GPIO_NUM_14
#define SD_DETECT_PIN       GPIO_NUM_10
#define SD_D0_PIN           GPIO_NUM_12
#define SD_D1_PIN           GPIO_NUM_11
#define SD_D2_PIN           GPIO_NUM_47
#define SD_D3_PIN           GPIO_NUM_21

#endif // SDCARD_ENABLE

Only tested with a breakout board and rather long cables, appeared to be unstable for me. Interested to hear if it works better with the board. And also if SDIO is significantly faster than SPI.

[luc-github]

excellent - thank you - I will give a try this week and feedback

Note

I am late on this sorry, busy with new year and other project that take more time than expected, still on it, will feedback once done

[luc-github]

Update:

• I have created a file for the board according the pins and settings : mks_dlc32_max_1_0_map.h

• I have switched to idf 4.4.4 because in 4.4.3 S3 is premium so I was afraid not stable

• I have enabled the board in my_machine.h and put basic settings

• To make it compile I had to :
  1 - rename sdkconfig to .old and rename sdkconfig.s3 to sdkconfig
  2 - reconfigure for S3 : idf.py set-target ESP32S3 reconfigure
  3 - be sure compilation follow USB_SERIAL_CDC so I have updated :

  • driver.c
  • usb_serial.c
    4 - I2C pins are currently not defined/ used so I disabled it https://github.com/luc-github/ESP32/blob/mks_dlc32_max/main/boards/mks_dlc32_max_1_0_map.h#L120
    Note: EEPROM is disabled because it seems linked to I2C
    5 - Updated sdkconfig to use Single large APP partition because current one if overflowed - that is temporary because the flash is 8MB so can define a better partitions file

Test Result:

?
<Idle|MPos:0.000,0.000,0.000|Bf:100,1022|FS:0,0|Pn:XYZS|Ov:100,100,100|A:M>
ok
$I
[VER:1.1f.20250104:]
[OPT:VNMDSL,100,1024,3,0]
[AXS:3:XYZ]
[NEWOPT:ENUMS,RT+,NOPROBE,ES,REBOOT,SED,RTC,WIFI,SD]
[FIRMWARE:grblHAL]
[SIGNALS:HSE]
[FREE MEMORY:205K]
[DRIVER:ESP32-S3]
[DRIVER VERSION:241222]
[DRIVER OPTIONS:v4.4.8-1-ga9d0f22193]
[WIFI MAC:64:e8:33:6e:a3:7c]
[IP:]
[PLUGIN:WebUI v0.24]
[PLUGIN:SDCARD v1.18]
[PLUGIN:FS macro plugin v0.13]
ok
$F
[FILE:/ANOLEX.NC|SIZE:18]
ok

SD is working and fast for what I can see, some comments:

• About SD

• if I remove the SD Card I got : error:62 (failed to open dir) but due to SD Detect pin you should get error:60 (failed to mount)
• if I insert SD Card I got debug output:

[0;32mI (789119) gpio: GPIO[13]| InputEn: 0| OutputEn: 0| OpenDrain: 0| Pullup: 1| Pulldown: 0| Intr:0 �[0m
�[0;32mI (789119) gpio: GPIO[14]| InputEn: 0| OutputEn: 0| OpenDrain: 0| Pullup: 1| Pulldown: 0| Intr:0 �[0m
�[0;32mI (789119) gpio: GPIO[12]| InputEn: 0| OutputEn: 0| OpenDrain: 0| Pullup: 1| Pulldown: 0| Intr:0 �[0m
�[0;32mI (789129) gpio: GPIO[11]| InputEn: 0| OutputEn: 0| OpenDrain: 0| Pullup: 1| Pulldown: 0| Intr:0 �[0m
�[0;32mI (789139) gpio: GPIO[47]| InputEn: 0| OutputEn: 0| OpenDrain: 0| Pullup: 1| Pulldown: 0| Intr:0 �[0m
�[0;32mI (789149) gpio: GPIO[21]| InputEn: 0| OutputEn: 1| OpenDrain: 0| Pullup: 0| Pulldown: 0| Intr:0 �[0m
�[0;32mI (789269) gpio: GPIO[21]| InputEn: 0| OutputEn: 0| OpenDrain: 0| Pullup: 1| Pulldown: 0| Intr:0 �[0m

• About Settings
  doing $73, $74, $75 to configure STA does not stay if board if reboot , this is due to EEPROM disabled because of I2C ? but on ESP32 you can use NVS - or I missed a setting or a command ? Sorry it has been a while I did not digged in grblHAL code

• About compilation
  you may add the sdkconfig path and partition path in CMakeList.txt, but I saw you moved out everything from it since last time I flashed grblHAL, so you may prefer people to rename sdkconfig instead

• About SERIAL1_PORT

for the moment it seems not communicating with the serial screen, I did not found how to set the baudrate on it - it may be the issue ?

The board is not on a full system for the moment - I am just starting so tests are limited , let me know if you want me to check something

Once done I can do a PR if you want

[terjeio]

Hmm, if you have USB_CDC enabled the MPG stream will be the primary UART since by default the first unclaimed stream is assigned to the MPG. This could be the reason you do not get any output.
To force the MPG stream to always use serial1 add #define MPG_STREAM 1 to the map file. For now you will have to add #define KEYPAD_STREAM 1 as well to have the MPG and keypad share the same stream. I'll change that to share the streams if the keypad stream is not explicitly defined to use a different stream.

When I did that MPG mode almost works... If I switch to MPG mode before connecting via Telnet or Websocket (WebUI) the resident guru awakens and reboots the controller, why I do not know. By delaying the transmission of the status report output on a switchover by 5ms I managed to not disturb him so I will add that change in the next commit. Switching MPG mode after the connection has been made does not require the delay so it might be related to RTOS task switching since I cannot remember seeing this with other MCUs.

FYI if MPG mode is active and a Telnet or Websocket connection is made MPG mode is switched off for the duration of the connection unless the MPG application rerequest MPG mode to regain control.

Tip: if you want you can change the 0x8B command character to a printable character for test purposes, this allows one to use a terminal program to simulate a MPG. I typically replace 0x8B with '&'.

[luc-github]

if you have USB_CDC enabled the MPG stream will be the primary UART since by default the first unclaimed stream is assigned to the MPG. This could be the reason you do not get any output.

no I did #define USB_SERIAL_CDC 0 , or you mean better no define at all ?

FYI if MPG mode is active and a Telnet or Websocket connection is made MPG mode is switched off for the duration of the connection unless the MPG application rerequest MPG mode to regain control.

Ok so the screen should be able to send such request / command, noted

Tip: if you want you can change the 0x8B command character to a printable character for test purposes, this allows one to use a terminal program to simulate a MPG. I typically replace 0x8B with '&'

Yes saw it but because I initially though that was a global command and not input specific I have added the command to WebUI, now I understand it is input specific as you explained:

The toggle command 0x8B cannot be used from the WebUI, it is only available for UART streams registered for MPG use

so I will test changing the char

Thank you for your help I will test and report

[luc-github]

@terjeio
I have updated source with latest code and followed your suggestions:
https://github.com/luc-github/ESP32/blob/mks_dlc32_max/main/my_machine.h#L75-L79

I have actually disabled keypad because I do not have one, I initially followed the documentation that wrote need to enable it.

Now it is working as expected :

• Serial is initially on control
• WebSocket take the control if webUI is open and release it when page is closed
• On Serial1 no control by default, only take control if use the toggle command - I have tested with & as you suggested
• If WebUI is connected, Serial and Serial 1 are read only, but using & on serial 1 still allow serial1 to send command
• Using & on Serial has no effect

About that last statement I have questions:

• Why Serial cannot send the toogle command ? it is a toogle of a state, origin of command should not matter
  If the screen is not responding for any reason when it is in control - currently there is no way for Serial to recover the control but restart the board or open and use WebUI- is that an acceptable behavior ? knowing that ? is always working what ever the origin and the state of MPG
• Is that normal the MPG report is only when MPG state change ? not periodicaly ?

>&
<Idle|MPos:0.000,0.000,0.000|Bf:100,1023|FS:0,0|Pn:XYZS|WCO:0.000,0.000,0.000|WCS:G54|A:M|Sc:|MPG:1|H:0|T:0|TLR:0|FW:grblHAL>
>?
<Idle|MPos:0.000,0.000,0.000|Bf:100,1022|FS:0,0|Pn:XYZS>
ok
>?
<Idle|MPos:0.000,0.000,0.000|Bf:100,1022|FS:0,0|Pn:XYZS>
ok
>?
<Idle|MPos:0.000,0.000,0.000|Bf:100,1021|FS:0,0|Pn:XYZS>
ok
>?
<Idle|MPos:0.000,0.000,0.000|Bf:100,1022|FS:0,0|Pn:XYZS>
ok
>?
<Idle|MPos:0.000,0.000,0.000|Bf:100,1023|FS:0,0|Pn:XYZS>
ok
>?
<Idle|MPos:0.000,0.000,0.000|Bf:100,1023|FS:0,0|Pn:XYZS>
ok
>?
<Idle|MPos:0.000,0.000,0.000|Bf:100,1023|FS:0,0|Pn:XYZS>
ok
>?
<Idle|MPos:0.000,0.000,0.000|Bf:100,1023|FS:0,0|Pn:XYZS>
ok
>?
<Idle|MPos:0.000,0.000,0.000|Bf:100,1023|FS:0,0|Pn:XYZS>
ok
>?
<Idle|MPos:0.000,0.000,0.000|Bf:100,1023|FS:0,0|Pn:XYZS|WCO:0.000,0.000,0.000>
ok
>?
<Idle|MPos:0.000,0.000,0.000|Bf:100,1023|FS:0,0|Pn:XYZS|Ov:100,100,100|A:M>
ok
>?
<Idle|MPos:0.000,0.000,0.000|Bf:100,1021|FS:0,0|Pn:XYZS>
ok
>?
<Idle|MPos:0.000,0.000,0.000|Bf:100,1023|FS:0,0|Pn:XYZS>
ok
>?
<Idle|MPos:0.000,0.000,0.000|Bf:100,1023|FS:0,0|Pn:XYZS>
ok
>?
<Idle|MPos:0.000,0.000,0.000|Bf:100,1023|FS:0,0|Pn:XYZS>
ok
>?
<Idle|MPos:0.000,0.000,0.000|Bf:100,1023|FS:0,0|Pn:XYZS>
ok
>?
<Idle|MPos:0.000,0.000,0.000|Bf:100,1023|FS:0,0|Pn:XYZS>
ok
>&
<Idle|MPos:0.000,0.000,0.000|Bf:100,1023|FS:0,0|Pn:XYZS|MPG:0>
>?
<Idle|MPos:0.000,0.000,0.000|Bf:100,1023|FS:0,0|Pn:XYZS>
>?
<Idle|MPos:0.000,0.000,0.000|Bf:100,1023|FS:0,0|Pn:XYZS|WCO:0.000,0.000,0.000>
>?
<Idle|MPos:0.000,0.000,0.000|Bf:100,1023|FS:0,0|Pn:XYZS|Ov:100,100,100|A:M>
>?
<Idle|MPos:0.000,0.000,0.000|Bf:100,1023|FS:0,0|Pn:XYZS>
>?
<Idle|MPos:0.000,0.000,0.000|Bf:100,1023|FS:0,0|Pn:XYZS>
>?
<Idle|MPos:0.000,0.000,0.000|Bf:100,1023|FS:0,0|Pn:XYZS>
>?
<Idle|MPos:0.000,0.000,0.000|Bf:100,1023|FS:0,0|Pn:XYZS>
>?
<Idle|MPos:0.000,0.000,0.000|Bf:100,1023|FS:0,0|Pn:XYZS>
>?
<Idle|MPos:0.000,0.000,0.000|Bf:100,1023|FS:0,0|Pn:XYZS>
>?
<Idle|MPos:0.000,0.000,0.000|Bf:100,1023|FS:0,0|Pn:XYZS>
>?
<Idle|MPos:0.000,0.000,0.000|Bf:100,1023|FS:0,0|Pn:XYZS>
>?
<Idle|MPos:0.000,0.000,0.000|Bf:100,1023|FS:0,0|Pn:XYZS|WCO:0.000,0.000,0.000>
>?
<Idle|MPos:0.000,0.000,0.000|Bf:100,1023|FS:0,0|Pn:XYZS|Ov:100,100,100|A:M>
>?
<Idle|MPos:0.000,0.000,0.000|Bf:100,1023|FS:0,0|Pn:XYZS>
>?
<Idle|MPos:0.000,0.000,0.000|Bf:100,1023|FS:0,0|Pn:XYZS>
>?
<Idle|MPos:0.000,0.000,0.000|Bf:100,1023|FS:0,0|Pn:XYZS>

[terjeio]

If WebUI is connected, Serial and Serial 1 are read only, but using & on serial 1 still allow serial1 to send command

Serial and Serial 1 can send real time commands at any time and they will be responded to. In addition Serial 1 can send single character commands to the keypad plugin that will be expandend and acted upon if the controller is in a state allowing that.

Why Serial cannot send the toogle command ? it is a toogle of a state, origin of command should not matter

Actually it can, but only to turn MPG mode off. Allowing any client to turn it on can be implemented but then the sender should somehow be made aware that a MPG is available?

Is that normal the MPG report is only when MPG state change ? not periodicaly ?

Yes, because the MPG should normally request status reports when in control. If not auto reporting can be enabled with $481 (auto report interval). There is a toggle command character for turning that on/off as well since it may mess with senders or even MPGs... Changes to the auto report mode will be added to the real-time report so it can be tracked.
FYI ioSender checks for auto reporting beeing enabled on connect and turns it off while connected since it has to be in control of when reports are needed.

So handling a MPG can be quite complicated - MPG implementations may be different and senders may or may not be able to handle switchover nicely etc. etc.

[luc-github]

thank for the clarifications, yes handling a MPG can be quite complicated, I agree

I will do the PR this week then and let you review it
I will check how to add the toogle command support to the MKS screen, but it will be next month , now here it is chinese new year, it is tonight actually, so all MKS teams are off, and I need to ask them because current FW source code is not yet available