search

luc-github / ESP3D-WEBUI

A Web UI for ESP8266 or ESP32 based boards connected to 3D printers / CNC
GNU General Public License v3.0
762 stars 305 forks source link

[BUG] SD Upload fails with BTT TFT connected ESP01s #158

Closed DaStivi closed 3 years ago

DaStivi commented 3 years ago

Describe the bug Uploading to (Mainboard) SD fails after a couple of seconds, webui displays error1, BTT TFT starts beeping and display "Error: No Checksum with line number, Last Line: 1"

gco file gets created on sd card but with 0byte

ESP3D-WEBui isn't able to communicate anylonger with TFT/Mainboard, i managed to send M29 once via TFT Terminal that stopped the beeping and constant error poping up on tft but else i've to reset tft to stop and bring back the connection

To Reproduce Steps to reproduce the behavior: upload gco file to mainboard sd

Expected behavior gco file uploaded successfully to sdcard

Screenshots

ESP3D Firmware: (please complete the following information)

Target Firmware: (please complete the following information)

Board used (please complete the following information):

Browser used (please complete the following information):

Additional context

set #define TX_BUFFER_SIZE 128 and #define RX_BUFFER_SIZE 1024 in marlin firmware, as i tried other solutions from past sd upload issues...

sending commands, getting temperatures works as expected

github-actions[bot] commented 3 years ago

Thank your for submiting, please be sure you followed template or your issue may be dismissed.

luc-github commented 3 years ago

do you have auto temperature send enabled ? or any auto command enabled ? if esp connected to BTT TFT ?

luc-github commented 3 years ago

to check if any autosend command is enabled just type in webUI M28 test.gco

if anything appear please share

Once done type M29 to end upload mode

DaStivi commented 3 years ago

do you have auto temperature send enabled ? or any auto command enabled ? if esp connected to BTT TFT ?

not that i'm aware of.... BUT, as you mention auto command... i do see ah constant M114 query gooing on in esp terminal and even in tft terminal..... but i don't know from where's this command is comming... at least on esp terminal i don't see the M114 command it self (only the answer) , but on tft terminal i see the M114 command sending...

luc-github commented 3 years ago

what is M115 TFT output ?

DaStivi commented 3 years ago

M115 TFT FIRMWARE_NAME: BIGTREE_TFT35_V3.0_B1.26.x SOURCE_CODE_URL:https://github.com/bigtreetech/BIGTREETECH-TouchScreenFirmware Cap:HOTEND_NUM:1 Cap:EXTRUDER_NUM:1 Cap:FAN_NUM:1 Cap:FAN_CTRL_NUM:0 ok

using latest build of tft firmware....

DaStivi commented 3 years ago

for the sake of completeness, its ah biqu B1 printer with an skr1.4 and ah special version of the btt tft35 ... and interesstingly the ESP port on the tft pcb has ah wrong pin layout... i did find this also happend on the older versions of btt tft35... and has been fixed with newer versions... interestingly in the b1 facebook group there is also ah user that reported on my post that his board has different print on it.... where mine shows esp01s his has only esp printed and on his tft he could just drop in the esp01s module... i had to pin it differently as the headers are "mirrored"

image image i did soldered the pins now out and put in onthe other side of the esp module to get rid of the cables... image

why do i mention this? i've read in another sd upload issue, it could also be an voltage supply problem, just not sure if the messed up anything else on the pcb maybe?! 🤔

luc-github commented 3 years ago

In theory the M28 test.gcode should make the tft stop polling which not the case - right ? Does auto check position uncheck in webUI still show the M114 feedback - just to be sure poilling come from TFT and not webUI

luc-github commented 3 years ago

no in this case it is not voltage issue as it is error 1 and check sum error - it is some polling that corrupt communication

DaStivi commented 3 years ago

In theory the M28 test.gcode should make the tft stop polling which not the case - right ?

i would say so yes!

Does auto check position uncheck in webUI still show the M114 feedback - just to be sure poilling come from TFT and not webUI

i am pretty sure its not webUI... auto check position is off yes... and even without the esp01 i did noticed this behavior in tft terminal...

no in this case it is not voltage issue as it is error 1 and check sum error - it is some polling that corrupt communication I'd tried to find out allready whats causing this m114 poll, but hadn't the idea thats maybe breaking the sd upload but make sense, ok...

so just need to find out whats causing this...

strange thing is that started with an marlin update... I'd might try another marlin fw to see if this stops... but using different tft firmwares didn't change it....

luc-github commented 3 years ago

weird Issue should be in TFT FW not in Marlin - when I did the PR to support upload I have added the stop polling feature It seems the TFT ignore it for M114 - so something changed in TFT FW after my PR and break the upload feature

luc-github commented 3 years ago

the poling consist in Temperature / Positions / Feedrate / Flow rate

do only positions answer appear after M28 test.gco or all answers ?

DaStivi commented 3 years ago

It seems the TFT ignore it for M114 - so something changed in TFT FW after my PR and break the upload feature

as i'm a noob with github i did checked the files manually and it differs from your code ... :(

the poling consist in Temperature / Positions / Feedrate / Flow rate

do only positions answer appear after M28 test.gco or all answers ?

currently i'm printing, but even when there is no print job running its constantly outputing these, of course without changing values...

T:245.31 /245.0000 B:70.04 /70.0000 @:37 B@:3 Count E:0 T:244.69 /245.0000 B:70.00 /70.0000 @:59 B@:15 Count E:0 Count E:0 T:245.00 /245.0000 B:70.00 /70.0000 @:46 B@:18 Count E:0 T:244.69 /245.0000 B:69.96 /70.0000 @:68 B@:36 Count E:0 Count E:0 T:245.00 /245.0000 B:70.00 /70.0000 @:43 B@:2 Count E:0 T:244.69 /245.0000 B:69.96 /70.0000 @:97 B@:20 Count E:0 Count E:0

luc-github commented 3 years ago

you cannot do upload when printing - Marlin will not handle that - no idea how he will behave to be honest....

Better wait your print is done to do the upload test I mentioned to be sure you get proper behavior

DaStivi commented 3 years ago

sending M28 test.gco "freeze" the webui and tft starts putting the error out... and i'm sorry, i got the error on tft wrong... its "no checksum with line number, last line: 0" image

sending M29 with tft terminal states file saved...

i tested it now again it behaves a little different, as i have just waited a longer time... and also tried to sent ah M29 with esp3d-webui, tft stilll beeping, but web-ui does react again and shows this now in terminal:

echo:E0 Flow: 100% Count E:0 X:-3.0000 Y:-2.0000 Z:0.0000 E:0.0000 Count X:-240 Y:-160 Z:50 FR:100% echo:E0 Flow: 100% T:29.32 /0.0000 B:31.15 /0.0000 @:0 B@:0 Count E:0 X:-3.0000 Y:-2.0000 Z:0.0000 E:0.0000 Count X:-240 Y:-160 Z:50 FR:100% echo:E0 Flow: 100% Count E:0 T:29.32 /0.0000 B:31.04 /0.0000 @:0 B@:0 M28 test.gco M29 echo:Now fresh file: test.gco Writing to file: test.gco //action:notification test.gcook Error:No Checksum with line number, Last Line: 0 Resend: 1 Error:No Checksum with line number, Last Line: 0 Resend: 1 Error:No Checksum with line number, Last Line: 0 Resend: 1 Error:No Checksum with line number, Last Line: 0 Resend: 1 T:29.32 /0.0000 B:31.04 /0.0000 @:0 B@:0 Error:No Checksum with line number, Last Line: 0 Resend: 1 Error:No Checksum with line number, Last Line: 0 Resend: 1 T:29.20 /0.0000 B:31.15 /0.0000 @:0 B@:0

interesstingly there is also a difference in the M114 output before and after trying to save ah file... before there is full x,y,z output... after m28/m29 i only see the temps...

btw: files empty... 0byte

DaStivi commented 3 years ago

another thing i noticed now... after the steps i can't even read from the tft sd... or sending any other command, it just puts the command in the esp-webui terminal but no answer to that command..

after restarting the esp3d from webui same behaviour... need to reset tft, so i really guess its a tft firmware issue :/

luc-github commented 3 years ago

M28 test.gco "freeze" the webui and tft starts putting the error out

It means TFT send commands when it should not, I can see it received T:29.32 /0.0000 B:31.04 /0.0000 @:0 B@:0 so it means M105 was sent by TFT the temperature is due to Marlin autosent as no command is actually working because after M28 test.gco polling should stop and any command sent but M29 are supposed to be writen in SD file also any command sent after M28 test.gco should follow N<number> <command + parameter>*<Checksum> and so any raw commands sent by TFT will generate error from Marlin, and of course no commands but M29 will work until a end of write command is sent M29 this one does not need line numbering / check sum to be able to handle error

It is definitly a TFT FW issue

luc-github commented 3 years ago

to Disable autotemp type M155 S0

I see you did M29 just after M28 and it open File after so I guess the M29 was ignored need to send only after it said file is open or has error You can try to enable Marlin DEBUG to echo each command, I am not sure it will echo anything when M28 is used but worth a try: M111 S1 about TFT FW issue may be it is due to : BTT TFT35v3 (variant of it) which may affect FW somehow but as M115 TFT is working I do not know - need to dig in code and I am busy right now due to personnal training I need to finish quickly - so need time

So you can raise issue in BTT github as issue come from it

DaStivi commented 3 years ago

to Disable autotemp type M155 S0

i try this later... indeed there is the option to have the "status" (temps) on tft screen all the time... but i don't believe thats the issue... i will raise an issue on btt tft git later too... need some time to recap everything there 😐

thx anyway for your massiv help sofar on this topic!

luc-github commented 3 years ago

yes it is not issue but clear a little the output to focus on unwished commands 😸

Thank you please update the status if any - I am curious what did broken the process 😎

emklap commented 3 years ago

to check if any autosend command is enabled just type in webUI M28 test.gco

if anything appear please share

Once done type M29 to end upload mode Same problem here on a stock BIQU B1 printer, SKR 1.4 main board, Marlin 2.0.6, TFT35_B1_V30 0.26.x ( I am the user mentioned earlier in this bug withe the ESP as text on the tft board :+1: ) I am the user with mentioned earlier in this bug with a TFT board with ESP next to the wifi connector a After sending M28 test.gco the empty file is created on the SD card with the following messages in the web UI

M28 test.gco echo:Now fresh file: test.gco Writing to file: test.gco //action:notification test.gco ok Error:No Checksum with line number, Last Line: 0 Resend: 1 Error:No Checksum with line number, Last Line: 0 Resend: 1

The M29 command doesn't stop the error message or beeps.

My command windows often shows the following line (once every second) X:0.00 Y:0.00 Z:0.00 E:0.00 Count X:0 Y:0 Z:0

and sometime the lines (once every second) ok T:20.94 /0.00 B:20.00 /0.00 @:0 B@:0 X:0.00 Y:0.00 Z:0.00 E:0.00 Count X:0 Y:0 Z:0 FR:100%

and probably not related, if verbose mode is enabled than the command M115 shows its output, but the M503 does not provide an output. To see the M503 output the verbose mode needs to be enabled.

emklap commented 3 years ago

I did a test by executing the M28 test2.gco command on the TFT of the printer itself, and it gives the same erroneous behavior. An empty file is created and the printer show the error message "Error:No Checksum with line number, Last Line: 0 Resend: 1"

This looks more like a TFT FW issue to me?

DaStivi commented 3 years ago

I did a test by executing the M28 test2.gco command on the TFT of the printer itself, and it gives the same erroneous behavior. An empty file is created and the printer show the error message "Error:No Checksum with line number, Last Line: 0 Resend: 1"

This looks more like a TFT FW issue to me?

what printer and especially tft do you have?

emklap commented 3 years ago

FW versions I have a BIQU B1 printer, SKR 1.4 main board, Marlin 2.0.6, TFT35_B1_V30 0.26.x. The Marlin and TFT FW is the one that came with the printer. main board + TFT Wifi module

luc-github commented 3 years ago

yes same behavior - after M28 command nothing should happen as no command should be sent, but obviously Marlin receive several commands

DaStivi commented 3 years ago

i am pretty sure with using one of lukes ABL Marlin firmware in the past i hadn't theissue of constant commands in the terminal... i've noticed this earlier but didn't had issue sofar... so i was wondering each update... but i'm not sure how this is all related to each other... because another issue with the UBL Firmware is thati don't receive the eeprom save notifications anymore... that was alsoworking with an older ABL version... i think to remember that this also not worked with the stockfirmware that came with the printer (where i think to rember have noticed the constat terminal stream too..)

luc-github commented 3 years ago

Ok just some feedback I have installed Marlin 2.0.7.2 on my SKR 1.3, I have installed latest ESP3D / latest WebUI - today git binary for my BTT TFT 35-E3

I can reproduce issue I have used Telnet to be sure no webUI related issue when I send M28 test.gco

definitly a TFT issue as this was not like that

I found the bug : https://github.com/bigtreetech/BIGTREETECH-TouchScreenFirmware/blob/master/TFT/src/User/API/interfaceCmd.c#L208

This block every commands from WebUI but let all commands from TFT when stop polling is enabled which is wrong

  if (!ispolling && !fromTFT)
  { //ignore any query from TFT
    purgeLastCmd();
    return;
  }

change the condition to correct one: if (!ispolling && fromTFT) Stop polling and allow upload like it should be

DaStivi commented 3 years ago

you're a genius... i did read exactly that if loop, but i had no idea how it supposed to be or if this is not correct :(

uploading looks like to work now!

luc-github commented 3 years ago

great thank you for the fix confirmation

DaStivi commented 3 years ago

i need to test tomorrow again, but one odd thing, after starting the print from esp webui (maiboard SD), its starts heating up... then i've aborted the print from the TFT and instead of aborting it starts printing, even nozzle&bed are not heated, also the tft then display busy processing and i can't do anything just reseting the tft to stop this!

output from webui terminal for that job

ile opened: KFACTOR.GCO Size: 7037 File selected //action:notification KFACTOR.GCO ok //action:resume//action:prompt_end//action:prompt_begin Resuming SD//action:prompt_button Dismiss//action:prompt_show//action:notification KFACTOR.GCO ok M876 Responding PROMPT_GCODE_INFO T:26.14 /0.0000 B:26.82 /0.0000 @:0 B@:0 T:26.14 /0.0000 B:26.70 /0.0000 @:0 B@:0 T:26.14 /0.0000 B:26.70 /0.0000 @:0 B@:0 T:26.25 /0.0000 B:26.82 /0.0000 @:0 B@:0 T:26.14 /0.0000 B:26.70 /0.0000 @:0 B@:0 T:26.14 /0.0000 B:26.70 /0.0000 @:0 B@:0 //action:notification Bed Heating... T:26.14 /245.0000 B:26.70 /70.0000 @:127 B@:127 T:27.61 /245.0000 B:27.16 /70.0000 @:127 B@:127 T:32.19 /245.0000 B:27.84 /70.0000 @:127 B@:127 T:38.50 /245.0000 B:28.86 /70.0000 @:127 B@:127 T:45.99 /245.0000 B:30.00 /70.0000 @:127 B@:127 T:53.81 /245.0000 B:30.94 /70.0000 @:127 B@:127 T:61.67 /245.0000 B:31.98 /70.0000 @:127 B@:127 T:69.57 /245.0000 B:33.13 /70.0000 @:127 B@:127 T:77.39 /245.0000 B:34.38 /70.0000 @:127 B@:127 T:85.03 /245.0000 B:35.25 /70.0000 @:127 B@:127 T:92.40 /245.0000 B:36.25 /70.0000 @:127 B@:127 T:99.73 /245.0000 B:37.25 /70.0000 @:127 B@:127 T:106.79 /245.0000 B:38.17 /70.0000 @:127 B@:127 T:113.82 /245.0000 B:39.08 /70.0000 @:127 B@:127 T:120.77 /245.0000 B:40.07 /70.0000 @:127 B@:127 T:127.46 /245.0000 B:41.03 /70.0000 @:127 B@:127 T:133.99 /245.0000 B:41.91 /70.0000 @:127 B@:127 M106 P0 S0 T:140.56 /245.0000 B:42.87 /70.0000 @:127 B@:127 T:146.85 /245.0000 B:43.75 /70.0000 @:127 B@:127 T:153.15 /245.0000 B:44.71 /70.0000 @:127 B@:127 T:159.35 /245.0000 B:45.66 /70.0000 @:127 B@:127

no sign from the abort command, but not even sure i should see it in esp-terminal?

luc-github commented 3 years ago

when press abort in web UI it send : image I do not know what TFT send but in all case if printing from Printer SD and already start heating you cannot send any command until heating is done unless emergency parser is enabled : https://github.com/MarlinFirmware/Marlin/blob/2.0.x/Marlin/Configuration_adv.h#L1940-L1947 And even so, only few command are supported at that time

luc-github commented 3 years ago

I guess issue can be closed now

jimhiggs commented 3 years ago

I was struggling the same problem with SD copy from ESP3D to the onboard SD with it failing with the error message: Error:No Checksum with line number, Last Line: 0 on the TFT on all but extremely short .gco files. I am using the latest ESP3D and TFT3.5 V2 2.0.26.x (loaded in December).

I just manually patched @luc-github fix of interfaceCmd.c. I am now getting successful SD copies from ESP3D to SD but the TFT display is still showing up the Error:No Checksum with line number, Last Line: 0 error shortly after the copy starts and then when the transfer is complete and I go into 'print' to list files on the SD it is stuck on the black loading screen and I have to reset the TFT to continue. The copied file then shows up on the print files list.

Edit: I just resynced with the master as I see @luc-github fix is included. Prblem still exists.

luc-github commented 3 years ago

I do not reproduce this issue but looks again TFT FW issue not esp3d neither webUI issue

jimhiggs commented 3 years ago

Sorry @luc-github I was jumping between github issues. I meant to post this in the BTT TFT firmware issues. I agree not a ESP3D or WebUI issue, will repost in the correct place.

DaStivi commented 3 years ago

I was struggling the same problem with SD copy from ESP3D to the onboard SD with it failing with the error message: Error:No Checksum with line number, Last Line: 0 on the TFT on all but extremely short .gco files. I am using the latest ESP3D and TFT3.5 V2 2.0.26.x (loaded in December).

I just manually patched @luc-github fix of interfaceCmd.c. I am now getting successful SD copies from ESP3D to SD but the TFT display is still showing up the Error:No Checksum with line number, Last Line: 0 error shortly after the copy starts and then when the transfer is complete and I go into 'print' to list files on the SD it is stuck on the black loading screen and I have to reset the TFT to continue. The copied file then shows up on the print files list.

Edit: I just resynced with the master as I see @luc-github fix is included. Prblem still exists.

and you're sure you applied and complied the fix correctly?? and also made shure you see the actual builddate and time then in the Info display??

jimhiggs commented 3 years ago

First I just changed the code in interfaceCmd.c from if (!ispolling && !fromTFT) to if (!ispolling && fromTFT). Recompiled and copied .bin to my Pro V1. Results were that the copy was now successful but still getting error on TFT. Then I updated the firmware on TFT to the latest version on git. Result is the same. I use git and I updated to the latest master of TFT35. I had to include the fonts, icons , language file and config.ini so I am pretty sure it is the latest. But I repeated just to be sure. info reads TFT35_V2.0.26.x Feb 4 2021. Yesterday it read TFT35_V2.0.26.x Feb 3 2021 so is the date the date the firmware is updated rather than a version date?

jimhiggs commented 3 years ago

I just started a copy of a large model file and it is transferring at a rate of 2% every 3 minutes. If that is normal it will only be useful to transfer files this way for fairly small prints

luc-github commented 3 years ago

https://github.com/luc-github/ESP3D/discussions/535

github-actions[bot] commented 3 years ago

This issue has been automatically locked since there has not been any recent activity after it was closed. Please open a new issue for related bugs.