[BUG] Broken TFT mode with BTT TFT35 E3 v3.0.1

[Carl393]

Did you test the latest bugfix-2.1.x code?

Yes, and the problem still exists.

Bug Description

I have a Big Tree Tech SKR 3 mainboard and a TFT 35 E3 touchscreen running latest Marlin release.

When I load the filament, I am getting this error after the filament loans

Printer paused, continue?

Marlin used to allow me to "Purge More, or Continue" right from the touchscreen...

This is a bug from like.. a few years ago where the Marlin is not sending this message to the touchscreen properly....
(I cannot locate the original thread in GitHub)

I figured this bug would be fixed by now?

I have to hold the button down and switch to Marlin mode, click continue, and then switch back to touch mode to continue using my printer

Bug Timeline

Started the other day but never had this problem with my other printers

Expected behavior

I expected the window to show that says "purge more, or continue?" which would allow me to continue using the printer in touchscreen mode. but this never happened.

Actual behavior

Printer Paused, Continue? message repeating over and over.. no amount of tapping OK makes it go away.

Steps to Reproduce

Go to touchscreen mode, heat nozzle, go to load filament menu which executes an M600 code.. filament loads, then extrudes 50mm.. afterwards immediately pauses the printer.

Version of Marlin Firmware

2.1.2.7 (latest stable)

Printer model

Ender 3 V1

Electronics

SKR 3 EZ (Had to compile my own configuration files in vscode...no example files to work with on marlin examples under Creality)

LCD/Controller

Big Tree Tech TFT 35 E3 V 3.0.1

Other add-ons

Added Z axis enabled the dual Z alignment feature, maybe this triggers the bug? it works on my other printer using SKR mini e3

Bed Leveling

UBL Bilinear mesh

Your Slicer

Cura

Host Software

None

Don't forget to include

• A ZIP file containing your Configuration.h and Configuration_adv.h.

Additional information & file uploads

Configfiles.zip

[thinkyhead]

My first impression of your configs has me curious. Why do you have REPRAP_DISCOUNT_FULL_GRAPHIC_SMART_CONTROLLER enabled in your configuration for this display? Is that the appropriate one for this particular BTT TFT?

[thinkyhead]

I actually looked it up. Interesting controller, there. It seems to me that it would need to be specially enabled as a TFT serial display in order for Marlin to know that it needs to send messages to it for its TFT mode. If it worked in the past just being enabled as a RRD Graphical Display then we'll need to figure out where the firmware differed in the past, or if that was a custom fork change.

Would you be able to test some bugfix-2.1.x commits to narrow this down? If there was some past version of Marlin where it worked, and then it got broken, we can track that down within a very short time using a bit of git bisect magic.

The strategy is for you to first find a commit from some point in the "distant" past where the feature works. Then by using git bisect we are able to test the commit from halfway between where we known it worked and where we know it doesn't work. After getting results from each test you tell git bisect good or git bisect bad and it picks the next commit half-way in between your "known to work" commit and your "known to be broken" commit. This process continues until you find the exact breaking change.

If you started from a point 256 commits in the past, it would take no more than 8 tests to find the exact commit that broke it.

The easiest way I found to bring the configs in for testing is to prepare them first, make a commit in another branch, and then cherry-pick that commit after each bisect step. Sometimes the configs need a little cleaning up before the test can proceed, but this is usually very easy. A fancy Git tool can help, but I find it best to resolve merge conflicts by hand in case something is weird.

[Carl393]

I don't know bro, I just asked ai what to do.. and also asked the ender 3 reddit group..maybe I have to set it to #StockCR10 since that is what marlin is designed to work with... Considering the TFT interface has its own firmware and is interacting with Marlin overtop.. maybe this is a workaround... ah shoot no it isnt. This only works with skr e3 mini as is sums the signal together as only uses one cable output from ext3 to the mainboard.... don't understand half of your message unfortunately... mostly when you start talking about git bisect onward. I'm not a programmer I'm new at all this stuff.

[pnt103]

(Replying to @thinkyhead) The TFT35 works in one of two modes.

In "Marlin" mode it uses one 10-way ribbon cable connected to its EXP3 port, exactly like a CR10_STOCKDISPLAY on an SKR Mini E3 or a Creality 1.1x or 4.2.x mainboard, or uses a pair of 10-way ribbon cables connected to its EXP1 and EXP2 ports on an SKR V1.4, SKR, etc. For that, it's normally configured as a REPRAP_DISCOUNT_FULL_GRAPHIC_SMART_CONTROLLER.

In "touchscreen" mode, which is evidently what @Carl393 is using, it behaves like an independent single-board computer, somewhat like OctoPrint does. For that, it uses a dedicated 5-pin serial ribbon cable connected between the mainboard's TFT port and its own RS232 port. That requires no configuration in Marlin, other than having #define SERIAL_PORT_2 uncommented and allocated correctly with the correct baud rate set, and then works exactly as would be the case for OctoPrint, or anything else that works by sending G-code (and M-code) commands over USB to the printer. The TFT sends appropriate commands to the printer, receives responses, and acts on those to update its condition/status accordingly.

So in touchscreen mode the TFT35 acts much more like an OctoPrint instance than like any serial-driven display.

If I initiate an M600 "Change filament" command from the (Marlin) menu on my Ender3Pro/SKR2/Marlin 2.1.2.1, it does exactly what I expect: unloads, prompts to insert new, loads new, purges, offers to purge more or continue. If I try to load filament from the TFT35's touchscreen (ie in touchscreen mode, it does the short slow load, fast long load, purge, and then pops up a dialog box with options "Purge more" and "Continue". So it works for me, at least with Marlin 2.1.2.1.

I suggest first trying that in Marlin mode, sending an M600 (or M701) command over a USB connection to see how Marlin responds.

Hope this helps a little.

[Carl393]

I was told that because I have a SKR 3 EZ with two inputs EXP1 and EXP2 I cannot use the EXP3 output like I do with the SKR mini e3... because only that board has only one input and one TFT input.

I'll check to see if EXT 2 is uncommented and report back

The purge more or Continue prompt shows up in marlin but not on the TFT35 touchscreen mode.. it simply pauses the printer. It used to work..but then again I USED to have the skr mini... which summed the cables into one output (EXT 3) so.... my guess is this is why the bug presents itself.. it works OK on that board but with this skr 3 board because of two inputs (EXT 1 and EXT 2) maybe it's interacting with Marlin differently somehow..im assuming this is a bug but not sure if a bug with Marlin or the TFT35 screen firmware... this screen is typically useful with skr mini however.. hence why I had no issues before... of course trying to contact btt is like trying to contact santa Claus. Will explore a bit more and report back.

[Carl393]

darn, the baud rate is not defined for some reason its
//#define BAUDRATE_2 250000
I'll uncomment and set the baud rate to the other 115200

What should each port be set at?

SERIAL_PORT 1
SERIAL_PORT2 -1 ??

[Carl393]

Im starring to think these two tft models are different and hence one is reprap screen tft35 v3.0 and maybe this one is stockcr10 given it's a different model tft35 e3 v3.0.1 talk about confusing. Ugh

[pnt103]

I was told that because I have a SKR 3 EZ with two inputs EXP1 and EXP2 I cannot use the EXP3 output like I do with the SKR mini e3... because only that board has only one input and one TFT input.

That's correct. For an SKR mainboard with EXP1 and EXP2, you use both, connected to the TFT35's EXP1 and EXP2 respectively. You set the display type in Marlin to REPRAP_DISCOUNT_FULL_GRAPHIC_SMART_CONTROLLER, not CR10_STOCKDISPLAY. But the EXP ports are only used in Marlin mode. They are not used at all in touchscreen mode and don't even need to be connected.

Im starring to think these two tft models are different and hence one is reprap screen tft35 v3.0 and maybe this one is stockcr10 given it's a different model tft35 e3

No, they're both set up as REPRAP_DISCOUNT_FULL_GRAPHIC_SMART_CONTROLLER if you're using the two EXP1 and EXP2 ports.

On an SKR2 and a TFT35 E3 (what I have) you set the serial ports and display as

#define BAUDRATE 250000
#define SERIAL_PORT_2 -1
//#define BAUDRATE_2 250000
    ...
#define REPRAP_DISCOUNT_FULL_GRAPHIC_SMART_CONTROLLER
//#define CR10_STOCKDISPLAY

unless you want a different baud rate for SERIAL_PORT_2 than for SERIAL_PORT. However, the rate set in Marlin for SERIAL_PORT_2 must match the rate set in the TFT35. If you set the wrong rate, the TFT35 won't communicate with the mainboard at all in touchscreen mode, and you'd just get a "No printer attached!" message. 250000 and 115200 are the most commonly used rates. I believe the settings for the serial ports are the same for an SKR 1.4, 2, and 3.

[Carl393]

My apologies, I rebooted the printer and the marlin suddenly loaded with the reprap setting. Will set filament change again to see if I'm still getting the error

[ellensp]

This screen requires a lot of marlin settings, see https://github.com/bigtreetech/BIGTREETECH-TouchScreenFirmware/blob/master/README.md#marlin-dependencies

[ellensp]

Marlin 2.1.2.7 is not what you think it is. It is a older branch of Marin + some patches. Bugfix code has several years worth of updates, fixes and new features that are not in 2.1.2.7

bugfix-2.1.x
... --- S (082fce5) --- B1 --- B2 --- B3 --- ... --- current bugfix
         \
          R1 --- R2 --- ... --- [tag: 2.1.2.4] --- ... --- [tag: 2.1.2.7]

The split point from bugfix-2.1.x is commit 082fce5 Thu Apr 29 00:34:04 2021 (according to git log)

[Carl393]

@thinkyhead Thanks for your response I truly appreciate your help.
I'll upgrade the Marlin Firmware to BugFix and we'll see if it helps
I didn't realize that the BugFix was so far ahead of the stable version.
When it's finished at some point, will there be another stable release?

This reminds me of Debian, where the stable release is often a few years behind the testing unstable and experimental.

Thanks again.

[thinkyhead]

Would you like to help me build the future of open source firmware?

I want to try something new. I'm doing a dry-run of a remote bisect system I've been working on -- basically, I handle all the heavy lifting (finding commits, building firmware, uploading binaries, tracking results), and all you need to do is flash each build and tell me whether the display works or not. It takes about 5-10 minutes per round, and we usually find the exact commit that broke things in 6 or fewer rounds.

This is something we've never done systematically before, and I think it could completely change how we track down firmware regressions. If we nail this down together, it becomes a workflow we can offer to every reporter with a regression on any issue going forward.

You already mentioned upgrading to bugfix -- if the issue persists there, would you be up for being our first test pilot?

-- Perkins