@gloomyandy said in RepRapDiscount display quite unresponsive in 3.4.0:

@3discnc said in RepRapDiscount display quite unresponsive in 3.4.0:

@dc42 @oliof @gloomyandy

Still think there's a bug in RRF3.4.x though. Different chip versions, I gather.

I thought you said that you went back to 3.3.0 and that you still had the same problem?

YAT+M115 showed me a version number of 3.3x, so it must have been properly installed. Hence my suspicion that 'something' got persistently stuck somewhere. Al these boards can have different chip versions or even close clones of them. I'm (not) guessing the Creality panel has a chip different from the BTT panel.

But, I had issues with the BTT before 3.4.x, the cursor often appeared to run backwards. Changes to the LCD part of the firmware perhaps started in 3.3.x already?

@dc42 @oliof @gloomyandy

Solved it (kind of) by replacing the RepRap LCD with the original one the Ender 3 came with. At the time of the Maestro swap I did not know the receptacles had their notches reversed and I ordered the almost similar BTT panel.

If you either remove the notches from the cable or cut notches in the opposite side of the receptacles the Ender panel seems to work flawlessly. Standard filament cutters are perfect for this.

The only difference seem to be the direction of rotation of the encoder. Easily solved per M918 in config.g

Still think there's a bug in RRF3.4.x though. Different chip versions, I gather.

From the 3.4.x release notes: "Improved the speed of display refresh on 12864 displays with ST7567 controllers". I have no clue which chip are on the BTT or Creality panels (these are epoxy blobs) but some instruction must be in the wrong place.

@oliof said in RepRapDiscount display quite unresponsive in 3.4.0:

@3discnc I don't think there is a way to program the 12864s. It sounds like the encoder knob has an issue.

Do you have another device to test the display with where it works better?

I have two of these panels and they behave the same. About 10% of the commands or so halt the menu system for a rather hit-and-miss experience.

@gloomyandy said in RepRapDiscount display quite unresponsive in 3.4.0:

@3discnc If you think something has changed between RRF 3.3 and 3.4 probably the best way to be sure is to revert to 3.3 and see if that fixes your problems. You may be able to get away without any config changes (especially if all you need to test is the LCD).

Also tried that, to no avail. Even reverted my latest menu changes back to januari. Same difference.

Two options come to mind:

• The recent changes to the 12864 LCD are persistantly programmed to the panel or the Maestro.
• The Maestro has developed a problem. Perhaps an aluminium shaving or a tiny piece of copper wire came loose while stripping and dropped on the PCB. I've seen no other signs of any issues.

Thing I noticed: If the cursor (white rectangle) seems to get stuck, I can 'loosen' it by 3-4 rotary clicks of the encoder. (also loosens by holding the knob down for a sec).

@oliof

Good plan. Tried it, but same problem.

@dc42

I've rebuild my Ender3/Maestro after five weeks of non-use and updated RRF to 3.4.0.

I have two LCD displays marked:

"RepRapDiscount, Full Graphic Smart Controller, www.bigtree-tech.com".

Neither panel works properly with 3.4.0 (did work up to, and including, 3.3). It kind of does all that is needed but is at places unresponsive. It often stops at menu options, it often does not listen to a selection press. If the latter happens, I need to hold down the rotary knob for 0.25-1 second to 'wake something' up. It seems to have issues with specific options; stops on them dead, often, but not always.

This could be my own doing as I've made countless changes to the macro, menu and system files. But, if I remember correctly, you said there were changes made to the 12864 mechanism and if the culprit is there, I'll never going to find it.

Before I go on a wild goose chase, I'd better ask first if the above rings a bell at Duet3d?

@dc42 said in N539/N538 not working in 3.4:

Sorry about the late response. I took the machine apart a month ago for some modifications and it took a month+ to receive the parts. I'm in the process of putting it together again and some reprogramming will be needed too.

@3discnc said in N539/N538 not working in 3.4:

1. I was using N539 in RFF3.3 on the 12864 LCD; in 3.4 it shows 0.

Thanks for reporting this. I have now fixed it. The fix will be in 3.4.0 stable.

2. I am designing a replacement G29 grid scanner and run into the limited string space of RFF. I cannot find any means to truncate a float and so cannot pack enough floats (17+?) into this limited 120 byte string space. Any advice on this?

Perhaps you could use multiple strings with fewer floats in each?

Already tried that but how do I put the two strings in the height map as a single sting (i.e. no CR/LF)?

3. Since there are no arrays I had to resort to 20 variables. Apparently these are not directly addressed or looked-up but searched for at each access, this makes them very slow. The further down, the slower. Perhaps something to look into. I've shortened the var's to two characters, that helped a bit

We will probably support array-valued variables in RRF 3.5.

Those will be welcomed with open arms!

4. I also have no clue how to use '#' to get a stringlen value. How is this supposed to work?

Trivial example:

echo #("hello"^"world")
10

That takes care of strlen() but, it would greatly help if I could truncate strings since I don't need fractional nanometre resolution (yet). ;o)

5. My Maestro manages to erase the CD card from time to time. I've seen this happen on several crashes into the X end-stop and a (probably) non ending loop.

Does this happen while your macros are writing to the SD card?

Don't really know and also it does not happen every 5 minutes. I home (or some process) the machine and it crashes as soon as X-stop is hit. The LCD is emptied and the Maesto does not respond of even reboot. On my PC the card is not recognized (Windows asks Reformat Y/N?). I reformat and copy a backup onto the card and all is well. How well depends on the amount of work lost. I was intensively modifying the system macros over the network, with the accompanying reboots and whatnot.

1. I was using N539 in RFF3.3 on the 12864 LCD; in 3.4 it shows 0.

2. I am designing a replacement G29 grid scanner and run into the limited string space of RFF. I cannot find any means to truncate a float and so cannot pack enough floats (17+?) into this limited 120 byte string space. Any advice on this?

3. Since there are no arrays I had to resort to 20 variables. Apparently these are not directly addressed or looked-up but searched for at each access, this makes them very slow. The further down, the slower. Perhaps something to look into. I've shortened the var's to two characters, that helped a bit

4. I also have no clue how to use '#' to get a stringlen value. How is this supposed to work?

5. My Maestro manages to erase the CD card from time to time. I've seen this happen on several crashes into the X endstop and a (probably) non ending loop.