Remanant residual text on Double_Height on 2x19 matrix

[angeoliver]

Hello,

Thanks for your impressive work.

I have an unsolved problem with the Double_Height feature, on a 2x19 matrix (38 modules). The result is good but there is a final frozen/remanant residual text on only zone 1, on the 6 last left modules. Maybe a buffer overflow with this large matrix ? All examples "Double_Height" scripts have this problem on my configuration, for example Parola_Double_Height_v2.ino

You can see my video of this issue : https://youtu.be/eFnl-EL3cHk

I can't resolve this issue by myself after many hours of research on the 2 libraries. On standard features (not Double_Height) this problem is not present (ex: Parola_Sprites_Library with one zone). Thank you very much for your advice or help please.

Regards,

Olivier

Libraries Versions: MD_Parola 3.7.1 / MD_MAX72XX 3.4.1

Arduino IDE version: 1.8.13 Host OS and Version: Linux Mint CPU Hardware model/type: i3-4170

Script of the test: Parola_Double_Height_v2.ino My tuning: 18: #define HARDWARE_TYPE MD_MAX72XX::ICSTATION_HW 19: #define MAX_DEVICES 19 38: #define SPEED_TIME 10

[MajicDesigns]

Looks like you are running an Arduino Uno? If you have 2 equally long strings of modules then your MAX_DEVICES should be an even number?

Also it looks like the text stops at a module boundary (6 modules if I can count correctly but it is difficult from the video). If this is the case then I would be looking at an electrical/connection problem.

• How are the matrices wired? Are the matrices in the top left of the display in the middle of the display modules or are they at the end?

• How are you distributing the digital signals (except for DAT which is module to module)?

[angeoliver]

Thanks for your reply. I use a Arduino Nano (on back of the matrix). New elements: with default scroll right left (#define SCROLL_LEFT 1) the problem is present but it's 100% OK with scroll right (#define SCROLL_LEFT 0). I confirm: 6 modules 8x8 are concerned by the problem: the 6 last of the chain. I use PCB of 4 modules 8x8 therefore no connection problem I think. A remark: the problem concern the 6 last modules on a total of 38 modules: (38-6)*8 = 256 columns. Pb if column > 256. It's a coincidence ? Or overflow of a uint8_t variable ? Thanks, Olivier

[MajicDesigns]

"I use PCB of 4 modules 8x8 therefore no connection problem I think." Are you able to rearrange the modules in a different order. Does the problem follow the modules or does it stay in the same place (ie, last 6)?

"A remark: the problem concern the 6 last modules on a total of 38 modules: (38-6)*8 = 256 columns. Pb if column > 256. It's a coincidence? Or overflow of a uint8_t variable?" Maybe, but scroll left/right code is the same code with different branching. Also, if this was the problem it would happen you have just one long single line display (which you say does not happen in the original post).

If you can confirm first point it will tell us whether the fault is software related.

Also can you confirm if the modules are joined as 4321 < top line order (incoming connection at 1) 4321 < bottom line order (incoming connection at 1, outgoing connection at 4 to 1 above) or 1234 4321 an also whether you are distributing the signals or just relying on daisy chain from first to last? (see this blog post)

[angeoliver]

Hello,

I found a correction for my issue :

Library MD_MAX72xx.h v3.4.1, line 661 replace param type 8b -> 16b : uint8_t getColumn(uint8_t c) { return getColumn((c / COL_SIZE), c % COL_SIZE); }; by uint8_t getColumn(uint16_t c) { return getColumn((c / COL_SIZE), c % COL_SIZE);};

The final frozen/remanant residual text for columns > 255 on Double_Height with scroll default left is corrected.

Thanks for you time and your work,

Olivier

[MajicDesigns]

Release v3.4.2