EVE intermittent display corruption

[LinusDillon]

I am running an EVE display (EVE3-50A-IPS) through a USB->SPI adapter (EVE2-USB2SPI-KIT-A) connected to a Windows PC. I am using software based on the basic_eve_demo code, however I have notixed some similar issues when pushing bitmaps to the display via the screen designer software as well.

Specifically, at random intervals it seems that a screen update is not entrely succesful - a character might be wrong, a color incorrect, a box positiioned in the wrong place or a bitmap shown partially or with lines through it. I have tried with the original screen timinings from the MatrixEve2Conf.h file, and also with some tweaked timings from another forum post, but neither has helped. I have also tried dropping the SPI clock down, and although that may have helped a little (unsure) the problem still occurs.

It SEEMS to be worse on more complex displays (e.g. bitmaps, etc) but that may just be due to more bytes being pushed and hence more scope for the problem to occur.

Also, I have two otherwise identical EVE3-50A-IPS displays, and one displays a noticable flickering while the other does not. Both have the same intermittent corruption issues however.

So, questions:

• Is this typical/expected bahaviour?

• Is this likely isolated to using dev-kit hardware rather than a dedicated PCB with integrated USB bridge? (if so, eagerly waiting on the updated model of the display that has the integrated USB bridge as an option)

• What mitigations can I apply?

• Is there any way in software of detecting a corrupted command so it can be re-done immediatly?

• What elements of my setup might be causing this?

These are for a production/commercial product, and so solving this is quite important to us.

[Raquel]

Hello,

Thank you for posting on the forum.

It sounds like your connection to the EVE is degraded. Power may also be a factor.
Can you please re-try with 5V supplied on the barrel jack (of the USB bridge).
Also, if you have another FFC cable you can use, please try.

The FFC cable contacts can easily be mangled; be careful when inserting/taking out from the connector.

I hope these help.
Thank you,

[LinusDillon]

I have two seperate displays set up (display + cable + USB bridge), and both show the same behaviour. I've bene careful with the cables and all appears to be connected correctly. One of the units is also being powered via the barrel jack in addition to the USB connection. So I don't think the cable is at fault unless this is typical of this hardware.

Are there any revised timings/parameters or updated libraries (libMPSEE.lib, ftd2xx.lib) for the sample code I mentioned?

I have considered seeing if there is some way of validating the sintructions sent to the display before telling it to execute them, however it seems that this would involve reading the data back OUT over SPI, and so would have a significant impact on performance.

[Raquel]

Hello,

You mentioned that "One of the units is also being powered via the barrel jack in addition to the USB connection" I imagine the one that is not powered is the one that is flickering.

"Is this typical/expected behaviour?" - no, there's definitely something corrupting the data.

Can you please confirm for me that the basic demo is working well?
What you are seeing may be coming from the EVE Screen Designer, or how it is implemented.

Also, if you please show me a short vid clip of what is happening, please email at support@matrixorbital.ca.

Thank you,
Raquel

[bmillier]

hello- I am driving my 5 inch EVE 3 board from the SPI port on a teensy 3.5 ARM MCU. I have experienced the same problem- if I repeatedly load one of the jpeg images included with Matrix Orbital sample programs, it will sometimes load fine, but more often than not, it will be badly rendered. It didn’t matter what SPI clock speed I used - from 1 MHz up to 12 Mhz. I found the problem disappeared when I hooked up my ‘scope to the SCK line.That adds 10-20 pdf of capacitance to the SCK signal and reduces ringing/overshoot . So, I added a 22 pf capacitor from SCK line to ground and everything is fine now.
I am not using the flex cable like you are from the USB2SPI. I am using about 6 inch of ribbon cable between the Teensy MCU and the PCB pads on the edge of the board which duplicate the flex cable pins.
May not be your problem but I figured it is good to get it out on the forum, in case it helps others
Cheers

[Rudolph]

When playing with the Teensy 4.1 I found it to be working only when I connected the logic-analyzer.

I was using one of my adapters:
https://github.com/RudolphRiedel/EVE_di ... 05_EVE.PDF

So this already is decoupled from the display itself.

I changed R3, R5 and R8 from 10R to 150R and added a 22pF between pin 1 of U2A and GND.
And now it works fine with 20MHz on the SPI.

None of the other boards I tried my adapters with needed this change so this is an "issue" with the controller on the Teensy 4.1.