LCD read issue on STM3240G-EVAL

[pmoore]

Posted on March 29, 2013 at 22:16

Does anyone have code that is able to successfully read the ID from the LCD on the STM3240G-EVAL board? 

Using the ST provide examples (without LCD reads), we can write to the LCD without any issues, but when trying to read instruction register 00h (driver code) from the Ampire LCD, we always get back whatever data we last wrote to the data register.  It's as if the LCD never writes fresh information into it's data register. Vastly extending the total NE3 period and the NOE active period (at least to 500nS, validated with a scope) does not affect the pattern, so we are given the LCD plenty of time to write new data to the register.  This system is running with the core at 168MHz, so HCLK ticks are slightly under 6 nanoseconds.

[emalund]

Posted on March 29, 2013 at 22:31

Does anyone have code that is able to successfully read the ID from the LCD on the STM3240G-EVAL board?  

I do not know the LCD you speak of, but have worked with several. If the LCD is open drain on output (very common) you can't read it without active pullups. Do you have that  (enabled)? also if you (accidentally) have switched the LCD to ''transmit mode'' (telling it you want to read) while the uC ports are configured as push-pull, the ability to read the LCD may be gone forever (been there, done that) Erik

[pmoore]

Posted on June 06, 2013 at 03:26

Erik:

It's the Ampire AM-240320D5TOQW-T00H(R) display ST puts on the large STM3240G-EVAL board. We can read other locations, so the LCD controller is writing the bus. I think this project is headed for a cheaper and smaller resolution monochrome pixel display, so it likely doesn't matter anymore...and we'll get the joy of learning how to talk to a new display. 🙂 -Patrick

[jj2]

Posted on June 06, 2013 at 17:15

Long ago - very similar QVGA TFT (same maker - on F103 Eval Bd) changed control chips on the display.  Can you read - and present here - any/all chip ID markings?

IIRC - F103 demo board ran code which successfully performed - just what you seek.  Unsure - but perhaps one key control bit ''flipped'' for just this Register read.  (insane yes - but that's my best recollection...)