Because I use other parts too. Here is a rough block diagram with the parts that I want to use.
If you can't run the whole system at 2.8V, you might check whether it's possible to just run the MCU and display at that voltage (possible problems: noise immunity and High logic level, need for >Vcc tolerant IOs).
Yet another possibility might be to configure the MCU pins as Open Drain, and use pull ups (but check the timing).
If none of the above applies, there are several level translators that can be used, e.g. TI
TXS0108E and the 4 bit version (that would cover the full display interface). They are not especially chip, though.
There's no need for a specific LCD interface, as the controller chip in this display is hardwired for an 8080-like 8 bit bus.
Moreover, the LCD peripheral of the STM32L4 is good for driving classic segment displays, not TFT ones with an integrated controller.
You are completely free to choose any MCU that:
- Can provide enough IOs, either directly or through I2C/SPI port extenders.
- Can work at the required 2.9V
One MCU peripheral that might come handy is a controller for external memory, which allows you to map the display simply as 2 memory positions (one for commands and one for data), but I don't think it should primarily drive your selection.
So, just choose what you are more comfortable with: Atmel, STM32xxxx, PSoC, etc.
or a MCU that has a parallel / display interface, whatever they want to call it. ST uses FSMC if i remember right
i recommend you READ the device datasheet before settling on the MCU. for example, lower pin count packages usually have only 8 bit data width available and/or no address lines available. Depending on your use case it could be a problem
Yes, FMC/FSMC for STM32s, EMIF component for PSoCs.
Still, even bit-banging the transaction would amount to 5-6 lines of C code as only 8 data bits, 3 control lines (WE,OE,CS) and one address line (command/data) are needed.
If the OP is familiar with a specific brand/family, the effort of learning a new set of peripherals/library or even a new toolchain or IDE may very well be much larger than the coding and testing of a couple of simple functions, and the trade-off in efficiency should be tolerable; that was the reason of my advice.