TheBackShed.com

Forum Index : Microcontroller and PC projects : ARMmiteb18: In-built TFT display handling

Page 1 of 2

Author

Message

matherp
Guru

Joined: 11/12/2012
Location: United Kingdom
Posts: 10246

Posted: 10:57pm 19 Mar 2015

Copy link to clipboard

Print this post

Attached is the latest beta release of the Micromite code for the STM32F407.

2015-03-20_083853_ARMmiteb18.zip

For details of how to get started with the ARMmite see this post.

Please note the default console on this release is the UART, Console RX is on pin PA3, TX is on pin PA2. You can use OPTION CONSOLE USB to swap the console to the USB port but please note some people are finding reliability problems with this. See this post for details of how to get started with USB.

This release includes full Cfunction capability and huge thanks are due to Peter Carnegie G8JCF for implementing this and helping with other "heavy" issues. Documentation on this will follow at a later date.

The other main change in this release is the implementation of in-built graphics routines. The following controllers are currently supported:
Parallel
ILI9341P : many 2.4" and 2.8" displays with an 8-bit parallel interface
ST7781 : other 2.4" displays with an 8-bit parallel interface
SSD1289 : 3.2" displays with a 16-bit parallel interface
SSD1963 : 5" and 7" displays with a 16-bit parallel interface
SPI
SSD1351 : Colour Oled displays with an SPI interface
ILI9341S : many 2.4" and 2.8" displays with a SPI interface
ST7735 : 1.8" displays with an SPI interface

Before using a display it must be initialized. The syntax for the parallel displays is:

DISPLAY OPEN displayname, command/data pin, WR pin, reset pin, databus (see below), screen width, screen height

The databus parameter can be:
0:use pins PD0-PD7 (use as required for ILI9341P, ST7735)
8:use pins PD8-PD15 (use as required for ILI9341P, ST7735)
16: use pins PD0-PD15 (mandatory for SSD1289 and SSD1963)

examples:
DISPLAY OPEN ILI9341P,1,2,3,0,240,320
DISPLAY OPEN SSD1963,1,2,3,16,800,480

In these examples I am using pins PE2, PE3, and PE4 as the signal pins.
Note that the chip-select pin on the display should be tied to GND, the RD pin on the display should be tied to VCC

The syntax for the serial displays is:

DISPLAY OPEN displayname, SPI port number, command/data pin, chip select pin, reset pin, screen width, screen height

The SPI port number can be 1, 2, or 3 corresponding to pins labelled SPI, SPI2, and SPI3 in the pinout. Note in general SPIIN pins need not be connected but should not be used for other purposes. SPIOUT from the ARMmite connects to SPIIN on the display (may be labelled MOSI)

e.g.
DISPLAY OPEN ILI9341S,2,1,2,3,240,320

If you have connected a display correctly, the OPEN command will initialize it and clear the display. If it doesn't you probably have one or more connections reversed.
DISPLAY CLOSE disconnects all the pins from the display and returns the pins for normal use

Once the display is open there are a range of drawing commands available:

DISPLAY font, xstart, ystart, foreground colour, background colour, textstring, (optional x-scale, or x and y if only this parameter specified),(Optional y-scale)
10 fonts are currently included as shown in the picture. The code includes wordwrap and protection against writing outside the screen area. All fonts can be scaled using the x-scale and y-scale parameters but note this does introduce a performance overhead.

PIXEL x-coordinate, y-coordinate, colour

LINE x-start-coordinate, y-start-coordinate, x-end-coordinate, y-end-coordinate, colour

CLS (optional parameter colour)

BOX x-coordinate, y-coordinate, width, height, colour, (optional parameter F to fill)

CIRCLE x-coordinate, y-coordinate, radius, colour, (optional parameter F to fill)

TRIANGLE x1-coordinate, y1-coordinate, x2-coordinate, y2-coordinate, x3-coordinate, y3-coordinate, colour, (optional parameter F to fill)

BITMAP x-coordinate, y-coordinate, foreground colour, background colour, bitmap (long long integer giving 8 x 8 bitmap). The code treats this as an array of type char[8]. This is read bit7 to bit0, array-index[0] to array-index[n]. The screen then updates the 8 x 8 box top left to bottom right.

RBOX x-coordinate, y-coordinate, width, height, radius of corners, colour, (optional parameter F to fill)

ROTATE orientation (only supported for screens with hardware rotate support, ILI9341S, ILI9341P, ST7735, SSD1963). Orientation can be:
0: Portrait
1: Portrait 180
2: Landscape
3: Landscape 180
For displays that support it, is quite OK to repeatedly switch orientation, this allows you to write text in any orientation and then switch back to your standard orientation.

In general all normal Micromite features should work as per the 4.6b manual. Other additional features include:
OPTION VCC n.nnn
If you measure accurately the supply voltage to the microprocessor, then using this option will calibrate both the ADCs and the DACs correctly. The option is stored in flash memory for subsequent restarts and power cycles.
The Discovery board has a diode between the 3.3V regulator and the chip supply which drops the voltage, "OPTION VCC 2.92" works for my board. You can then confirm the setting by using "print pin(0)" which should then report 1.2V.
Valid values are 1.8V to 3.6V - the allowable range for the chip.

PWM and SERVO both support up to 8 channels with 4 on each timebase i.e.
PWM 1,10000,20,40,60,80
SERVO 2,1.44,1.66,1.88,2.22

The STM32F407 has two independent onboard 12-bit DACs. The format of the Basic commands is:
DAC number,voltage ' number can be 1 or 2, voltage is a float between 0 and 3.3v
DAC number,STOP ' turn off the DAC

There are dedicated Basic commands and functions for using pins PD0 to PD15 as a parallel databus
SETPORTD databus , MODE, OPTIONS
The databus parameter can be:
0:use pins PD0-PD7 (use as required for ILI9341P, ST7735)
8:use pins PD8-PD15 (use as required for ILI9341P, ST7735)
16: use pins PD0-PD15 (mandatory for SSD1289 and SSD1963)
MODE can be DIN, or DOUT, or OFF to disable the port and return the pins for general use
OPTIONS can be OC for output or PULLUP, PULLDOWN for input

PORTD(databus)=nnn
Sets the pins specified by databus to the value nnn. If databus is 0 or 8 this should be a number between 0 and 255. If databus is 16 this can be a number between 0 and 65535

variable =PORTD(databus)
Reads the pins defined by databus. If databus is 0 or 8 this will return a number between 0 and 255. If databus is 16 this will return a number between 0 and 65535

The pinout is:

Edited by matherp 2015-03-21

G8JCF

Guru

Joined: 15/05/2014
Location: United Kingdom
Posts: 676

Posted: 04:34am 20 Mar 2015

Copy link to clipboard

Print this post

Working with Peter Mather is exhausting

The effort PeterM puts in is out of this world far in excess of anything I've contributed to the ARMMite port.

The thing is that with all the memory, and I/O capabilities of these ST ARM MCUs there is so much more that can be done - as evidenced by the huge amount of devices PeterM keeps adding to ARMMite for example !

Peter

The only Konstant is Change

WhiteWizzard
Guru

Joined: 05/04/2013
Location: United Kingdom
Posts: 2933

Posted: 09:25am 20 Mar 2015

Copy link to clipboard

Print this post

Great work Peter (and Peter!) - lots of useful things to keep everyone here happy.

So dare I ask; would it be easy to incorporate into this list the SSD1306 (which is the 'original' I2C 0.96" OLED you got working) OR would it be a right 'pain'? If doable then I believe that you have captured all the display controllers that have been mentioned here on the Shed.

Am getting a SSD1963 controller board next week to test this stuff (along with the cheap ARM 'kit')

Great stuff . . . .

WW

boss

Senior Member

Joined: 19/08/2011
Location: Canada
Posts: 268

Posted: 11:09am 20 Mar 2015

Copy link to clipboard

Print this post

@matherp

Hi Peter,

this release is awesome progress indeed.
Maybe I found the glitch - embedded EDITOR doesn't work for me although via MMedit I everything seems be O.K.

Btw. I have SSD1963 shield that allows you connect the ordinary PC monitor via DVI, but unfortunately the it use 8bit bus. It is tested and works fine. I would like to ask you if this is interesting enough to write ARMite driver or I have use the own driver with CFunction.

Regards
Bo

matherp
Guru

Joined: 11/12/2012
Location: United Kingdom
Posts: 10246

Posted: 12:22pm 20 Mar 2015

Copy link to clipboard

Print this post

Quote Maybe I found the glitch - embedded EDITOR doesn't work for me

Works for me using the hex file above. make sure you erase all before programming and put a link between boot0 and VDD while programming

boss

Senior Member

Joined: 19/08/2011
Location: Canada
Posts: 268

Posted: 01:08pm 20 Mar 2015

Copy link to clipboard

Print this post

@matherp

Boot x VDD connected, the chip was completely erased, then flashed and verified. The result is the same. Everything looks fine, even flash x file verification (see picture). No other problem was found yet, but embedded Editor.

G8JCF

Guru

Joined: 15/05/2014
Location: United Kingdom
Posts: 676

Posted: 01:19pm 20 Mar 2015

Copy link to clipboard

Print this post

The inbuilt MMBasic editor should of course work, BUT, is anybody who is serious about developing an embedded application doing so directly on the target MCU ?

MMEdit is a really, really, great IDE for MMBasic

Peter

The only Konstant is Change

boss

Senior Member

Joined: 19/08/2011
Location: Canada
Posts: 268

Posted: 02:24pm 20 Mar 2015

Copy link to clipboard

Print this post

Hi Peter,

first I'd like say that ARMite was my dream for years (and Cfunction is above the dream). You and Peter did awesome progress with MM(Mk)basic. Thank you.

Now regarding the Editor - actually I'm personally prefer MMEdit as well, but I'm always trying to test as many MMbasic features as I can. Because if somewhere is a bug it is always wise to find and correct it.

Regards
Bo

Grogster

Admin Group

Joined: 31/12/2012
Location: New Zealand
Posts: 9593

Posted: 03:31pm 20 Mar 2015

Copy link to clipboard

Print this post

I'll just add my 2c here - I never use the editor for more then minor changes to test things. All writing of code is in MMEDIT.

Now, this is NOT to take away from the built-in editor.
I think this concept was a SUPERB idea for the original series of chips, as you hook the chip up to a computer and Tera-Term or similar, and you can write and edit your program directly on the chip - that is a great feature for newcomers and beginning programming, as you don't need to worry about learing one single jot of any kind of computer-based IDE development program.

On the ARMmite?

Probably not much use for developing code, but still a useful tool to have there in the background, IMHO. Having said that, I would not want to develop a program for the ARMmite using only the editor, as the coloured syntax of an IDE is really, really useful for bigger sized code listings.

...again - my 2c only, as others are commenting on this at the moment....

Smoke makes things work. When the smoke gets out, it stops!

matherp
Guru

Joined: 11/12/2012
Location: United Kingdom
Posts: 10246

Posted: 10:27pm 20 Mar 2015

Copy link to clipboard

Print this post

Quote No other problem was found yet, but embedded Editor.

What terminal are you using?

It works with teraterm but I think the dataflow may be too fast for the MMEDIT VT100

boss

Senior Member

Joined: 19/08/2011
Location: Canada
Posts: 268

Posted: 04:44am 21 Mar 2015

Copy link to clipboard

Print this post

@matherp

Terraterm 4.86 and W 8.1 pro. It behaves madly. Greeting message prints 9 times, the reset button seems not working now. I'll install one more disk and try it again with fresh installed W7. What kind od Windows you have installed? Could virus infect hex file? I don't know. b12 seems working fine.

Bo

matherp
Guru

Joined: 11/12/2012
Location: United Kingdom
Posts: 10246

Posted: 05:57am 21 Mar 2015

Copy link to clipboard

Print this post

Quote Greeting message prints 9 times, the reset button seems not working now

That sounds like the reset pull-up resistor is missing/dry joint

I'm using W7 64-bit, Teraterm 4.82, FTDI (possibly) adapter

Also, make sure you are only using three wires from the USB/TTL adapter, GND, RX, TX - you don't want to connect VCC

Attached is a version of b19 that should default to USB rather than UART for the console

2015-03-21_160543_ARMmiteb19USB.zip Edited by matherp 2015-03-22

boss

Senior Member

Joined: 19/08/2011
Location: Canada
Posts: 268

Posted: 07:18am 21 Mar 2015

Copy link to clipboard

Print this post

@matherp

All mine FTDI went to the garbage can as I decided upgrade to 8.1 and FTDI doesn't support it under W8. I switched to CP210x and they working fine.

there are some picture, the Discovery board is firmly inserted in big prototype board, there could be absolutely no Discovery board damage

Hope that iPhone didn't flip picture upside down as happened last time

Will continue with W7

Thanks
Bo

matherp
Guru

Joined: 11/12/2012
Location: United Kingdom
Posts: 10246

Posted: 08:04am 21 Mar 2015

Copy link to clipboard

Print this post

I can only suspect something about your proto board is shorting out the NRST and GND pins - this is definitely not a software issue

boss

Senior Member

Joined: 19/08/2011
Location: Canada
Posts: 268

Posted: 08:24am 21 Mar 2015

Copy link to clipboard

Print this post

@matherp

nope, tested NRST is working fine, fortunately I have another Discovery on its way. I plan to play with this mad one on W7 just determine what is wrong (HW or SW).

Thanks
Bo

matherp
Guru

Joined: 11/12/2012
Location: United Kingdom
Posts: 10246

Posted: 09:06am 21 Mar 2015

Copy link to clipboard

Print this post

Can you unplug the Discovery from the protoboard? Then remove the shorting link SB11 on the bottom, also the two ST-LINK jumpers on the top. That will disconnect the st-link from the 407. Then just connect the RX, TX, and GND from the CP210x and the USB into CN1 to power it up. That eliminates all outside influences

boss

Senior Member

Joined: 19/08/2011
Location: Canada
Posts: 268

Posted: 01:03pm 21 Mar 2015

Copy link to clipboard

Print this post

@matherp

I did as recommended and the result was the same then reconnect the jumpers back, erased, re-flashed and the result was same.
I believe it happened when I put jumper between boot0 x VDD fleshed then probably forgot remove this jumper.

SteveP
Newbie

Joined: 21/03/2013
Location: United States
Posts: 19

Posted: 08:02pm 29 Mar 2015

Copy link to clipboard

Print this post

I am unable to produce any display on a ILI9341S device, TJCTM24024-SPI, a 2.4 TFT SPI 240*3202 using ARMmite B23.
Display Open ILI9341S,1,1,2,3,240,320 makes the display dim but then commands such as CLS (white) or CLS (blue) or line (10,10,50,50,white) or DISPLAY 5,2,2,white,black,"test" do nothing. Using built-in editor, no errors. If I intentionally include bad syntax in these commands, an error does result.
Display Close returns the display to its brighter intensity.

Connections from the display to STM32F407 Discovery board are :
VCC to 5 V
GND to GND
CS to PE3 (pin 2)
RESET to PE4 (pin 3)
D/C to PE2 (pin 1)
SDI (MOSI) to PB5 (pin 91)
SCK to PB3 (pin 89)
LED to 5 V through 150 ohm resistor

Tried 2 different display units.

matherp
Guru

Joined: 11/12/2012
Location: United Kingdom
Posts: 10246

Posted: 09:20pm 29 Mar 2015

Copy link to clipboard

Print this post

The LED pin on my display (same part number) is actually the drive to the backlight not a logic signal. I run the display at 3.3V. Use VDD from the Discovery board to connect to VCC on the display (make the shorting link J1 to bypass the regulator) and use an external power supply + 3.3V with common ground to connect to LED.

I've just downloaded the b23 hex and run your exact configuration and it works perfectly for me.

Just tried connecting LED direct to 3V or 5V on the Discovery - both work fine.

Also try checking/replacing the D/C connection wire. I can reproduce your results if I disconnect D/C completely
Edited by matherp 2015-03-31

matherp
Guru

Joined: 11/12/2012
Location: United Kingdom
Posts: 10246

Posted: 04:38am 30 Mar 2015

Copy link to clipboard

Print this post

Another thought:

Are you using literals "white", "red" etc?

These aren't defined unless you define them e.g.
const red = &Hf800

otherwise the interpreter is using these as variables that are all set to 0 i.e "black"

Use OPTION EXPLICIT in every program and this will never happen!!!!!!!!!!!!!!

Page 1 of 2

Print this page