Display TFT ST7789 (OshonSoft Basic).

[DogFlu66]

I leave the library to use ST7789 TFT display, I have translated one that is in C language. But unfortunately I have not been able to get it to work yet. I leave the files in case someone with more experience with these displays can review it.

 

[camerart]

That sounds a very strange thing to do?, the problem is almost certainly something you're doing wrong.

First off, have you disabled slew rate control? (SLRCON registers), the default is ON, and this cripples SPI speed - it 'might' work at 8MHz, it might not, or it might work 'sometimes' - it's something you absolutely need to turn OFF. However, if it's not working at 2MHz, that's unlikely to be the issue - but it still wants turning OFF (it's up there with ANSEL, the defaults are stupid).

Hi N,
SLRCON =0.
C

 

[camerart]

Hi Nigel, The power supply will need to be concreate for this device to work on a 3V3 supply
I have had issues before running these chinesium LCD's, as have you. On the Fantastic "Geoff's projects" pages, they used the same screen on Pico's and Micromites (Pic32) but they supply 5V for the screen's power.

When I tested this on my breadboard, it works with a pic18f46k22, but not a pic18f46k20. the latter is 3v3 only... I think this is the problem The PSU on the breadboard is a 3V3 switch mode 7803 replacement.

Hi I,
I use 12V batteries, then 7805, then 3.3v LDO to get the 3.3v.
I've put the Regs that you suggest in the Farnell Basket.
C.

 

[camerart]

Hi,
Today, I took the PIC off one of the working boards, and swapped it with the PIC on the none working board, and 'in this case' with PICs swapped, the boards gave the same result, in the the previously working board still works.

So in this case it is the board that causes faults, so whether it's the Voltage regulation or capacitance something is happening beyond my understanding.

I've changed all of the other components on this board previously.
C.

 

[Nigel Goodwin]

Hi N,
SLRCON =0.
C

Good

 

[Nigel Goodwin]

Hi I,
I use 12V batteries, then 7805, then 3.3v LDO to get the 3.3v.
I've put the Regs that you suggest in the Farnell Basket.
C.

Are they cheaper at Farnell than at RS?, often things are - but not always, and sometimes they are more expensive?. It's also worth checking CPC (a division of Farnell) as they are often considerably cheaper.

For example, last week I was re-ordering some LED mounting clips, which we normally source from Farnell, they were £6.50 odd for 25, but out of stock, expected delivery date 10 weeks.

So I checked RS, in stock, but similar price for only FIVE!! - so I googled the manufacturers (Kingbright) part number, and up popped CPC, in stock, and only £4 for 25 - happy days!

On a similar theme, we use a LOT of four core cable, originally from RS, but it was always out of stock, then the P/N changed, more out of stock, and going from the number in stock when it was, we were the only people buying it Eventually they discontinued it entirely, and we ended up using a slightly inferior American cable from Farnell.

It's available on 100ft or 1000ft drums (because it's American), and rather bizarrely it was considerably cheaper to buy 10x100ft drums, than one 1000 foot drum (no problem, because we cut it into short lengths) - which made no sense. Anyway, we've been happily buying the little drums. But the other week I had to order more, so (as always) I checked the 1000ft drums - it was now cheaper than ten 100ft drums - so I ordered five drums, as there's a price break at five, they only had three in stock, but you get the price break anyway, and the other two will arrive later (they haven't yet).

It's often an adventure ordering parts!

 

[camerart]

Are they cheaper at Farnell than at RS?, often things are - but not always, and sometimes they are more expensive?. It's also worth checking CPC (a division of Farnell) as they are often considerably cheaper.

For example, last week I was re-ordering some LED mounting clips, which we normally source from Farnell, they were £6.50 odd for 25, but out of stock, expected delivery date 10 weeks.

So I checked RS, in stock, but similar price for only FIVE!! - so I googled the manufacturers (Kingbright) part number, and up popped CPC, in stock, and only £4 for 25 - happy days!

On a similar theme, we use a LOT of four core cable, originally from RS, but it was always out of stock, then the P/N changed, more out of stock, and going from the number in stock when it was, we were the only people buying it Eventually they discontinued it entirely, and we ended up using a slightly inferior American cable from Farnell.

It's available on 100ft or 1000ft drums (because it's American), and rather bizarrely it was considerably cheaper to buy 10x100ft drums, than one 1000 foot drum (no problem, because we cut it into short lengths) - which made no sense. Anyway, we've been happily buying the little drums. But the other week I had to order more, so (as always) I checked the 1000ft drums - it was now cheaper than ten 100ft drums - so I ordered five drums, as there's a price break at five, they only had three in stock, but you get the price break anyway, and the other two will arrive later (they haven't yet).

It's often an adventure ordering parts!

Hi N,
I don't buy much stuff really, and I do shop around.

Sometimes when I need something sooner, I compare baskets, and with 'no postage' one is cheaper, although other times, it's just easier to pay more than shop for hours.
Also I check the basket and have changed my mind as I may have 'gone off' a project. Retirement ay
C

 

[camerart]

Hi,
Ian suggested that the power supply could be a problem!

The PIC is normally powered by a 12V battery, though a 7805, then through a 3.3V LDO.
I tried a 5V battery before the LDO and a 3V cell after the LDO, and the PIC works powered by 3V after the LDO, and sometimes can be jumped to work 5V before the LDO.
There are capacitors, all around!
The crystal always works, but in this case, not working means a TFT screen using SPI.
It doesn't work when the 3.3V comes from my rework station.

So my PCB is affecting the 18F46K20! and in this case the weakest link is the 3.3V LDO
This problem is so intermittent, that it's difficult to figure out exactly how to proceed.
C

 

[DogFlu66]

That sounds a very strange thing to do?, the problem is almost certainly something you're doing wrong.

First off, have you disabled slew rate control? (SLRCON registers), the default is ON, and this cripples SPI speed - it 'might' work at 8MHz, it might not, or it might work 'sometimes' - it's something you absolutely need to turn OFF. However, if it's not working at 2MHz, that's unlikely to be the issue - but it still wants turning OFF (it's up there with ANSEL, the defaults are stupid).

I have verified that the Oshonsoft compiler leaves it at 1 by default, in my case as I am using the Pic18F46K22 the value is SLRCON = %11111.

In my case, since all my projects work just fine, I keep it enabled. I still build a few that get installed in places with loads of electronics around. Honestly, I had totally forgotten about this register — I checked once that it was enabled by default and haven’t thought about it since.

 

[Ian Rogers]

@ Nigel post #380.. But That's good.. As most devices run up to 3.6V with no issue I use those batteries on a lil wireless unit we supply. The board gets up to 4.2V and I lick it down to 3.3V THEN kick a small bit up to 12v and a bit to 5v... Those screens will run at 3.6v I have tried it.

 

[camerart]

I have verified that the Oshonsoft compiler leaves it at 1 by default, in my case as I am using the Pic18F46K22 the value is SLRCON = %11111.

In my case, since all my projects work just fine, I keep it enabled. I still build a few that get installed in places with loads of electronics around. Honestly, I had totally forgotten about this register — I checked once that it was enabled by default and haven’t thought about it since.

Hi D,
All of my projects worked ok, until they didn't!
On the REMOTE CONTROL project I worked on for years, all of the peripherals worked, then started mysteriously not working as I consolidated them into the final units.

This project soon highlighted the problems.
C

 

[camerart]

Hi,
I've been looking at the 18F46K20 ERRATA D/Ss, and many are related to SPI.
The previous project using these boards, had 5 or 6 peripherals, all using SPI, and set to a particular SPI MODE. I haven't found yet a definitive setting for these ST7789 TFT screens yet, or whether they need to be set to 'my' settings, but suspect they are slightly different than my settings perhaps the CKL at REST setting causing these misfires?

These misfires, could of course be triggered by Voltage regulators or capacitance etc
C.

 

[camerart]

Hi D,
I've been looking though your CODEs as best as I can, trying to understand how you set up the SPI settings.
I think I've found which MODE you use: MODE 2 = CKP=1 and CKE=1
So far I haven't been able to find the SMP setting?
SSPSTAT.7

Can you clarify these 3x settings please?
Cheers, C.

'Example: Call _SPI1_Init(_SPI_MODE0, _SPI_MASTER_SPEED1 | _SPI_ENABLE) 'CLK 4Mhz (64Mhz)
Proc _SPI1_Init(mode As Byte, speed As Byte)
    ' Configure SPI pins
    ConfigPin _SPI_SCK = Output ' SCK as output
    ConfigPin _SPI_SDI = Input  ' SDI as input
    ConfigPin _SPI_SDO = Output ' SDO as output

    ConfigPin _SPI_RST = Output ' RESET as output
    ConfigPin _SPI_DC = Output  ' DC (SS) Data/Command pin as output
    ConfigPin _SPI_CS = Output  ' CS Chip Select pin

    ' Estado inicial de los pin
    _SPI_RST = 1    ' RST RESET pin
    _SPI_DC  = 1    ' DC (SS) Data/Command pin
    _SPI_CS  = 1    ' CS Chip Select pin
    '_SPI_SCK = 1    ' SCK as output
    '_SPI_SDO = 1    ' SDO as output -> SDI

    ' Configure SPI operation mode
    Select Case mode
        Case 0
            SSP1CON1.4 = 0 ' CKP = 0
            SSP1STAT.6 = 1 ' CKE = 1
        Case 1
            SSP1CON1.4 = 0 ' CKP = 0
            SSP1STAT.6 = 0 ' CKE = 0
        Case 2
            SSP1CON1.4 = 1 ' CKP = 1
            SSP1STAT.6 = 1 ' CKE = 1
        Case 3
            SSP1CON1.4 = 1 ' CKP = 1
            SSP1STAT.6 = 0 ' CKE = 0
    EndSelect

    SSP1CON1 = SSP1CON1 | speed

End Proc

 

[DogFlu66]

Hi D,
I've been looking though your CODEs as best as I can, trying to understand how you set up the SPI settings.
I think I've found which MODE you use: MODE 2 = CKP=1 and CKE=1
So far I haven't been able to find the SMP setting?
SSPSTAT.7

Can you clarify these 3x settings please?
Cheers, C.

'Example: Call _SPI1_Init(_SPI_MODE0, _SPI_MASTER_SPEED1 | _SPI_ENABLE) 'CLK 4Mhz (64Mhz)
Proc _SPI1_Init(mode As Byte, speed As Byte)
    ' Configure SPI pins
    ConfigPin _SPI_SCK = Output ' SCK as output
    ConfigPin _SPI_SDI = Input  ' SDI as input
    ConfigPin _SPI_SDO = Output ' SDO as output

    ConfigPin _SPI_RST = Output ' RESET as output
    ConfigPin _SPI_DC = Output  ' DC (SS) Data/Command pin as output
    ConfigPin _SPI_CS = Output  ' CS Chip Select pin

    ' Estado inicial de los pin
    _SPI_RST = 1    ' RST RESET pin
    _SPI_DC  = 1    ' DC (SS) Data/Command pin
    _SPI_CS  = 1    ' CS Chip Select pin
    '_SPI_SCK = 1    ' SCK as output
    '_SPI_SDO = 1    ' SDO as output -> SDI

    ' Configure SPI operation mode
    Select Case mode
        Case 0
            SSP1CON1.4 = 0 ' CKP = 0
            SSP1STAT.6 = 1 ' CKE = 1
        Case 1
            SSP1CON1.4 = 0 ' CKP = 0
            SSP1STAT.6 = 0 ' CKE = 0
        Case 2
            SSP1CON1.4 = 1 ' CKP = 1
            SSP1STAT.6 = 1 ' CKE = 1
        Case 3
            SSP1CON1.4 = 1 ' CKP = 1
            SSP1STAT.6 = 0 ' CKE = 0
    EndSelect

    SSP1CON1 = SSP1CON1 | speed

End Proc

The SMP bit is not currently implemented in the function, by default it is = 0.
Please note that I am expanding or completing the function as needed, for the moment all the devices I have tested SMP = 0 is correct.

Details of the 3 mentioned parameters:

CKP (Clock Polarity) – bit in the SSPCON1 register
Controls the idle state level of the clock (SCK).

CKP = 1 → the clock is high when idle.

CKP = 0 → the clock is low when idle.

CKE (Clock Edge) – bit 6 of the SSPSTAT register
Determines on which clock edge the data is transmitted.

CKE = 1 → Data is transmitted on the falling edge (from 1 to 0).

CKE = 0 → Data is transmitted on the rising edge (from 0 to 1).

The direction of transmission (rising or falling) also depends on the value of CKP. Both CKP and CKE bits must be considered together to define the SPI mode.

 

[DogFlu66]

You can use this function to easily view the status of a register. It sends the value to a serial console. I usually use it to debug code and avoid having to constantly use a calculator. The group of functions is attached.

UART1_Write "SSPSTAT:", ToBinary(SSPSTAT, 8), CrLf

 

[camerart]

The SMP bit is not currently implemented in the function, by default it is = 0.
Please note that I am expanding or completing the function as needed, for the moment all the devices I have tested SMP = 0 is correct.

Details of the 3 mentioned parameters:

CKP (Clock Polarity) – bit in the SSPCON1 register
Controls the idle state level of the clock (SCK).

CKP = 1 → the clock is high when idle.

CKP = 0 → the clock is low when idle.

CKE (Clock Edge) – bit 6 of the SSPSTAT register
Determines on which clock edge the data is transmitted.

CKE = 1 → Data is transmitted on the falling edge (from 1 to 0).

CKE = 0 → Data is transmitted on the rising edge (from 0 to 1).

The direction of transmission (rising or falling) also depends on the value of CKP. Both CKP and CKE bits must be considered together to define the SPI mode.

Hi D,
Ok, thanks.
I've tried all of the 3x setting combinations now, and still find SPI errors.

I wonder if it depends on the source of the TFT screens, whether or not there are errors or not?

C.

 

[DogFlu66]

Hi D,
Ok, thanks.
I've tried all of the 3x setting combinations now, and still find SPI errors.

I wonder if it depends on the source of the TFT screens, whether or not there are errors or not?

C.

I don't know, as I haven't been able to duplicate the problem.

 

[camerart]

I don't know, as I haven't been able to duplicate the problem.

Hi D,
I think that you use 18F46K22, but do you also use 18F46K20, with no problems?
Also are they SMD or TTH?
C.

 

[Ian Rogers]

Hi C.. here are two screen shots one on the ILI9341 datasheet and one from the pick

pic

ILI9341 datasheet

It looks to me like mode 2 CKP = 0 and CKE = 1 is what they want on the ILI. The sample is in the middle of data

The st7789 however is sample after clock thats why the ST7789 is better on mode 3..

 

[camerart]

Hi C.. here are two screen shots one on the ILI9341 datasheet and one from the pick

pic
View attachment 149471

ILI9341 datasheet

View attachment 149472

It looks to me like mode 2 CKP = 0 and CKE = 1 is what they want on the ILI. The sample is in the middle of data

The st7789 however is sample after clock thats why the ST7789 is better on mode 3..

Hi I,
I think I've tried all of the modes now, but I'll try again with 3.

NOT IMPORTANT:
This will mean that 'say' for the RADIO CONTROL project, that the board was designed for, uses a different SPI set-up, so if I were to use the TFT with all of the other peripherals, how would that work.
Thanks C.

 

[DogFlu66]

Hi D,
I think that you use 18F46K22, but do you also use 18F46K20, with no problems?
Also are they SMD or TTH?
C.

There shouldn't be any difference in the microcontrollers because the packaging is different.

I've tested both types of displays with the 46K22 and 46K20, and they worked fine.