Logic Level shifter for 12864 display on Duet 2 Wifi

oliof · 7 Oct 2020, 17:46

If the encoder turning causes a reset, the wires might be turned 180 degrees.

seraser · 7 Oct 2020, 18:16

@oliof I used dc42 pin asignement.

Schmart · 8 Oct 2020, 14:54

@oliof said in Logic Level shifter for 12864 display on Duet 2 Wifi:

If the encoder turning causes a reset, the wires might be turned 180 degrees.

Oliof might be right. Just to rule out a wiring problem, can you provide a photo of the connections you made between the display and the mainboard so we can check?

You may have done so already, but for troubleshooting purposes, I would also suggest to only connect the wires that are absolutely necessary for a working display.

Just checked the source code and the pin mapping that David mentioned hasn't changed. You compiled the firmware yourself with #define SUPPORT_12864_LCD?

The only change made to the source code recently was to the default contrast (changed to C30) for a ST7565 display. I believe David has used the Fysetc Mini12864 display to test the code with, so the default settings for the P2 display type should be already optimized for that model.

seraser · 8 Oct 2020, 15:42

@Schmart Firmware is 3.2beta2, public, not compiled for me.

Wiring is simple, exp1 of lcd to explcd on duet and exp2 of lcd to expsd on duet, checked pin asignement but maybe I miss some thing.

With fysetc mini v1.2 backlight and leds light but with fysetc v2.1 nothing light and with both displays rotate the encoder result restart.

From github sources:

Pinout fysetc:

Edit:

My bad, checked under board pinout and my wiring is bad for bad initial comprension, I will redo the wiring with board pinout, sorry my bad english Im spanish.

Schmart · 8 Oct 2020, 16:42

@seraser, yes, I think the one-to-one pin mapping of standard ribbon cables will not cut it in this case. For this purpose, I made a custom cable using Dupont connectors and twisted wires from a UTP cable (not solid core but the stranded kind).

Pinouts for your board are also here BTW.

seraser · 9 Oct 2020, 16:48

@Schmart doesn't work, conected from connlcd2 to connlcd10 and supply 5v and only light with 1.2 version and no lights with 2.1 version tested with many M918 P2 configs.

I will test both LCD with ramps and if work ok I will stop because best choice is wait other guy trying.

Best regards.

dc42 · 9 Oct 2020, 16:52

The version 2.1 Fysetc display won't work on the Duet WiFi (yet) because RRF for Duet 2 doesn't support the Neopixel backlight.

I will investigate the other problem with ST7567-based 12864 display and Duet 2.

seraser · 9 Oct 2020, 17:19

@dc42 Ok, thank you very much.

pixelpieper · 9 Oct 2020, 21:49

I am trying to get the Mini 12864 working with my maestro, but it seems like I may have gotten the pinout wrong - the display lights up, but does not show anything... - is the final pinout somewhere documented?

I tried to follow the Maestro side identically to this post from Schmart i.e. Exp0 -> A0, Exp1 ->CS, remaining SPI pins from the EXP1 connector. For the display I followed the pinout from github. My config.g contains "M918 P2 E2". Issuing M112 yields "RepRapFirmware for Duet 2 Maestro version 3.2-beta2 running on Duet Maestro 1.0"

Did I miss anything?

A Former User · 9 Oct 2020, 22:19

@pixelpieper said in Logic Level shifter for 12864 display on Duet 2 Wifi:

, but does not show anything...

you have added menu files to the sd card?

is the final pinout somewhere documented?

best source will always be the source on github / Pins_DuetNG.h for the version in question

pixelpieper · 9 Oct 2020, 22:27

@bearer yes, I did. Just checked the maestro header file:

 constexpr Pin LcdCSPin = PortCPin(9);
constexpr Pin LcdA0Pin = PortAPin(21);    // EXP_0
constexpr Pin LcdCSAltPin = PortAPin(22); // EXP_1

The chip select is defined twice, do I need to enable something else for the “alt” pin to work?

A Former User · 9 Oct 2020, 22:29

@pixelpieper said in Logic Level shifter for 12864 display on Duet 2 Wifi:

The chip select is defined twice,

once for lcd and once for sd card presumably?

Schmart · 10 Oct 2020, 10:02

@pixelpieper said in Logic Level shifter for 12864 display on Duet 2 Wifi:

@bearer yes, I did. Just checked the maestro header file:
 constexpr Pin LcdCSPin = PortCPin(9);
constexpr Pin LcdA0Pin = PortAPin(21);    // EXP_0
constexpr Pin LcdCSAltPin = PortAPin(22); // EXP_1
The chip select is defined twice, do I need to enable something else for the “alt” pin to work?

Using an alternative pin for CS instead of the original CS pin is unfortunately necessary on a Duet 2 Maestro when using a ST7567. Here's some background. The electronics on the Maestro were originally designed with a ST7920-based display in mind. That display requires an active high CS. On the Maestro, the CS, clock (SCK) and data output (MOSI) signals are buffered by a 74HCT08 chip (from 3.3V to 5V) . Additionally, the SCK and MOSI are passed through (gated) only when CS is high. Now, the ST7567 display has an active low CS. Which means that if CS is low, SCK and MOSI will be blocked and useless. So therefore, the original CS is set high permanently, and the alternative pin EXP_1 (unbuffered 3.3V) is used for the active low CS to the ST7567.

The ST7567 also requires an additional pin to indicate if data sent to the display should be considered a command or display data. That is called the A0 and on the Maestro, EXP_0 is used for that.

Getting a ST7567 display to work successfully is based on the right wiring, configuration (display type, display contrast and resistor ratio), and a compatible display that accepts 3.3V levels, so there are quite some variables to get right. At least, the schema in my post that you mention is correct on the Maestro side, I'm using that right now.

Schmart · 10 Oct 2020, 10:44

@pixelpieper I've taken a look at the picture you provided, and as far as I can tell the wiring seems in order (based on the Fystec Wiki article).

Schmart · 10 Oct 2020, 10:53

@pixelpieper in the Wiki for this display I do see mentioned that dynamic contrast may not be supported. Did you play with the contrast parameter (e.g. try C100) already?

pixelpieper · 10 Nov 2020, 06:23

@Schmart Yes, I did, however there was no difference :(. Did pull out my scope for a change, there is traffic on A0 and CS as well as on MOSI/SCK. Will try to do a little bit more debugging later the week.

pixelpieper · 10 Oct 2020, 21:26

Well, it is alive. Turns out the display likes a nonactive reset to work - go figure.

Seems we can use the high active ST7920 chip select for this purpose.

seraser · 11 Oct 2020, 07:24

@pixelpieper can this solution be applied to duetwifi?

pixelpieper · 11 Oct 2020, 10:08

@seraser this is not maestro specific - the lcd reset pin of the mini 12864 has to be somehow pulled high in order to have the display leave its reset.

@dc42 from the firmware side: when configured as M918 P2, are we guaranteed that the then not functional active high chip select on the maestro display connector stays always high? I.e. is it safe to use as the reset pin?

Schmart · 11 Oct 2020, 11:34

@pixelpieper said in Logic Level shifter for 12864 display on Duet 2 Wifi:

nonactive reset

Congratulations, great find! It's nice to see you also have the right tooling at hand to do the troubleshooting Apologies that I couldn't help you with this one, I don't have any Fysetc displays, and the Creality display I tested with doesn't need a reset.

As for your question to use the CS as RESET; does your display also work correctly when the reset pin is permanently tied to Vcc or does the reset signal need a delay?

	constexpr Pin LcdCSPin = PortCPin(9);
	constexpr Pin LcdA0Pin = PortAPin(21); // EXP_0
	constexpr Pin LcdCSAltPin = PortAPin(22); // EXP_1