I feel I now have a good grasp on how I would repurpose the OLED dot matrix display from a broken FormLabs Form 1+ laser resin 3D printer. I felt I could have figured out enough to play back commands captured by my logic analyzer, interspersed with my own data, similar to how I controlled a salvaged I2C LCD. But this exploration was much easier because a user on FormLabs forums recognized the SSD1305-based display module. Thanks to that information, I had a datasheet to decipher the commands, and I could go searching to see if anyone has written code to interface with a SSD1305. Adafruit, because they are awesome, published an Arduino library to do exactly that.
Adafruit’s library was written to support several of their products that used an SSD1305, including product #2675 Monochrome 2.3″ 128×32 OLED Graphic Display Module Kit which looks very similar to the display in a Form 1+ except not on a FormLabs custom circuit board. Adafruit’s board has 20 pins in a single row, much like the Newhaven Display board but visibly more compact. Adafruit added level shifters for 5V microcontroller compatibility as well as an extra 220uF capacitor to help buffer power consumption.
Since the FormLabs custom board lacked such luxuries, I need to use a 3.3V Arduino-compatible microcontroller. The most convenient module at hand (because it was used in my most recent project) happened to be an ESP32. The
ssd1305test example sketch of Adafruit’s library compiled and uploaded successfully but threw the ESP32 into a reset loop. I changed the Arduino IDE Serial Monitor baud rate to 115200 and saw this error message repeating endlessly every few seconds.
ets Jun 8 2016 00:22:57 rst:0x8 (TG1WDT_SYS_RESET),boot:0x13 (SPI_FAST_FLASH_BOOT) configsip: 0, SPIWP:0xee clk_drv:0x00,q_drv:0x00,d_drv:0x00,cs0_drv:0x00,hd_drv:0x00,wp_drv:0x00 mode:DIO, clock div:1 load:0x3fff0030,len:1344 load:0x40078000,len:13516 load:0x40080400,len:3604 entry 0x400805f0 SSD1305 OLED test
Three letters jumped out at me: WDT, the watchdog timer. Something in this example sketch is taking too long to do its thing, causing the system to believe it has locked up and needs a reset to recover. One unusual aspect of
ssd1305test code is that all work live in
setup() leaving an empty
loop(). As an experiment, I moved majority of code to
loop(), but that didn’t fix the problem. Something else is wrong but it’ll take more debugging.
To see if it’s the code or if it is the hardware, I pulled out a different 3.3V microcontroller: an Arduino Nano 33 BLE Sense. I chose this hardware because its default SPI communication pins are those already used in the sample sketch, making me optimistic it is a more suitable piece of hardware. The sketch ran without triggering its watchdog dimer, so there’s an ESP32 incompatibility somewhere in the Adafruit library. Once I saw the sketch was running, I connected the OLED and immediately saw the next problem: screen resolution. I see graphics, but only the lower half. To adjust, I changed the height dimension passed into the constructor from 64 to 32. (Second parameter.)
Adafruit_SSD1305 display(128, 32, &SPI, OLED_DC, OLED_RESET, OLED_CS, 7000000UL);
Most of the code gracefully adjusted to render at 32 pixel height, but there’s a visual glitch where pixels are horizontally offset: the entire image has shifted to the right by 4 pixels, and what’s supposed to be the rightmost 4 pixels are shown on the left edge instead.
The third problem I encountered is this sketch only runs once, immediately after successful uploading to the Nano 33 BLE Sense. If I press the reset button or perform a power cycle, the screen never shows anything again.
Graphics onscreen prove this OLED responds to an SSD1305 library, but this behavior warrants a closer look into library code.