Hello guys,
I have a question for an upcoming project. I really like the DZ65RGB in my Tofu acrylic case, but sadly it doesn't have unerglow, only per-key RGB, and there are only solder PCBs with underglow. So I thought, why not adding underglow myself.
Idea 1: Piggyback only power for LED strip
This will be the easiest option, as power can be directly drawn from the USB pins. Somebody already did this with a remote controllable strip, but in the perfect case I want QMK control.
Idea 2: Finding empty pins to connect some WS2812Bs
At first glance I couldn't find any exposed pins on the PCB, so this idea might not be viable.
Idea 3: Using existing per-key LEDs
This is my favorite idea so far, as it seems quite "simple". I thought about adding the same type of LEDs that are already mounted as per-key LED and just "steal" the signal for the underglow. This way the lighting can be reasonably controlled via "VIA" (pun intended).
So my question is, what's the best way to identify the type of LED used? The QMK firmware is configured to use the is31fl3731 LED matrix controller, but I have no idea what kind of LEDs this thing is driving.
And if it's some kind of WS2812B, can I just add a second one parallel to the same data line?
DZ65RGB underglow mod
-
- Location: Stockholm, Sweden
- DT Pro Member: 0011
What you're describing is called "underglow", "side-glow" or "side-lighting".
"Backlighting" is what the PCB already has.
I wonder if there exists a type of LED in the right size with reverse pinout to the LEDs already installed.
Then it should be possible to desolder some backlight LEDs and replace those with that LED type mounted upside-down.
(only the GND pin needs to be different, as shuffling R,G and B could be done in software).
"Backlighting" is what the PCB already has.
I wonder if there exists a type of LED in the right size with reverse pinout to the LEDs already installed.
Then it should be possible to desolder some backlight LEDs and replace those with that LED type mounted upside-down.
(only the GND pin needs to be different, as shuffling R,G and B could be done in software).
-
- Location: Germany
Thanks, I always confuse those terms , updated the post.Findecanor wrote: ↑18 Aug 2023, 12:03What you're describing is called "underglow", "side-glow" or "side-lighting".
"Backlighting" is what the PCB already has.
That would also be an idea. There is a high chance, that the current LEDs are already reverse mount, as this seems to be standard for backlighting. So I just need to find the specific type of LED.Findecanor wrote: ↑18 Aug 2023, 12:03I wonder if there exists a type of LED in the right size with reverse pinout to the LEDs already installed.
Then it should be possible to desolder some backlight LEDs and replace those with that LED type mounted upside-down.
(only the GND pin needs to be different, as shuffling R,G and B could be done in software).
-
- Location: Germany
Little update for those who are interested.
Was analyzing the LEDs a bit and checking the pins with a multimeter, but the results are "interesting". If they were WS2812B or SK6812, I expected around 5V on the Vcc and Vss pins, but it was only 1.9V on each LED pin (except Vss) and a global ground pin.
Does anyone have an idea? I now think that they aren't even WS2812Bs but something different.
Was analyzing the LEDs a bit and checking the pins with a multimeter, but the results are "interesting". If they were WS2812B or SK6812, I expected around 5V on the Vcc and Vss pins, but it was only 1.9V on each LED pin (except Vss) and a global ground pin.
Does anyone have an idea? I now think that they aren't even WS2812Bs but something different.
-
- Location: Stockholm, Sweden
- DT Pro Member: 0011
Are they individually addressable LEDs or are they part of a LED matrix, driven by a LED driver chip?
If a LED driver, you could find the chip on the board (or the QMK source for the DZ64RGB) and then look up its documentation to see what the voltages is produces.
LEDs in LED matrices are typically driven not with a direct steady current but with pulse-width modulation. In simpler terms: they are blinking very very fast. And they could do that even at "100% brightness" setting in the keyboard.
I suspect that a multimeter ~ expecting direct current — measuring a LED in a matrix therefore could be reporting an average of the voltage over time.
If a LED driver, you could find the chip on the board (or the QMK source for the DZ64RGB) and then look up its documentation to see what the voltages is produces.
LEDs in LED matrices are typically driven not with a direct steady current but with pulse-width modulation. In simpler terms: they are blinking very very fast. And they could do that even at "100% brightness" setting in the keyboard.
I suspect that a multimeter ~ expecting direct current — measuring a LED in a matrix therefore could be reporting an average of the voltage over time.
-
- Location: Germany
Looked it up in my original post and it should be a is31fl3731 LED matrix controller. So the LEDs are without an integrated IC? This would explain why I didn't get a steady 5V but rather some PWMed voltage. I checked it with a board that has known WS2812Bs and I did indeed get a ~4.5-5V supply voltage (only on the supply pins, not data pins).Findecanor wrote: ↑18 Aug 2023, 23:27Are they individually addressable LEDs or are they part of a LED matrix, driven by a LED driver chip?
If a LED driver, you could find the chip on the board (or the QMK source for the DZ64RGB) and then look up its documentation to see what the voltages is produces.
UPDATE:
These are indeed "dumb" RGB LEDs with no IC. Got a little test setup to work, where I put an RGB LED to one of the backlight LED and it worked. But it drew to much power, to that the existing LED dimmed (and thus changed color due to one color getting less power). It might be possible to limit the current with some resistors or even go as far as creating a little amplifier with transistors, but good luck fitting that into the case.