Issue with driving the backlight of 2 LCD screens

[nicco82]

Hey, I've just received a PCB I created with 2 LCD screens. Unfortunately the LCD were black and I think it's due to a wrong schematic for converting BL to LEDK. If I bridge LEDK and GND the screens work fine at what I guess it's full brightness.
Same code works with 2 lcds in the breadboard (each takes BL and converts it to LEDK).
My schematics is attached and also here is the inspiration schematic: https://www.waveshare.net/w/upload/f/f5/1.47inch_LCD_Module_SchDoc.pdf
and these are the LCDs I'm using:
https://vi.aliexpress.com/i/1005003771379232.html?gatewayAdapt=glo2vnm

Thanks for the help
Nicco

[cowana]

The issue is likely the driving voltage across the LED being less than the forward voltage.

If your supply is 3.3v, and the voltage drop over the NPN transistor is 0.6v; that just leaves 2.7v for the LED which might not be enough depending on the colour.

Bypassing the transistor avoids that 0.6v drop (providing the LED with a full 3.3v) allowing it to light.

[nicco82]

Is it because it's 2 screens instead of one? My assumption was that running them in parallel the voltage would still be the same.... Would it be better if I split the LEDK instead of BG?

Thanks

[Buriedcode]

Is it because it's 2 screens instead of one? My assumption was that running them in parallel the voltage would still be the same.... Would it be better if I split the LEDK instead of BG?

Thanks

Unlikely. Simply having two backlights (with their own current limiting resistors/driving circuitry) in parallel will just draw twice the current.

I'll second what cowana said - the backlight for those displays is white.  If it was two LEDs in series they wouldn't light up at all, so it must just be one LED (or LEDs in parallel). White LED's generally require > 3.2V, usually about 3.4-3.6V to function, so 3.3V with the burden voltage across the NPN just isn't enough to light it.

[nicco82]

Since the LCD backlight needs between 2.9 and 3.3 volts to operate could it be a solution to do LEDA connected to 3.3v and then instead connect LEDK directly to BL and code BL to have a voltage between 0v for full brightness and 0.4v for minimum brightness? Would LEDK have any load or the load would only be on LEDA?


Thanks!

[thm_w]

Are you using the bare LCD or the LCD module from that aliexpress listing?
If its the module, it already has a transistor on the backlight pin, its not clear if you are adding a second transistor for your own circuitry or not. But it would only need one.

[nicco82]

In the breadboard, to test the system I use the board. Then in the manufactured PCB I connect directly to the 12 pins of the 2 LCDs hence cen need to drive LEDK with my own schematics.

[Buriedcode]

Since the LCD backlight needs between 2.9 and 3.3 volts to operate could it be a solution to do LEDA connected to 3.3v and then instead connect LEDK directly to BL and code BL to have a voltage between 0v for full brightness and 0.4v for minimum brightness? Would LEDK have any load or the load would only be on LEDA?


Thanks!

No LED will light with 0.4V, you vary the current in an LED to vary brightness - or more commonly pulse width modulation PWM.  The fact the backlight lights when LEDK is connected directly to ground indicates a problem with your NPN transistor.  Meaure the actual voltage of your 3.3V supply, and with LEDK connected to ground, the voltage across the backlight (or the voltage across the 10 ohm resistor).  That will tell you both the forward voltage of the LED used in the backlight, and the current (I = V/R = voltage across 10R / 10).

[Professzore]

First is first,
If you don't have any, make a datasheet.
At the first step, I highly recommend making some measurements (best with a current limited bench power supply), to find out the exact characteristics of the led blacklight. By the Aliexpress link, it seems to be (presumably true), it works out with three white LEDs, all connected parallel. In this configuration, I suggest using an 8-10 mA current limit and a maximum voltage of 3.3 volts. Check the backlight (if there is nothing, it's also a good idea to make a simple code to make a full white screen (or use a screen testing firmware). Current limit may be up to 30 mA as an absolute maximum value. But do NOT go over this! Theoretically, near the Vf value, there is no resistor needed in series, but, due to manufacturing variance, it may be a good idea to have one in the final design (10 R may be fairly enough). For most white LEDs, Vf is around 3.1 V.
With measurements, you must fund out:

• the minimum voltage where the LEDs lite up,
• the optimal voltage where you find the LEDs bright enough for your needs,
• the current in mA, where you find the LEDs bright enough for your needs.

These three numbers are all mandatory to follow in your design. I'm afraid, as many others mentioned before, this configuration is not capable to produce the minimum voltage value (Vf or forward voltage) to light up the LEDs.
Based on these numbers, and the capabilities of your logic (MCU), there are many solutions. But without the actual data, it's just a trial-and-error game (with a risk of burning your MCU) to go anywhere forward.
Of course, there are many more steps after the measurements.

[thm_w]

In the breadboard, to test the system I use the board. Then in the manufactured PCB I connect directly to the 12 pins of the 2 LCDs hence cen need to drive LEDK with my own schematics.

So it works on the breadboard with their module, but your PCB doesn't work.

I can't see any difference in the schematics, but there must be something you've missed.