RGBCCT LED Driver Final Version [2] (RGBCW AL8861 WS2805)

I will try best to document this project even though the end result wasn't as expected. Stay tuned for the next version.

Graphics: Excuse the way I do my silkscreen. I'm open to feedback. https://redd.it/1frrw2r Feedback is hosted here for ease of access and reachablity. 

1. Design Information

Design an LED Driver that can be used with the WS2805 Protocol for ease of control, have redudency and protections while keeping the cost low (avoid the poorly designed high power WS2811 on Ali Express). The LED Driver has at least 5 colour control channels: RGB + Cool White + Warm White. This ensures that the design is future proof. The LED Driver should be capiable of driving up to 1A per channel. 

Specifications:

1. 12~24 VDC @ 1.5A
2. Output ${led_voltage} at up to 1A per channel (1.5A is absolute max)
3. Reverse polarity protection, TVS and 1.5A Fused.
4. Reinforced mounting holes and solder pads

2. Design Explaination

An AL8860 AL8861 is used to drive the LED and has an output current of up to 1.2A. The AL8861 has a DIM pin, leave floating to turn on, pull down to turn off. The state of the LED Driver is controlled by a 2n7002 mosfet to ground that is normally pulled up to 5v with a 100k to shut down the driver and pulled down by the WS2805 to turn on. The 2n7002 is connected to the AL8861 through a small 4.7k to dampen the control. It is also clamped with 5.1 zener as the DIM pin cannot exceed 5v or it will let out the magic smoke. As the DIM pin isn't pulled up sufficently internally, it is biased with a 47k resistor to VCC which is clamped by the zener diode. 

2.2 Disclaimer

You are free to replicate this circuit but keep in mind that the AL8860 is replaced with the AL8861 and that there is a mild high pitch sound, removing the output capaictors to stop inductor resonance doesn't fix this.

https://www.lcsc.com/datasheet/lcsc_datasheet_2108071630_Diodes-Incorporated-AL8861QMP-13_C2678638.pdf

If you'd like a fully functional LED Driver, see this page: https://oshwlab.com/hexawiz/rgbcct-addressable-led-driver-with-led-heatsink-community-edition

3. Circuit Explanation

3.1 Input protection -  the circuit polarity is protected by a SOP-8 PMOS that can handle up to 30V. The mosfet gate voltage is regulated by a zener, resistor and decoupled by a 100nF MLCC Capaictor. This provides for a cheap insurnace considering that it's possible that someome can break into the control box and mess up the wiring to damage the lighting system.

The whole circuit is protected by a 1.5A 1812 PPTC or 3A 1206 Single blow fuse. The Fuse footprint can accomidate both types. a TVS diode is placed after the fuse so excessive trasients will disconnect the circuit temporally. 

3.2 Control - After power makes it through the protection, it is regulated to 5v to power the WS2805 IC and LED Driver pwm inverter. The bulk capaictors of the 7805 are feed with a 4007 diode to ensure stability. For more stability, the WS2805 is powered through a 100 ohm resistors (according to the datasheet) and decoupled with a 100nF on both sides. The data lines are protected by 100 ohm resistors, this seems to be the sweet spo

3.3 LED Driver stages - The LED Driver consists of the AL8860 AL8861 LED driver from diodes incorporated.

WS2805 -> Grounded NMOS pulled high 100k -> Slew Rate +4.7K -> AL8861 <- CTRL Bias 47K to VCC

CTRL clamped to GND with 5.1V Zener.

Normally when the WS2805 is off, the NMOS that is pulled high pulls the CTRL Pin down to turn off the driver. When the WS2805 outputs a signal (low), the NMOS stops conducting the CTRL line to ground allowing the 47k to charge it up to 5v and turn the driver on.

47~68uH inductor can be used. This IC works best with 47uH.

A super barrier diode has been used for the most efficiency and minimal reverse leakage at high temperatures. Double 10uf 1210 MLCC capaictors for good decoupling. 

Note for colour balancing at hardware level assuming max brightness equates to pure white:

RED: 330mR
GREEN: 160mR

BLUE: 220mR
CCT (Warm/Cold White): 160mR

4. PCB Design and layout

4.1 Outline - The PCB is layed out so it can be mounted onto the LED heatsink like a backpack and accept KF128 screw terminals for external use. Over 61x66mm, holes are 48x53mm. The small daughter board is for attatching a 2.54mm through hole header to the aluminum PCB of the LED heatsink.

4.2 Power Routing - The power delivery is done on the bottom layer so signal traces can be kept on the same layer. Expose copper is used to increase current. Optional capaictors can be added.

4.3 LED Driver Layout - The driver circuitry is layed out in the most compact way possible. The switch loop has been kept as small as possible. I use rounded edges as it looks nicer to route and doesn't leave sharp angles for EMI radiation. Dispite my best efforts, the CTRL pin must be routed to the other side with a VIA as the LED_VCC ? colour channel name is taking up space.

Side note after completion: It is not recomended to put traces under shunt resistors as it can breach the solder mask and short. This will be addressed in the next revision.

[future me add image]

4.4 PCB Design Considerations - Power traces are made as thicc as possible with plenty of vias while general signal traces range from 0.6 to 0.75mm. General power traces are 1~1.5mm and short runs are 0.5mm like the data lines of the WS2805. Signal traces that go to soldering pads are set to 1mm.

4.5 PCB solderablity - I designed the multilayer power pad to allow high power connections to the PCB. A single multilayer pad with a via on top and bottom. You can also find my EasyEDA Pro resources here.

5. Reflection

All testing done, this LED driver emits a rather loud high pitch sound so I will reconsider the PT4115 IC which is more robust.

I would add resistance to the output MLCC to reduce ringing, align the silkscreen text better and use a PMOS rated for more voltage and current. The OLD PT4115 version has a long inductor trace but it's the most optimal design for layout.

Name: High power RGBCW addressable LED Driver FINAL REVISION v2.0r5 ~2024-08-02 t22-49-04

FV = final version (maybe not anymore)