Other Parts Discussed in Thread: TPS22918
Hello everybody, this is my first post on this board as we are trying to integrate a Texas Instruments LED Driver into our design and have come across an issue that we haven´t been able to solve. We are in a hurry to get the design finalised and would greatly appreciate any assistance given.
We have 12 LEDs modules, each of them consists of 2 PCBs (just now for testing purposes), one for the driver and one for the LEDs.
Both PCBs are one sided aluminium boards.
For the driver board we follow the typical design, which consists in just the driver IC, a 0.1uF capacitor between VDD and GND, a 187ohm resistor for fixing the current flow on the LEDs close to 100mA and a micro switch for I2C addressing (by using it we can control 12 modules on the same I2C bus).
The LEDs board simply hosts the 11 required LEDs (8 white, 1 deep red, 1 royal blue and finally 1 far red).
* All white LEDs are exactly the same model (LM301H from samsung MPN:SPMWHD32AMH5XAV5SK) which can draw more than 100mA, so we use one driver channel per LED.
* The deep red one (LH351 from samsung MPN:SPHRD1L3DH00A4R2D2) is able to draw more than 300mA, so we´re using 3 channels from the driver board.
* Far red (L1SP-FRD00035R0000 from lumileds) and royal blue (L1SP-RYL0003500000) are expected 180mA and 200mA, so 2 channels per LED this time.
By now, on our first tests, for all colour LEDs are only using one channel, so 100mA maximum per LED.
LED Module: 11 x LEDs with total max demand per module 1500mA (or 1100mA with all on 1 channel).
Power Supply: Mean Well RPS-75-5 rated at 5VDC 14A (link)
Supply should be enough for at least 8 modules fully on (8*1500mA=12A).
Micro Controller: ESP32 (Dev Kit V1)
Problem: when powered on several LED modules everything seems to be ok for a few seconds, and then some of the modules go off completely, but not all the modules.
The problem occurs when using more than 3 modules. It seems to be that the driver resets. We think that because we can do a software reset and send the I2C command again to turn the module on as normal but when some time passes it fails again.
Test 1: The ESP32 and the drivers board were powered by using the mentioned PSU (everything powered at 5VDC).
Test 2: Connect he ESP32 and the driver board to another Power Supply (with common GND) and letting the full Mean Well Power Supply just for the LED Modules, but the same result.
I'm wondering if the fault could come from the power supply. Think that what the drivers really do is to quickly open and close the current flow through the LEDs, I'm working on pwm mode for controlling the brightness, so perhaps it's causing some weird behaviour on the power supply, of perhaps it's affecting to the VDD at the driver input, or even to the reset pin (which is connected to the VDD line).
Later we found an Evaluation Module (link) for the drivers and I see that they are using a 1.0uF between VDD and GND at driver IC side (remember that I'm using a 0.1uF)
Another important thing is that you are using a multilayer design for the PCB and mine is one sided (aluminium).
I can't see more great differences between my design and the Evaluation Module.
I also checked the I2C line from the ESP32 (I'm using pullup resistors, 2*4k7 in parallel for SDA and SCL) thinking that perhaps it could be a bus issue but I don't think that because when resetting the bus everything comes normal.
Anyway I also tried to use an Adafruit I2C extender (link) in order to increase the quality on the I2C comms, and the result was the same as described above.
Could it be an EMI or noise issue?
Some additional notes:
It is not always the same module that is resetting.
The output current is fine on all driver channels until the LEDs turn off, at which point the current goes to 0.
The LEDs are being held at steady state when the problem occurs (using PWM to control intensity).