Analysis of LED controller (so I can create an alternative controller that's connected to the internet)

Thread Starter

smj-edison

Joined Oct 17, 2022
4
Disclaimer: I am still a beginner with electronics, but I'm happy to read articles.

Hi! I recently took apart a full-spectrum light I have (verilux vt31), as I'd like to redesign it as a IoT device for a sunrise alarm. However, I'm not sure how I'd go about figuring out what voltage/current to drive the LEDs with. I posted photos of the driver board, as well as a schematic below.

I was thinking of dimming it with a PWM signal from an esp8266, a RC filter, and a transistor. Would that be sufficient? Would I want to get a new power supply, or reuse the existing one? Also, has someone already created an LED controller that I could use?

Thank you for your time!
Mason

Front:
View attachment 278695
Back:
View attachment 278696
LED strip:
View attachment 278697
Schematic (as best as I could figure):
View attachment 278698
Power supply (19.2 VDC):
View attachment 278699
 

bassbindevil

Joined Jan 23, 2014
641
Some LED driver chips have a control pin that can dim using an analog voltage or PWM. Your attachments aren't visible at the moment, so I don't know if that applies.
 

Thread Starter

smj-edison

Joined Oct 17, 2022
4
Hmm, that's strange that the attachments didn't show up—I'll reupload them here:

Front:
PXL_20221017_182629451.jpg
Back:
PXL_20221017_182707854.jpg
LED strip:
PXL_20221017_182841557 (1).jpg
Schematic (as best as I could figure):
verilux-schematic.jpg
Power supply:
PXL_20221017_184109029.jpg
 

Thread Starter

smj-edison

Joined Oct 17, 2022
4
Actually, I might be making this way more complicated than it needs to be. I just educated myself on how LEDs work, and what an LED driver actually does. Which means now all I need to do is figure out the forward voltage of the LED strip, and get an adjustable current LED driver (I'm looking at this driver currently).

I tried using my voltmeter to get the forward voltage of the strip, but it didn't provide enough current to register (unsurprisingly). How would I go about figuring out the forward voltage of the strip?
 

bassbindevil

Joined Jan 23, 2014
641
OK, the power supply appears to be fixed voltage 19.2V, not a constant current LED driver. I found a data sheet for the CSS102A, but it's not very informative if you can't read Chinese. I suspect it just turns Q1 on, which is probably a power MOSFET (but searching on AOQ5K and A0Q5K doesn't find anything). Is there a surface mount resistor on the LED strip? If not, that might mean they've just relied on the internal resistance of those LEDs.
As for the forward voltage of the LEDs, the best way is to measure it with them illuminated. There could be multiple LEDs in series hiding inside each of those surface-mount LEDs, so voltage could be some multiple of 3(ish) volts. And the strip will have them wired in some series/parallel arrangement to get to 19.2 V.

Additional comments: there's chips like the PT4115 which drive a string of LEDs with a constant current, and have the dim/pwm input pin. If you have any of the little MR11/MR16 LED track light lamps that run from 12VAC, those nearly always have a chip like that.
https://www.electroschematics.com/high-power-dimming-led-driver-pt4115/
You can buy just the boards for those lamps:
https://www.ebay.com/itm/253819461077
... or there's PT4115 modules without the pins and full-wave rectifier...
https://www.ebay.com/itm/304486653322
There's a resistor that sets the constant current, which should be somewhat close to what your strip is currently drawing.
 
Last edited:

Thread Starter

smj-edison

Joined Oct 17, 2022
4
Apologies for an obvious question, but how would I go about measuring the forward voltage? I've found a lot of conflicting advice on how to do it. Also, as far as I can tell there is no resistor on the LED strip.

Also, let me check my understanding of LED drivers: If I can figure out the current draw and forward voltage of the LED strip, I should be able to use a 24V driver, as long as I properly limit the current, correct? I wasn't sure as the 19.2V was oddly specific.

Thanks for pointing out those chips! I'll definitely take a look at them.
 

bassbindevil

Joined Jan 23, 2014
641
If you can power up the strip with the stock power supply, measure the voltage across individual LEDs. That'll tell you if they're a single white LED or several in series in a single SMD package. Looking more closely at the picture, you might have 3 there, so about 9V across each SMD LED. That suggests the strip could be wired with two LEDs in series, so not all that far from 19.2V (and 3.2V per LED). The power supply clearly says constant voltage, but possibly it current-limits at 650 mA.
An ad for the VT31 says there are 74 LED lights, so if two are in series, that's 37 in parallel. That works out to about 18 mA per LED if the supply limits at 650 mA. Target says power is 15.3 watts; 0.206 watts/LED. If LEDs are 9.6V, that's 21.5 mA. If they're single LEDs in each SMD, that's about 64 mA.
Description says 3 power levels, so the driver chip must be dimming via PWM.
So, yes, if the driver has current limiting of some sort, the supply voltage doesn't matter. But to match the performance of the stock lamp, you really need to measure the voltage at the overall LED terminals, and the total current at full power. I'm not sure how well a meter will respond to measuring current when the lamp is dimmed via PWM. The best would be to use an oscilloscope to measure the duty cycle and frequency if you wanted to exactly duplicate the 2 dimmed levels.
 
Top