Fran ๐ฆ๐ทThings got really cramped when fitting the larger board with the extra protections into the space of the previous smaller board ๐.
๐งต #PortuariaLamp #lamp #LED #driver
I implemented a protection for the remaining board, with a resistor and a zener diode as explained in a comment here: https://forum.arduino.cc/t/difficulties-with-pt4115-dim-pin/210874
And I also added a Schottky diode to VIN just in case the problem was a polarity reversion between CSN and VIN pins when the lamp is turned off and VIN goes back to 5V while the flicker-filtering capacitor in parallel with the LED is still charged at about 17V ๐ค.
Difficulties with PT4115 dim pin
Hi ! My project is to build an RGB light like the Philips Living colors but cheaper and smaller while brighter. The lamps will be connected together with 2.4 gHz modules (using nrf24l01 chips as they are incredibly cheap !! (As it is my first post on don't hesitate to correct me if I did anything wrong on this post ! ) I've google a lot about mood lamp, high power leds (RGB, RGBA, RGBW...) and on led drivers. I found many very interesting projects. As I have limited number of instruments (un...
Found some unexpected problems with the LED driver in the USB-PD version of the lamp ๐คจ. It seems the PT4115 chip I'm using is VERY sensitive to ESD on the DIM input, and I already toasted one of the two boards I have.
It would have been nice to know that the PT4115 was such an unreliable chip before ordering the 20 boards with the same chip that are now coming from China ๐คฆโโ๏ธ.
But the problem now is that everything looks very woody and very brown ๐. I liked the contrast of materials that the ceramic made. So, continuing with my materials hysteria, I now started trying to make the same shape of the wooden screen (held with screws from below) but in ceramic, which would need only two holes that are easier to make than the one with cutouts.
Making the wooden screen is a more precise and predictable process, because it does not shrink between 5% and 9% like ceramic. It takes a lot of sanding work to get an acceptable finish, but eventually the rough cutting part of the piece could be sped up with CNC ๐ค.
Experiments with an alternative wooden lampshade. The problem with the ceramic screen I was using is that it is very difficult to manufacture, because it deforms when it dries due to the cutouts it has. The only way to avoid that that I found is to make the cutouts after the piece is already baked... and that takes a looooong time with saws and diamond micro-files ๐คฏ.
The first attempt with the IP2721 chip didn't work because that chip can't renegotiate another voltage after power on, but I found another chip that can do it: the HUSB238! ๐ฅณ
The trick was to put a double switch and another cable to the HUSB238 so that when turning off the lamp it asks the charger to go down to 5V, and the result was quite interesting: 0.19W in standby, 80% less than before ๐ช.
More efficient USB PD chargers with GaN chips are already on the way...
๐งต #PortuariaLamp
But not everything is pure pleasure in the world of USB PD. The USB PD charger I'm using is not a big deal, and has a standby power consumption of 0.125W when it's working at 5V (unpresentable for 2024 ๐), but when it's put to work at 20V the standby consumption goes to 1W! ๐ฑ. A 5W lamp that consumes 1W when turned off?!?! ๐ฉ
After a couple of days pondering this depressing result, I thought of trying something: Make the charger go back to 5V when the lamp is off!
๐งต #PortuariaLamp #USB #USBPD
The solution? Use the new USB PD chargers, which can deliver different voltages if we ask them nicely ๐. I'm using a board with the HUSB238 to ask the charger to give me 20V, which is enough to turn on the 17V LED and makes everything work more efficient and cooler because the current in the cable is only 0.25A, converting from 220V to 20V is more efficient, and the step-down to regulate the LED current is more efficient than the step-up of the previous version.
๐งต #PortuariaLamp #USB #USBPD
The problem with the previous circuit, which worked with the 5V of the common USB, is that it had a lot of power losses due to the inefficiency of working at such a low voltage. The charger got very hot due to the low conversion efficiency from 220V to 5V, power was lost in the cable due to the 1.4A current, and the step-up worked hard if the voltage dropped much below 5V due to using a charger or cable of dubious quality ๐.