EEVblog Electronics Community Forum
Electronics => Repair => Topic started by: Bud on January 03, 2023, 12:17:48 am
-
There had been a few days of rain and wet weather here, so my string fading Christmas lights that I have outside eventually quit. It is a fairly nice color LED lights (not that eye candy one but softer light similar to what incandescent bulbs used to be) so I decided to give it a look instead of throwing away. The thing has three parts - the power supply, two strings of LED lights connected in opposite direction, one yellow white and the other color ones, and a little controller with a button in between. Under normal conditions when powered on, the white LED string comes solid on. You push the button repetitively and the lights switch to white slow fading in and out, next button push gives color LED string solid, next the color LED string fading in/out, and with the next button press the lights switch alternatively white fading/color fading. It is rather nice and calm, I liked it.
(https://www.eevblog.com/forum/repair/repair-of-a-christmas-fading-string-lights/?action=dlattach;attach=1679596;image)
To troubleshoot I first removed the controller from the circuit, connected the LED strings to the power supply directly and got the solid white light. This indicated that the LEDs were OK and that the problem was with the controller. It was glued shut so I used a small hack saw to cut it along the perimeter and there it was.
(https://www.eevblog.com/forum/repair/repair-of-a-christmas-fading-string-lights/?action=dlattach;attach=1679602;image)
It had a 7805 SMD regulator, an 8-pin microcontroller and 4 transistors. Apparently moisture has crept inside and made conductive oxides between the parts of the schematics in the center of the PCB that were not suppose to be connected, as I found the two PNP transistors Q3 and Q6 blown.
(https://www.eevblog.com/forum/repair/repair-of-a-christmas-fading-string-lights/?action=dlattach;attach=1679608;image)
I reverse engineered the schematics. Only the power part is shown here as there is nothing special about the microcontroller part. The microcontroller was unmarked and only had a quartz crystal connected across two pins, the mode switching button to the other pin, and two outputs that control the lights via Q1 and Q2. When there is a logic 1 on the base of Q2 and logic 0 on the base of Q1, Q2 conducts and pulls line L4 to ground. At the same time Q2 through a resistor pulls the base of Q3 to ground, making it condict, so 21V power supply voltage through Q3 appears on the line L3. In this condition L3 is positive and L4 is negative, and the respective LED string becomes forward biased and goes on, whereas the other LED string is reverse biased and stays off.
Then when logic levels on terminals 1 and 2 flip, Q1 and Q6 conduct, L3/L4 switch polarity and the second LED string goes on and the first one goes off. Now if the microcontroller begins to pulse modulate its control signals on the terminals 1 or 2, the respective LED string lights fade in and out.
The restriction of this design is the microcontroller can only alternate Q1 and Q2 and must not turn on Q1 and Q2 at the same time, otherwise the transistors state may become unpredictable and may cause shoot-thru currents through transistors which would certainly damage them. Apparently this was what happened when the board became wet.
I found a pair of suitable SMD PNP transistors in my parts box and replaced Q3 and Q6. Checked the voltages were changing on L3/L4 output terminals as I pressed the button and glued the case back using waterproof glue for shoes :box:. The controller is now back in place and the lights work great as before. :-+