LED driver repair

Thread Starter

HulkHogan

Joined May 17, 2021
8
Hi everyone, this is a refrigerator light system 36V, main power board with 4 LEDs, 2 additional slave boards each having also 4 LEDs. Each board is running on 12V in series connection totaling 36V and 50 mA each LED if I am correct. 200mA is the current marked on 1 of those slave boards. Found tiny transistor burnt/shorted and 2 its related resistors burnt but still conducting/responding to ohm meter. I guess but not sure that it has a constant voltage driver to maintain 36V DC. 1 LED marked NR270U - (0 is in shape of football slanted 45 degrees) is damaged on 1 of the slave boards . The LED’s lens is 5mm x 3mm and has 5 pins, each corner one pin and in the middle, 2 pins are connected together, couldn’t cross-reference it yet and I tried. All other LEDs are good tested by ohm meter and they produce dim light triggered by the ohmmeter.

I replaced the shorted transistor and replaced burnt resistors 91 and 127 ohms. The closest I found in my supply was 100 ohms for the burnt 91 ohms and 131 ohms for the burnt 127 ohms. Electrolytic caps were still good. I had only the main board with 4 LEDs so I hooked to the output connector 250 ohm resistor advised by youtube clip as a dummy load to test it. It comes on but the light intensity is about 1/3rd and the voltage on the 36 volts side is only over 1 volt. If I leave it running little bit longer, some of the resistors I replaced starts to smoke so I turned it off to find the culprit.

Other semiconductors like diodes and transistors tested good so I am thinking about the perhaps mosfet driver its self but not sure. The boards have what I think is a LED diode protectors PL18, 1 diode protector for each board.
One of the slave board is marked as 11V-14V 200mA which would total of 600mA for the light setup requirement.
The driver transistor tested good but perhaps under load due to getting 120V dc maybe not responding correctly after the torture short of taking it out and test bench it. Driver transistor marked as ON (firm) RMB 1014B, package SOT 223.
I and my electronic store tried to cross it to N.A. even on the web.

A guide how to properly match driver for existing set of lights will help me to pick a small drivers on the web for a few of bucks and modify the main board so I don’t have to deal with this issue but I prefer to fix the existing main board so I can learn and perhaps to modify the current to little bit of current less by changing some resistor so these LEDs wouldn’t be at full speed/load. There is not too many components so it should be easy to fix it but matching and cross-referencing of these parts is a cookie to me now. I'll check if I can get to the ON firm for support.

0.1% resistors to match correctly are on the way from Digikey so I'll see if that will help.
I just guess that these driver transistors for these applications have same principles and differ in current or voltage regulation, voltage/current amount where some of that is achieved by sizing resistors in that circuit if I am correct. I can't find detailed education how these things work. These fridges have frequently issues with lights and my friend already once replaced all of it for it to fail again 2 years later so I want to fix it and learn the circuits. 150 bucks for new lights to be "On" for short time and it won't last, hmmm.... we used to just change cheap bulbs in the past. Shopjimmy has these boards and are cheaper but still made same way and will fail again. https://www.shopjimmy.com/w10515058-main-led-light-and-driver-for-whirlpool-kenmore-maytag/

Specs Summary:
120V AC in, 36V DC out, constant voltage or current driver?, no dimming, damaged LED NR270U 5 pins 50mA each, suspect driver transistor ON (firm) RMB 1014B - package SOT 223/mosfet or not? This bellow is end portion of the driver circuit, is not including rectification and filtering with the suspect driver transistor. Q3 was shorted so I replace also Q2. R15 and R17 tend to burn on these boards when the thing hits the fan. Full schematic is not available.Svetylka.pngNot sure if current for replacement driver project has to be exactly 600mA or can be lower or higher.
Thank you for looking and responding.
 

Thread Starter

HulkHogan

Joined May 17, 2021
8
In regard those smoking resistors or one, it is hard to figure out the wattage of the pile I have at home and that circuit is pulling on the R15 side 0.6 amps and it may be the issues as well that I soldered in a low wattage resistor and the thing perhaps can not get full power due to the wattage?
Q3 is a miniature transistor so R17 will be a low wattage I guess.
 

sghioto

Joined Dec 31, 2017
5,376
Q3 is a miniature transistor so R17 will be a low wattage I guess.
Where did you find this partial schematic? Q3 is a small signal transistor rated 200 ma and Q2 is shown as a voltage regulator. The base of Q3 is grounded and the emitter collector leads are reversed. Doesn't make sense.
What did you replaced Q2 with?
 
Last edited:

Thread Starter

HulkHogan

Joined May 17, 2021
8
Here is the clip:
the producer claims that he made the diagram. True that I am not a wizard in these things so a pro will see quickly some flaws in these diagrams due to experience. I want to be there as well but it takes time. Q2 is now PZT2907AE signal pnp transistor 600mA and I didn't realized that it is a regulator in the diagram but Q stands for transistor but device on diagram looks like a regulator so I took the part number ON P2F and my store gave me whatever they gave me PZT2907AE and as a stupid sheep I soldered it in w/o verification , Q3 is now C49 and was originally 6KN. Digikey helped me to ID 6KN to "CMPT491E" TR PBFREECT-ND because I gave them 2N3904 from the diagram which can be wrong. How is regulator marked in diagrams?
Nice catch, thank you so much.
 

sghioto

Joined Dec 31, 2017
5,376
Yes, transistors are usually marked with a "Q" and chips with a "U".
A voltage regulator chip is drawn like it shows on the schematic IN GND OUT.
Q2 is most likely a PNP transistor and the PZT2907 is probably fine.
The Q3 replacement is pretty close to a 2N3904 but the circuit schematic is definitely incorrect.
I suspect that Q2 and Q3 form a constant current regulator to drive the LEDs. However all the circuits I've seen using 2 transistor they are both NPN or PNP. The Q2 or Q3 replacement may not be correct.
Not sure if all the LEDs shown in that schematic are in series though. Your description of the LED board marked 200ma suggest the 4 LEDs are in parallel and each board is in series with the main board so the total current is still 200 ma.
Can you verify if the LEDs are in parallel or series on the boards?
A constant current regulator should not exceed the limit set (say 200 ma) even if the output is shorted but that's probably a design flaw on the current
boards.
An example of a 2 transistor current limiter with PNP transistors.
1643743627442.png
 
Last edited:

Thread Starter

HulkHogan

Joined May 17, 2021
8
I agree, once there is 1 mistake whether intentional or by lack of knowledge, more or the whole pie is spoiled. Picture shows how I traced connections between LEDs and final traces to J2 connector. X means no connections. Colored marks show which LED pins are connected together, matching colors.
2 resistors of low ohms and 1 diode protector on the slave board other side is responding to J2 connector pin 1 with low or no ohms. It is hard/impossible to see those traces they painted it white and clear coated it. I'll try more to inspect it how are LED's interconnected.
That would be perfect if these LED's are not 50 mA because I can get much easier lesser mAmps LEDs and manage somehow to place it there even with just 2 pins if there is just some interconnections under the LED with 4 pins I have to replace.
Smaller LED drivers are available so with less mAps it should be easy to find another driver.
Thank you again for your assistance. IMG_6201.JPG
 

Thread Starter

HulkHogan

Joined May 17, 2021
8
Blue, Red, Green and Yellow are hoked together by corresponding colors, each marked trace color dot is interconnected.
 

Thread Starter

HulkHogan

Joined May 17, 2021
8
So I checked it more and it is all in serial connection. Made another diagram more specific. When the lights went out, 1 bad LED on a slave board out of 12 LEDs was identified plus the main power board's 2 burnt resistors and 1 shorted transistor. Just received Digikey shipment with these 2 accurate resistors and will replace them in couple of days. The LED on the picture is the back side of it soldered to the board, had to use re-flow to remove it.IMG_6227.JPG
 

sghioto

Joined Dec 31, 2017
5,376
So I checked it more and it is all in serial connection.
Makes sense.
What kind of concerns me is that LED protection circuit wired directly across the series LEDs. You show it as a normal diode but it's marked U11 on the pc board like an IC chip. I sort of see what might happen. If one of the LEDs would open on the board that circuit would maintain continuity so the other two boards would still operate. Is that what happened when the board first failed?
 

Thread Starter

HulkHogan

Joined May 17, 2021
8
That is the only thing I can think off shunting the failed circuit through the ressitor and the LED protector for the failed slave board so 2 remaing slave lights will still funciton and someowhere I read it as well. The LED protector shown on the first picture of the board is marked as a diode accompanied be the U11 and by ohmeter it doesn't act as a diode. Perhaps acts only at certain conditions and switches "On" to shunt the current? But it wasn't good enough I gues for the main board when the LED shorted and burnt off internally while in the mean time took out components on the main board due to high current. I am not a pro but this is the only think crossed my mind. The protection maybe failed to protect the board's components? All those 3 lights failed once the issue arised and no other board was lit.
 
Top