220VAC COB LED lamp - how does it work?

Thread Starter

Marley

Joined Apr 4, 2016
502
Time to ask a question...
I have recently obtained some 5W LED lamp circuit boards.
They are "COB" - chip on board. Built on aluminum backed printed circuit boards.
I was skeptical at first, so connected them via an isolation transformer using a low current fuse in the circuit.
But they work very well, direct from a 240VAC supply. They have very few components on the board: a bridge rectifier, a couple of SMT resistors and a mystery integrated circuit. No inductors or capacitors.
What I want to know is: how does it work? What is the operating principal? Presumably not switchmode because there are no energy storage components. I realise the COB led is actually a lot of LEDs - about 24 I think. So in series that's about 72V. But how is the 240V dropped to this level? Or are they only switched on during the part of the cycle when the voltage is approximately this level?
Surely that would be a very low duty cycle?
Anyone got any data on the mystery IC?
Picture attached.220V_COB_LED.png
 

ronsimpson

Joined Oct 7, 2019
2,988
I can not find the IC I wanted to but here is a 120V one that we will pretend is the right one.
Input 220vac, LED voltage 72V. LED current 30mA. 24 LEDs.
There is no capacitors to store energy curing the low voltage portions of the 60hz power line cycle.

When the power line voltage is below 72 volts there is no LED current. No light. This is happening at 120hz frequency and your eye can not see that fast. The IC shorts pin-1 to pins 5,6,7,8 but current can not flow.
As the voltage reaches 65 volts current starts to flow. At (maybe 72V) the current reaches 30mA and the IC opens up a little to keep the current at 30mA. At 100V the LEDs have 72V and the IC has 28V. At 200V the IC has 128V. The current stays at 30mA.
On some of these ICs, at about the peak of the power line voltage the (voltage X current) on the IC is too much. So the IC opens up. Full voltage and no current on the IC for a short time.
1638797699706.png
Bottom line is that the IC is a constant current device. The LEDs might be on 50% of the time. (don't know in your case) You can rum 30mA LEDs at 60mA at 50% of the time.

Hope this helps
RonSimpson
 

Thread Starter

Marley

Joined Apr 4, 2016
502
The problem with that circuit (correct me if I am wrong) - it's a linear regulator for some part of the cycle.
Presumably only a small portion of the cycle so the dissipation is not too high.
I will have to hook up this LED lamp up to the 'scope to see how the current draw varies over the cycle.
 
Last edited:

Thread Starter

Marley

Joined Apr 4, 2016
502
Just looked up the data on the NUD4011. Think it must be something like this. In the 120V example given on the datasheet, the IC dissipates about 0.9W. Quite a lot for a little SO8. I will investigate further!
 

ronsimpson

Joined Oct 7, 2019
2,988
I will have to hook up this LED lamp up to the 'scope to see how the current draw varies over the cycle.
The circuit is on the power line. What looks like ground is not! You may have a hard time seeing what is happening. By connecting the scope ground clip you will short out some part of the circuit.
 

ronsimpson

Joined Oct 7, 2019
2,988
I will have to hook up this LED lamp up to the 'scope to see how the current draw varies over the cycle.
To trouble shoot power line supplies I use a isolation transformer. Mine has inputs (100,110,120,130,200,222, 240) and outputs of the same but the outputs are not connected to the power line. It allows you to connect the scope ground to anywhere and not have sparks. (some people "float" or isolate the scope with a transformer but I find that dangerous. I sometimes use a battery powered scope to see power line voltages.

Keep safe.
RonS.
 

Thread Starter

Marley

Joined Apr 4, 2016
502
I have an isolation transformer. Don't like floating the scope. Might take some measurements today!
 

Ian0

Joined Aug 7, 2020
9,670
There are a lot of LEDs in series, and a control IC that switches different quantities of LEDs into the series chain, depending on the instantaneous mains voltage.
[edit] here's the datasheet.
 
Last edited:

MisterBill2

Joined Jan 23, 2018
18,176
I suggest not digging into this because, first, it is not repairable, and second, if it is working now, probing will probably change that to non-working. It undoubtedly is a switching type of circuit with some linear components and it certainly needs adequate heat sinking.
What you would learn from probing this device is how esily they can be damaged and how difficult they would be to even diagnose the problem would be. Several companies make LED drivers and their websites contain descriptions of how the systems work.
 

Jon Chandler

Joined Jun 12, 2008
1,029
Please, as I suggested, look for Big Clive on YouTube. He is a far more knowledgeable source than many of the commenter here.

One of the tricks with the bulbs with "active resistor" chips is changing a sense resistor to reduce heating of the bulb. Without much impact to brightness, the bulb will run cooler and has a better chance to reach the claimed 50,000 life.

And yes, I have replied LED bulbs with good effect. Why? Because I CAN. Common failure modes are blown electrolytic filter caps or LEDs failed the the black spot of death.
 

MisterBill2

Joined Jan 23, 2018
18,176
I just upgraded the lighting in a walk-in freezer. The cobb-lite had failed, and the replacement was available at a price of $81.49. And not currently in stock locally. So off to the local home improvement stor, $3 for a pins to edison socket adapter and the a 75 watt equivalent LED bulb, dimmable, 3 for $9.49. So now they are not cold and in the dark.
The cob light has seven strings of 36 LEDs in series, all tied in parallel. The electronics board is in the center, lots of parts, heavy opaque solder mask, and double sides plated thru holes. No obviously failed parts, and the fuse is good. So it will take some additional examination.
 

Thread Starter

Marley

Joined Apr 4, 2016
502
OK, I have powered this COB lamp via an isolation transformer and a variac so I could view the current draw on an oscilloscope.
The current was measured using a series 10R resistor.
The current only flows when the supply voltage is above about ±200V. So there must be quite a few more than 24 LEDs in series.
The current is constant at about 36mA. So the IC is basically a linear current regulator and must dissipate significant power.
If I vary the supply voltage the current ON time varies as expected so the light will dim until the above ±200V time reduces to zero.
Looking very closely I have identified the IC as a MX2082S. It only has 3 active pins. Could not find a pdf data sheet so grabbed some images off a Chinese web site.
View of the scope screen and IC data attached.
MX2082p1.pngMX2082p2.pngMX2082p3.pngMX2082p4.pngscopescreenshot.jpg
 
Top