A simple driver for powering LEDs with your own hands. Drivers for LEDs: types, characteristics and criteria for selecting devices. Examples of LED lamp repairs
Recently, a friend asked me to help with a problem. He is developing LED lamps, selling them along the way. He has accumulated a number of lamps that are not working correctly. Externally, this is expressed as follows: when turned on, the lamp flashes for a short time (less than a second), goes out for a second, and so repeats endlessly. He gave me three such lamps to study, I solved the problem, the fault turned out to be very interesting (just in the style of Hercule Poirot) and I want to tell you about the way to find the fault.
The LED lamp looks like this:
Fig 1. Appearance of a disassembled LED lamp
The developer has used an interesting solution - the heat from the operating LEDs is taken by a heat pipe and transferred to a classic aluminum radiator. According to the author, this solution allows for the correct thermal conditions for LEDs, minimizing thermal degradation and ensuring the longest possible service life of the diodes. At the same time, the service life of the diode power driver increases, since the driver board is removed from the thermal circuit and the board temperature does not exceed 50 degrees Celsius.
This solution - to separate the functional zones of light emission, heat removal and power current generation - made it possible to obtain high performance characteristics of the lamp in terms of reliability, durability and maintainability.
The disadvantage of such lamps, oddly enough, directly follows from its advantages - manufacturers do not need a durable lamp :). Does everyone remember the story about the conspiracy among incandescent lamp manufacturers about the maximum service life of 1000 hours?
Well, I can’t help but note the characteristic appearance products. My “state control” (wife) did not allow me to put these lamps in the chandelier where they are visible.
Let's return to the driver problems.
This is what the driver board looks like:
Fig 2. Appearance of the LED driver board from the surface mount side
And on the reverse side:
Fig 3. Appearance of the LED driver board from the power parts side
Studying it under a microscope made it possible to determine the type of control chip - it is MT7930. This is a flyback converter control chip (Fly Back), hung with various protections, like a Christmas tree with toys.
The MT7930 has built-in protection:
From excess current of the key element
supply voltage reduction
increasing supply voltage
short circuit in the load and load break.
from exceeding the temperature of the crystal
Declaring protection against short circuit in the load for a current source is rather of a marketing nature :)
It was not possible to obtain a schematic diagram for just such a driver, but a search on the Internet yielded several very similar diagrams. The closest one is shown in the figure:
Fig 4. LED Driver MT7930. Electrical circuit diagram
Analysis of this circuit and thoughtful reading of the manual for the microcircuit led me to the conclusion that the source of the blinking problem is the activation of the protection after the start. Those. the initial start-up procedure goes through (the lamp flashes - that’s what it is), but then the converter turns off due to one of the protections, the power capacitors are discharged and the cycle begins again.
Attention! The circuit contains life-threatening voltages! Do not repeat without proper understanding of what you are doing!
To study signals with an oscilloscope, you need to decouple the circuit from the network so that there is no galvanic contact. For this I used an isolation transformer. On the balcony, two Soviet-made TN36 transformers, dated 1975, were found in the reserves. Well, these are timeless devices, massive, covered in completely green varnish. I connected it according to the scheme 220 – 24 – 24 -220. Those. First I lowered the voltage to 24 volts (4 secondary windings of 6.3 volts each), and then increased it. Availability of several primary windings with taps gave me the opportunity to play with different supply voltages - from 110 volts to 238 volts. This solution is, of course, somewhat redundant, but quite suitable for one-time measurements.
Fig 5. Photo of the isolation transformer
From the description of the start in the manual it follows that when power is applied, capacitor C8 begins to charge through resistors R1 and R2 with a total resistance of about 600 kohms. Two resistors are used for safety reasons, so that if one breaks down, the current through this circuit does not exceed the safe value.
So, the power capacitor slowly charges (this time is about 300-400 ms) and when the voltage on it reaches 18.5 volts, the converter start procedure starts. The microcircuit begins to generate a sequence of pulses to the key field-effect transistor, which leads to the appearance of voltage on the Na winding. This voltage is used in two ways - to generate feedback pulses to control the output current (circuit R5 R6 C5) and to generate the operating supply voltage of the microcircuit (circuit D2 R9). At the same time, a current arises in the output circuit, which leads to the ignition of the lamp.
Why does the protection work and by what parameter?
First guess
Triggering of protection when output voltage is exceeded?
To test this assumption, I unsoldered and tested the resistors in the divider circuit (R5 10 kohm and R6 39 kohm). You can't check them without soldering them, since they are paralleled through the transformer winding. The elements turned out to be fine, but at some point the circuit started working!
I checked the shapes and voltages of the signals at all points of the converter with an oscilloscope and was surprised to see that they were all completely certified. No deviations from the norm...
I let the circuit run for an hour - everything was OK.
What if you let it cool? After 20 minutes in the off state it does not work.
Very good, apparently it’s a matter of heating some element?
But which one? And what element parameters can float away?
At this point I concluded that there was some kind of temperature sensitive element on the converter board. Heating this element completely normalizes the operation of the circuit.
What is this element?
Second guess
Suspicion fell on the transformer. The problem was thought of as follows: the transformer, due to manufacturing inaccuracies (say, the winding is under-wound by a couple of turns), operates in the saturation region, and due to a sharp drop in inductance and a sharp increase in current, the current protection of the field switch is triggered. This is a resistor R4 R8 R19 in the drain circuit, the signal from which is supplied to pin 8 (CS, apparently Current Sense) of the microcircuit and is used for the current feedback circuit and, when the setting of 2.4 volts is exceeded, turns off generation to protect the field-effect transistor and transformer from damage. On the board under study there are two resistors R15 R16 in parallel with an equivalent resistance of 2.3 ohms.
But as far as I know, the parameters of the transformer deteriorate when heated, i.e. The behavior of the system should be different - turn on, work for 5-10 minutes and turn off. The transformer on the board is quite massive and its thermal constant is no less than a few minutes.
Maybe, of course, there is a short-circuited turn in it that disappears when heated?
Resoldering the transformer to a guaranteed working one was impossible at that moment (they had not yet delivered a guaranteed working board), so I left this option for later, when there were no versions left at all :). Plus the intuitive feeling is not it. I trust my engineering intuition.
At this point, I tested the hypothesis about the operation of the current protection by reducing the current resistor by half by soldering the same one in parallel to it - this did not affect the blinking of the lamp in any way.
This means that everything is normal with the current of the field-effect transistor and there is no excess current. This was clearly visible from the signal shape on the oscilloscope screen. The peak of the sawtooth signal was 1.8 volts and clearly did not reach the value of 2.4 volts, at which the microcircuit turns off generation.
The circuit also turned out to be insensitive to changes in load - neither connecting the second head in parallel, nor switching a warm head to a cold one and back changed anything.
Third guess
I examined the supply voltage of the microcircuit. When operating in normal mode, all voltages were absolutely normal. In flashing mode too, as far as one could judge from the waveforms on the oscilloscope screen.
As before, the system blinked in a cold state and began to work normally when the transformer leg was warmed up with a soldering iron. Warm it up for 15 seconds and everything starts up fine.
Warming up the microcircuit with a soldering iron did nothing.
And the short heating time was very confusing... what could change in 15 seconds?
At some point, I sat down and methodically, logically cut off everything that was guaranteed to work. Once the lamp lights up, it means the starting circuits are working.
Once heating the board manages to start the system and it works for hours, it means the power systems are working properly.
It cools down and stops working - something depends on the temperature...
Is there a crack on the board in the feedback circuit? It cools and contracts, the contact is broken, it heats up, expands and the contact is restored?
I climbed a cold board with a tester - there are no breaks.
What else can interfere with the transition from startup mode to operating mode?!!!
Out of complete hopelessness, I intuitively soldered a 10 uF 35 volt electrolytic capacitor in parallel to power the same microcircuit.
And then happiness came. It's working!
Replacing the 10 uF capacitor with a 22 uF capacitor completely solved the problem.
Here it is, the culprit of the problem:
Figure 6. Capacitor with incorrect capacitance
Now the mechanism of the malfunction has become clear. The circuit has two power circuits for the microcircuit. The first, triggering, slowly charges capacitor C8 when 220 volts are supplied through a 600 kΩ resistor. After it is charged, the microcircuit begins to generate impulses for the field operator, starting the power part of the circuit. This leads to the generation of power for the microcircuit in operating mode on a separate winding, which is supplied to the capacitor through a diode with a resistor. The signal from this winding is also used to stabilize the output current.
Until the system reaches operating mode, the microcircuit is powered by the stored energy in the capacitor. And it was missing a little - literally a couple or three percent.
The voltage drop was enough for the microcircuit protection system to trigger due to low power and turn off everything. And the cycle began again.
It was not possible to detect this drop in the supply voltage with an oscilloscope - it was too rough an estimate. It seemed to me that everything was fine.
Warming up the board increased the capacitor capacity by the missing percentage - and there was already enough energy for a normal start-up.
It is clear why only some of the drivers failed despite the elements being fully functional. A bizarre combination of the following factors played a role:
Low power supply capacitance. The tolerance for the capacitance of electrolytic capacitors (-20% +80%) played a positive role, i.e. capacitances with a nominal value of 10 microfarads in 80% of cases have a real capacity of about 18 microfarads. Over time, the capacity decreases due to the drying out of the electrolyte.
Positive temperature dependence of the capacitance of electrolytic capacitors on temperature. Increased temperature at the output control point - just a couple of degrees is enough and the capacity is enough for normal startup. If we assume that at the exit control site it was not 20 degrees, but 25-27, then this turned out to be enough for almost 100% passing of the exit control.
The driver manufacturer saved money, of course, by using capacitors with a lower nominal value compared to the reference design from the manual (22 µF is indicated there), but fresh capacitors at elevated temperatures and taking into account the +80% spread allowed the batch of drivers to be delivered to the customer. The customer received seemingly working drivers, but over time they began to fail for some unknown reason. It would be interesting to know whether the manufacturer’s engineers took into account the peculiarities of the behavior of electrolytic capacitors with increasing temperature and the natural scatter, or did this happen by chance?
The article is devoted to the repair of LED spotlight drivers. I remind you that I recently already had an article on, I recommend you read it.
Article on LED driver circuits and their repair
Sasha, hello.
In particular, on the topic of lighting - diagrams of two modules from automotive LED spotlights with a voltage of 12V. At the same time, I want to ask you and the readers a few questions about the components of these modules.
I am not good at writing articles; I write about my experience in repairing some electronic devices (this is mainly power electronics) only on forums, answering questions from forum participants. There I also share diagrams that I copied from devices that I had to repair. I hope the LED driver diagrams I drew will help readers with repairs.
I paid attention to the circuits of these two LED drivers because they are simple, like a scooter, and very easy to repeat with your own hands. If there were no questions with the YF-053CREE-40W module driver, then there are several of them regarding the circuit topology of the second module of the TH-T0440C LED spotlight.
LED driver circuit for YF-053CREE-40W LED module
The appearance of this spotlight is shown at the beginning of the article, but this is what this lamp looks like from behind, the radiator is visible:
The LED modules of this spotlight look like this:
I have a lot of experience in copying circuits from real complex devices, so I copied the circuit of this driver easily, here it is:
YF-053 CREE LED spotlight driver, electrical circuit
Schematic diagram of LED driver TH-T0440C
What does this module look like (this is a car LED headlight):
Electrical diagram:
There is more incomprehensibility in this scheme than in the first one.
Firstly, due to the unusual switching circuit of the PWM controller, I was not able to identify this microcircuit. In some connections it is similar to the AL9110, but then it is not clear how it works without connecting its pins Vin (1), Vcc (Vdd) (6) and LD (7) to the circuit?
The question also arises about connecting MOSFET Q2 and its entire wiring. After all, it has an N-channel, but is connected in reverse polarity. With such a connection, only its antiparallel diode works, and the transistor itself and its entire “retinue” are completely useless. It was enough to replace it with a powerful Schottky diode, or a “accordion” of smaller ones.
What's new in the VK group? SamElectric.ru ?
Subscribe and read the article further:
LEDs for LED drivers
I couldn't decide on LEDs. They are the same in both modules, although their manufacturers are different. There are no inscriptions on the LEDs (on the reverse side either). I searched from different sellers under the line “Ultra-bright LEDs for LED spotlights and LED chandeliers.” They sell a bunch of different LEDs there, but all of them are either without lenses or with lenses at 60º, 90º and 120º.
I have never met one similar in appearance to mine.
Actually, both modules have the same malfunction - partial or complete degradation of the LED crystals. I think the reason is the maximum current from the drivers, set by the manufacturers (Chinese) for marketing purposes. Like, look how bright our chandeliers are. And the fact that they shine for at most 10 hours does not bother them.
If there are complaints from buyers, they can always answer that the spotlights are out of order due to shaking, because such “chandeliers” are mainly bought by the owners of jeeps, and they drive not only on the highway.
If I can find LEDs, I will reduce the driver current until the brightness of the LEDs noticeably decreases.
It is better to look for LEDs on AliExpress, there is a large selection there. But this is roulette, depending on your luck.
Datasheets (technical information) for some high-power LEDs will be at the end of the article.
I think the main thing for long-term operation of LEDs is not to chase brightness, but to set the optimal operating current.
See you later, Sergey.
P.S. I’ve been a fan of electronics since 1970, when I assembled my first detector receiver during a physics lesson.
More driver circuits
Below I will post some information on diagrams and repairs from me (author of the SamElectric.ru blog)
LED floodlight Navigator, discussed in the article (the link was already given at the beginning of the article).
The circuit is standard, the output current varies due to the ratings of the piping elements and the power of the transformer:
LED Driver MT7930 Typical. Typical electrical circuit diagram for an LED spotlight
The circuit is taken from the datasheet for this chip, here it is:
/ Description, typical switching circuit and microcircuit parameters for drivers of LED modules and matrices., pdf, 661.17 kB, downloaded: 1674 times./
The datasheet describes in detail what needs to be changed and how to get the desired output current of the driver.
Here is a more detailed driver diagram, closer to reality:
Do you see the formula to the left of the diagram? It shows what the output current depends on. First of all, from the resistor Rs, which is located at the source of the transistor and consists of three parallel resistors. These resistors, and at the same time the transistor, burn out.
Having the diagram, you can begin repairing the driver.
But even without a diagram, we can immediately say that first of all we need to pay attention to:
- input circuits,
- diode bridge,
- electrolytes,
- power transistor,
- soldering
I myself have repaired just such drivers several times. Sometimes the only thing that helped was a complete replacement of the microcircuit, transistor and almost the entire wiring. This is very labor-intensive and economically unjustified. As a rule – it’s much easier and cheaper – I bought and installed a new Led Driver, or refused repairs altogether.
Download and buy
Here are the datasheets (technical information) for some high-power LEDs:
/ Technical information on high-power LEDs for headlights and spotlights, pdf, 689.35 kB, downloaded: 725 times./
/ Technical information on high-power LEDs for headlights and spotlights, pdf, 1.82 MB, downloaded: 906 times./
Special thanks to those who have circuits of real LED drivers for the collection. I will publish them in this article.
In the photo you can see many LED lamps. I got them as a gift. It became possible to study the design of these lamps, electrical circuits, as well as repair these lamps. The most important thing is to find out the reasons for failure, since the service life indicated on the box does not always coincide with the service life.
MR-16 type lamps can be disassembled without any effort.
Judging by the label, the lamp is model MR-16-2835-F27. Its body contains 27 SMD LEDs. They emit 350 lumens. This lamp is suitable for network connection alternating current 220-240 V. Power consumption is 3.5 W. Such a lamp glows white, the temperature of which is 4100 degrees Kelvin and creates a narrowly directed flow due to the flow angle of 120 degrees. The type of base used is “GU5.3”, which has 2 pins, the distance between which is 5.3 mm. The body is made of aluminum, the lamp has a removable base, which is secured with two screws. The glass that protects the lamp from damage is glued at three points.
How to disassemble LED lamp MR-16
To identify the cause of the breakdown, it is necessary to disassemble the lamp housing. This is done without much effort.
As you can see in the photo, a ribbed surface is visible on the body. It is designed for better heat dissipation. We insert a screwdriver into one of the ribs and try to lift the glass.
Happened. You can see the printed circuit board, it is glued to the case. Prying it with a screwdriver, it separates.
Repair of LED light bulb MR-16
One of the first to be disassembled was the lamp, the LED inside of which had burned out. The printed circuit board, which is made of fiberglass, burned through.
This lamp will be suitable as a “donor”; the necessary spare parts will be taken from it to repair other lamps. The LEDs on the remaining 9 lamps also burned out. Since the driver is intact, the cause of the breakdown is the LEDs.
Electrical circuit of the LED lamp MR-16
To reduce lamp repair time, it is necessary to create its electrical circuit. It's pretty simple.
Attention! The circuit is connected to the network phase by galvanic means. It is prohibited to use it to power any devices.
How does the scheme work? A voltage of 220 V is supplied to the diode bridge VD1-VD4 through capacitor C1. Then it is supplied to the LEDs HL1-HL27, which are connected in series in the circuit. The number of LEDs can be about 80 pieces. Capacitor C2 (the larger the capacitance, the better) is a smoother for rectified voltage ripples. It eliminates the flicker of light having a frequency of 100 Hz. R1 was set to discharge C1. This is necessary in order to prevent electric shock when replacing the lamp. C2 is protected from breakdown of R2 in the event of an open circuit. R1, R2 do not accept work as such in the circuit.
C1 - red, C2 - black, diode bridge - housing with four legs.
Classic driver circuit for LED lamps up to 5 W
The electrical circuit of the lamps does not have protection elements. You will need a 100-200 ohm resistor, or better yet two. One will be installed in the connection circuit, the second will serve as protection against current surges.
Above is a circuit with protective resistors. R3 protects the LEDs and C2 capacitor, R2 in turn protects the diode bridge. This driver is perfect for lamps whose power is less than 5 W. It will easily power a lamp with 80 SMD3528 LEDs. If you need to reduce or increase the current, manipulate capacitor C1. To eliminate flickering, increase capacity C2.
The efficiency of such a driver is less than 50%. For example, the MR-16-2835-F27 lamp requires a 6.1 kOhm resistor with a power of 4 W. Then the driver will consume power that exceeds the power consumption of the LEDs. Due to the large release of thermal energy, it will not be possible to place it in a small lamp body. In this case, you can separately make a housing for this driver.
It should be remembered that the efficiency of the lamp directly depends on the number of LEDs.
Finding faulty LEDs
After the protective glass has been removed, you can inspect the LEDs. If the slightest black speck is detected on the surface of the LED, it has failed. Inspect the soldering areas and check the quality of the leads. 4 poorly soldered LEDs were found in one of the lamps
LEDs with black dots were marked with a cross. Upon external inspection, the LEDs may be intact. Therefore, you need to call them with a tester. To check, you will need a voltage of a little more than 3 V. A battery, battery, or power supply will do. A current-limiting resistor with a nominal value of 1 kOhm is connected in series behind the power source.
We touch the LED with the probes. In one direction the resistance should be small (the LED can glow), in the other it should be equal to tens of megaohms.
During the test, the lamp must be secured. A bank can come to the rescue.
You can check the LED without special instruments if the device driver is intact. Voltage is applied to the lamp base, the LED leads are short-circuited with tweezers or a piece of wire.
If all LEDs are visible, the shorted one is faulty. But this method is suitable if 1 LED in the circuit fails.
If a failure of several LEDs is detected in the circuit, the lamp will light. Only its luminous flux will decrease. Just short-circuit the pads to which the LEDs were soldered.
Other malfunctions of LED lamps
If upon inspection it turns out that the LEDs are working properly, then the problem is in the driver or the soldering area.
Cold soldering of the conductor was detected in this lamp. Soot, which appeared due to poor soldering, settled on the board tracks. To remove the soot, you needed a cloth moistened with alcohol. The wire was desoldered, tinned and soldered. This lamp worked.
Of all the lamps, one had a driver failure. The diode bridge was replaced by 4 “IN4007” diodes, which are rated for a current of 1 A and a reverse voltage of 1000 V.
Soldering SMD LEDs
To replace a faulty LED, you need to unsolder it without damaging the printed conductors. This can be done with difficulty with a regular soldering iron; it is better to put a tip made of copper wire on the soldering iron.
When soldering the LED, you must pay attention to the polarity. Install the LED at the soldering site, take a 10-15 W soldering iron and heat its ends.
If the LED is burnt and the board is charred, this area should be cleaned. Because it is a conductor. If the pad is delaminated, solder the mono LED to the “neighbors”. This is done if the paths lead exactly to them. Just take a piece of wire, fold it two or three times and solder it.
Analysis of the causes of failure of LED lamps MR-16-2835-F27
According to the table, we can conclude that lamp failures often occur due to failure of LEDs. The reason for this is the lack of protection in the circuit. Although there is space for a varistor on the board.
Repair of LED lamp series “LL-CORN” (corn lamp) E27 4.6 W 36x5050SMD
The technology for repairing a corn lamp differs from the repair of the lamp shown above.
Repairing such a lamp is simple, since the LEDs are located on the body. And dialing does not require any extra steps. This lamp was disassembled purely out of interest.
The technique for checking “corn” is no different from that described above. Only in the body of these lamps there are 3 LEDs. When ringing, all 3 should light up.
If one of the LEDs is found to be broken, short-circuit it or solder in a new one. This will not affect the life of the lamp. The lamp driver does not have an isolating transformer. Therefore, any touching of the LED tracks is unacceptable.
If the LEDs are intact, the problem is in the driver. In order to inspect it, it is necessary to disassemble the body.
To get to the driver, you need to remove the bezel. Pry it with a screwdriver at the weakest point, it should come off.
The driver has the same circuit as our first lamp with the difference that C1-1µF, C2- 4.7 µF. The wires are long, so the driver can be pulled out without effort. After work on replacing the LED, the rim was installed with Moment glue.
Repair of LED lamp “LL-CORN” (corn lamp) E27 12 W 80x5050SMD
Repairing a 12 W lamp is done according to the same scheme. No burnt-out LEDs were found on the case, so I had to open the case to inspect the driver.
There are problems with this lamp. The driver wires were too short and the base had to be removed.
The base is made of aluminum. It was attached to the body using a core. Therefore, it was necessary to drill out the fastening points with a drill whose diameter is 1.5 mm. Next, the base was pryed off with a knife and removed. The wires inside had to be cut.
Inside there were 2 identical drivers, each of which powered 43 diodes.
The driver is wrapped in a heat-shrinkable tube, which had to be cut.
After troubleshooting, the same tube is placed on the driver and crimped with a plastic tie.
The driver circuit includes protection. C1 protects against pulse surges, R2, R3 against current surges. During verification work R2 breaks were noticed. Most likely, a voltage exceeding the norm was applied to the lamp. There was no 10 ohm resistor, so a 5.1 ohm resistor was soldered in. The lamp lit up. Next we needed to connect the driver to the socket.
First of all, the short wires were replaced with longer ones. The drivers were connected by supply voltage. To attach the wires to the threaded part of the base, you need to clamp them between the plastic housing and the base.
How to connect to the central contact? Aluminum cannot be soldered, so the wire was soldered to a brass plate in which a hole was drilled for M 2.5. A similar hole was drilled in the contact. The whole thing was screwed together. Next, the base was put on and secured to the lamp body with a cap. The lamp was operational.
Repair of LED lamp series “LLB” E27 6 W 128-1
The design of the lamp is ideal for repairs. The housing is easy to disassemble.
You should hold the base with one hand and turn the protective shade counterclockwise with the other.
Under the body there are five rectangular boards on which LEDs are soldered. The rectangle is soldered to a round board on which the driver circuit is located.
To gain access to the LED terminals, you need to remove one of the covers. To make work easier, it is better to remove the board located at the driver voltage supply points. The photo shows that this wall is parallel to the capacitor body and is separated from it at the maximum distance.
To remove the board, you need to warm up the soldering areas with a soldering iron. Then, to remove it, we heat up the soldering on the round board and it disconnects.
Access to check damage is open. The driver is designed according to a simple design. Checking its rectifier diodes, as well as all the LEDs (there are 128 of them in this lamp) did not show a problem.
When I inspected the solder joints, I discovered that they were missing at some points. These places were soldered; in addition, I connected the printed circuit board tracks at the corners.
When you look at the light, these paths are clearly visible and you can easily determine which path is which.
Before assembling the lamp, it was necessary to test it. To do this, a jumper was installed on the board, and the soldered part of the lamp was connected to the power source with two temporary wires.
The lamp lit up. All that remains is to solder the board in its original place and assemble the lamp.
Repair of LED lamp series “LLB” LR-EW5N-5
In appearance, the lamp is made with high quality. The body is aluminum and the design is beautiful.
The lamp is assembled securely. Therefore, to disassemble it, you need to remove the protective glass. To do this, insert the end of a screwdriver between the radiator. The glass is fixed here without glue, with a collar. You need to rest the screwdriver on the end of the radiator and lift the glass up, using the screwdriver as a lever.
The tester did not show any failure of the LEDs. So it's all about the driver. To get to it, you need to unscrew 4 screws.
But failure overtook me. Behind the board there was a radiator plane. It is lubricated with a paste that conducts heat. I had to collect everything I had unwound. I decided to disassemble the lamp from the base side.
In order to remove the base, I had to drill out the core points. But he didn't act. As it turned out, it was fastened to plastic with a threaded connection.
The radiator had to be separated from the plastic adapter. To do this, I cut with a hacksaw in the place where the plastic was attached to the radiator. Then, by turning the screwdriver, the parts were separated from one another.
The pins were unsoldered from the LED board, which made it possible to work with the driver. Its circuit was more complex compared to other drivers. Upon inspection, a swollen capacitor 400 V 4.7 µF was found. It has been replaced.
The Schottky diode "D4" type SS110 was damaged. It's at the bottom left of the photo. It was replaced by the analogue "10 BQ100", which has 1 A and 100 V. The light bulb lit up.
Repair of LED lamp series “LLB” LR-EW5N-3
The lamp is similar to the "LLB" LR-EW5N-5, but its design has been changed.
The protective glass is secured with a ring. If you pick up the junction of the ring and the glass, it can be easily removed.
The printed circuit board is made of aluminum. There are nine crystal LEDs on it, numbering 3 pieces. The board is secured with 3 screws to the heatsink. The check did not reveal any problems with the LEDs. So it's a driver issue. Experience in repairing a similar lamp has shown that it is better to immediately unsolder the wires that come from the driver. The lamp was disassembled from the base side.
The ring connecting the base and the radiator was removed with great effort. At the same time, a piece broke off. And all because it was screwed with 3 screws. The driver has been removed.
The screws are located under the driver; you can reach them with a Phillips screwdriver.
This driver is based on a transformer circuit. The check showed the serviceability of all parts except the microcircuit. I didn't find any information about her. The lamp was set aside as a donor.
Repair of LED lamp series "LLC" E14 3W1 M1
This lamp is similar to an incandescent lamp. The first thing you notice is the wide metal ring.
I started disassembling the lamp. The first step was to remove the lampshade. As it turned out, it was placed on the base with an elastic compound. After I took it off, I realized that it was in vain.
The lamp contained 1 LED, the power of which was 3.3 W. It could be checked from the base side.
The standard RT4115 LED driver circuit is shown in the figure below: 
The supply voltage should be at least 1.5-2 volts higher than the total voltage across the LEDs. Accordingly, in the supply voltage range from 6 to 30 volts, from 1 to 7-8 LEDs can be connected to the driver.
Maximum supply voltage of the microcircuit 45 V, but operation in this mode is not guaranteed (better pay attention to a similar microcircuit).
The current through the LEDs has a triangular shape with a maximum deviation from the average value of ±15%. The average current through the LEDs is set by a resistor and calculated by the formula:
I LED = 0.1 / R
The minimum permissible value is R = 0.082 Ohm, which corresponds to a maximum current of 1.2 A.
The deviation of the current through the LED from the calculated one does not exceed 5%, provided that resistor R is installed with a maximum deviation from the nominal value of 1%.
So, to turn on the LED at constant brightness, we leave the DIM pin hanging in the air (it is pulled up to the 5V level inside the PT4115). In this case, the output current is determined solely by resistance R.
If we connect a capacitor between the DIM pin and ground, we get the effect of smooth lighting of the LEDs. The time it takes to reach maximum brightness will depend on the capacitor capacity; the larger it is, the longer the lamp will light up.
For reference: Each nanofarad of capacitance increases the turn-on time by 0.8 ms. 
If you want to make a dimmable driver for LEDs with brightness adjustment from 0 to 100%, then you can resort to one of two methods:
- First way assumes that a constant voltage in the range from 0 to 6V is supplied to the DIM input. In this case, brightness adjustment from 0 to 100% is carried out at a voltage at the DIM pin from 0.5 to 2.5 volts. Increasing the voltage above 2.5 V (and up to 6 V) does not affect the current through the LEDs (the brightness does not change). On the contrary, reducing the voltage to a level of 0.3V or lower leads to the circuit turning off and putting it into standby mode (the current consumption drops to 95 μA). Thus, you can effectively control the operation of the driver without removing the supply voltage.
- Second way involves supplying a signal from a pulse-width converter with an output frequency of 100-20000 Hz, the brightness will be determined by the duty cycle (pulse duty cycle). For example, if the high level lasts 1/4 of the period, and the low level, respectively, 3/4, then this will correspond to a brightness level of 25% of the maximum. You must understand that the driver operating frequency is determined by the inductance of the inductor and in no way depends on the dimming frequency.
The PT4115 LED driver circuit with constant voltage dimmer is shown in the figure below: 
This circuit for adjusting the brightness of the LEDs works great due to the fact that inside the chip the DIM pin is “pulled up” to the 5V bus through a 200 kOhm resistor. Therefore, when the potentiometer slider is in its lowest position, a voltage divider of 200 + 200 kOhm is formed and a potential of 5/2 = 2.5V is formed at the DIM pin, which corresponds to 100% brightness.
How the scheme works
At the first moment of time, when the input voltage is applied, the current through R and L is zero and the output switch built into the microcircuit is open. The current through the LEDs begins to gradually increase. The rate of current rise depends on the magnitude of the inductance and supply voltage. The in-circuit comparator compares the potentials before and after resistor R and, as soon as the difference is 115 mV, a low level appears at its output, which closes the output switch.
Thanks to the energy stored in the inductance, the current through the LEDs does not disappear instantly, but begins to gradually decrease. The voltage drop across the resistor R gradually decreases. As soon as it reaches a value of 85 mV, the comparator will again issue a signal to open the output switch. And the whole cycle repeats all over again.
If it is necessary to reduce the range of current ripples through the LEDs, it is possible to connect a capacitor in parallel with the LEDs. The larger its capacity, the more the triangular shape of the current through the LEDs will be smoothed out and the more similar it will become to a sinusoidal one. The capacitor does not affect the operating frequency or efficiency of the driver, but increases the time it takes for the specified current through the LED to settle. 
Important assembly details
An important element of the circuit is capacitor C1. It not only smoothes out ripples, but also compensates for the energy accumulated in the inductor at the moment the output switch is closed. Without C1, the energy stored in the inductor will flow through the Schottky diode to the power bus and can cause a breakdown of the microcircuit. Therefore, if you turn on the driver without a capacitor shunting the power supply, the microcircuit is almost guaranteed to shut down. And the greater the inductance of the inductor, the greater the chance of burning the microcontroller.
The minimum capacitance of capacitor C1 is 4.7 µF (and when the circuit is powered with a pulsating voltage after the diode bridge - at least 100 µF).
The capacitor should be located as close to the chip as possible and have the lowest possible ESR value (i.e. tantalum capacitors are welcome).
It is also very important to take a responsible approach to choosing a diode. It must have a low forward voltage drop, short recovery time during switching and stability of parameters with increasing temperature p-n junction to prevent an increase in leakage current.
In principle, you can take a regular diode, but Schottky diodes are best suited to these requirements. For example, STPS2H100A in SMD version (forward voltage 0.65V, reverse - 100V, pulse current up to 75A, operating temperature up to 156°C) or FR103 in DO-41 housing (reverse voltage up to 200V, current up to 30A, temperature up to 150 °C). The common SS34s performed very well, which you can pull out of old boards or buy a whole pack for 90 rubles.
The inductance of the inductor depends on the output current (see table below). An incorrectly selected inductance value can lead to an increase in the power dissipated on the microcircuit and exceeding the operating temperature limits.
If it overheats above 160°C, the microcircuit will automatically turn off and remain in the off state until it cools down to 140°C, after which it will start automatically.
Despite the available tabular data, it is permissible to install a coil with an inductance deviation greater than the nominal value. In this case, the efficiency of the entire circuit changes, but it remains operational.
You can take a factory choke, or you can make it yourself from a ferrite ring from a burnt motherboard and PEL-0.35 wire.
If maximum autonomy of the device is important (portable lamps, lanterns), then, in order to increase the efficiency of the circuit, it makes sense to spend time carefully selecting the inductor. At low currents, the inductance must be larger to minimize current control errors resulting from the delay in switching the transistor.
The inductor should be located as close as possible to the SW pin, ideally connected directly to it.
And finally, the most precision element of the LED driver circuit is resistor R. As already mentioned, its minimum value is 0.082 Ohms, which corresponds to a current of 1.2 A.
Unfortunately, it is not always possible to find a resistor of a suitable value, so it’s time to remember the formulas for calculating the equivalent resistance when resistors are connected in series and in parallel:
- R last = R 1 +R 2 +…+R n;
- R pairs = (R 1 xR 2) / (R 1 +R 2).
By combining different connection methods, you can obtain the required resistance from several resistors at hand.
It is important to route the board so that the Schottky diode current does not flow along the path between R and VIN, as this can lead to errors in measuring the load current.
The low cost, high reliability and stability of driver characteristics on the RT4115 contribute to its widespread use in LED lamps. Almost every second 12-volt LED lamp with an MR16 base is assembled on PT4115 (or CL6808). 
The resistance of the current-setting resistor (in Ohms) is calculated using exactly the same formula:
R = 0.1 / I LED[A]
A typical connection diagram looks like this: 
As you can see, everything is very similar to the circuit of an LED lamp with a RT4515 driver. Description of operation, signal levels, features of elements used and layout printed circuit board exactly the same as those, so there is no point in repeating them.
CL6807 sells for 12 rubles/pcs, you just need to be careful that they don’t slip soldered ones (I recommend taking them).
SN3350
SN3350 is another inexpensive chip for LED drivers (13 rubles/piece). It is almost a complete analogue of PT4115 with the only difference being that the supply voltage can range from 6 to 40 volts, and the maximum output current is limited to 750 milliamps (continuous current should not exceed 700 mA).
Like all the microcircuits described above, the SN3350 is a pulsed step-down converter with an output current stabilization function. As usual, the current in the load (and in our case, one or more LEDs act as the load) is set by the resistance of the resistor R:
R = 0.1 / I LED
To avoid exceeding the maximum output current, resistance R should not be lower than 0.15 Ohm.
The chip is available in two packages: SOT23-5 (maximum 350 mA) and SOT89-5 (700 mA).
As usual, by applying a constant voltage to the ADJ pin, we turn the circuit into a simple adjustable driver for LEDs.
A feature of this microcircuit is a slightly different adjustment range: from 25% (0.3V) to 100% (1.2V). When the potential at the ADJ pin drops to 0.2V, the microcircuit goes into sleep mode with a consumption of around 60 µA.
Typical connection diagram: 
For other details, see the specifications for the microcircuit (pdf file).
ZXLD1350
Despite the fact that this chip is another clone, there are some differences in technical specifications do not allow their direct replacement with each other.
Here are the main differences:
- the microcircuit starts at 4.8V, but reaches normal operation only with a supply voltage of 7 to 30 Volts (up to 40V can be supplied for half a second);
- maximum load current - 350 mA;
- resistance of the output switch in the open state is 1.5 - 2 Ohms;
- By changing the potential at the ADJ pin from 0.3 to 2.5V, you can change the output current (LED brightness) in the range from 25 to 200%. At a voltage of 0.2V for at least 100 µs, the driver goes into sleep mode with low power consumption (about 15-20 µA);
- if the adjustment is carried out by a PWM signal, then at a pulse repetition rate below 500 Hz, the range of brightness changes is 1-100%. If the frequency is above 10 kHz, then from 25% to 100%;
The maximum voltage that can be applied to the ADJ input is 6V. In this case, in the range from 2.5 to 6V, the driver produces the maximum current, which is set by the current-limiting resistor. The resistor resistance is calculated in exactly the same way as in all of the above microcircuits:
R = 0.1 / I LED
The minimum resistor resistance is 0.27 Ohm.
A typical connection diagram is no different from its counterparts: 
It is IMPOSSIBLE to supply power to the circuit without capacitor C1!!! At best, the microcircuit will overheat and produce unstable characteristics. In the worst case, it will fail instantly.
More detailed characteristics ZXLD1350 can be found in the datasheet for this chip.
The cost of the microcircuit is unreasonably high (), despite the fact that the output current is quite small. In general, it’s very much for everyone. I wouldn't get involved.
QX5241
QX5241 is Chinese equivalent MAX16819 (MAX16820), but in a more convenient package. Also available under the names KF5241, 5241B. It is marked "5241a" (see photo).
In one well-known store they are sold almost by weight (10 pieces for 90 rubles).
The driver operates on exactly the same principle as all those described above (continuous step-down converter), but does not contain an output switch, so operation requires the connection of an external field-effect transistor.
You can take any N-channel MOSFET with suitable drain current and drain-source voltage. For example, the following are suitable: SQ2310ES (up to 20V!!!), 40N06, IRF7413, IPD090N03L, IRF7201. In general, the lower the opening voltage, the better.
Here are some key features of the LED driver on the QX5241:
- maximum output current - 2.5 A;
- Efficiency up to 96%;
- maximum dimming frequency - 5 kHz;
- maximum operating frequency of the converter is 1 MHz;
- accuracy of current stabilization through LEDs - 1%;
- supply voltage - 5.5 - 36 Volts (works normally at 38!);
- output current is calculated by the formula: R = 0.2 / I LED
Read the specification (in English) for more details.
The LED driver on the QX5241 contains few parts and is always assembled according to this scheme: 
The 5241 chip comes only in the SOT23-6 package, so it’s best not to approach it with a soldering iron for soldering pans. After installation, the board should be thoroughly washed to remove flux; any unknown contamination can negatively affect the operation of the microcircuit.
The difference between the supply voltage and the total voltage drop across the diodes should be 4 volts (or more). If it is less, then some glitches in operation are observed (current instability and inductor whistling). So take it with reserve. Moreover, the greater the output current, the greater the voltage reserve. Although, perhaps I just came across a bad copy of the microcircuit.
If the input voltage is less than the total drop across the LEDs, then generation fails. In this case, the output field switch opens completely and the LEDs light up (of course, not at full power, since the voltage is not enough).
AL9910
Diodes Incorporated has created one very interesting LED driver IC: the AL9910. It is curious in that its operating voltage range allows it to be connected directly to a 220V network (via a simple diode rectifier).
Here are its main characteristics:
- input voltage - up to 500V (up to 277V for alternating);
- built-in voltage stabilizer for powering the microcircuit, which does not require a quenching resistor;
- the ability to adjust brightness by changing the potential on the control leg from 0.045 to 0.25V;
- built-in overheating protection (triggered at 150°C);
- operating frequency (25-300 kHz) is set by an external resistor;
- an external field-effect transistor is required for operation;
- Available in eight-legged SO-8 and SO-8EP packages.
The driver assembled on the AL9910 chip does not have galvanic isolation from the network, so it should be used only where direct contact with the circuit elements is impossible.
Modern high-power LEDs are great for providing bright and efficient lighting. The power supply for such LEDs is somewhat difficult - powerful DC sources and current-stabilizing drivers are required. At the same time, in any room there is a socket with an alternating voltage of 220V. And, of course, I would really like to organize the operation of powerful LEDs from the network at minimal cost. Nothing is impossible - let's look at the driver circuit for an LED from a 220V network.
Before we begin to discuss specific circuits, I would like to remind you that the work will be carried out with a potentially life-threatening alternating voltage of 220V. Designing and calculating the circuit will require at least a general understanding of the electrical processes taking place; the likelihood that if you make a mistake you could suffer damage or damage is very high. We strongly discourage carrying out high voltage work if you feel unsafe and will not be held responsible for any damage or damage you may suffer while working on the proposed circuits. In fact, it is quite possible that it will be easier and cheaper to purchase and use a ready-made driver or even the entire lamp. The choice is yours.
Typically the voltage drop across an LED is between 3 and 30V. The difference with the mains voltage of 220V is very large, so the step-down driver will certainly be pulsed. There are several specialized chips for the manufacture of such drivers - HV9901, HV9961, CPC9909. They are all very similar and differ from other microcircuits in that they have a very wide range of permissible input voltage - from 8 to 550V - and a very high efficiency - up to 85-90%. However, it is expected that the total LED voltage drop in the finished device will be at least 10-20% of the power supply voltage. You should not try to power, for example, one or two 3-6 volt LEDs from 220V. Even if they do not burn out immediately, the efficiency of the circuit will be low.
Let's consider a driver based on the CPC9909 chip, since it is newer than the others and is quite affordable. In general, all of these microcircuits are interchangeable and pin-to-pin compatible (but you will need to recalculate the parameters of the inductor and resistors).
The basic driver circuit is as follows:
Driver circuit for LEDs based on the CPC9909 chip
The alternating mains voltage must first be rectified; a diode bridge is used for this. C1 and C2 are smoothing capacitors. C1 – electrolyte with a capacity of 22 µF and a voltage of 400V (when using a 220V network), C2 – a ceramic capacitor with a capacity of 0.1 µF, 400V. Capacitor C1 – ceramic 0.1 µF, 25 V. The CPC9909 chip, during operation, generates pulses that open and close power transistor Q1, thereby controlling the flow of current through the LEDs. The switching frequency, the inductance of the inductor L, the parameters of the mosfet Q1 and the diode D1 are closely interrelated and depend on the required voltage drop across the LEDs and their operating current. Let's try to calculate the necessary parameters of the key parts of the circuit using a specific example.
I have a mighty LED. 50 watts of power, voltage 30-36V, operating current up to 1.4A. 4-5 THOUSAND lumens! The light output of a good spotlight.
COB LED 50 watt
To cool it, I used thermal paste and superglue to place it on a video card cooler.
We will limit the maximum LED current to 1A. Means
LED voltage drop –
Let us take the current ripple equal to +-15%:
I D = 1 * 0.15 * 2 = 0.3A
With an AC mains voltage of 220V, the voltage after the rectifier bridge and smoothing capacitors will be
The driver current is regulated by resistor Rs, the resistance of which is calculated by the formula
Rs = 0.25 / I LED = 0.25 / 1 = 0.25 Ohm.
We use a 0.5W 0.22 Ohm resistor in a 2512 SMD package:
which will give a current of 1.1A. At this current, the resistor will dissipate approximately 0.2 W of heat and will not get particularly hot.
The CPC9909 chip generates control pulses. The total pulse duration is the sum of the “high level” time when the mosfet is open and the duration of the pause when the transistor is closed. We can only strictly fix the duration of the pause. The resistor Rt is responsible for it. Its resistance is calculated by the formula:
Rt = (tp - 0.8) * 66, where tp is the pause in microseconds. Resistance Rt is obtained in kiloohms.
The duration of the “high level” - this is the time during which the operating current reaches the required value - is regulated by the CPC9909 microcircuit. The standard frequency range is in the range of 30-120KHz. Moreover, the higher the frequency, the lower the inductor inductance will ultimately be required. But the more the power transistor will heat up. Since the inductance of the inductor (and its associated dimensions) is more important for us, we will try to stay at the upper part of the permissible frequency range.
Let's calculate the allowable pause time. The ratio of the duration of the “high level” to the total duration of the pulse - the duty cycle of the pulse - is calculated by the formula:
D = V LED / V IN = 30 / 310 = 0.097
The switching frequency is calculated as follows:
F = (1 - D) / tp, which means tp = (1 - D) / F
Let the frequency be 90KHz. In this case
tp = (1 - 0.097) / 90,000 = 10µs
Accordingly, resistor Rt will be required
Rt = (10 - 0.8) * 66 = 607.2KOhm
The closest available rating is 620KOhm. Any resistor with this resistance will do, preferably with an accuracy of 1%. We specify the pause time with a resistor rated 620KOhm:
tp = Rt / 66 + 0.8 = 620 / 66 + 0.8 = 10.19µs
The minimum inductance of the inductor L is calculated by the formula
Lmin = (V LED * tp) / I D
Using the refined tp value, we obtain
Lmin = (30 * 10.19) / 0.3 = 1mH
The operating current of the inductor, at which it is guaranteed not to go into saturation, is 1.1 + 15% = 1.3A. It's better to take one and a half extra. Those. not less than 2A.
I have not found a ready-made throttle with such parameters for sale. You need to do it yourself. In general, the calculation of inductors is a large separate topic. Here I will only leave a link to the thorough work of Kuznetsov A.
I used a choke soldered from the non-working ballast of an ordinary energy-saving lamp. Its inductance is 2 mH; there is a gap of about 1 mm in the core. We calculate the operating current, we get up to 1.3 - 1.5A. Not enough, but for a test build it will do.
All that remains is the power transistor and diode. It’s simpler here - both must be designed for a voltage of at least 400V and a current of 4-5A. A fast Schottky diode could be, for example, STTH5R06. The mosfet must be N-channel. For it, the minimum resistance in the open state and the minimum gate charge are extremely important - less than 25 nC. An excellent choice for the current we need is FDD7N60NZ. In a DPAK case and with a current of up to 1A, it will not get particularly hot. It will be possible to do without a radiator.
When laying out a printed circuit board, you need to pay attention to the length of the conductors and correct location"earth". The conductor between the CPC9909 and the gate of the FET should be as short as possible. The same applies to the conductor from the touch resistor. The area of the “land” should be as large as possible. It is highly advisable to place one layer of the printed circuit board completely on the ground. Resistor Rt should be kept away from inductance and other conductors operating at high frequencies.
The LD pin of the microcircuit can be used to smoothly adjust the brightness of the LED, the PWMD pin can be used for dimming using PWM.
Here are examples from technical documentation who implement this.
In this circuit, the current strength, and accordingly, the brightness of the LEDs, is smoothly adjusted from zero to 350 mA by variable resistor RA1. Also on the diagram there are ratings and names of key elements for powering a line of bright LEDs with a current of up to 350mA.
A circuit that involves controlling brightness using PWM looks like this:
The permissible dimming frequency is up to 500Hz. Pay attention to the very desirable electrical isolation of the control pulse generator (usually a microcontroller) and the power part of the circuit. Isolation is performed using an optocoupler.
I assembled a circuit with continuously adjustable variable resistor. The result was a board 60x30mm.
The driver worked immediately and as needed. Using a variable resistor, the current is regulated from 0.1 to calculated 1.1A. The cooler fan where the LED is installed is powered by 3 volts. It rotates completely silently, while the radiator heats up weakly. On the board, after 5 test minutes of operation at maximum current, the inductor heated up to 50C. Its operating current, as expected, turned out to be insufficient. The field-effect transistor also gets noticeably hot. The remaining parts heat up slightly.
The heart of the future powerful lamp in a test run
The board layout in the Sprint-Layout 6.0 program can be taken
After some time, the LED and driver took over workplace in aquarium lighting. They work 15 hours a day at a current of 0.7A. In my opinion, there is enough light for an aquarium with a volume of 140 liters. The radiator is equipped with a thermistor and a simple circuit - the cooler turns on automatically and cools the entire structure.
The driver for an LED from a 220V network requires attention during design and assembly. I repeat - 220V voltage is dangerous to life, and on the driver circuit, almost all parts are under this and higher voltage.
However, with careful assembly, you will get a fairly miniature and efficient driver capable of powering one or more powerful LEDs from a 220V household network.
