DIY EPR power supply. DIY power supply from an energy-saving lamp Power supply from a lamp

How to make a switching power supply from a burnt-out light bulb in an hour?

In this article you will find a detailed description of the process of manufacturing switching power supplies of different powers based on the electronic ballast of a compact fluorescent lamp.

You can make a switching power supply for 5...20 Watts in less than an hour. It will take several hours to make a 100-watt power supply. https://site/

Building a power supply won't be much more difficult than reading this article. And certainly, it will be easier than finding a low-frequency transformer of suitable power and rewinding its secondary windings to suit your needs.

The most interesting videos on Youtube

Introduction.

Compact Fluorescent Lamps (CFLs) are now widely used. To reduce the size of the ballast choke, they use a high-frequency voltage converter circuit, which can significantly reduce the size of the choke.

If the electronic ballast fails, it can be easily repaired. But when the bulb itself fails, the light bulb is usually thrown away.

However, the electronic ballast of such a light bulb is an almost ready-made switching power supply unit (PSU). The only way the electronic ballast circuit differs from a real pulse power supply is the absence of an isolation transformer and a rectifier, if necessary. https://site/

At the same time, modern radio amateurs experience great difficulty in finding power transformers to power their homemade products. Even if a transformer is found, its rewinding requires the use of a large amount of copper wire, and the weight and dimensions of products assembled on the basis of power transformers are not encouraging. But in the vast majority of cases, the power transformer can be replaced with a switching power supply. If you use ballast from faulty CFLs for these purposes, the savings will amount to a significant amount, especially if we are talking about transformers of 100 watts or more.

The difference between a CFL circuit and a pulse power supply.

This is one of the most common electrical circuits for energy-saving lamps. To convert a CFL circuit into a switching power supply, it is enough to install just one jumper between the points A – A’ and add a pulse transformer with a rectifier. Elements that can be deleted are marked in red.

And this is a complete circuit of a switching power supply, assembled on the basis of a CFL using an additional pulse transformer.

To simplify, the fluorescent lamp and several parts were removed and replaced with a jumper.

As you can see, the CFL circuit does not require major changes. Additional elements introduced into the scheme are marked in red.

What power power supply can be made from CFLs?

The power of the power supply is limited by the overall power of the pulse transformer, the maximum permissible current of the key transistors and the size of the cooling radiator, if used.

A small power supply can be built by winding the secondary winding directly onto the frame of an existing inductor.

If the choke window does not allow winding the secondary winding or if it is necessary to build a power supply with a power significantly exceeding the power of the CFL, then an additional pulse transformer will be needed.

If you need to get a power supply with a power of over 100 Watts, and you are using a ballast from a 20-30 Watt lamp, then, most likely, you will have to make small changes to the electronic ballast circuit.

In particular, you may need to install more powerful diodes VD1-VD4 in the input bridge rectifier and rewind the input inductor L0 with a thicker wire. If the current gain of the transistors turns out to be insufficient, then you will have to increase the base current of the transistors by reducing the values ​​of resistors R5, R6. In addition, you will have to increase the power of resistors in the base and emitter circuits.

If the generation frequency is not very high, then it may be necessary to increase the capacitance of the isolation capacitors C4, C6.

Pulse transformer for power supply.

A feature of half-bridge switching power supplies with self-excitation is the ability to adapt to the parameters of the transformer used. And the fact that the feedback circuit will not pass through our homemade transformer completely simplifies the task of calculating the transformer and setting up the unit. Power supplies assembled according to these schemes forgive errors in calculations of up to 150% or more. Tested in practice.

Input filter capacitance and voltage ripple.

In the input filters of electronic ballasts, to save space, small capacitors are used, on which the magnitude of voltage ripple with a frequency of 100 Hz depends.

To reduce the level of voltage ripple at the power supply output, you need to increase the capacitance of the input filter capacitor. It is advisable that for every watt of PSU power there is one microfarad or so. An increase in capacitance C0 will entail an increase in the peak current flowing through the rectifier diodes at the moment the power supply is turned on. To limit this current, a resistor R0 is needed. But, the power of the original CFL resistor is small for such currents and it should be replaced with a more powerful one.

If you need to build a compact power supply, you can use electrolytic capacitors, which are used in film flash lamps. For example, Kodak disposable cameras have miniature capacitors without identification marks, but their capacity is as much as 100µF at a voltage of 350 Volts.

Power supply 20 Watt.

A power supply with a power close to the power of the original CFL can be assembled without even winding a separate transformer. If the original inductor has enough free space in the magnetic circuit window, then you can wind a couple of dozen turns of wire and get, for example, a power supply for a charger or a small power amplifier.

The picture shows that one layer of insulated wire was wound over the existing winding. I used MGTF wire (stranded wire in fluoroplastic insulation). However, in this way you can get a power of only a few watts, since most of the window will be occupied by the wire insulation, and the cross-section of the copper itself will be small.

If more power is required, then ordinary varnished copper winding wire can be used.

Attention! The original inductor winding is under mains voltage! When making the modification described above, be sure to take care of reliable inter-winding insulation, especially if the secondary winding is wound with ordinary varnished winding wire. Even if the primary winding is covered with a synthetic protective film, an additional paper gasket is necessary!

As you can see, the winding of the inductor is covered with a synthetic film, although often the winding of these chokes is not protected by anything at all.

We wrap two layers of electrical cardboard 0.05 mm thick or one layer 0.1 mm thick over the film. If there is no electrical cardboard, we use any paper of suitable thickness.

We wind the secondary winding of the future transformer on top of the insulating gasket. The wire cross-section should be selected as large as possible. The number of turns is selected experimentally, fortunately there will be few of them.

Thus, I managed to obtain power at a load of 20 Watts at a transformer temperature of 60ºC, and a transistor temperature of 42ºC. It was not possible to obtain even more power at a reasonable temperature of the transformer due to the too small area of ​​the magnetic circuit window and the resulting wire cross-section.

The picture shows a working power supply model.

The power supplied to the load is 20 watts. The frequency of self-oscillations without load is 26 kHz. Self-oscillation frequency at maximum load – 32 kHz Transformer temperature – 60ºС Transistor temperature – 42ºС

To increase the power of the power supply, we had to wind a TV2 pulse transformer. In addition, I increased the capacitance of the mains voltage filter capacitor C0 to 100µF.

Since the efficiency of the power supply is not 100%, we had to attach some radiators to the transistors.

After all, if the efficiency of the unit is even 90%, you will still have to dissipate 10 Watts of power.

I was unlucky; my electronic ballast was equipped with transistors 13003 pos. 1 of a design that was apparently designed to be attached to a radiator using shaped springs. These transistors do not need gaskets, since they are not equipped with a metal platform, but they also transfer heat much worse. I replaced them with transistors 13007 pos. 2 with holes so that they could be screwed to the radiators with ordinary screws. In addition, 13007 have several times higher maximum permissible currents.

If you wish, you can safely screw both transistors onto one radiator. I checked it works.

Only, the housings of both transistors must be insulated from the radiator housing, even if the radiator is located inside the electronic device housing.

It is convenient to fasten with M2.5 screws, onto which you must first put insulating washers and sections of an insulating tube (cambric). It is allowed to use heat-conducting paste KPT-8, since it does not conduct current.

Attention! Transistors are under mains voltage, so insulating gaskets must ensure electrical safety conditions!

The drawing shows a sectional view of the connection of the transistor to the cooling radiator.

1. Screw M2.5.
2. Washer M2.5.
3. Insulating washer M2.5 – fiberglass, textolite, getinax.
4. Transistor housing.
5. The gasket is a piece of tube (cambric).
6. Gasket – mica, ceramics, fluoroplastic, etc.
7. Cooling radiator.

And this is a working 100-watt switching power supply.

The load equivalent resistors are placed in water because their power is insufficient.

The power released at the load is 100 watts.

The frequency of self-oscillations at maximum load is 90 kHz.

The frequency of self-oscillations without load is 28.5 kHz.

Transistor temperature – 75ºC.

The area of ​​the radiators of each transistor is 27 cm².

Throttle temperature TV1 – 45ºC.

TV2 – 2000NM (Ø28 x Ø16 x 9mm)

(CFLs or “energy saving lights”) appeared in everyday life quite a long time ago, but still hold, if not primacy among lighting devices, then one of the leading places. They are compact, economical, and can work instead of a conventional incandescent light bulb. But these devices also have disadvantages. Despite the service life declared by the manufacturer, CFLs often fail before they even reach the end of their service life.

This is most often caused by unstable supply voltage and frequent “clicking” of the switch. Is it possible to somehow use a burnt-out device that costs quite a lot of money? Of course you can! In this article we will try to assemble a power supply from an energy-saving lamp with our own hands.

Design and principle of operation of electronic ballasts

Before we begin remaking the electronic ballast for compact fluorescent lamps, let’s take a closer look at this unit and the principle of its operation. The main task of ballast:

• start the gas discharge tube of the lamp;
• maintain the current and voltage necessary for the tube to operate.

Let's take a look at the classic circuit of an electronic ballast or, to call it correctly, an electronic ballast (Electronic Start Control Apparatus).

Electronic ballast circuit (electronic ballast) for energy-saving lamps

In essence, this is a regular switching power supply with minor differences, but more on them later. The network voltage is supplied to the bridge rectifier VD1-VD4, smoothed by capacitor C1 and supplied to a high-frequency (self-oscillation frequency 10-60 kHz) generator assembled on transistors VT2, VT3. Generation in it occurs due to the positive feedback provided by transformer T1; startup when power is applied occurs thanks to the symmetrical dinistor DB1.

The pulse voltage is supplied through a current-limiting inductor T2 to an energy-saving lamp made in the form of a curved tube. Capacitor C8 is needed to create a high-voltage pulse that ignites the tube. As soon as a breakdown of the gas section occurs in the lamp, a choke comes into operation, limiting the current to the level necessary for the lamp to operate. Since the voltage frequency is relatively high, the inductor is very compact.

Important! Manufacturers of energy-saving lamps use different ballast designs in their products, but their operating principle is the same.

Differences between the lamp design and the pulse unit

What is the difference between a CFL electronic ballast and a switching power supply (UPS)? First of all, there is a current-limiting choke at the ballast output. Further, the circuit does not have galvanic isolation of the mains voltage from the output voltage, therefore all elements of the circuit that the electronic ballast powers are under life-threatening voltage. Now let’s try to make a switching power supply from an energy-saving lamp.

In addition to these differences, the voltage at the output of the electronic ballast is pulsed, while the power supply usually produces a constant voltage.

Scheme for converting electronic ballasts into a UPS

To convert electronic ballasts into a power supply, it is necessary to solve three problems:

1. Ensure electrical safety by creating galvanic isolation.
2. Reduce the output voltage of the converter, since at its output it is quite high - about 100-150 V.
3. Rectify the output voltage.

If you need a low-power power supply - up to 15 W, then no special modifications to the electronic ballasts will be required. A dozen centimeters of winding wire, four diodes and a pair of capacitors are enough. And, of course, you will need an electronic ballast from a 40 W lamp. Let's take a look at the modified diagram:

A simple 12 V switching power supply from an electronic ballast of a fluorescent lamp

Here the inductor plays the role of an isolating and at the same time step-down transformer of the power supply, and the rectifier (diodes VD8-VD11) turns the pulse voltage into a constant one. Capacitors C8 and C9 are smoothing. Otherwise, the operation of the power supply is no different from the electronic ballast circuit.

We will convert the electronic ballasts into a power supply in the following sequence:

1. We remove the fluorescent tube and capacitor C8.
2. We connect the terminals of capacitors C6, C7 and inductor T2, which previously went to the lamp, to each other. The easiest way to do this is to simply short-circuit all the terminals of the lamp.

Now our inductor is the load of the converter. All that remains is to wind the secondary winding onto it. Since the conversion frequency is quite high, only a few turns of winding wire with a diameter of 0.5-0.8 mm are needed. The gap between the core and the inductor winding is small, but it is quite sufficient for several turns, the number of which is selected experimentally.

Expert opinion

Alexey Bartosh

Ask a question to an expert

Important! For greater reliability of the power supply, it is better to use not the usual winding wire in enamel insulation, but a mounting wire in fluoroplastic. This will prevent breakdown between the windings due to careless winding and the appearance of dangerous voltage in the secondary circuit.

The winding method is as follows. We wind about 10 turns as a secondary, connect a diode bridge with smoothing capacitors to it and load the future power supply with a resistor with a power of about 30 W and a resistance of 5-6 Ohms. We measure the voltage across the resistor with a DC voltmeter. Then we divide the resulting voltage by the number of turns, and the voltage obtained from one turn comes out. Now we divide the voltage we need (12-13 V) by the last value and get the required number of turns of the secondary winding.

Suppose, having wound 10 turns, we get a voltage of 8 V. 8/10 = 0.8. This means that one turn produces 0.8 volts. We need 12. Divide 12 by 0.8, we get 15. So, we need to wind 15 turns.

Standard and modified power supply choke from electronic ballasts

In the diode bridge, you can use any rectifier diodes with a reverse voltage of at least 25 V and a current of 1A. It is better to use Schottky diodes for these purposes - they have a lower forward voltage drop and work better in pulse mode, increasing the efficiency of the power supply. Instead of C8, a ceramic capacitor with a capacity of 0.1 µF can work, C9 - an electrolytic capacitor with a capacity of 10-50 µF and an operating voltage of at least 25 V.

Everyone likes the circuit of such a power supply, but the voltage at its output is not stabilized. That is, it will fluctuate along with changes in the network. It’s quite easy to get out of this situation by installing a 12-volt stabilizer in the power supply circuit. Ideal for this purpose would be the integrated stabilizer KR142EN8B or the foreign analogue L1812. In this case, the output fragment of the circuit will look like this:

Power supply circuit with stabilized output voltage

Capacitors C10 and C11 must be taken of the same ratings as C8, C9.

Expert opinion

Alexey Bartosh

Specialist in repair and maintenance of electrical equipment and industrial electronics.

Ask a question to an expert

Important! If a stabilizer is used in the power supply circuit, then the number of turns must be increased until the voltage across the load resistor (see the calculation method above) is 15-16 V. This is the normal input voltage for a linear 12-volt stabilizer.

How to increase power

Typically, CFL power is relatively low and ranges from 10-40 W. Not bad in theory, but in practice the current-limiting choke spoils the whole thing. It does not allow the homemade power supply to develop maximum power, firstly, due to its current-limiting properties, and secondly, due to its own low power. As the current increases, the magnetic circuit begins to operate in saturation mode, reducing the efficiency of the power supply and overloading the key transistors, and overloading in vain.

How to make a relatively powerful power supply from? The task is not as difficult as it seems at first glance. To do this, it is enough to replace the inductor with a relatively powerful pulse transformer. Of course, this will require more in-depth knowledge of radio engineering, but it’s worth it.

The transformer can be taken, for example, from an unnecessary power supply from a computer or other office equipment (printer, scanner, small-sized TV, etc.). You will also need a resistor with a power of 3 W and a resistance of 5 Ohms, as well as a new high-voltage capacitor with a nominal value of 100 μF and an operating voltage of at least 350 V. Let's take a look at the modified circuit:

Power supply circuit with increased output power

Here, instead of a choke, a pulse transformer is installed, and the primary winding is the one that was connected to the converter (high-voltage), and the secondary is a step-down winding. In addition, resistor R1 was chosen to be of higher power, and the capacity of smoothing capacitor C1 (according to the modified circuit C0) was increased to 100 μF. Otherwise, the circuit has remained virtually unchanged, but now it is quite capable of delivering a current of 5-8 A to the load at a voltage of 12 V. Such power supplies can already be used for a screwdriver and similar 12-volt tools.

1. At the first start-up, it is better to connect the modified power supply to the network through a 220 V 60-100 W incandescent lamp. If everything is in order, the lamp will barely glow. If there is an error in the circuit, the lamp will burn quite brightly. This will save the transistors from breakdown due to installation errors.
2. Before putting the power supply into long-term operation, it is necessary to “drive” it with a load resistor. In this case, the transformer and transistors should not heat above 60 degrees Celsius.
3. If the transformer gets very hot, you will have to wind the step-down winding with a thicker wire.
4. If transistors get very hot, they need to be equipped with small radiators.
5. You should not use such a power supply to charge and power expensive gadgets. It is much safer to buy a factory power supply. This will cost much less than repairing, for example, a laptop or smartphone.

This is where we can probably end the conversation about converting electronic ballasts for compact fluorescent lamps into a switching power supply. If you carefully read the article and have at least a slight understanding of radio engineering, then you can handle this simple modification yourself.

Currently, so-called energy-saving fluorescent lamps are becoming increasingly widespread. Unlike conventional fluorescent lamps with electromagnetic ballast, energy-saving lamps with electronic ballast use a special circuit.

Thanks to this, such lamps can be easily installed in a socket instead of a conventional incandescent light bulb with a standard E27 and E14 base. It is about household fluorescent lamps with electronic ballast that will be discussed further.

Distinctive features of fluorescent lamps from conventional incandescent lamps.

It is not for nothing that fluorescent lamps are called energy-saving, since their use can reduce energy consumption by 20–25%. Their emission spectrum is more consistent with natural daylight. Depending on the composition of the phosphor used, it is possible to produce lamps with different shades of glow, both warmer tones and colder ones. It should be noted that fluorescent lamps are more durable than incandescent lamps. Of course, a lot depends on the quality of the design and manufacturing technology.

Compact fluorescent lamp (CFL) device.

A compact fluorescent lamp with electronic ballast (abbreviated CFL) consists of a bulb, an electronic board and an E27 (E14) socket, with which it is installed in a standard socket.

Inside the case there is a round printed circuit board on which the high-frequency converter is assembled. The converter at rated load has a frequency of 40 - 60 kHz. As a result of the fact that a fairly high conversion frequency is used, the “blinking” characteristic of fluorescent lamps with electromagnetic ballast (based on a choke), which operate at a power supply frequency of 50 Hz, is eliminated. The schematic diagram of a CFL is shown in the figure.

According to this concept, mostly fairly cheap models are assembled, for example, those produced under the brand Navigator And ERA. If you use compact fluorescent lamps, then most likely they are assembled according to the above diagram. The spread of the values ​​of the parameters of resistors and capacitors indicated on the diagram actually exists. This is due to the fact that lamps of different wattages use elements with different parameters. Otherwise, the circuit design of such lamps is not much different.

Let's take a closer look at the purpose of the radioelements shown in the diagram. On transistors VT1 And VT2 a high-frequency generator has been assembled. Silicon high-voltage transistors are used as transistors VT1 and VT2 n-p-n MJE13003 series transistors in TO-126 package. Typically, only the digital index 13003 is indicated on the housing of these transistors. MPSA42 transistors in a smaller TO-92 format or similar high-voltage transistors can also be used.

Miniature symmetrical dinistor DB3 (VS1) serves to autostart the converter at the moment of power supply. Externally, the DB3 dinistor looks like a miniature diode. An autostart circuit is necessary because the converter is assembled according to a circuit with current feedback and therefore does not start on its own. In low-power lamps, the dinistor may be absent altogether.

Diode bridge made on elements VD1 – VD4 serves to rectify alternating current. Electrolytic capacitor C2 smoothes out the ripples of the rectified voltage. The diode bridge and capacitor C2 are the simplest network rectifier. From capacitor C2, constant voltage is supplied to the converter. The diode bridge can be made using separate elements (4 diodes), or a diode assembly can be used.

During its operation, the converter generates high-frequency interference, which is undesirable. Capacitor C1, choke (inductor) L1 and resistor R1 prevent the spread of high-frequency interference through the electrical network. In some lamps, apparently to save money :) a wire jumper is installed instead of L1. Also, many models do not have a fuse FU1, which is indicated in the diagram. In such cases, the breaking resistor R1 also plays the role of a simple fuse. If the electronic circuit malfunctions, the current consumed exceeds a certain value, and the resistor burns out, breaking the circuit.

Throttle L2 usually assembled at Sh-figurative ferrite magnetic core and looks like a miniature armored transformer. On the printed circuit board this inductor takes up quite an impressive amount of space. The inductor winding L2 contains 200 - 400 turns of wire with a diameter of 0.2 mm. You can also find a transformer on the printed circuit board, which is indicated on the diagram as T1. Transformer T1 is assembled on a ring magnetic core with an outer diameter of about 10 mm. The transformer has 3 windings wound with mounting or winding wire with a diameter of 0.3 - 0.4 mm. The number of turns of each winding ranges from 2 - 3 to 6 - 10.

The fluorescent lamp bulb has 4 leads from 2 spirals. The leads of the spirals are connected to the electronic board using the cold twist method, that is, without soldering and are screwed onto rigid wire pins that are soldered into the board. In low-power lamps with small dimensions, the leads of the spirals are soldered directly into the electronic board.

Repair of household fluorescent lamps with electronic ballast.

Manufacturers of compact fluorescent lamps claim that their lifespan is several times longer than that of conventional incandescent lamps. But despite this, household fluorescent lamps with electronic ballast fail quite often.

This is due to the fact that they use electronic components that are not designed to withstand overloads. It is also worth noting the high percentage of defective products and low quality workmanship. Compared to incandescent lamps, the cost of fluorescent lamps is quite high, so repairing such lamps is justified at least for personal purposes. Practice shows that the cause of failure is mainly a malfunction of the electronic part (converter). After a simple repair, the performance of the CFL is completely restored and this allows you to reduce financial costs.

Before we start talking about CFL repairs, let’s touch on the topic of ecology and safety.

Despite their positive qualities, fluorescent lamps are harmful to both the environment and human health. The fact is that there are mercury vapors in the flask. If it is broken, dangerous mercury vapors will enter the environment and, possibly, the human body. Mercury is classified as a substance 1st hazard class .

If the flask is damaged, you must leave the room for 15–20 minutes and immediately forcefully ventilate the room. You must be careful when using any fluorescent lamps. It should be remembered that mercury compounds used in energy-saving lamps are more dangerous than ordinary metallic mercury. Mercury can remain in the human body and cause harm to health.

In addition to this disadvantage, it should be noted that the emission spectrum of a fluorescent lamp contains harmful ultraviolet radiation. If you stay close to a fluorescent lamp for a long time, skin irritation is possible, as it is sensitive to ultraviolet radiation.

The presence of highly toxic mercury compounds in the bulb is the main motive of environmentalists who call for reducing the production of fluorescent lamps and switching to safer LED lamps.

Disassembling a fluorescent lamp with electronic ballast.

Despite the ease of disassembling a compact fluorescent lamp, you should be careful not to break the bulb. As already mentioned, there are mercury vapors inside the flask that are hazardous to health. Unfortunately, the strength of glass flasks is low and leaves much to be desired.

In order to open the housing where the electronic circuit of the converter is located, it is necessary to release the plastic latch that holds the two plastic parts of the housing together with a sharp object (a narrow screwdriver).

Next, you should disconnect the leads of the spirals from the main electronic circuit. It is better to do this with narrow pliers, picking up the end of the spiral wire output and unwinding the turns from the wire pins. After this, it is better to place the glass flask in a safe place to prevent it from breaking.

The remaining electronic board is connected by two conductors to the second part of the housing, on which a standard E27 (E14) base is mounted.

Restoring the functionality of lamps with electronic ballast.

When restoring a CFL, the first thing you should do is check the integrity of the filaments (spirals) inside the glass bulb. The integrity of the filaments can be easily checked using a regular ohmmeter. If the resistance of the threads is low (a few ohms), then the thread is working. If during measurement the resistance is infinitely high, then the filament has burned out and it is impossible to use the flask in this case.

The most vulnerable components of an electronic converter made on the basis of the circuit already described (see circuit diagram) are capacitors.

If the fluorescent lamp does not turn on, then capacitors C3, C4, C5 should be checked for breakdown. When overloaded, these capacitors fail because the applied voltage exceeds the voltage for which they are designed. If the lamp does not turn on, but the bulb glows in the area of ​​the electrodes, then capacitor C5 may be broken.

In this case, the converter is working properly, but since the capacitor is broken, a discharge does not occur in the bulb. Capacitor C5 is included in an oscillatory circuit, in which, at the moment of startup, a high-voltage pulse occurs, leading to the appearance of a discharge. Therefore, if the capacitor is broken, the lamp will not be able to switch to operating mode normally, and a glow caused by heating of the spirals will be observed in the area of ​​the spirals.

Cold And hot mode starting fluorescent lamps.

There are two types of household fluorescent lamps:

With cold start

With hot start

If the CFL lights up immediately after switching on, then it has a cold start. This mode is bad because in this mode the cathodes of the lamp are not preheated. This can lead to burnout of the filaments due to the flow of a current pulse.

For fluorescent lamps, hot starting is preferable. During a hot start, the lamp lights up smoothly within 1-3 seconds. During these few seconds, the filaments heat up. It is known that a cold filament has less resistance than a heated filament. Therefore, during a cold start, a significant current pulse passes through the filament, which can eventually cause it to burn out.

For conventional incandescent lamps, a cold start is standard, so many people know that they burn out just the moment they are turned on.

To implement hot start in lamps with electronic ballast, the following circuit is used. A posistor (PTC - thermistor) is connected in series with the filaments. In the circuit diagram, this posistor will be connected in parallel with capacitor C5.

At the moment of switching on, as a result of resonance, a high voltage appears on the capacitor C5, and, consequently, on the electrodes of the lamp, necessary for its ignition. But in this case, the filaments are poorly heated. The lamp turns on instantly. In this case, a posistor is connected in parallel with C5. At the moment of startup, the posistor has a low resistance and the quality factor of the L2C5 circuit is significantly lower.

As a result, the resonance voltage is below the ignition threshold. Within a few seconds the posistor heats up and its resistance increases. At the same time, the filaments also heat up. The quality factor of the circuit increases and, consequently, the voltage at the electrodes increases. A smooth hot start of the lamp occurs. In operating mode, the posistor has a high resistance and does not affect the operating mode.

It is not uncommon that this particular posistor fails, and the lamp simply does not turn on. Therefore, when repairing lamps with ballast, you should pay attention to it.

Quite often, the low-resistance resistor R1 burns out, which, as already mentioned, plays the role of a fuse.

Active elements such as transistors VT1, VT2, rectifier bridge diodes VD1 - VD4 are also worth checking. As a rule, the cause of their malfunction is an electrical breakdown. p-n transitions. Dinistor VS1 and electrolytic capacitor C2 rarely fail in practice.

I bought 10 W 900 lm warm white LEDs on AliExpress to try. The price in November 2015 was 23 rubles per piece. The order arrived in a standard bag, I checked everything was in good order.


To power LEDs in lighting devices, special units are used - electronic drivers, which are converters that stabilize the current rather than the voltage at their output. But since the drivers for them (I also ordered on AliExpreess) were still on the way, I decided to power them from ballast from energy-saving lamps. I've had several of these faulty lamps. whose filament in the bulb burned out. As a rule, the voltage converter for such lamps is working properly, and it can be used as a switching power supply or LED driver.
We disassemble the fluorescent lamp.

For the conversion, I took a 20 W lamp, the choke of which can easily deliver 20 W to the load. For a 10W LED, no further modifications are required. If you plan to power a more powerful LED, you need to take a converter from a more powerful lamp, or install a choke with a larger core.
Installed jumpers in the lamp ignition circuit.

I wound 18 turns of enamel wire around the inductor, solder the terminals of the wound winding to the diode bridge, apply mains voltage to the lamp and measure the output voltage. In my case, the unit produced 9.7V. I connected the LED through an ammeter, which showed a current passing through the LED of 0.83A. My LED has an operating current of 900mA, but I reduced the current to increase the resource. I assembled the diode bridge on the board using a hinged method.

Remodeling scheme.

I installed the LED using thermal paste on a metal lampshade of an old table lamp.

I installed the power board and diode bridge into the body of a table lamp.

When working for about an hour, the LED temperature is 40 degrees.

To the eye, the illumination is like that of a 100-watt incandescent lamp.

I'm planning to buy +128 Add to favorites I liked the review +121 +262

How to convert a housekeeper converter into a switching power supply?

If you have a housekeeper lamp with a faulty bulb lying around, do not rush to throw it away. Inside the base there is a high-frequency converter circuit, which replaces the large and heavy ballast choke, as in the connection circuits of conventional LDS. Based on this converter, you can make a 20-watt switching power supply, and with a more careful approach you can squeeze out more than a hundred.

Below is one of the most common options for housekeeper converter circuits:

This is a diagram of a 25-watt Vitoone energy-saving lamp. The red color on it indicates those elements that we do not need, so we exclude them from the diagram, and put a jumper between points A and A’. The only thing left to do is screw a pulse transformer and a rectifier to the output.

A version of the already converted “energy saving” circuit into a switching power supply is shown in the figure below:

As can be seen from the diagram, R0 was set to 2 times less than the nominal value, but its power was increased, C0 was replaced by 100.0 mF, and TV2 was added at the output with a rectifier for VD14, VD15, C9 and C10. Resistor R0 serves as a fuse and charge current limiter when turned on. Select the nominal capacity C0 so that it is (approximately) numerically equal to the power of the power supply unit you are making.

Regarding capacitor C0: it can be “torn out” from an old Kodak-type film camera, or any other film soap dish; in the flash lamp circuit there is exactly the one we need, 100mF at 350V.

TV2 is a pulse transformer; the power of the power supply itself depends on its overall power, as well as on the maximum permissible current of the key transistors. To make a low-power pulsed power supply, it is enough to wind a secondary winding around the existing inductor, as shown in the following diagram:

To power any low-voltage charger or a not very powerful amplifier, wind 20 turns on top of the existing winding L5, this will be enough.

The picture above shows a working version of the power supply without a 20-watt rectifier. At idle, the frequency of self-oscillations is 26 kHz, under load 20W 32 kHz, the transformer heats up to 60 ºС, transistors up to 42 ºС.

Important!!! Mains voltage is present on the primary winding when the converter is operating, so be sure to lay a layer of paper insulation that will separate the primary and secondary windings, even if there is already a synthetic protective film on the primary.

But it also happens that in the window of the existing choke there is not enough space for winding the secondary winding, or in the case when we have to create a power supply with much greater power than the power of the “energy saving” being converted - here we cannot do without the use of an additional pulse trance (see second outline of the article).

For example, we make a switching power supply with more than 100W power, and use ballast from a 20-watt light bulb. In this case, you will need to replace VD1 - VD4 with more “current” diodes, and wind the inductor L0 with a thicker wire. If the current gain of VT1 and VT2 is insufficient, increase the base current of the transistors by reducing the ratings of R5 and R6, as well as increasing the power of the resistances in the base and emitter circuits.

If the generation frequency is insufficient, increase the ratings of capacitors C4 and C6.

Practical tests have shown that half-bridge pulse power supplies are not critical to the parameters of the output transformer, because the OS circuit does not pass through it, therefore calculation errors of up to 150 percent are allowed.

Switching power supply 100 Watt.

As already written above, in order to get a powerful power supply, an additional pulse transformer TV2 is wound, R0 is replaced, C0 is replaced by 100 mF, it is advisable to replace transistors 13003 with 13007, they are designed for higher current, and it is better to put them on small radiators through insulating gaskets (mica for example).

A cross-section of the connection of transistors with radiators is shown in the figure below:

The current model of a switching power supply operating at a load of 100 W is shown in the picture below:

The transformer is wound on a 2000HM ring, outer diameter 28mm, inner diameter 16mm, ring height 9mm.
Due to the insufficient power of the load resistors, they are placed in a saucer of water.
Generation without load 29 kHz, under load 100 W - 90 kHz.

Regarding the rectifier.

To prevent the magnetic circuit of transformer TV2 from entering saturation, make the rectifiers in half-bridge pulse power supplies full-wave, that is, they must be bridged (1), or with a zero point (2). See picture below.

With a bridge circuit, a little less wire is required per winding, but at the same time 2 times more energy is dissipated on VD1-VD4. The second fragment of the figure shows a version of the rectifier circuit with a zero point; it is more economical, but the windings in this case must be absolutely symmetrical, otherwise the magnetic circuit will go into saturation. The second option is used when, with a small output voltage, you need to have a significant current. To minimize losses, silicon diodes are replaced with Schottky diodes; the voltage across them drops less than 2-3 times.

Let's look at an example:

At P=100W, U=5V, TV1 with midpoint, 100 / 5 * 0,4 = 8 , i.e. Schottky diodes dissipate 8 W of power.
At P=100W, U=5V, TV1 with a bridge rectifier and conventional diodes, 100 / 5 * 0,8 * 2 = 32 , i.e. power will be dissipated on VD1-VD4 of about 32 W.

Keep this in mind, and don’t look for half of the missing power later.

Setting up a pulse power supply.

Connect the UPS to the network according to the diagram below (fragment 1). Here HL1 will act as a ballast, which has a nonlinear characteristic and will protect your device if an emergency situation arises. The power of HL1 should be approximately equal to the power of the power supply you are testing.

When the power supply is turned on without load, or is operating at a low load, the HL1 filament has little resistance, so it does not have any effect on the operation of the power supply. When some problems occur, the currents VT1 and VT2 increase, the lamp begins to glow, the resistance of the filament increases, thereby reducing the current in the circuit.

If you are constantly repairing and adjusting switching power supplies, it would be a good idea to assemble a special stand (figure above, fragment 2). As you can see, there is an isolation transformer (galvanic isolation between the power supply and the household network), and there is also a toggle switch that allows you to supply voltage to the power supply, bypassing the lamp. This is necessary in order to test the converter when operating under a powerful load.

Powerful glass-ceramic resistors can be used as a load; they are usually green (see figure below). The red numbers in the figure indicate their power.

During long-term tests, when you need to check the thermal conditions of the elements of the power supply circuit, and the load resistors do not have sufficient power, the latter can be lowered into a saucer of water. During operation, the equivalent load gets very hot, so do not grab the resistors with your hands to avoid burns.

If you did everything carefully and correctly, and at the same time used a known-good ballast from an energy-saving lamp, then there is nothing special to adjust. The scheme should work immediately. Connect the load, supply power, and figure out whether your power supply is capable of delivering the required power. Monitor the temperatures of VT1, VT2 (should be no higher than 80-85 ºС) and the output transformer (should be no more than 60-65 ºС).

If the transformer heats up high, increase the cross-section of the wire, or wind the transformer on a magnetic core with a larger overall power, or you may have to do both the first and second.

When heating transistors, place them on a radiator (through insulating gaskets).

If you invented a low-power UPS, and at the same time you wound up the existing choke, and during operation it heats up above the permissible norm, try how it works on a lower-power load.

You can download programs for calculating pulse transformers in the article:

Happy remodeling.