Chinese LED lamps: everything is bad. The light of communism - a multi-review of Chinese LED light bulbs (LED lamps that you have long wanted, but were afraid to buy)

I decided to change the bulbs in the DRLs. There are incandescent lights there, the light is good, but the tint is yellow. Which is dissonant with the bluish-white xenon and LED dimensions.

I rummaged around on Ali and dug up these bulbs:

Well, I also saw reviews on them from one of my DS clubmates on Drive2. They look good. But the main thing here is the diode assemblies and the housing. The rest can be finished on the spot. I'm ordering...

Got it... Looks good. Include:

I covered it with a plastic cap on top to worsen the cooling conditions, bringing them closer to those that await it in the headlight. Well, it hurts your eyes less. It worked for an hour, it warmed up to about 80 degrees. Not very good, but maybe it can be improved a little. And then we'll see how they behave. Voltage, by the way, does not affect brightness. Those. There is some kind of driver inside. We need to open it up and take a look.

We pull the base with a bend to the side and it comes off. It is not glued, just pressed in. A contact is clamped between the case and the base and two wires are drawn from the case. We unsolder them.

Inside is a PT4115 driver, a normal pulse generator. Here is its diagram from the datasheet. It was repeated one after the other, even a diode bridge was inserted :)

Cheap, but not a bad solution. They could have put in a ruler or even a couple of resistors. And here everything is human. There is a diode bridge at the input, which makes it possible to plug it in any way you like. Reduced brightness (double-filament lamp) is achieved by switching it on through a ballast resistor. Which of course is lame. You can't use it as a two-filament one - other flashlights might go crazy. Although it depends on the car’s electrics. I only use one thread. Therefore, I will throw away this resistor.

Let's disassemble the diode assembly. The diode modules here are kind of their own, terribly Chinese. The crystals stand haphazardly, but thanks to the generous amount of phosphor they glow very brightly. The structure appears to be sealed tightly, but you can tinker with it. We solder the cross, you need to unsolder the wire from it going to the end module:

After which the end can be pulled back and its second end can be desoldered:

And pull out the end diode:

Then everything is simply soldered into parts. And so attention! There is a problem. The fact is that the modules have polarity markings. The minus is stamped on the ass. BUT! He's knocked out of his mind. It may or may not coincide with reality! You can't trust him! Therefore, we mark the minus ourselves.

The assemblies are connected in parallel, each with three crystals in series. Not the best option, the death of one assembly leads to the cascading death of the others, but it is cheap and practical. Let's hope that the spread of parameters across diodes is not very fatal.

Judging by the driver current shunt rating (0.3Ohm), the driver injects 300mA into the assembly. It turns out to be approximately 2.7 W for all the corn. Not very much, but in terms of lumens it is not weaker than the native 5W light bulb that is supposed to be there. In general, this driver can output up to 1.5A.

I decided to test how it behaves on noodles. Put everything back together:

I checked it experimentally... exactly, 300mA. The driver itself does not heat up. Well, great.

Now let's put everything back together. But collecting everything was bad manners. It would be necessary to put all the assemblies on hot-melt adhesive so that, firstly, they better dissipate heat to the case, and secondly, they are equalized with each other, which will facilitate the thermal regime and stability of the crystal’s operating parameters, there is less chance that one thread will die and drag the rest with it to the grave.

Somehow I found hot glue in the city. Alsil-5 was everywhere, but upon inspection it turned out to be dried out. And then I accidentally came across the right hot melt adhesive: Arctic Alumina Thermal Adhesive:

Unlike Alsil-5, it is two-component. You squeeze the snot out of two syringes and after 5 minutes they polymerize to the state of hard rubber.

It grabs quickly and tightly, but at the same time is quite elastic. Rulez!

Next, we carefully check the assembly again so as not to mess up the plus and minus. Otherwise, I already collected one here, I relied on the narrow-eyed ones, but they are foolish enough to create polarity. And then we apply glue. We first assemble the crosspiece, and then, in the reverse order, insert and glue the end one:

Let's take a look at the driver. Everything was ok with it, so I didn’t do anything there, I just got rid of the ballast resistor by short-circuiting the lamp spots from the inside with a wire. This is not critical to me, but you see for yourself.

I soldered it to the assembly.

I stretched the heat shrink on top and dripped a little silicone sealant to prevent it from loosening:

I soldered one wire to the nickels of the plinth, and the second, drilling a hole in the side panel for the glass of the plinth. After which I put on the base, having previously coated it with the same hot-melt adhesive. At the same time, the heat will also be transferred to the base and further.

Screwed into the headlight:

The result is cool white DRLs.

Here the Sun also plays on the side of the lamp, illuminating the right side. A normal photo I’ll show you when you’re not too lazy to leave the house in the evening. Because You can't really tell in the sun. In general, it shines no worse than an incandescent lamp, and focuses well in the reflector too. If you turn off the headlight and look directly into it, the entire reflector becomes uniformly yellow - the phosphor is reflected.

ZY
I'm on Drive2, but it's more about exploitation and that's all.

It is easy for an inexperienced buyer to get confused among the colored packages with 220 V LED lamps, which are presented in dozens different manufacturers. Trying to figure it out on the spot and asking the seller: “Which light bulb from the inexpensive series is better?” usually end with the standard answer: “Yes, they are all from China.”

Are they all equally bad?

Are all cheap LED lamps from China of approximately the same quality? Not at all! Some Chinese companies, specializing in the production of LED light sources, have been supplying Russia with products of acceptable quality for more than 10 years. affordable prices. Among them it is worth noting, for example, Camelion and Supra, which operate under a certificate of conformity. Also in Russia, companies that assemble LED lamps from high-quality semi-finished products of Chinese origin are successfully developing their business.

However, the Russian market is still overcrowded with low-quality 220-volt LED lamps from little-known Chinese companies. They continue to fill our stores with light bulbs and lamps that cannot last even six months. However, people buy them mainly because they are attractive. appearance and low price.

What's wrong with cheap Chinese LED lamps and why is it better to refuse to buy them? Let's figure it out.

Electronic filling

It is impossible to identify a low-quality product in a store, and especially on a website. At first glance, the no name LED bulb is no worse than others and shines brightly, and the parameters indicated on the packaging are simply impressive. But once you disassemble the housing, your opinion about the purchased lighting device will change. First, LEDs. Enterprising Chinese continue to mount SMD crystals of previous generations, which leading manufacturers are getting rid of for pennies. Such LEDs have a lower power to luminous flux, and therefore less effective. To increase brightness, more than the rated current is deliberately passed through them and, thus, attract the attention of buyers at the time of purchase.
Secondly, instead of an LED driver, the lamps are equipped with a simple power source, assembled on an RC circuit, a diode bridge, a capacitive filter and a limiting resistor. In total, we get 7–8 radio components costing less than $0.5 at retail. Any overvoltage in the network will lead to degradation of the crystals and failure of the lamp as a whole.

Third, soldering all components electrical circuit leaves much to be desired. In some lamp models you can see crooked SMD elements, which indicates that the lamps were assembled without the use of automated equipment.
To connect the boards to each other and to the base, they use the thinnest wires, which often fall off on their own.

Housing materials

If you simultaneously pick up a branded LED lamp like a lamp from Osram and a cheap no name type, you can immediately feel the difference in the materials from which the body is made. A quality product will have most of the body made of aluminum alloy or heat-conducting plastic, usually with oblong holes around the circumference.
In cheap LED light bulbs from China, the housing is made of thin plastic, which can turn yellow over time due to heat. There are either no holes in it or they are ineffective. When pressed, it can become deformed and even crack if the lamp is accidentally dropped. Same low quality has an insulator on the base that separates the phase from zero. The matte diffuser and metal base, as a rule, cope with their functional load.

Build quality

For an LED light bulb to be truly cheap, it must be assembled quickly and without quality control. This is where ridiculous malfunctions arise. For example, there may be no spot welding or soldering where the wires are attached to the base. In this case, one of the wires is clamped between the base and the plastic.

Do not be surprised if, when screwing the light bulb into the socket, the diffuser becomes loose and remains in your hand. In the simplest case, it is attached to glue, sometimes to silicone, which, after heating, loses its adhesive properties.

The board, the so-called driver, inside the plastic case is often fixed with it with a piece of adhesive tape. This simple fastening is enough for a couple of shakes, after which the board is held solely by soldered wires.

Avoid buying cylindrical shaped LED bulbs, which are popularly called “corn” bulbs. This technology has not proven its worth and is considered a relic of the past.

Video review of a Chinese lamp

Results

At the beginning of 2018, the cost of 7 watt LED lamps of normal quality decreased to 200 rubles. Of course, this only applies to products whose products are assembled on the territory of the Russian Federation and differ stable work. Now middle-income families can gradually switch to LED lighting your homes and experience real energy savings.

It turns out that it is better not to mess with Chinese LED lamps from unknown brands. Buying a “pig in a poke” through AliExpress will end up being quote-unquote cheap.

Read also

) I immediately want to take it apart and look inside, to see how it all works and works. Apparently, this is what distinguishes scientists from ordinary people. Agree, what normal person would disassemble a light bulb for 1000 rubles, but what can you do - the party said: it’s necessary!

Theoretical part

Why do you think everyone is so concerned about replacing incandescent lamps, which have become a symbol of an entire era, with gas-discharge and LED lamps?

Of course, firstly, it is energy efficiency and energy saving. Unfortunately, a tungsten filament emits more “thermal” photons (i.e. light with a wavelength greater than 700-800 nm) than it produces light in the visible range (300-700 nm). It’s hard to argue with this - the graph below will tell everything for itself. Taking into account the fact that the power consumption of gas-discharge and LED lamps is several times lower than that of incandescent lamps at the same illumination, which is measured in lux. Thus, we see that this is truly beneficial for the end consumer. Another thing is industrial facilities (not to be confused with offices): lighting may be an important part, but still the main energy costs are associated precisely with the operation of machines and industrial installations. Therefore, all the gigawatts generated are spent on pipe rolling, electric furnaces, etc. That is real savings within the entire state is not so great.

Secondly, the service life of the lamps that replaced the “Ilyich bulbs” is several times longer. For an LED lamp, the service life is almost unlimited if the heat dissipation is properly organized.

Thirdly, these are innovations/modernizations/nanotechnologies (underline as appropriate). Personally, I don’t see anything innovative in either mercury or LED lamps. Yes, this is a high-tech production, but the idea itself is just a logical application in practice of knowledge about semiconductors, which is 50-60 years old, and materials known for about two decades.

Since the article is devoted to LED lamps, I will dwell in more detail on their design. It has long been known that the conductivity of an illuminated semiconductor is higher than the conductivity of an unlit one (Wiki). In some unknown way, light causes electrons to travel through the material with less resistance. Photon, if its energy more width band gap of a semiconductor (E g), is capable of knocking an electron out of the so-called valence band and throwing it into the conduction band.

Diagram of the arrangement of bands in a semiconductor. E g - band gap, E F - Fermi energy, numbers indicate the distribution of electrons across states at T>0 ()

Let's complicate the task. Let's take two semiconductors with different types conductivity and and connect together. If in the case of one semiconductor we simply observed an increase in the current flowing through the semiconductor, now we see that this diode (which is another name for the p-n junction that appears at the interface of semiconductors with different types conductivity) became a mini-source DC, and the magnitude of the current will depend on the illumination. If you turn off the light, the effect will disappear. By the way, this is the principle of operation of solar panels.

Now let's return to LEDs. It turns out that you can do the opposite: connect a p-type semiconductor to the positive on the battery, and an n-type to the negative, and... And nothing will happen, there will be no radiation in the visible part of the spectrum, since the most common semiconductor materials (for example , silicon and germanium) are opaque in the visible region of the spectrum. The reason for this is that Si or Ge are not direct-gap semiconductors. But there is a large class of materials that have semiconductor properties and are at the same time transparent. Prominent representatives- GaAs (gallium arsenide), GaN (gallium nitride).

In total, to get an LED, we just need to make a p-n junction from a transparent semiconductor. I’ll probably stop here, because the further we go, the more complex and incomprehensible the behavior of LEDs becomes.

Let me just say a few words about modern technologies LED production. The so-called active layer is a very thin 10-15 nm thick alternating layers of p- and n-type semiconductors, which consist of elements such as In, Ga and Al. Such layers are grown epitaxially using the MOCVD (metal-oxide chemical vapor deposition or chemical vapor deposition) method.

For interested readers, I can suggest getting acquainted with the physics underlying the operation of LEDs. Besides this interesting work, carried out within the walls of their native Moscow State University, Svetlana and Optogan have a wonderful galaxy of scientific teams in St. Petersburg itself. For example, PhysTech. You can also read.

Methodological part

All measurements of the lamp spectra were made within 30 minutes (i.e., the background signal changed slightly) in a darkened room using an Ocean Optics QE65000 spectrometer. You can read about the structure of the spectrometer. In addition to 10 dependences for each type of lamp, the dark spectrum was measured, which was then subtracted from the spectra of the lamps. All 10 dependencies for each sample were summed up and averaged. Additionally, each final spectrum was normalized to 100%.

SEM image of individual LEDs on a substrate after removing the polymer layer

The polymer layer itself has a rather interesting structure. It consists of small (diameter ~10 µm) balls:

Optical micrographs of the “underside” of the polymer layer

It happened by chance that one diode cut with a microtome remained in the polymer layer. It is worth noting that the diode itself is truly transparent and the contacts on the other side of the chip are visible through it:

Optical micrographs of the LED from the back: excellent transparency for this type of product

The polymer layer is so firmly glued both to the copper substrate itself and to individual chips that after its removal, a thin layer of polymer still remains on the surface of the diodes. Below, in the images obtained using an electron microscope, you can see in all its glory the “chip” of the very active layer of the diode in which electrons are “degenerated” into photons:

SEM images of the light-emitting layer of a separate LED (arrows indicate the location of the active layer)

And here is the textured buffer layer, take a closer look at the bottom right image - it will be useful to us later (the arrows indicate the buffer layer)

After careless handling of the chip, some contacts were damaged, while others remained intact.

And the last lamp is “SvetaLED”. The first thing that surprises is the substrate with LED modules - attention! - screwed onto a hefty bolt to the rest of the lamp (just like they did in China). When I took it apart, I thought that it might be in the way of “tearing” it away from the rest of the lamp, and then I saw a bolt... By the way, there was a marker on the back of this aluminum substrate! some number is written. It seems that at Svetlana’s factory near St. Petersburg there are migrant workers who assemble these lamps by hand. Although no, wait, light bulbs are produced by the military... ...

Note: I was able to see how the individual chips in the module from Svetlana are connected. Consistently, to my great disappointment. Thus, if at least 1 LED burns out, the entire module will stop working.

SEM images of a light-emitting diode from the Svetlana company (arrows indicate the active region). In the top left picture, an image of the proposed contacts has been added as they should have been routed in the module (4 x3 diodes).

1 light bulb. The Svetlana module has dimensions of 5 by 5 mm, 2 corners on the “lid” are cut at 45 degrees, etc. - much coincides with the Optogan specification. The ongoing effect of déjà vu is not tormenting?! Or maybe everything is just purchased in Taiwan?!

And, of course, conclusions

Are you ready to be a patriot and call a “domestic” (for example, Optogan’s chips are made in Germany) lamp the best in terms of the combination of all factors?! Probably not. Honestly, I was pleasantly pleased with the Chinese-made LED lamp: the relative simplicity of the diode power supply circuit, simple materials, successful placement of the LEDs on the substrate. Problem with color temperature can be solved, but the only negative that confuses me as a buyer is the durability of the light bulb from the Middle Kingdom.

Lamps of “domestic” production, and in particular, “Optogan”, as always, “please” with their price. I’m more than sure that it would be possible to start with “handicraft” design, cheap materials (glass instead of polycarbonate) and fill the niche of budget light sources (it seems like there aren’t that many rich people in Russia, or am I missing something?! ). But this is not even the main thing; there are quite a few who are ready to invest 1000 rubles in a light bulb and not think about buying them for several years. Let's leave aside the striking external similarity between the modules; I am more concerned with something else - the similarity between individual LED chips (geometric dimensions, location, contacts, etc.). It seems that they were made using equipment from the same company, only the versions of this equipment differ as v.1.0 and v.1.1. Of course, I understand that the most important thing in an LED is the internal structure of the active zone, but, you see, it is difficult to get 1 chip measuring 160 by 500 microns (the thickness of a human hair is 50-80 microns) and compare the emission spectra of the Optogan and Svetlana chips "

However, if the Optogan company improves the base, removes expensive materials (polycarbonate), reduces the size, replaces 1 powerful chip with several simpler ones and optimizes the driver (in short, you understand - completely redesigns the lamp), then such a light bulb will have everything chances to conquer the Russian market, since in addition to the indicated disadvantages, there are also a lot of advantages such as proper connection of diodes in the module, smart “driver”, etc. Thanks to the technical documentation.

As for “Svetlana”, apart from the simplest driver, which should influence the price downwards, the location of the light-emitting modules on the substrate, there are practically no advantages. Technical documentation cloudy, the LEDs are connected in series, which if 1 diode “burns out” disables the entire module (i.e., in our case, it reduces the luminous flux by 12.5%), thermal paste smeared everywhere - all this does not add confidence. But this was just a prototype, maybe industrial designs will be better.

This article is not intended to denigrate or, conversely, exalt the products of some manufacturers over others. I present only the facts, and let you draw the conclusion! As they say, think for yourself, decide for yourself...

Video section

Thank you very much OSRAM for preparing such detailed video about how it produces LEDs (although this company makes LEDs using a slightly different technology than all the light bulbs we have studied):

If there are enthusiasts ready to help with writing Russian subtitles, I will gladly accept help

The process of transferring LED chips inside the plastic case:

And so in Taiwan they “package” led chips in plastic modules with dye application and packaging in reels:

P.S. It will start on Wednesday (26.10), and the Optogan company will be widely represented. I hope that my microphone will not be turned off at the press conference and I will be able to ask uncomfortable questions... The main thing is to get out alive later...
P.P.S. In light of the latest personal problems I'm not sure I'll find the strength to finish the work I started. Namely, to get even with flash memory and displays (E-Ink and LCD). There were also plans to write a publication on biological objects, but apparently they will have to be shelved...

THANK YOU! Everyone for reading and commenting...