The saga of LED lamps. Part 4
I want to say right away that I did not forget about the survey . His results are taken into account, the lamps are ordered and go to me. Apparently, they are coming from afar, but nonetheless. So this stuff will be. In the meantime, so that no one (including myself) gets bored, I suggest looking at two more bulbs from my bins.
The first will be the Philips 5W. Specifically, a friend of mine was very interested in this lamp, and recently, being in a store, I saw it and bought it for the sake of interest. No code or article was found either on it or on the package, so I’ll give a photo of the box:

The measured ripple coefficient was 13%, the power factor was 0.64. Not bad. Let's see how this is achieved.

The plastic diffuser is easy to remove.

Below it we see a matrix of eight half-watt diodes, an aluminum part bolted to a millimeter thickness.

I must say that getting to the driver turned out to be incredibly difficult, since this same aluminum part was pressed into an aluminum glass in the manner of a tin can (and welded?). Yes, in fact the case is also almost entirely aluminum, plastic is only the top layer. I had to use an abrasive disc and scissors for metal.


In general, the design is very strong, strong, and, I would say, vandal resistant. Thick aluminum probably dissipates heat well.
The driver was finally taken out:


It looks very complicated. I figured out how it works, and for a long time I could not believe my eyes. Which of the topologies discussed above corresponds to this driver?
The correct answer is no. A very unusual solution for LED lamps is used here - in fact, it is a ballast for compact fluorescent lamps, adapted for LEDs. Apparently, this was done to reduce the cost of reconfiguring the production line and changing the procurement scheme of components - they used what is already bought by millions.
The solution is very witty from the point of view of the compromise "production optimization / parameters", but still in the sense of LED mode and efficiency it is not the best. Well, from the point of view of the philosophy of circuitry, it is somehow strange.
To keep the article from being too short, let's look at something else. In general, from a more or less similar I still had only Gauss 3W. Judging by the inscriptions on the package, this lamp is also positioned as the equivalent of a 40 W incandescent lamp. The item on the box is HA105201103.

Formally, putting them side by side, in general, is not entirely correct - this three-watt Gauss has an E14 base and is generally declared as decorative. Nevertheless, if you follow the link written on the box, you can read that the options for this lamp for the E27 base also seem to be available .
The measured ripple factor is about 1%, the power factor is about 0.6.
To disassemble the lamp, I barbarously poked the base.


As it turned out, doing this was completely optional. See the thread on the plastic part? She's getting out. That is, the lamp can not only be disassembled, but also assembled painlessly. She is maintainable!
There are no surprises regarding the driver. Inside there is a flyback converter based on MT7952 (just the case when current control is carried out on the primary side).


The inclusion seems to be typical.

An electrolytic capacitor of the brand BERYL is visible on the board . Not Rubycon, but not bad anyway . Failure of such capacitors, as claimed, - about 10,000 hours at the temperatures of the LED lamp. What pleases, it is located in the colder part, closer to the base. In general, electrolytic capacitors are one of the determining factors in the durability of LED lamps, so if you believe the numbers given in the link above, this lamp has every chance of working happily ever after. The only thing that can cause concern is the heat sink from the LEDs. Nevertheless, judging by the heating of the case, it is also more or less in order with it (the "stamens" on which the LEDs are sealed, aluminum, this is MCPCB ).
In general, today we saw two worthy lamps. Philips, however, was amazed by the driver’s solution, but it has good parameters (ripple of 13%, cos (φ) - about 0.6), and the quality of the design itself is excellent, as it should be for a brand of this level.
The Gauss lamp is just a good lamp (ripple of 1%, cos (φ) is also about 0.6). Standard circuitry, moderately high-quality components. But for my taste, its main feature is maintainability. This is in my memory the first lamp that can not only be disassembled, but also assembled. For example, it is possible to re-solder the capacitor as part of the DIY, when it nevertheless dries.
The first will be the Philips 5W. Specifically, a friend of mine was very interested in this lamp, and recently, being in a store, I saw it and bought it for the sake of interest. No code or article was found either on it or on the package, so I’ll give a photo of the box:

The measured ripple coefficient was 13%, the power factor was 0.64. Not bad. Let's see how this is achieved.

The plastic diffuser is easy to remove.

Below it we see a matrix of eight half-watt diodes, an aluminum part bolted to a millimeter thickness.

I must say that getting to the driver turned out to be incredibly difficult, since this same aluminum part was pressed into an aluminum glass in the manner of a tin can (and welded?). Yes, in fact the case is also almost entirely aluminum, plastic is only the top layer. I had to use an abrasive disc and scissors for metal.


To the question of why, when using an abrasive disk, protective glasses should be worn

In general, the design is very strong, strong, and, I would say, vandal resistant. Thick aluminum probably dissipates heat well.
The driver was finally taken out:


It looks very complicated. I figured out how it works, and for a long time I could not believe my eyes. Which of the topologies discussed above corresponds to this driver?
The correct answer is no. A very unusual solution for LED lamps is used here - in fact, it is a ballast for compact fluorescent lamps, adapted for LEDs. Apparently, this was done to reduce the cost of reconfiguring the production line and changing the procurement scheme of components - they used what is already bought by millions.
Technical details for those interested
According to the power requirements, fluorescent lamps are quite similar to LEDs - they also require current stabilization. Canonically, an inductor is used as the ballast for them. However, in a previous article, we already wondered that in the general case, the inductance of the inductor for the range of currents, frequencies, and voltages of interest to us turns out to be indecently large. However, it, in principle, can be reduced due to the fact that the reactance of the coil increases with increasing operating frequency. Thus, if we convert the network 50 Hz into several tens of kilohertz, it will be possible to dispense with the inductance in the region of several milligenes. This is exactly what they do in compact fluorescent lamps.
The traditional driver scheme for a compact fluorescent lamp, with some variations encountered in 90% of cases, consists of a rectifier, to which a converter is connected, which is an auto-generating half-bridge. A gas discharge tube is connected to this design via a throttle, and everything works happily.
The driver of the lamp in question is made exactly the same, only instead of the gas discharge tube there is a block with LEDs (indicated by the number 1).

I didn’t draw the whole scheme, for illustration, I just took the basis from the reference design from Philips . Judging by the details present on the board, the coincidence in this part, if not 100%, is at least very large.
Here is the drawn part of the diagram corresponding to block “1”:

On the thyristor D6, a protection unit is assembled in case of a break in the LED chain (so that C2 does not explode). One could simply put C2 at a higher voltage (400 V), but this is, firstly, more expensive, and, secondly, a capacitor of the same capacitance for the desired voltage has much larger dimensions.
The traditional driver scheme for a compact fluorescent lamp, with some variations encountered in 90% of cases, consists of a rectifier, to which a converter is connected, which is an auto-generating half-bridge. A gas discharge tube is connected to this design via a throttle, and everything works happily.
The driver of the lamp in question is made exactly the same, only instead of the gas discharge tube there is a block with LEDs (indicated by the number 1).

I didn’t draw the whole scheme, for illustration, I just took the basis from the reference design from Philips . Judging by the details present on the board, the coincidence in this part, if not 100%, is at least very large.
Here is the drawn part of the diagram corresponding to block “1”:
On the thyristor D6, a protection unit is assembled in case of a break in the LED chain (so that C2 does not explode). One could simply put C2 at a higher voltage (400 V), but this is, firstly, more expensive, and, secondly, a capacitor of the same capacitance for the desired voltage has much larger dimensions.
The solution is very witty from the point of view of the compromise "production optimization / parameters", but still in the sense of LED mode and efficiency it is not the best. Well, from the point of view of the philosophy of circuitry, it is somehow strange.
To keep the article from being too short, let's look at something else. In general, from a more or less similar I still had only Gauss 3W. Judging by the inscriptions on the package, this lamp is also positioned as the equivalent of a 40 W incandescent lamp. The item on the box is HA105201103.

Formally, putting them side by side, in general, is not entirely correct - this three-watt Gauss has an E14 base and is generally declared as decorative. Nevertheless, if you follow the link written on the box, you can read that the options for this lamp for the E27 base also seem to be available .
Speaking of decorative
Since there are point sources of light in this lamp, being screwed into the ceiling with a pattern, it, unlike matte ones, will give curly shadows loved by many, such as these:

(photo courtesy of my friend).
Who knows, maybe it was declared as decorative just for that reason - its form is completely traditional, and it seems that it does not seem to have any other bias in decorativeness.

(photo courtesy of my friend).
Who knows, maybe it was declared as decorative just for that reason - its form is completely traditional, and it seems that it does not seem to have any other bias in decorativeness.
The measured ripple factor is about 1%, the power factor is about 0.6.
To disassemble the lamp, I barbarously poked the base.


As it turned out, doing this was completely optional. See the thread on the plastic part? She's getting out. That is, the lamp can not only be disassembled, but also assembled painlessly. She is maintainable!
There are no surprises regarding the driver. Inside there is a flyback converter based on MT7952 (just the case when current control is carried out on the primary side).


The inclusion seems to be typical.

An electrolytic capacitor of the brand BERYL is visible on the board . Not Rubycon, but not bad anyway . Failure of such capacitors, as claimed, - about 10,000 hours at the temperatures of the LED lamp. What pleases, it is located in the colder part, closer to the base. In general, electrolytic capacitors are one of the determining factors in the durability of LED lamps, so if you believe the numbers given in the link above, this lamp has every chance of working happily ever after. The only thing that can cause concern is the heat sink from the LEDs. Nevertheless, judging by the heating of the case, it is also more or less in order with it (the "stamens" on which the LEDs are sealed, aluminum, this is MCPCB ).
In general, today we saw two worthy lamps. Philips, however, was amazed by the driver’s solution, but it has good parameters (ripple of 13%, cos (φ) - about 0.6), and the quality of the design itself is excellent, as it should be for a brand of this level.
The Gauss lamp is just a good lamp (ripple of 1%, cos (φ) is also about 0.6). Standard circuitry, moderately high-quality components. But for my taste, its main feature is maintainability. This is in my memory the first lamp that can not only be disassembled, but also assembled. For example, it is possible to re-solder the capacitor as part of the DIY, when it nevertheless dries.