Saga of LED lamps. Part 2 - about what is not written on the boxes

Last time, we briefly recalled the history of artificial lighting, and also talked a little about what the main parameters are for energy-saving lamps in general and LED lamps in particular. Today, as promised, we will move on to measurements and comparisons (however, so far without unwinding).

Is it worth it?

First of all, I was worried about the obvious question - nevertheless, is it true that ordinary LED lamps, which can be bought in a store, are really fabulous in real conditions? To answer it, I decided to measure the illumination created in my room by different bulbs, screwed into the same (my) chandelier. Initially, there were three twenty- ovative ELL CFLs ; for comparison, I took three 12W Gauss LED lamps (claimed to be an analogue of a 100W incandescent lamp) and, for the purity of the experiment, three ordinary incandescent lamps of 95W each. The measurements were carried out in the center of the room, that is exactly where the brightness of the lighting is most interesting and necessary for me. I will say right away - from the point of view of photometry, this is probably not entirely correct; but from the point of view of ordinary life, such a comparison, it seems to me, is of primary interest, since it reflects the behavior of a light bulb not in the integrating sphere, but in the ordinary chandelier.

The measurements were carried out with a luxometer Mastech MS6610 . I ruled out outside illumination with thick curtains (with the lights turned off, the device showed zero lux). Since the luminous flux of fluorescent and LED lamps depends on their temperature, the luminance values ​​were taken twice - immediately after switching on and after a ten-minute warm-up (it was empirically found that after ten minutes of operation the illumination changes very little). Of course, incandescent bulbs do not need to warm up, so for them the measurement was carried out only once, immediately after switching on, so as not to spoil the chandelier, designed, if memory serves me, a maximum of 40 watts (for an incandescent lamp) in each horn. The results of this experiment can be observed in the table below.
')

Lamp type	condition	Measured illumination, lux
ELL CFL 20 W 2700 K	cold start	thirty
	warmed up	131
Incandescent lamp 95 W	-	101
Gauss LED 2700 K 12 W	cold start	146
	warmed up	137

Well, it is clear that this test LED lamps (at least those that I had) really surpass everything that can now be screwed into a regular E27 cartridge (with the exception, perhaps, of some exotic things). With incandescent bulbs everything is clear - I already guessed that the result would not be very impressive. It is more interesting to compare LED lamps and still popular CFLs.

Immediately striking that in the first ten minutes of the CFL brightness change almost five times. In practice, this means that for the domestic scenario “I went into the room (storage room) for two minutes to find something” they are the worst of all - by the time they reach the operating mode, they will most likely be turned off. This is in addition to the fact that gas-discharge lamps and so poorly tolerate frequent inclusions, although, suppose, in the closet, they may not be so frequent, but, nevertheless, short. LED lamps, on the contrary, slightly reduce the brightness as warming up - the voltage drop, and, consequently, the power (at constant current) on the heated LED is less. However, the difference in brightness here is not as stunning as in the case of CFLs (which indirectly indicates a fairly good heat sink specifically in these lamps). By the way, it is clear that after warming up, the difference is still in favor of the LEDs, although its size is such that the illumination created by those and others can be considered approximately equal. However, we are talking about approximately equal illumination produced by a twenty-watt CFL and a twelve-watt LED-lamp - almost twice the power savings. You can not even talk about incandescent lamps - with many times more power consumption by the created illumination, they lose both CFLs and LEDs. In addition, as I mentioned above, ninety-five-watt lamps cannot be screwed into my chandelier at all, so in reality with incandescent lamps I would not even get one hundred lux. Of course, such a limitation is associated with heating.

Incandescent bulbs, obviously, have gone the distance, so let's compare the CFL and the LED lamp for heating.



These images were also taken after ten minutes of warm-up. It can be seen that the CFL heats up to one hundred degrees or more, while the maximum temperature of the LED lamp is only about sixty. That is, the opportunity to get burned about CFLs, in principle, exists (the protein begins to collapse at eighty degrees Celsius), while with an LED lamp it is impossible in principle. Trifle, but nice.

More measurements

So, we figured out that in terms of those characteristics that come to mind first, the LEDs are clearly better. Time to talk about more subtle matters, such as power factor and ripple factor. For some reason, they rarely remember these hack characters, and, of course, they (for the time being?) Are never written on their packages, but in vain.

The ripple factor is a very important indicator. Despite the fact that changes in brightness with a frequency of more than 16 - 20 Hz, our brain consciously does not process, the effect of them is quite noticeable. Significant pulsations of general illumination can lead to increased fatigue, migraines, depressions and other unpleasant things in the psyche. This indicator is normalized in SNiP 23-05-95 . There are a lot of different tables there, but, in general, one can take out of them that the ripple factor of the general illumination should not exceed 20%. It is worth making a reservation that talking about all this makes sense up to a frequency of about 300 Hz, because the retina itself no longer has time to react to changes in light, and therefore in this case an irritating signal simply does not come to the brain.

The power factor for the end user, in principle, is unimportant. This parameter shows the ratio of the active power consumed by the device to the total power that takes into account the reactive part, which does not produce useful work, but, in particular, the heating wire. The name “cosine phi” is also common - this is all because the quantity of interest to us can be entered as the cosine of some conventional angle. The maximum, ideal value of the power factor is 1. Household meters take into account only active power, they write it on packages; for the consumer in this sense, there is no problem. However, if we are talking about global scale (for example, a millionth city, entirely illuminated by LED lamps), a low power factor can create big problems for power engineers. Therefore, its rating is the rating of the lamp in the sense of a bright LED future.

I measured the power and power factor with the muRata ACM20-2-AC1-RC head. The ripple factor was measured with an Uni-Trend UTD2052CL oscilloscope , to which the following circuit was connected:



Who cares, this is a classic frequency-compensated converter "current-voltage" on the operational amplifier, supplemented with an artificial midpoint. It feeds, to eliminate interference, from the battery. The BPW21R diode is a photometric class device with a characteristic compensated according to the sensitivity of the human eye. The documentation guarantees the linearity of the current depending on the illumination in the photovoltaic mode, so that the circuit produces a voltage directly proportional to the illuminance of the photodiode and is quite suitable for measuring the ripple factor. It is determined, by the way, as the ratio of the range of pulsations to twice the average value. Both the scope and average value are included in the standard automatic measurements of any modern digital oscilloscope, so there is no problem with this - it only remains to double and divide. Comparisons of the measurement results of this improvised construction with the values ​​given by the TKA-PULSE device (State Register) showed that the measured pulsation coefficient did not exceed one percent.

So, the measurement results for the lamps that were at my fingertips:

With E27 socket:

Lamp type	Measured power, W (cold start)	cos (φ)	K p	Generally
ASD 11W	9	0.82	one%	Very good
Gauss 12 W	12	0.62	one%	Good
Gauss 6.5 W	6	0.50	one%	Acceptable
SUPRA 11 W	9	0.95	35%	poorly
ASD 7 W	four	0.45	100%	Disgusting

With E14 socket:

Lamp type	Measured power, W (cold start)	cos (φ)	K p	Generally
Gauss 3W	2	0.60	one%	Good
Gauss 6.5W	6	0.95	49%	Very bad
Wolta 5W	2.2	0.40	68%	Disgusting

As we see, here at LED lamps not everything is so rosy. Immediately it turns out a very interesting fact - it seems that the quality of LED lamps can not be judged only by the manufacturer; the same brands, generally speaking, set both quality records and anti-records. It should be noted that I passed the general verdict presented in the table, attaching greater importance to the ripple coefficient than to the power factor, for the reasons stated above. But even the ripple factor of 1% cannot fully justify the power factor of 0.5 in the case of an industrial product sold in millions of copies. However, for the home it is better to take such a lamp than a product with a unit power factor and a ripple level of 50%.

Of course, lamps with a ripple factor of more than 20% are categorically not suitable for general lighting (there is no need to twist it into a six-piece chandelier). By the way, for the “Era” CFLs mentioned by me, it is slightly less than 10%, and for the classic incandescent lamp - about 13%.

The last parameters that can be casually discussed are the color temperature and the color rendering index. Despite the fact that they are formalized, at the household level it all comes down to “like / dislike”. I must say that all the lamps tested in this regard pleased me - none had a clear bias or excessive yellowness, all had a pleasant white tint. But this, of course, for my taste, and nothing more.

In the following articles, we will finally see what's inside the lamps, and try to figure out which internal causes make good lamps good and bad ones bad.

Note:

The choice of lamps for tests is due solely to the consideration of "what was." If (when) there will be other lamps - I will measure and lay out.