Interpretation of GE, Philips, Osram LED Bulb Test Data Analysis

From left to right, from top to bottom are:

Self-made test lamp holder 1, oscilloscope , precision thermometer 2, digital thermometer, illuminance meter, power meter, tape measure, homemade photoresistor test circuit and power supply battery, self-made test lamp holder 2

4. Error factors existing during the test

There are many error factors in the test process, the most important error is the nonlinearity of the photosensitive device. Such as the frequency response curve of a digital illuminometer

The non-linear characteristics of this tool lead to different sensitivity to different spectra. The spectral characteristics of each LED bulb are difficult to test under amateur conditions, and the manufacturer, except OSRAM, does not provide the spectrum, so the error caused by the spectral characteristics Not only can't be measured, it's even difficult to evaluate. Based on this situation, almost all of the following tests can be said to be untrustworthy.

But I still do all the tests, the main reason is that except for GE 3W (94101), all other bulbs to be tested have a color temperature of 2700K-3300K. The color temperature is close, and the spectral characteristics and sensitivity of the photosensitive device are not too large. Moreover, this test is mainly based on horizontal comparison, and does not insist on accurate quantification, so it still has some reference value.

Other errors, such as instrument accuracy, manual operation uncertainty, ambient temperature, ambient brightness, etc., are not too large, and should be paid attention to during the test process or data processing.

5, power consumption test

Power consumption testing is a more accurate test throughout the process.

The significant difference between LED lamps and incandescent lamps is that LED lamps have drive circuits. The power consumption of the entire LED luminaire consists of the insertion loss of the circuit plus the power consumption of the LED itself.

Although various LED driver circuits are currently quite efficient, it is obvious that the efficiency of any circuit is not 100%. The few watts we buy are rated power, not the power consumption of the LED itself. A little bit different from common sense. Although the circuit efficiency cannot be obtained without disassembling, even if the circuit efficiency is assumed to be 100%, it can be seen from the comparison of the rated power and the measured power that which products have virtual standard power.

The blue part of the above figure is the rated power and the red is the measured power. From the figure we can see that the products of each family, the rated power and the actual power are not much different. It should be noted that the part of the product whose measured power is less than the rated power, there must be a phenomenon of power virtual standard, but most of the virtual marks are not much.

While measuring power consumption, the power factor was also measured. Although the state does not require the power factor of residential electrical equipment, it will not affect the actual electricity cost, but the low power factor is always a burden for the power grid, so we should try to choose a power factor when possible. The product is better.

From the point of view of power factor, it is generally relatively large in power and relatively high in power factor. This may be related to its circuit design: low power consumption estimates direct capacitance buck, high power may be driven by a dedicated chip - this is only speculation.

Here I like GE 3W (94101), which is the only product that identifies the power factor on the package. As you can see, it has been mentioned many times from top to bottom. This is definitely a wonderful product, not only different from other manufacturers' products, but also unique with GE's own products. There will be a lot of performances below.

6, electro-optical conversion efficiency test

This parameter cannot be accurately measured. We can only obtain the photoelectric conversion efficiency of each bulb by means of the twist back comparison method. The conversion efficiency of the luminaire should be lumens/watt, but the measurement of the luminous flux requires the use of an integrating sphere. Without this condition, without this precision, the luminous intensity is simply used violently instead of the luminous flux for lateral comparison. The specific method is as follows:

The power consumption of the bulb is measured, the distance at which the bulb produces a certain illuminance is measured, the luminous intensity is determined, and the contribution of the power consumption per watt to the light intensity is obtained.

In the actual test, the power consumption of each bulb is obtained by using a power meter; a point of 40 Lux illumination is obtained for each bulb by a illuminance meter, and the distance D between this point and the center point of the bulb is measured. The formula (40*D^2)/P is the contribution of power consumption per watt to the light intensity, which is the photoelectric conversion efficiency we want. This efficiency measurement error is very large, mainly affected by the size of the bulb and the angle of illumination.

But for our users, illuminance is an indicator of the actual use of the luminaire, so this test is not accurate but very practical. The test results are as follows

From this chart we can see that incandescent lamps are significantly less efficient than LED lamps. Between the LED lights, the higher the power, the higher the efficiency. This is easy to understand: the lower the overall power consumption, the greater the energy consumed by the LED driver circuit as a percentage of the overall power consumption, and the lower the efficiency. After the overall power consumption reaches a certain level, the efficiency is reduced. The personal speculation is that the internal space of the bulb is too small, the higher the power, the more serious the heat, and the temperature will seriously affect the circuit efficiency.

Here we can once again see the wonderful GE 3W (94101), which is far more efficient than the same brand of GE 3W (11453). Mainly because GE 3W (94101) is 6500K cool white light, cold white LED itself is more efficient than warm white light; of course there may be other unknown reasons, such as the sensitivity of the illuminometer.

7, temperature test

The temperature test is also a relatively accurate test. This test focuses on two properties: temperature control and heat dissipation efficiency. The temperature control test is very simple, let the bulb light for a period of time, and measure the maximum temperature. The heat dissipation efficiency is not so easy to measure. The method I think of is: look at the time when the temperature of the casing reaches a steady state after lighting. The shorter the time, the lower the thermal resistance between the internal circuit and the outer casing, and the better the heat dissipation performance.

Just like the end of a fired iron stick, the other end will soon feel hot; while the one that burns the stick with fire, it is more difficult to burn the other end. This method is only a qualitative comparison of the heat dissipation efficiency, which is actually caused by the error factors such as the emission coefficient of the shell material, the ambient temperature, and the air flow.

The test method is as follows: firstly select two T-type thermocouples with good consistency by using a precision thermometer; after correcting the deviation; simultaneously collect the ambient temperature and the surface temperature of the lamp with a digital thermometer, and collect at least 10000 points of data for each. The obtained data is processed to obtain the required conclusions.

In actual testing, all fixtures are in an open space (no wind room environment) with the fixture screw facing down the ball. The ambient temperature is difficult to control (the time span of the test is very large, and the test time of each luminaire is about two hours). Therefore, some of the ambient temperature in this test is high, which is unfair to some luminaires. for reference. The overall ambient temperature is between 28 and 29 degrees. The incandescent lamp is tested late and the ambient temperature is around 27 degrees. A K-type thermocouple is used.

It can be seen that the operating temperature of the incandescent lamp is much higher than the operating temperature of the LED lamp. This is good in the winter, which will cause the temperature in the room to rise in the summer and the cost of air conditioning to cool down.

Due to the excessive temperature of the incandescent lamp, the temperature rise curve of the LED lamp part is too dense, and a temperature rise curve showing only the LED lamp is provided below.

The legend on the right side of this picture is arranged from highest to lowest according to the highest temperature. This is the temperature rise diagram, let's take a look at the relationship between the highest ambient temperature, the maximum temperature of the lamp body, and the temperature difference (ignoring incandescent lamps)

This picture purple is the highest temperature in the environment. It can be seen that the difference between the tests is not too big; yellow is the highest temperature on the surface of the lamp. It can be seen that the surface temperature of most high-power lamps is nearly 80 degrees; red is the temperature rise. It is obvious that the higher the power, the greater the temperature rise. However, Philips3W is an exception, actually higher than the temperature of Philips 7.5W.

Temperature is the biggest killer of LED life, the temperature outside the casing is so high, and the internal temperature is definitely higher. Is the heat dissipation particularly good, which will cause the temperature outside the casing to be extremely high? The heat dissipation is done well, and the external temperature is not higher than the internal temperature. However, if the heat dissipation efficiency is low, the internal temperature will be much higher than the external temperature.

Let's take a look at the cooling efficiency. This test is to find the maximum temperature of each bulb from the previous test data, and then find the time it takes for the bulb to reach 95% of the maximum temperature for the first time. Why is it 95% of the maximum temperature, not directly the highest temperature? One 95% is very close to the maximum temperature; the second is that after the bulb reaches a certain temperature, its surface temperature will fluctuate up and down, and the maximum temperature will not be accurate.

This time should be as short as possible. There are two points to note here: one is the settling time of the incandescent lamp, and the other is the settling time of the high power LED.

We can see that in all luminaires, the time spent on incandescent lamps is the shortest. Is this the best heat dissipation for incandescent lamps? No. But the incandescent lamp heat dissipation mainly depends on heat radiation, and the heat conduction of LED lamps depends on the essential difference between them, so it cannot be compared.

High-power LED lights have a long time consuming due to their high power consumption and temperature rise.

Written here, I suddenly found this test to be flawed: the time to reach a stable temperature inside the fixture will affect the external stable temperature. What is more meaningful should be: the same rated power of the luminaire, its time to reach a stable temperature.

Pinyuan 3W looks like a good heat sink; GE 2W looks very bad, it may be related to its ceramic shell; Philips 3W seems to be able to dissipate heat, but its maximum temperature is too high. I don't understand why this is the case. Checking the internal circuit is just a simple capacitor current limit. It is said that the circuit efficiency is high. If the constant current performance is not good, the LED overload is not working. After all, the power test is actually measured. The power has not yet reached the nominal power. I found a disassembly diagram on the Internet and found that there is an LED3 not soldered. Is it really super power? -- This is speculation!

8, stroboscopic control

Early ballast fluorescent lamps were criticized for their severe stroboscopic light. Later, electronic fluorescent lamps were used and stroboscopic improvements were made. There is also an eye-protection lamp on the market, there is no stroboscopic at all. Although the stroboscopic sound is not visible to the naked eye, it is easy to cause eye fatigue. There have been many studies to prove this, and the IEEE also has risk warnings.

LED lights originally had a special drive circuit, which should be easy to achieve without stroboscopic. Mains-powered products, reduce or eliminate strobos mainly rely on capacitors or inductors, the inductor volume is too large, ordinary electrolytic capacitors are prone to problems at high temperatures inside the bulbs, tantalum capacitors are expensive, so how do I control strobes on LED lamps? Very uneasy. In fact, this test, the protagonist is stroboscopic control.

To measure the strobe of a luminaire, only photodiodes can be used in amateur conditions. I chose the PT333-3C, and its typical rise and fall time is 15uS, which is several orders of magnitude worse than the 10mS of the power frequency. It can be ignored. What affects the accuracy of the test is its frequency response curve.

In the figure, we can see that it is different in sensitivity to incident light of different wavelengths, which leads to different light sources, and the measured stroboscopic is different from the actual stroboscopic. As mentioned earlier, qualitative testing is not precise – it can only be so.

The circuit is very simple, just pay attention to battery power. As shown below

At the 40Lux illuminance distance of each luminaire, the photosensitive tube D1 is aligned with the center of the luminaire, and the oscilloscope measures the waveform at both ends of D1.

I am quite speechless when I see this picture: the same brand, the same series, different power LED lamps, the stroboscopic difference can be so big. For example, OSRAM 10w and OSRAM 9W, one is only 120mV, one is up to 1320mV, the power is only 1w, the strobe amplitude is 11 times! The exaggeration between GE 3W (11452) and GE 3W (94101) is 14.5 times!

With regard to strobes, one thing is often overlooked: large chandeliers typically use a large number of bulbs of the same type, which have the same phase, and the stroboscopic amplitudes are superimposed, resulting in a more severe strobe than a single bulb.

What does the same phase mean? The waveform of a sine wave is known. The so-called same phase means that when a certain bulb is at the peak, the other bulbs are also at the peak; when one bulb is out of the trough, the other bulbs are also in the trough, and the same phase. Since there is a problem with this stroboscopic stacking, then measure the phase of each bulb.

The phase test uses two upper photosensitive circuit diagrams, one of which is aligned with the incandescent lamp, one is aligned with the light source to be tested, and the oscilloscope simultaneously acquires two outputs to obtain the phase referenced to the incandescent lamp.

The phase test results are as follows

Since the frequency of the power frequency after full-wave rectification is 100 Hz and the period is 10 mS, 0 mS and 10 mS, 1 mS and 9 mS in the above figure, and so on, are equivalent phases.

It should be noted on this figure that GE 2W and GE 3W (11452) are not missing data. The GE 2W has changed its output mode via the drive circuit, independent of the power frequency; the strobe output of the GE 3W (11452) is almost a straight line, and its phase is negligible. The phase diagram of GE 2W is as follows

The yellow curve in the figure is an incandescent lamp and the blue curve is GE 2W.

In order to verify the overlapping effect of multiple identical bulb stroboscopic amplitudes on large luminaires, for example, two luminaires with the same bulb are installed. The stroboscopic output is measured at 2 meters from two Philips 3W LED lamps.

In the same position, when the two lights are turned on at the same time, the stroboscopic output is measured as follows

It can be seen that the peak-to-peak value of the strobe has risen from 480mV, 460mV to 900mV, and its stroboscopic amplitude is superimposed.

Conversely, two different phase lamps, if their phase difference is 5mS or close to 5mS, their stroboscopic amplitudes will cancel each other out. For example, Philips3W LED lights and incandescent lights, the phase difference is close to 5mS, let us look at the effect of their two stroboscopic output superimposed

The left picture above shows the stroboscopic light measured 1.5 meters from the Philips3W LED light; the right picture shows the 25W incandescent strobe measured at the same position. After both are turned on, the strobe measured at the same position is as shown below.

It can be clearly seen that the stroboscopic amplitude is from 760mV and 680mV, and after superposition, it is reduced to 320mV, and the waveform after superposition is also different. The reason why it is not completely offset is because the two lamps are not only different in phase difference, but the other parameters are not completely identical.

These two tests give us good advice on how to assemble large luminaires: large luminaires should avoid using the same type of LED bulbs as much as possible - in fact, incandescent and energy-saving lamps are equally applicable, just because of incandescent lamps. There is no drive circuit, and how to change the model, the phase is the same.

9, appearance, workmanship, price

The appearance is similar, 2W, 3W are mainly small lamps, and others are mainly headlights. More special is GE 2W, the shell is ceramic (?), I do not know if this affects its heat dissipation performance.

Pinyuan 3W has designed many small holes in the lamp body. Although these small holes help to dissipate heat, it is estimated that it is easy to get into the ash, and it may be more unfavorable for heat dissipation over time. In addition, there is no word/logo on this lamp, and the cost saving is a little overdone.

Let's talk about work. What I have to say about workmanship is OSRAM. First, Taobao bought three LED bulbs of his home, which are 4W 6W and 10W respectively. As a result, there are obvious flaws in the three 4W and 6W.

Look at 4W first, this side looks like the font printing is obviously from light to deep to light

On the other hand, one has a very light logo and the other has white glue on the screw.

Look at the 6W bulb, there is a clear glue overflow between the lampshade and the lamp body - the picture is not very good, the object is very obvious.

Only 10W quality seems to be a success, but the ratio with Philips and GE is still a little bit different. I don't think I bought a fake on Taobao, but I went to B&Q and bought a 9W. The result. . . . . . .See below

This 9W, actually left a metal ball at the junction of the screw and the cup. If all of the above are just affecting the perception, then this is a safety flaw: Will it cause an accidental electric shock?

At this point, if the character of the non-deaf is the worst, there is a problem with the quality control of the OSRAM. On Weibo, I sent a photo of the problem product @Osram China Lighting. After a few days, they contacted me and said that I could change the quality of the lamp without any problem. I also gave Taobao sellers and more detailed pictures to the past to identify the true and false. I haven't responded to the past ten days. I guess Taobao sellers are fine. Things are also real. That is, quality control is really a problem.

Even more strange, there is no product information for 4W, 6W, 9W bulbs on the OSRAM homepage.

Speaking of the homepage, GE's homepage is even more strange. The Chinese homepage can't find any information about the lamps I bought. In the North American version, there is some information. The 11W products are actually discontinued and the inventory is firesale. . . . . . .

In terms of price, the product garden 3W is less than 5 yuan, very close to the people; the other is a little expensive, as shown below

Among them, OSRAM 9W was purchased from B&Q, so the price is more expensive than other Taobao. Basically, the higher the power, the more expensive the price. The transformation of a bunch of light bulbs at home is also a huge expense.

10, summary

From such a bunch of test results, there is no one that meets all expectations - of course, this is the norm.

In terms of the products being tested, GE's products, except GE 3W (94101), are not bad, but the price is high; Philips's products are relatively modest in price and performance; OSRAM is difficult to comment on, and the difference is very good. And the quality control seems to have problems, buy it to spell the character? Domestically, there are not many measurements, there are prices to win, and there is no cost-effective.

It can only be said that as a consumer, it is too difficult: I know a little about hardware, there are so many tools, I spend a lot of time and energy, and I can only get a limited amount of reference information, not to mention the average consumer. How to choose if there is not enough information.