New on lamptest.ru
For three years I have been testing LED lamps and a half year running site lamptest.ru . At the end of January, I made a small act of heroism and in a week I tested almost 250 light bulbs, so now the site contains data on 1244 tested lamps.
And even at lamptest there were quite significant changes.
First of all, the display format of the results table has changed.
Now there are two modes of displaying the results table - regular and full (enabled by the “full mode” checkbox in the parameter selection panel). In normal mode, only the most basic lamp parameters are displayed, and all the declared parameters are collected in one cell of the “description” table. It seems to me that this is easier and clearer.
In full mode, all the available parameters are displayed, including those that duplicate the parameters in the "description" cell (this is done so that you can sort the table by them). Even those parameters that previously were only in the lamp card (barcode, date of manufacture) appeared. In full mode, the table is very wide, but this is not the main mode, and those who wanted to look at all the parameters at once I think will put up with it.
The columns in the table are re-sorted so that the declared and measured parameters in full mode are always nearby.
As before, by default, the table is sorted in the order in which the lamps were tested and on the last page you can always see the lamps that were tested last.
New parameters were added to the site:
No - the lamp number according to the list (this number did not appear on the site
before , but I decided to display it first of all for myself);
CQS - Color Quality Scale - an alternative color rendering index, more suitable for LED lamps than CRI and allowing to detect lamps with an unpleasant glow color. So far, the CQS is not filled, but there is data on it and we will gradually fill this parameter for all tested lamps;
Umin- the measured minimum voltage at which the lamp operates stably and gives at least 90% of the light. Many manufacturers write the operating voltage range for lamps 220-240 V, while lamps with a high-quality driver can operate at much lower voltages. At the same time, lamps with a linear driver reduce the brightness when the voltage drops and in some lamps the minimum voltage at which they give 90% of the light is even higher than 220 volts. By this parameter it is possible to judge which driver is used in the lamp;
drv- driver type. By the behavior of the lamp when the voltage drops, you can judge the type of driver used. Now for the lamp one of the four types of driver can be specified: 1 - the brightness does not change in a wide voltage range, but below a certain level it begins to decrease sharply (as a rule, between the beginning of the decrease in brightness and its decrease by 10%, the voltage decreases by only 10 volts) , 2 - the brightness does not change over a wide voltage range, when a certain level is reached the lamp turns off; 3 - the brightness does not change in a wide voltage range, when the voltage drops to a certain level, the lamp starts flashing; 4 - linear driver, the brightness decreases linearly with decreasing voltage. I will definitely write an article, what drivers are used in LED lamps.
tmax- The maximum temperature of the lamp housing when used in an open ceiling and a vertical installation of the base down;
dim - the minimum level of dimming. Earlier in the "dim" appeared the letter "D", if the lamp supports brightness control. Now, if the lamp is not dimmable, this field is empty, and if dimmable there indicates the minimum level of dimming in percent. So far, many lamps have a unit there, the remainder of the old display, but soon all the data will be entered.
Weight - now the lamps are weighed, because by weight you can indirectly determine how full-fledged cooling system is used in the lamp, as well as determine when manufacturers begin to produce other lamps under the same part numbers and bar codes.
In connection with the appearance of new parameters, the format of the data file lamptest.ru/led.csv has been changed - new fields have been added to it.
The quality of LED lamps can not be judged by any one parameter. I measure and publish at lamptest.ru more than a dozen different parameters, many of which affect the overall quality of light and the quality of the lamp itself. But many visitors to the site do not want to understand the parameters and choose the lamp solely by the overall assessment. In the old formula, the estimate was most influenced by how much the measured luminous flux differs from the declared one, and the actual quality of light (color rendering index and ripple level) did not affect the final grade so much.
I had long thought that the formula needed to be changed somehow, but the final impetus for change was a letter from a reader who, choosing from two lamps, gave preference to a lamp with CRI 72 due to the fact that the lamp with CRI 85 had a lower rating.
The new formula allows you to evaluate the lamp primarily on the quality of light. The idea of the formula is as follows:
Excellent (rating 5) the lamp should have a CRI of at least 90 and a ripple of no more than 5%.
A good (grade 4) lamp should have a CRI of at least 80 and a ripple of no more than 20%.
The average (grade 3) lamp should have a CRI of at least 70 and a ripple of no more than 35%.
For the inconsistency of the measured luminous flux to the declared and measured color temperature of the declared, the final estimate decreases, but in general the lamp with CRI 70 will not be able to have a higher rating than the lamp with CRI 80.
As a result, the final estimate is calculated as follows:
1. Calculate the color rendering quality Kcri:
Kcri 5 if CRI is 90 or more
Kcri 4.5 if CRI is 85 or more
Kcri 4 if CRI is 80 or more
Kcri 3.5 if CRI is 75 or more
Kcri 3 if CRI is 70 or more
Kcri 2 if CRI is 60 or more
Kcri 1 if CRI is less than 60
2. We calculate the quality of the absence of ripple Kfl:
Kfl 5 if KP is less than 5
Kfl 4.5 if KP is less than 10
Kfl 4 if KP is less than 20
Kfl 3.5 if KP is less than 25
Kfl 3 if KP is less than 35
Kfl 2 if KP is less than 50
Kfl 1 if KP is 50 or more
3. Take the smaller of the two calculated coefficients. The idea is that a lamp with a CRI of 90 and a ripple of 70% is still bad, just like a lamp with a ripple of 0%, but a CRI of 60.
4. Calculate the coefficient of conformity of the measured luminous flux to the declared Klm:
Klm = 0 if the measured luminous flux (Lm) is 0.95 from the declared (Lm ') or more
Klm = 1-Lm' / Lm if Lm '/ Lm is 0.9 or more
Klm = 1-Lm' / Lm * 0.9 if Lm '/ Lm 0.8 or more
Klm = 1-Lm '/ Lm * 0.6 if Lm' / Lm less than 0.8
Klm is equal to zero and does not affect the final estimate if the measured luminous flux is 0.95 from the stated or more. If it is less, the coefficient increases, but cannot exceed one.
5. We calculate the compliance coefficient of the measured color temperature of the declared Kk:
Kk 0 if the difference of the measured color temperature from the declared 6% or less than
Kk 0.2 if the difference of the measured color temperature from the declared more than 6%.
6. Take the smaller Kcri or Kfl and subtract the decreasing indices Klm and Kk from it.
As a result, the formula for assessing the quality of the lamp looks like this: LEAST (Kcri: Kfl) -Klm-Kk
After changing the formula, most of the final ratings dropped and many were afraid of ratings of lamps at 2.5. So that there is no “school” deuce, you can multiply the estimate by two, and then you will get a 10-point system, but I’m not sure yet what it is worth doing.
I want to thank Matthew Ivanov for the implementation of the new formula in Excel.
The site has a forum where you can discuss everything related to LED lighting.
In the near future we are planning to rework the page of the card for each lamp and add the possibility to leave feedback there.
ps For the development of the project lamptest.ru I am looking for:
1. Volunteers who are ready to start adding CQS values to all 1244 lamps to the database;
2. Assistants willing to engage in the purchase and return of lamps in stores in Moscow.
3. “DIY” and Arduino experts, ready to help with the implementation of several technical ideas: a simple goniophotometer from a stepper motor and a light sensor for measuring the angle of illumination and the luminous flux of lamps, a light pulsation meter, an analyzer of the form of current consumed.
© 2017, Alexey Nadyozhin
And even at lamptest there were quite significant changes.
First of all, the display format of the results table has changed.
New results table
Now there are two modes of displaying the results table - regular and full (enabled by the “full mode” checkbox in the parameter selection panel). In normal mode, only the most basic lamp parameters are displayed, and all the declared parameters are collected in one cell of the “description” table. It seems to me that this is easier and clearer.
In full mode, all the available parameters are displayed, including those that duplicate the parameters in the "description" cell (this is done so that you can sort the table by them). Even those parameters that previously were only in the lamp card (barcode, date of manufacture) appeared. In full mode, the table is very wide, but this is not the main mode, and those who wanted to look at all the parameters at once I think will put up with it.
The columns in the table are re-sorted so that the declared and measured parameters in full mode are always nearby.
As before, by default, the table is sorted in the order in which the lamps were tested and on the last page you can always see the lamps that were tested last.
New options
New parameters were added to the site:
No - the lamp number according to the list (this number did not appear on the site
before , but I decided to display it first of all for myself);
CQS - Color Quality Scale - an alternative color rendering index, more suitable for LED lamps than CRI and allowing to detect lamps with an unpleasant glow color. So far, the CQS is not filled, but there is data on it and we will gradually fill this parameter for all tested lamps;
Umin- the measured minimum voltage at which the lamp operates stably and gives at least 90% of the light. Many manufacturers write the operating voltage range for lamps 220-240 V, while lamps with a high-quality driver can operate at much lower voltages. At the same time, lamps with a linear driver reduce the brightness when the voltage drops and in some lamps the minimum voltage at which they give 90% of the light is even higher than 220 volts. By this parameter it is possible to judge which driver is used in the lamp;
drv- driver type. By the behavior of the lamp when the voltage drops, you can judge the type of driver used. Now for the lamp one of the four types of driver can be specified: 1 - the brightness does not change in a wide voltage range, but below a certain level it begins to decrease sharply (as a rule, between the beginning of the decrease in brightness and its decrease by 10%, the voltage decreases by only 10 volts) , 2 - the brightness does not change over a wide voltage range, when a certain level is reached the lamp turns off; 3 - the brightness does not change in a wide voltage range, when the voltage drops to a certain level, the lamp starts flashing; 4 - linear driver, the brightness decreases linearly with decreasing voltage. I will definitely write an article, what drivers are used in LED lamps.
tmax- The maximum temperature of the lamp housing when used in an open ceiling and a vertical installation of the base down;
dim - the minimum level of dimming. Earlier in the "dim" appeared the letter "D", if the lamp supports brightness control. Now, if the lamp is not dimmable, this field is empty, and if dimmable there indicates the minimum level of dimming in percent. So far, many lamps have a unit there, the remainder of the old display, but soon all the data will be entered.
Weight - now the lamps are weighed, because by weight you can indirectly determine how full-fledged cooling system is used in the lamp, as well as determine when manufacturers begin to produce other lamps under the same part numbers and bar codes.
In connection with the appearance of new parameters, the format of the data file lamptest.ru/led.csv has been changed - new fields have been added to it.
New formula for assessing the quality of lamps
The quality of LED lamps can not be judged by any one parameter. I measure and publish at lamptest.ru more than a dozen different parameters, many of which affect the overall quality of light and the quality of the lamp itself. But many visitors to the site do not want to understand the parameters and choose the lamp solely by the overall assessment. In the old formula, the estimate was most influenced by how much the measured luminous flux differs from the declared one, and the actual quality of light (color rendering index and ripple level) did not affect the final grade so much.
I had long thought that the formula needed to be changed somehow, but the final impetus for change was a letter from a reader who, choosing from two lamps, gave preference to a lamp with CRI 72 due to the fact that the lamp with CRI 85 had a lower rating.
The new formula allows you to evaluate the lamp primarily on the quality of light. The idea of the formula is as follows:
Excellent (rating 5) the lamp should have a CRI of at least 90 and a ripple of no more than 5%.
A good (grade 4) lamp should have a CRI of at least 80 and a ripple of no more than 20%.
The average (grade 3) lamp should have a CRI of at least 70 and a ripple of no more than 35%.
For the inconsistency of the measured luminous flux to the declared and measured color temperature of the declared, the final estimate decreases, but in general the lamp with CRI 70 will not be able to have a higher rating than the lamp with CRI 80.
As a result, the final estimate is calculated as follows:
1. Calculate the color rendering quality Kcri:
Kcri 5 if CRI is 90 or more
Kcri 4.5 if CRI is 85 or more
Kcri 4 if CRI is 80 or more
Kcri 3.5 if CRI is 75 or more
Kcri 3 if CRI is 70 or more
Kcri 2 if CRI is 60 or more
Kcri 1 if CRI is less than 60
2. We calculate the quality of the absence of ripple Kfl:
Kfl 5 if KP is less than 5
Kfl 4.5 if KP is less than 10
Kfl 4 if KP is less than 20
Kfl 3.5 if KP is less than 25
Kfl 3 if KP is less than 35
Kfl 2 if KP is less than 50
Kfl 1 if KP is 50 or more
3. Take the smaller of the two calculated coefficients. The idea is that a lamp with a CRI of 90 and a ripple of 70% is still bad, just like a lamp with a ripple of 0%, but a CRI of 60.
4. Calculate the coefficient of conformity of the measured luminous flux to the declared Klm:
Klm = 0 if the measured luminous flux (Lm) is 0.95 from the declared (Lm ') or more
Klm = 1-Lm' / Lm if Lm '/ Lm is 0.9 or more
Klm = 1-Lm' / Lm * 0.9 if Lm '/ Lm 0.8 or more
Klm = 1-Lm '/ Lm * 0.6 if Lm' / Lm less than 0.8
Klm is equal to zero and does not affect the final estimate if the measured luminous flux is 0.95 from the stated or more. If it is less, the coefficient increases, but cannot exceed one.
5. We calculate the compliance coefficient of the measured color temperature of the declared Kk:
Kk 0 if the difference of the measured color temperature from the declared 6% or less than
Kk 0.2 if the difference of the measured color temperature from the declared more than 6%.
6. Take the smaller Kcri or Kfl and subtract the decreasing indices Klm and Kk from it.
As a result, the formula for assessing the quality of the lamp looks like this: LEAST (Kcri: Kfl) -Klm-Kk
After changing the formula, most of the final ratings dropped and many were afraid of ratings of lamps at 2.5. So that there is no “school” deuce, you can multiply the estimate by two, and then you will get a 10-point system, but I’m not sure yet what it is worth doing.
I want to thank Matthew Ivanov for the implementation of the new formula in Excel.
Forum
The site has a forum where you can discuss everything related to LED lighting.
Immediate plans
In the near future we are planning to rework the page of the card for each lamp and add the possibility to leave feedback there.
ps For the development of the project lamptest.ru I am looking for:
1. Volunteers who are ready to start adding CQS values to all 1244 lamps to the database;
2. Assistants willing to engage in the purchase and return of lamps in stores in Moscow.
3. “DIY” and Arduino experts, ready to help with the implementation of several technical ideas: a simple goniophotometer from a stepper motor and a light sensor for measuring the angle of illumination and the luminous flux of lamps, a light pulsation meter, an analyzer of the form of current consumed.
© 2017, Alexey Nadyozhin