Versão atual de: Michael
Texto:
| + | == introduction == |
|---|---|
| + | |
| + | Before you choose this solution, keep in mind, there are many problems that cause no backlight. For example the mosfet, coil and diode components may be broken. In this case the driver detects a failure (with the feedback resistor) and turns off. In my case everything was right. The microcontroller doesn't provide the PWM signal. But it took a long time to get this information. First I measured the signals on the board. Is the driver enabled? What is the input value of the PWM (S1 and S2) pin. I also checked (with the help of the data sheet) the mosfets, coils and diodes: Which pin of the driver is connected to the gate of what mosfet and what kind of function does it have? In the end I found out, that the external components are ok, so I began to make some tests with the driver! |
| + | |
| In the image | |
| [image|729723|size=medium|align=center|caption=100% pwm signal] | |
| you can see the little wire that put the 3.3 voltage as input for the pwm signal. The backlight turns on and the cinema display is alive again! | |
| == the long story == | |
| After four years of using the backlight of the LED Cinema Display 27" was turned off. First the backlight was very low or was flickering while trying to dim the display via the system panel. I turned the display off and on and then it was dark. | |
| With the help of a torch light I saw a faint image of the desktop. USB and magsafe were working as well. So the display is working without the backlight. | |
| First of all I teardowned the display like it shown in [guide|6525|Teardown Apple Display] | |
| The interesting parts are | |
| * the led driver HV9982 | |
| * the mosfets and coils of the driver (producing 80 voltage for the led stripes) | |
| * the microcontroller | |
| The led driver (called HV9982) is responsible to drive the leds stripes with current. It cares about the right current, because LEDs require more precise current management. The data sheet of the HV9982 you can find some basic information about the driver and this function: | |
| HV9982 is a three-channel, closed loop, peak-current mode PWM controller designed to drive a constant out- put current. It can be used for driving either RGB LEDs or multiple channels of white LEDs. | |
| === identifying the components === | |
| In the next image you can see the led driver with its external components: | |
| [image|729733|size=medium|align=center|caption=the led driver] | |
| * the mosfet, coils and feedback resistors (orange) | |
| * the connector for the led stripes (green) | |
| * the led driver chip (blue) | |
| Using the datasheet of the led driver we get some interesting facts about the pins: | |
| [image|729725|size=medium|align=center|caption=the pins of the led driver] | |
| The led driver is controlled by the microntroller in the next picture: | |
| [image|729732|size=medium|align=center|caption=the microcontoller] | |
| The controller turns the led driver on if the computer is connected via thunderbolt cable. With a multimeter you can measure a 3.3 volt signal on pin 10 (enabled). If the computer is unplugged then the signal is bound to ground. But be aware the is only litte space between the pins! Keep an eye on it to avoid short circuits! (To be honest, I did three short circuits, one hurts a bit and one does a nice spark.) | |
| === identifying the connections === | |
| With the multimeter you can track the connections from pin to pin. I found out that the microntroller is connected to the pin 10 (enabled) and the pwm pins (17,18,19). The S1 and S2 pin are bound to ground and that means that the led driver is driven in the PWM mode. With the help of an oscilloscope I found out that the microcontroller does not send a PWM signal. He turns the led driver on but "forget" the create the PWM signal. In this case the backlight stays dark without any light. And you think the display is broken. (see [https://en.wikipedia.org/wiki/Pulse-width_modulation|for PWM]) | |
| But we have some luck! We can bridge the signal to 100%. We just connect the PWM pins to logical one and that is 3.3 volt. I found a pad that give me the 3.3 voltage and the pwm pins from the microcontoller are bridget by three 0 ohm resistor (labeled with R1132, R1133 and R1135). You cannot just put 3.3 volt to the pwm pins because they are connected with the microcontroller. This would cause a short circuit. In the following image you can see how the microntroller is connected to the pwm pins of the led driver: | |
| [image|729740|size=medium|align=center|caption=the connection the the pwm pins] | |
| The "R1" is a pull down resistor which is used the put a logical 0 to the pwm pins if the microncontroller does nothing. Now we make the following connection: | |
| [image|729743|size=medium|align=center|caption=the modified schema with constant 100% PWM] | |
| The led driver is a three channel pwm controller and the board contains all components for three channels and the connector is prepared for three channel but channel three is not used! So don't put 3.3 ohn R1133! In this case the led driver detects an error and switch off. The led cinema display has only two led stripes! Therefore connect only R1135 and R1132 with 3.3 volt. | |
| So what do you think now? The problem is not a hardware problem, it is a software problem. The software within the microcontroller won't create the pwm signal. Unfortunately we don't have the source code, so we are not able to patch the firmware. We just cut the connections from the microcontroller and send now a pwm signal about 100%. You aren't able to dim any more and you needn't buy a new one! Is this a software error or just a hint that you should buy a new display? | |
| [image|729751|size=medium|align=center|caption=the 3.3 volt pad and the 0 ohm resistors] | |
| Good luck! Please keep in mind: | |
| * Avoid short circuit! The spaces between the components are very small! | |
| * The led driver creates up to 80 voltage!! That hurt's! | |
| * The board has 24 voltage! | |
| * You need two solder irons to unsolder the resistors. | |
| === next steps === | |
| * Disassamble the firmware and find the interesting code | |
| * Maybe we find a code that is used to disable pwm after reaching a curtain value? | |
| * Maybe we find a code that simulates flickering and abnormal dimming to give the user some impression that the cinema display is "going" to get out of order next time? | |
| * Or it is just an error and the firmware thinks, that the display is in sleeping mode. | |
| PS: It was the **third display** where the backlight turns out. Unfortunately I didn't check them and they are trashed, because the local apple dealer said, that repairing is too expensive than buying a new one! | |
| [image|729791|size=medium|align=center|caption=full backlight again] |
Status:
open
Editado por: Michael
Texto:
| In the image | |
| [image|729723|size=medium|align=center|caption=100% pwm signal] | |
| you can see the little wire that put the 3.3 voltage as input for the pwm signal. The backlight turns on and the cinema display is alive again! | |
| == the long story == | |
| After four years of using the backlight of the LED Cinema Display 27" was turned off. First the backlight was very low or was flickering while trying to dim the display via the system panel. I turned the display off and on and then it was dark. | |
| With the help of a torch light I saw a faint image of the desktop. USB and magsafe were working as well. So the display is working without the backlight. | |
| First of all I teardowned the display like it shown in [guide|6525|Teardown Apple Display] | |
| The interesting parts are | |
| * the led driver HV9982 | |
| * the mosfets and coils of the driver (producing 80 voltage for the led stripes) | |
| * the microcontroller | |
| The led driver (called HV9982) is responsible to drive the leds stripes with current. It cares about the right current, because LEDs require more precise current management. The data sheet of the HV9982 you can find some basic information about the driver and this function: | |
| HV9982 is a three-channel, closed loop, peak-current mode PWM controller designed to drive a constant out- put current. It can be used for driving either RGB LEDs or multiple channels of white LEDs. | |
| === identifying the components === | |
| In the next image you can see the led driver with its external components: | |
| [image|729733|size=medium|align=center|caption=the led driver] | |
| * the mosfet, coils and feedback resistors (orange) | |
| * the connector for the led stripes (green) | |
| * the led driver chip (blue) | |
| Using the datasheet of the led driver we get some interesting facts about the pins: | |
| [image|729725|size=medium|align=center|caption=the pins of the led driver] | |
| The led driver is controlled by the microntroller in the next picture: | |
| [image|729732|size=medium|align=center|caption=the microcontoller] | |
| The controller turns the led driver on if the computer is connected via thunderbolt cable. With a multimeter you can measure a 3.3 volt signal on pin 10 (enabled). If the computer is unplugged then the signal is bound to ground. But be aware the is only litte space between the pins! Keep an eye on it to avoid short circuits! (To be honest, I did three short circuits, one hurts a bit and one does a nice spark.) | |
| === identifying the connections === | |
| With the multimeter you can track the connections from pin to pin. I found out that the microntroller is connected to the pin 10 (enabled) and the pwm pins (17,18,19). The S1 and S2 pin are bound to ground and that means that the led driver is driven in the PWM mode. With the help of an oscilloscope I found out that the microcontroller does not send a PWM signal. He turns the led driver on but "forget" the create the PWM signal. In this case the backlight stays dark without any light. And you think the display is broken. (see [https://en.wikipedia.org/wiki/Pulse-width_modulation|for PWM]) | |
| But we have some luck! We can bridge the signal to 100%. We just connect the PWM pins to logical one and that is 3.3 volt. I found a pad that give me the 3.3 voltage and the pwm pins from the microcontoller are bridget by three 0 ohm resistor (labeled with R1132, R1133 and R1135). You cannot just put 3.3 volt to the pwm pins because they are connected with the microcontroller. This would cause a short circuit. In the following image you can see how the microntroller is connected to the pwm pins of the led driver: | |
| [image|729740|size=medium|align=center|caption=the connection the the pwm pins] | |
| The "R1" is a pull down resistor which is used the put a logical 0 to the pwm pins if the microncontroller does nothing. Now we make the following connection: | |
| [image|729743|size=medium|align=center|caption=the modified schema with constant 100% PWM] | |
| The led driver is a three channel pwm controller and the board contains all components for three channels and the connector is prepared for three channel but channel three is not used! So don't put 3.3 ohn R1133! In this case the led driver detects an error and switch off. The led cinema display has only two led stripes! Therefore connect only R1135 and R1132 with 3.3 volt. | |
| So what do you think now? The problem is not a hardware problem, it is a software problem. The software within the microcontroller won't create the pwm signal. Unfortunately we don't have the source code, so we are not able to patch the firmware. We just cut the connections from the microcontroller and send now a pwm signal about 100%. You aren't able to dim any more and you needn't buy a new one! Is this a software error or just a hint that you should buy a new display? | |
| [image|729751|size=medium|align=center|caption=the 3.3 volt pad and the 0 ohm resistors] | |
| Good luck! Please keep in mind: | |
| * Avoid short circuit! The spaces between the components are very small! | |
| * The led driver creates up to 80 voltage!! That hurt's! | |
| * The board has 24 voltage! | |
| * You need two solder irons to unsolder the resistors. | |
| === next steps === | |
| * Disassamble the firmware and find the interesting code | |
| * Maybe we find a code that is used to disable pwm after reaching a curtain value? | |
| * Maybe we find a code that simulates flickering and abnormal dimming to give the user some impression that the cinema display is "going" to get out of order next time? | |
| * Or it is just an error and the firmware thinks, that the display is in sleeping mode. | |
| + | PS: It was the **third display** where the backlight turns out. Unfortunately I didn't check them and they are trashed, because the local apple dealer said, that repairing is too expensive than buying a new one! |
| + | |
| [image|729791|size=medium|align=center|caption=full backlight again] |
Status:
open
Editado por: Michael
Texto:
| In the image | |
| [image|729723|size=medium|align=center|caption=100% pwm signal] | |
| you can see the little wire that put the 3.3 voltage as input for the pwm signal. The backlight turns on and the cinema display is alive again! | |
| == the long story == | |
| After four years of using the backlight of the LED Cinema Display 27" was turned off. First the backlight was very low or was flickering while trying to dim the display via the system panel. I turned the display off and on and then it was dark. | |
| With the help of a torch light I saw a faint image of the desktop. USB and magsafe were working as well. So the display is working without the backlight. | |
| First of all I teardowned the display like it shown in [guide|6525|Teardown Apple Display] | |
| The interesting parts are | |
| * the led driver HV9982 | |
| * the mosfets and coils of the driver (producing 80 voltage for the led stripes) | |
| * the microcontroller | |
| The led driver (called HV9982) is responsible to drive the leds stripes with current. It cares about the right current, because LEDs require more precise current management. The data sheet of the HV9982 you can find some basic information about the driver and this function: | |
| HV9982 is a three-channel, closed loop, peak-current mode PWM controller designed to drive a constant out- put current. It can be used for driving either RGB LEDs or multiple channels of white LEDs. | |
| === identifying the components === | |
| In the next image you can see the led driver with its external components: | |
| [image|729733|size=medium|align=center|caption=the led driver] | |
| - | * the mosfet (orange) |
| - | * the connector the led stripes (green) |
| + | * the mosfet, coils and feedback resistors (orange) |
| + | * the connector for the led stripes (green) |
| * the led driver chip (blue) | |
| Using the datasheet of the led driver we get some interesting facts about the pins: | |
| [image|729725|size=medium|align=center|caption=the pins of the led driver] | |
| The led driver is controlled by the microntroller in the next picture: | |
| [image|729732|size=medium|align=center|caption=the microcontoller] | |
| - | The controller turns the led driver on if the computer is connected via thunderbolt cable. With a multimeter you can measure a 3.3 volt signal on pin 10 (enabled). If the computer is unplugged then the signal is bound to ground. But be aware the is only litte space between the pins! Keep an eye to avoid short circuits! |
| + | The controller turns the led driver on if the computer is connected via thunderbolt cable. With a multimeter you can measure a 3.3 volt signal on pin 10 (enabled). If the computer is unplugged then the signal is bound to ground. But be aware the is only litte space between the pins! Keep an eye on it to avoid short circuits! (To be honest, I did three short circuits, one hurts a bit and one does a nice spark.) |
| === identifying the connections === | |
| With the multimeter you can track the connections from pin to pin. I found out that the microntroller is connected to the pin 10 (enabled) and the pwm pins (17,18,19). The S1 and S2 pin are bound to ground and that means that the led driver is driven in the PWM mode. With the help of an oscilloscope I found out that the microcontroller does not send a PWM signal. He turns the led driver on but "forget" the create the PWM signal. In this case the backlight stays dark without any light. And you think the display is broken. (see [https://en.wikipedia.org/wiki/Pulse-width_modulation|for PWM]) | |
| But we have some luck! We can bridge the signal to 100%. We just connect the PWM pins to logical one and that is 3.3 volt. I found a pad that give me the 3.3 voltage and the pwm pins from the microcontoller are bridget by three 0 ohm resistor (labeled with R1132, R1133 and R1135). You cannot just put 3.3 volt to the pwm pins because they are connected with the microcontroller. This would cause a short circuit. In the following image you can see how the microntroller is connected to the pwm pins of the led driver: | |
| [image|729740|size=medium|align=center|caption=the connection the the pwm pins] | |
| The "R1" is a pull down resistor which is used the put a logical 0 to the pwm pins if the microncontroller does nothing. Now we make the following connection: | |
| [image|729743|size=medium|align=center|caption=the modified schema with constant 100% PWM] | |
| The led driver is a three channel pwm controller and the board contains all components for three channels and the connector is prepared for three channel but channel three is not used! So don't put 3.3 ohn R1133! In this case the led driver detects an error and switch off. The led cinema display has only two led stripes! Therefore connect only R1135 and R1132 with 3.3 volt. | |
| So what do you think now? The problem is not a hardware problem, it is a software problem. The software within the microcontroller won't create the pwm signal. Unfortunately we don't have the source code, so we are not able to patch the firmware. We just cut the connections from the microcontroller and send now a pwm signal about 100%. You aren't able to dim any more and you needn't buy a new one! Is this a software error or just a hint that you should buy a new display? | |
| [image|729751|size=medium|align=center|caption=the 3.3 volt pad and the 0 ohm resistors] | |
| Good luck! Please keep in mind: | |
| * Avoid short circuit! The spaces between the components are very small! | |
| * The led driver creates up to 80 voltage!! That hurt's! | |
| * The board has 24 voltage! | |
| * You need two solder irons to unsolder the resistors. | |
| === next steps === | |
| * Disassamble the firmware and find the interesting code | |
| * Maybe we find a code that is used to disable pwm after reaching a curtain value? | |
| * Maybe we find a code that simulates flickering and abnormal dimming to give the user some impression that the cinema display is "going" to get out of order next time? | |
| * Or it is just an error and the firmware thinks, that the display is in sleeping mode. | |
| [image|729791|size=medium|align=center|caption=full backlight again] |
Status:
open
Editado por: Michael
Texto:
| In the image | |
| [image|729723|size=medium|align=center|caption=100% pwm signal] | |
| you can see the little wire that put the 3.3 voltage as input for the pwm signal. The backlight turns on and the cinema display is alive again! | |
| == the long story == | |
| After four years of using the backlight of the LED Cinema Display 27" was turned off. First the backlight was very low or was flickering while trying to dim the display via the system panel. I turned the display off and on and then it was dark. | |
| With the help of a torch light I saw a faint image of the desktop. USB and magsafe were working as well. So the display is working without the backlight. | |
| First of all I teardowned the display like it shown in [guide|6525|Teardown Apple Display] | |
| The interesting parts are | |
| * the led driver HV9982 | |
| - | * the mosfets and coils of the driver ( |
| + | * the mosfets and coils of the driver (producing 80 voltage for the led stripes) |
| * the microcontroller | |
| The led driver (called HV9982) is responsible to drive the leds stripes with current. It cares about the right current, because LEDs require more precise current management. The data sheet of the HV9982 you can find some basic information about the driver and this function: | |
| HV9982 is a three-channel, closed loop, peak-current mode PWM controller designed to drive a constant out- put current. It can be used for driving either RGB LEDs or multiple channels of white LEDs. | |
| === identifying the components === | |
| In the next image you can see the led driver with its external components: | |
| [image|729733|size=medium|align=center|caption=the led driver] | |
| * the mosfet (orange) | |
| * the connector the led stripes (green) | |
| * the led driver chip (blue) | |
| Using the datasheet of the led driver we get some interesting facts about the pins: | |
| [image|729725|size=medium|align=center|caption=the pins of the led driver] | |
| The led driver is controlled by the microntroller in the next picture: | |
| [image|729732|size=medium|align=center|caption=the microcontoller] | |
| The controller turns the led driver on if the computer is connected via thunderbolt cable. With a multimeter you can measure a 3.3 volt signal on pin 10 (enabled). If the computer is unplugged then the signal is bound to ground. But be aware the is only litte space between the pins! Keep an eye to avoid short circuits! | |
| === identifying the connections === | |
| With the multimeter you can track the connections from pin to pin. I found out that the microntroller is connected to the pin 10 (enabled) and the pwm pins (17,18,19). The S1 and S2 pin are bound to ground and that means that the led driver is driven in the PWM mode. With the help of an oscilloscope I found out that the microcontroller does not send a PWM signal. He turns the led driver on but "forget" the create the PWM signal. In this case the backlight stays dark without any light. And you think the display is broken. (see [https://en.wikipedia.org/wiki/Pulse-width_modulation|for PWM]) | |
| But we have some luck! We can bridge the signal to 100%. We just connect the PWM pins to logical one and that is 3.3 volt. I found a pad that give me the 3.3 voltage and the pwm pins from the microcontoller are bridget by three 0 ohm resistor (labeled with R1132, R1133 and R1135). You cannot just put 3.3 volt to the pwm pins because they are connected with the microcontroller. This would cause a short circuit. In the following image you can see how the microntroller is connected to the pwm pins of the led driver: | |
| [image|729740|size=medium|align=center|caption=the connection the the pwm pins] | |
| The "R1" is a pull down resistor which is used the put a logical 0 to the pwm pins if the microncontroller does nothing. Now we make the following connection: | |
| [image|729743|size=medium|align=center|caption=the modified schema with constant 100% PWM] | |
| The led driver is a three channel pwm controller and the board contains all components for three channels and the connector is prepared for three channel but channel three is not used! So don't put 3.3 ohn R1133! In this case the led driver detects an error and switch off. The led cinema display has only two led stripes! Therefore connect only R1135 and R1132 with 3.3 volt. | |
| So what do you think now? The problem is not a hardware problem, it is a software problem. The software within the microcontroller won't create the pwm signal. Unfortunately we don't have the source code, so we are not able to patch the firmware. We just cut the connections from the microcontroller and send now a pwm signal about 100%. You aren't able to dim any more and you needn't buy a new one! Is this a software error or just a hint that you should buy a new display? | |
| [image|729751|size=medium|align=center|caption=the 3.3 volt pad and the 0 ohm resistors] | |
| Good luck! Please keep in mind: | |
| * Avoid short circuit! The spaces between the components are very small! | |
| * The led driver creates up to 80 voltage!! That hurt's! | |
| * The board has 24 voltage! | |
| * You need two solder irons to unsolder the resistors. | |
| === next steps === | |
| * Disassamble the firmware and find the interesting code | |
| * Maybe we find a code that is used to disable pwm after reaching a curtain value? | |
| * Maybe we find a code that simulates flickering and abnormal dimming to give the user some impression that the cinema display is "going" to get out of order next time? | |
| * Or it is just an error and the firmware thinks, that the display is in sleeping mode. | |
| [image|729791|size=medium|align=center|caption=full backlight again] |
Status:
open
Editado por: Michael
Texto:
| In the image | |
| [image|729723|size=medium|align=center|caption=100% pwm signal] | |
| you can see the little wire that put the 3.3 voltage as input for the pwm signal. The backlight turns on and the cinema display is alive again! | |
| == the long story == | |
| After four years of using the backlight of the LED Cinema Display 27" was turned off. First the backlight was very low or was flickering while trying to dim the display via the system panel. I turned the display off and on and then it was dark. | |
| With the help of a torch light I saw a faint image of the desktop. USB and magsafe were working as well. So the display is working without the backlight. | |
| First of all I teardowned the display like it shown in [guide|6525|Teardown Apple Display] | |
| The interesting parts are | |
| * the led driver HV9982 | |
| * the mosfets and coils of the driver (producting 80 voltage for the led stripes) | |
| * the microcontroller | |
| The led driver (called HV9982) is responsible to drive the leds stripes with current. It cares about the right current, because LEDs require more precise current management. The data sheet of the HV9982 you can find some basic information about the driver and this function: | |
| HV9982 is a three-channel, closed loop, peak-current mode PWM controller designed to drive a constant out- put current. It can be used for driving either RGB LEDs or multiple channels of white LEDs. | |
| === identifying the components === | |
| In the next image you can see the led driver with its external components: | |
| [image|729733|size=medium|align=center|caption=the led driver] | |
| * the mosfet (orange) | |
| * the connector the led stripes (green) | |
| * the led driver chip (blue) | |
| Using the datasheet of the led driver we get some interesting facts about the pins: | |
| [image|729725|size=medium|align=center|caption=the pins of the led driver] | |
| The led driver is controlled by the microntroller in the next picture: | |
| [image|729732|size=medium|align=center|caption=the microcontoller] | |
| The controller turns the led driver on if the computer is connected via thunderbolt cable. With a multimeter you can measure a 3.3 volt signal on pin 10 (enabled). If the computer is unplugged then the signal is bound to ground. But be aware the is only litte space between the pins! Keep an eye to avoid short circuits! | |
| === identifying the connections === | |
| With the multimeter you can track the connections from pin to pin. I found out that the microntroller is connected to the pin 10 (enabled) and the pwm pins (17,18,19). The S1 and S2 pin are bound to ground and that means that the led driver is driven in the PWM mode. With the help of an oscilloscope I found out that the microcontroller does not send a PWM signal. He turns the led driver on but "forget" the create the PWM signal. In this case the backlight stays dark without any light. And you think the display is broken. (see [https://en.wikipedia.org/wiki/Pulse-width_modulation|for PWM]) | |
| But we have some luck! We can bridge the signal to 100%. We just connect the PWM pins to logical one and that is 3.3 volt. I found a pad that give me the 3.3 voltage and the pwm pins from the microcontoller are bridget by three 0 ohm resistor (labeled with R1132, R1133 and R1135). You cannot just put 3.3 volt to the pwm pins because they are connected with the microcontroller. This would cause a short circuit. In the following image you can see how the microntroller is connected to the pwm pins of the led driver: | |
| [image|729740|size=medium|align=center|caption=the connection the the pwm pins] | |
| The "R1" is a pull down resistor which is used the put a logical 0 to the pwm pins if the microncontroller does nothing. Now we make the following connection: | |
| [image|729743|size=medium|align=center|caption=the modified schema with constant 100% PWM] | |
| The led driver is a three channel pwm controller and the board contains all components for three channels and the connector is prepared for three channel but channel three is not used! So don't put 3.3 ohn R1133! In this case the led driver detects an error and switch off. The led cinema display has only two led stripes! Therefore connect only R1135 and R1132 with 3.3 volt. | |
| So what do you think now? The problem is not a hardware problem, it is a software problem. The software within the microcontroller won't create the pwm signal. Unfortunately we don't have the source code, so we are not able to patch the firmware. We just cut the connections from the microcontroller and send now a pwm signal about 100%. You aren't able to dim any more and you needn't buy a new one! Is this a software error or just a hint that you should buy a new display? | |
| [image|729751|size=medium|align=center|caption=the 3.3 volt pad and the 0 ohm resistors] | |
| Good luck! Please keep in mind: | |
| * Avoid short circuit! The spaces between the components are very small! | |
| * The led driver creates up to 80 voltage!! That hurt's! | |
| * The board has 24 voltage! | |
| * You need two solder irons to unsolder the resistors. | |
| === next steps === | |
| * Disassamble the firmware and find the interesting code | |
| - | * Maybe we find a count that is used to disable pwm after reaching a curtain value? |
| - | * Maybe we find code that simulates flickering and abnormal dimming to give the user some impression that the cinema display is "going" to get out of order? |
| + | * Maybe we find a code that is used to disable pwm after reaching a curtain value? |
| + | * Maybe we find a code that simulates flickering and abnormal dimming to give the user some impression that the cinema display is "going" to get out of order next time? |
| + | * Or it is just an error and the firmware thinks, that the display is in sleeping mode. |
| [image|729791|size=medium|align=center|caption=full backlight again] |
Status:
open
Editado por: Michael
Texto:
| In the image | |
| [image|729723|size=medium|align=center|caption=100% pwm signal] | |
| you can see the little wire that put the 3.3 voltage as input for the pwm signal. The backlight turns on and the cinema display is alive again! | |
| == the long story == | |
| After four years of using the backlight of the LED Cinema Display 27" was turned off. First the backlight was very low or was flickering while trying to dim the display via the system panel. I turned the display off and on and then it was dark. | |
| With the help of a torch light I saw a faint image of the desktop. USB and magsafe were working as well. So the display is working without the backlight. | |
| First of all I teardowned the display like it shown in [guide|6525|Teardown Apple Display] | |
| The interesting parts are | |
| * the led driver HV9982 | |
| * the mosfets and coils of the driver (producting 80 voltage for the led stripes) | |
| * the microcontroller | |
| The led driver (called HV9982) is responsible to drive the leds stripes with current. It cares about the right current, because LEDs require more precise current management. The data sheet of the HV9982 you can find some basic information about the driver and this function: | |
| HV9982 is a three-channel, closed loop, peak-current mode PWM controller designed to drive a constant out- put current. It can be used for driving either RGB LEDs or multiple channels of white LEDs. | |
| === identifying the components === | |
| In the next image you can see the led driver with its external components: | |
| [image|729733|size=medium|align=center|caption=the led driver] | |
| * the mosfet (orange) | |
| * the connector the led stripes (green) | |
| * the led driver chip (blue) | |
| Using the datasheet of the led driver we get some interesting facts about the pins: | |
| [image|729725|size=medium|align=center|caption=the pins of the led driver] | |
| The led driver is controlled by the microntroller in the next picture: | |
| [image|729732|size=medium|align=center|caption=the microcontoller] | |
| The controller turns the led driver on if the computer is connected via thunderbolt cable. With a multimeter you can measure a 3.3 volt signal on pin 10 (enabled). If the computer is unplugged then the signal is bound to ground. But be aware the is only litte space between the pins! Keep an eye to avoid short circuits! | |
| === identifying the connections === | |
| - | With the multimeter you can track the connections from pin to pin. I found out that the microntroller is connected to the pin 10 (enabled) and the pwm pins (17,18,19). The S1 and S2 pin are bound to ground and that means that the led driver is driven in the PWM mode. With the help of an oscilloscope I found out that the microcontroller does not send a PWM signal. He turns the led driver on but "forget" the create the PWM signal. In this case the backlight stays dark without any light. And you think the display is broken. (see [https://en.wikipedia.org/wiki/Pulse-width_modulation|for PWM] |
| + | With the multimeter you can track the connections from pin to pin. I found out that the microntroller is connected to the pin 10 (enabled) and the pwm pins (17,18,19). The S1 and S2 pin are bound to ground and that means that the led driver is driven in the PWM mode. With the help of an oscilloscope I found out that the microcontroller does not send a PWM signal. He turns the led driver on but "forget" the create the PWM signal. In this case the backlight stays dark without any light. And you think the display is broken. (see [https://en.wikipedia.org/wiki/Pulse-width_modulation|for PWM]) |
| But we have some luck! We can bridge the signal to 100%. We just connect the PWM pins to logical one and that is 3.3 volt. I found a pad that give me the 3.3 voltage and the pwm pins from the microcontoller are bridget by three 0 ohm resistor (labeled with R1132, R1133 and R1135). You cannot just put 3.3 volt to the pwm pins because they are connected with the microcontroller. This would cause a short circuit. In the following image you can see how the microntroller is connected to the pwm pins of the led driver: | |
| [image|729740|size=medium|align=center|caption=the connection the the pwm pins] | |
| The "R1" is a pull down resistor which is used the put a logical 0 to the pwm pins if the microncontroller does nothing. Now we make the following connection: | |
| [image|729743|size=medium|align=center|caption=the modified schema with constant 100% PWM] | |
| The led driver is a three channel pwm controller and the board contains all components for three channels and the connector is prepared for three channel but channel three is not used! So don't put 3.3 ohn R1133! In this case the led driver detects an error and switch off. The led cinema display has only two led stripes! Therefore connect only R1135 and R1132 with 3.3 volt. | |
| So what do you think now? The problem is not a hardware problem, it is a software problem. The software within the microcontroller won't create the pwm signal. Unfortunately we don't have the source code, so we are not able to patch the firmware. We just cut the connections from the microcontroller and send now a pwm signal about 100%. You aren't able to dim any more and you needn't buy a new one! Is this a software error or just a hint that you should buy a new display? | |
| [image|729751|size=medium|align=center|caption=the 3.3 volt pad and the 0 ohm resistors] | |
| Good luck! Please keep in mind: | |
| * Avoid short circuit! The spaces between the components are very small! | |
| * The led driver creates up to 80 voltage!! That hurt's! | |
| * The board has 24 voltage! | |
| * You need two solder irons to unsolder the resistors. | |
| === next steps === | |
| * Disassamble the firmware and find the interesting code | |
| * Maybe we find a count that is used to disable pwm after reaching a curtain value? | |
| * Maybe we find code that simulates flickering and abnormal dimming to give the user some impression that the cinema display is "going" to get out of order? | |
| - | [image| |
| + | [image|729791|size=medium|align=center|caption=full backlight again] |
Status:
open
Editado por: Michael
Texto:
| In the image | |
| [image|729723|size=medium|align=center|caption=100% pwm signal] | |
| you can see the little wire that put the 3.3 voltage as input for the pwm signal. The backlight turns on and the cinema display is alive again! | |
| == the long story == | |
| After four years of using the backlight of the LED Cinema Display 27" was turned off. First the backlight was very low or was flickering while trying to dim the display via the system panel. I turned the display off and on and then it was dark. | |
| With the help of a torch light I saw a faint image of the desktop. USB and magsafe were working as well. So the display is working without the backlight. | |
| First of all I teardowned the display like it shown in [guide|6525|Teardown Apple Display] | |
| The interesting parts are | |
| * the led driver HV9982 | |
| * the mosfets and coils of the driver (producting 80 voltage for the led stripes) | |
| * the microcontroller | |
| The led driver (called HV9982) is responsible to drive the leds stripes with current. It cares about the right current, because LEDs require more precise current management. The data sheet of the HV9982 you can find some basic information about the driver and this function: | |
| HV9982 is a three-channel, closed loop, peak-current mode PWM controller designed to drive a constant out- put current. It can be used for driving either RGB LEDs or multiple channels of white LEDs. | |
| === identifying the components === | |
| In the next image you can see the led driver with its external components: | |
| [image|729733|size=medium|align=center|caption=the led driver] | |
| * the mosfet (orange) | |
| * the connector the led stripes (green) | |
| * the led driver chip (blue) | |
| Using the datasheet of the led driver we get some interesting facts about the pins: | |
| [image|729725|size=medium|align=center|caption=the pins of the led driver] | |
| The led driver is controlled by the microntroller in the next picture: | |
| [image|729732|size=medium|align=center|caption=the microcontoller] | |
| The controller turns the led driver on if the computer is connected via thunderbolt cable. With a multimeter you can measure a 3.3 volt signal on pin 10 (enabled). If the computer is unplugged then the signal is bound to ground. But be aware the is only litte space between the pins! Keep an eye to avoid short circuits! | |
| === identifying the connections === | |
| With the multimeter you can track the connections from pin to pin. I found out that the microntroller is connected to the pin 10 (enabled) and the pwm pins (17,18,19). The S1 and S2 pin are bound to ground and that means that the led driver is driven in the PWM mode. With the help of an oscilloscope I found out that the microcontroller does not send a PWM signal. He turns the led driver on but "forget" the create the PWM signal. In this case the backlight stays dark without any light. And you think the display is broken. (see [https://en.wikipedia.org/wiki/Pulse-width_modulation|for PWM] | |
| But we have some luck! We can bridge the signal to 100%. We just connect the PWM pins to logical one and that is 3.3 volt. I found a pad that give me the 3.3 voltage and the pwm pins from the microcontoller are bridget by three 0 ohm resistor (labeled with R1132, R1133 and R1135). You cannot just put 3.3 volt to the pwm pins because they are connected with the microcontroller. This would cause a short circuit. In the following image you can see how the microntroller is connected to the pwm pins of the led driver: | |
| [image|729740|size=medium|align=center|caption=the connection the the pwm pins] | |
| The "R1" is a pull down resistor which is used the put a logical 0 to the pwm pins if the microncontroller does nothing. Now we make the following connection: | |
| [image|729743|size=medium|align=center|caption=the modified schema with constant 100% PWM] | |
| The led driver is a three channel pwm controller and the board contains all components for three channels and the connector is prepared for three channel but channel three is not used! So don't put 3.3 ohn R1133! In this case the led driver detects an error and switch off. The led cinema display has only two led stripes! Therefore connect only R1135 and R1132 with 3.3 volt. | |
| So what do you think now? The problem is not a hardware problem, it is a software problem. The software within the microcontroller won't create the pwm signal. Unfortunately we don't have the source code, so we are not able to patch the firmware. We just cut the connections from the microcontroller and send now a pwm signal about 100%. You aren't able to dim any more and you needn't buy a new one! Is this a software error or just a hint that you should buy a new display? | |
| [image|729751|size=medium|align=center|caption=the 3.3 volt pad and the 0 ohm resistors] | |
| Good luck! Please keep in mind: | |
| * Avoid short circuit! The spaces between the components are very small! | |
| * The led driver creates up to 80 voltage!! That hurt's! | |
| * The board has 24 voltage! | |
| * You need two solder irons to unsolder the resistors. | |
| === next steps === | |
| * Disassamble the firmware and find the interesting code | |
| * Maybe we find a count that is used to disable pwm after reaching a curtain value? | |
| - | * Maybe we find code that simulates flickering and abnormal dimming to give the user some impression that the cinema display is "going" to get out of order? |
| + | * Maybe we find code that simulates flickering and abnormal dimming to give the user some impression that the cinema display is "going" to get out of order? |
| + | |
| + | [image| 729791 |size=medium|align=center|caption=full backlight again] |
Status:
open
Editado por: Michael
Texto:
| In the image | |
| [image|729723|size=medium|align=center|caption=100% pwm signal] | |
| you can see the little wire that put the 3.3 voltage as input for the pwm signal. The backlight turns on and the cinema display is alive again! | |
| == the long story == | |
| - | After four years of using the backlight of the LED Cinema Display 27" was turned off. First the backlight was very low or was flickering while trying to dim the display. I turned the display off and on and then it was dark. |
| + | After four years of using the backlight of the LED Cinema Display 27" was turned off. First the backlight was very low or was flickering while trying to dim the display via the system panel. I turned the display off and on and then it was dark. |
| With the help of a torch light I saw a faint image of the desktop. USB and magsafe were working as well. So the display is working without the backlight. | |
| First of all I teardowned the display like it shown in [guide|6525|Teardown Apple Display] | |
| The interesting parts are | |
| * the led driver HV9982 | |
| - | * the mosfets of the driver |
| - | * the |
| + | * the mosfets and coils of the driver (producting 80 voltage for the led stripes) |
| + | * the microcontroller |
| The led driver (called HV9982) is responsible to drive the leds stripes with current. It cares about the right current, because LEDs require more precise current management. The data sheet of the HV9982 you can find some basic information about the driver and this function: | |
| HV9982 is a three-channel, closed loop, peak-current mode PWM controller designed to drive a constant out- put current. It can be used for driving either RGB LEDs or multiple channels of white LEDs. | |
| === identifying the components === | |
| In the next image you can see the led driver with its external components: | |
| [image|729733|size=medium|align=center|caption=the led driver] | |
| * the mosfet (orange) | |
| * the connector the led stripes (green) | |
| * the led driver chip (blue) | |
| Using the datasheet of the led driver we get some interesting facts about the pins: | |
| [image|729725|size=medium|align=center|caption=the pins of the led driver] | |
| The led driver is controlled by the microntroller in the next picture: | |
| [image|729732|size=medium|align=center|caption=the microcontoller] | |
| The controller turns the led driver on if the computer is connected via thunderbolt cable. With a multimeter you can measure a 3.3 volt signal on pin 10 (enabled). If the computer is unplugged then the signal is bound to ground. But be aware the is only litte space between the pins! Keep an eye to avoid short circuits! | |
| - | |
| + | === identifying the connections === |
| - | But we have some luck! We can bridge the signal to 100%. We just connect the PWM pins to logical one and that is 3.3 volt. I found a pad that give me the 3.3 voltage and the pwm pins from the microcontoller are bridget by three 0 ohm resistor (labled with R1132, R1133 and R1135). You cannot just put 3.3 volt to the pwm pins because they are connected with the microcontroller. This would cause a short circuit. In the following image you can see how the microntroller is connected to the pwm pins of the led driver: |
| + | With the multimeter you can track the connections from pin to pin. I found out that the microntroller is connected to the pin 10 (enabled) and the pwm pins (17,18,19). The S1 and S2 pin are bound to ground and that means that the led driver is driven in the PWM mode. With the help of an oscilloscope I found out that the microcontroller does not send a PWM signal. He turns the led driver on but "forget" the create the PWM signal. In this case the backlight stays dark without any light. And you think the display is broken. (see [https://en.wikipedia.org/wiki/Pulse-width_modulation|for PWM] |
| + | |
| + | But we have some luck! We can bridge the signal to 100%. We just connect the PWM pins to logical one and that is 3.3 volt. I found a pad that give me the 3.3 voltage and the pwm pins from the microcontoller are bridget by three 0 ohm resistor (labeled with R1132, R1133 and R1135). You cannot just put 3.3 volt to the pwm pins because they are connected with the microcontroller. This would cause a short circuit. In the following image you can see how the microntroller is connected to the pwm pins of the led driver: |
| [image|729740|size=medium|align=center|caption=the connection the the pwm pins] | |
| The "R1" is a pull down resistor which is used the put a logical 0 to the pwm pins if the microncontroller does nothing. Now we make the following connection: | |
| [image|729743|size=medium|align=center|caption=the modified schema with constant 100% PWM] | |
| The led driver is a three channel pwm controller and the board contains all components for three channels and the connector is prepared for three channel but channel three is not used! So don't put 3.3 ohn R1133! In this case the led driver detects an error and switch off. The led cinema display has only two led stripes! Therefore connect only R1135 and R1132 with 3.3 volt. | |
| - | So what do you think now? The problem is not a hardware problem, it is a software problem. The software within the microcontroller won't create the pwm signal. We just cut the connections from the microcontroller and send now a pwm signal about 100%. You aren't able to dim any more and you needn't buy a new one! Is this a software error or just a hint that you should buy a new display? |
| + | So what do you think now? The problem is not a hardware problem, it is a software problem. The software within the microcontroller won't create the pwm signal. Unfortunately we don't have the source code, so we are not able to patch the firmware. We just cut the connections from the microcontroller and send now a pwm signal about 100%. You aren't able to dim any more and you needn't buy a new one! Is this a software error or just a hint that you should buy a new display? |
| - | [image| |
| + | [image|729751|size=medium|align=center|caption=the 3.3 volt pad and the 0 ohm resistors] |
| Good luck! Please keep in mind: | |
| * Avoid short circuit! The spaces between the components are very small! | |
| - | * The led driver creates |
| + | * The led driver creates up to 80 voltage!! That hurt's! |
| * The board has 24 voltage! | |
| - | * You need two solder irons for the resistors. |
| + | * You need two solder irons to unsolder the resistors. |
| + | |
| + | === next steps === |
| + | |
| + | * Disassamble the firmware and find the interesting code |
| + | * Maybe we find a count that is used to disable pwm after reaching a curtain value? |
| + | * Maybe we find code that simulates flickering and abnormal dimming to give the user some impression that the cinema display is "going" to get out of order? |
Status:
open
Editado por: Michael
Texto:
| In the image | |
| [image|729723|size=medium|align=center|caption=100% pwm signal] | |
| you can see the little wire that put the 3.3 voltage as input for the pwm signal. The backlight turns on and the cinema display is alive again! | |
| == the long story == | |
| After four years of using the backlight of the LED Cinema Display 27" was turned off. First the backlight was very low or was flickering while trying to dim the display. I turned the display off and on and then it was dark. | |
| With the help of a torch light I saw a faint image of the desktop. USB and magsafe were working as well. So the display is working without the backlight. | |
| First of all I teardowned the display like it shown in [guide|6525|Teardown Apple Display] | |
| The interesting parts are | |
| * the led driver HV9982 | |
| * the mosfets of the driver | |
| * the micro controller | |
| The led driver (called HV9982) is responsible to drive the leds stripes with current. It cares about the right current, because LEDs require more precise current management. The data sheet of the HV9982 you can find some basic information about the driver and this function: | |
| HV9982 is a three-channel, closed loop, peak-current mode PWM controller designed to drive a constant out- put current. It can be used for driving either RGB LEDs or multiple channels of white LEDs. | |
| === identifying the components === | |
| In the next image you can see the led driver with its external components: | |
| [image|729733|size=medium|align=center|caption=the led driver] | |
| * the mosfet (orange) | |
| * the connector the led stripes (green) | |
| * the led driver chip (blue) | |
| Using the datasheet of the led driver we get some interesting facts about the pins: | |
| [image|729725|size=medium|align=center|caption=the pins of the led driver] | |
| The led driver is controlled by the microntroller in the next picture: | |
| [image|729732|size=medium|align=center|caption=the microcontoller] | |
| The controller turns the led driver on if the computer is connected via thunderbolt cable. With a multimeter you can measure a 3.3 volt signal on pin 10 (enabled). If the computer is unplugged then the signal is bound to ground. But be aware the is only litte space between the pins! Keep an eye to avoid short circuits! | |
| With the multimeter you can track the connections from pin to pin. I found out that the microntroller is connected to the pin 10 (enabled) and the pwm pins (17,18,19). The S1 and S2 pin are bound to ground and that means that the led driver is driven in the PWM mode. With the help of an oscilloscope I found out that the microcontroller does not send a PWM signal. He turns the led driver on but "forget" the create the PWM signal. In this case the backlight stays dark without any ligt. And you think, the display is broken. | |
| But we have some luck! We can bridge the signal to 100%. We just connect the PWM pins to logical one and that is 3.3 volt. I found a pad that give me the 3.3 voltage and the pwm pins from the microcontoller are bridget by three 0 ohm resistor (labled with R1132, R1133 and R1135). You cannot just put 3.3 volt to the pwm pins because they are connected with the microcontroller. This would cause a short circuit. In the following image you can see how the microntroller is connected to the pwm pins of the led driver: | |
| [image|729740|size=medium|align=center|caption=the connection the the pwm pins] | |
| The "R1" is a pull down resistor which is used the put a logical 0 to the pwm pins if the microncontroller does nothing. Now we make the following connection: | |
| [image|729743|size=medium|align=center|caption=the modified schema with constant 100% PWM] | |
| The led driver is a three channel pwm controller and the board contains all components for three channels and the connector is prepared for three channel but channel three is not used! So don't put 3.3 ohn R1133! In this case the led driver detects an error and switch off. The led cinema display has only two led stripes! Therefore connect only R1135 and R1132 with 3.3 volt. | |
| So what do you think now? The problem is not a hardware problem, it is a software problem. The software within the microcontroller won't create the pwm signal. We just cut the connections from the microcontroller and send now a pwm signal about 100%. You aren't able to dim any more and you needn't buy a new one! Is this a software error or just a hint that you should buy a new display? | |
| + | [image| 729751 |size=medium|align=center|caption=the 3.3 volt pad and the 0 ohm resistors] |
| + | |
| Good luck! Please keep in mind: | |
| * Avoid short circuit! The spaces between the components are very small! | |
| * The led driver creates update 80 voltage!! That hurt's! | |
| * The board has 24 voltage! | |
| * You need two solder irons for the resistors. |
Status:
open
Postagem original de: Michael
Texto:
In the image [image|729723|size=medium|align=center|caption=100% pwm signal] you can see the little wire that put the 3.3 voltage as input for the pwm signal. The backlight turns on and the cinema display is alive again! == the long story == After four years of using the backlight of the LED Cinema Display 27" was turned off. First the backlight was very low or was flickering while trying to dim the display. I turned the display off and on and then it was dark. With the help of a torch light I saw a faint image of the desktop. USB and magsafe were working as well. So the display is working without the backlight. First of all I teardowned the display like it shown in [guide|6525|Teardown Apple Display] The interesting parts are * the led driver HV9982 * the mosfets of the driver * the micro controller The led driver (called HV9982) is responsible to drive the leds stripes with current. It cares about the right current, because LEDs require more precise current management. The data sheet of the HV9982 you can find some basic information about the driver and this function: HV9982 is a three-channel, closed loop, peak-current mode PWM controller designed to drive a constant out- put current. It can be used for driving either RGB LEDs or multiple channels of white LEDs. === identifying the components === In the next image you can see the led driver with its external components: [image|729733|size=medium|align=center|caption=the led driver] * the mosfet (orange) * the connector the led stripes (green) * the led driver chip (blue) Using the datasheet of the led driver we get some interesting facts about the pins: [image|729725|size=medium|align=center|caption=the pins of the led driver] The led driver is controlled by the microntroller in the next picture: [image|729732|size=medium|align=center|caption=the microcontoller] The controller turns the led driver on if the computer is connected via thunderbolt cable. With a multimeter you can measure a 3.3 volt signal on pin 10 (enabled). If the computer is unplugged then the signal is bound to ground. But be aware the is only litte space between the pins! Keep an eye to avoid short circuits! With the multimeter you can track the connections from pin to pin. I found out that the microntroller is connected to the pin 10 (enabled) and the pwm pins (17,18,19). The S1 and S2 pin are bound to ground and that means that the led driver is driven in the PWM mode. With the help of an oscilloscope I found out that the microcontroller does not send a PWM signal. He turns the led driver on but "forget" the create the PWM signal. In this case the backlight stays dark without any ligt. And you think, the display is broken. But we have some luck! We can bridge the signal to 100%. We just connect the PWM pins to logical one and that is 3.3 volt. I found a pad that give me the 3.3 voltage and the pwm pins from the microcontoller are bridget by three 0 ohm resistor (labled with R1132, R1133 and R1135). You cannot just put 3.3 volt to the pwm pins because they are connected with the microcontroller. This would cause a short circuit. In the following image you can see how the microntroller is connected to the pwm pins of the led driver: [image|729740|size=medium|align=center|caption=the connection the the pwm pins] The "R1" is a pull down resistor which is used the put a logical 0 to the pwm pins if the microncontroller does nothing. Now we make the following connection: [image|729743|size=medium|align=center|caption=the modified schema with constant 100% PWM] The led driver is a three channel pwm controller and the board contains all components for three channels and the connector is prepared for three channel but channel three is not used! So don't put 3.3 ohn R1133! In this case the led driver detects an error and switch off. The led cinema display has only two led stripes! Therefore connect only R1135 and R1132 with 3.3 volt. So what do you think now? The problem is not a hardware problem, it is a software problem. The software within the microcontroller won't create the pwm signal. We just cut the connections from the microcontroller and send now a pwm signal about 100%. You aren't able to dim any more and you needn't buy a new one! Is this a software error or just a hint that you should buy a new display? Good luck! Please keep in mind: * Avoid short circuit! The spaces between the components are very small! * The led driver creates update 80 voltage!! That hurt's! * The board has 24 voltage! * You need two solder irons for the resistors.
Status:
open