We received two iPhone 6s logic boards yesterday from a new B2B customer. Both had the complaint of No Backlight. David was worried that his replacement screen was actually causing the damage, so he sent us the screen to look at as well. With the help of the iBridge Flex Set for iPhone 6s, we quickly determined exactly what had happened.
For phone A, we checked the diode mode reading at the screen connector pin 33 for PP_LCM_BL_ANODE_CONN. The normal value of the backlight line, PP_LCM_BL_ANODE, is around 500 (diode mode units-DMU). Our multimeter reading was very high at 1.8 DMU. This tells us that there is open line fault somewhere on the line. The common fault is a failure of the backlight filter, FL4211. A brief look at the schematic shows us that if the battery is not disconnected, then live battery voltage will always be present at the backlight pin--even if the device is powered off! Any transient short to ground during the screen connection process will instantly create a path for electricity to rush through, unregulated. This burst of current will overheat the backlight filter, FL4211, and damage it. Indeed, that was the case with Phone A. We replaced the backlight filter and measured the relative resistance of the line in diode mode---back to normal. So that tells us that the problem in Phone A was indeed technician damage due to performing a screen replacement without disconnecting the battery.
After replacing the backlight filter, it was time to test. But what about the connector on the board--it LOOKS pretty mangled--but is it functional? If possible, we'd like to save David the extra cost of an add-on connector job if we can. This is where the iBridge really shines. By connecting the iBridge into the damaged connector, we could measure THROUGH the connector. We found that despite the damage, this connector has normal relative resistance on all lines and indeed was functional. Phone A got a new screen installed and booted up working.
We then turned our attention to Phone B. We used a multimeter to check the backlight anode line in this phone. There was no short or open at PP_LCM_BL_ANODE. In our experience, this tells us that the phone probably does not have a backlight problem at all. Next, we checked to make sure the phone could boot and be detected by iTunes (it could) so we looked for image. There was no image. This is pretty common, backlight and image can be independent of each other. When a phone isn't detecting the screen and producing image, it is often also not producing backlight for the same reason. Pro Tip: Always troubleshoot image before backlight!
To do this, we installed the iBridge Flex to assist us in checking for all the signature faults that cause no image in iPhone 6s.
Using diode mode, we checked for the common short on the image power rail PP5V7_LCM_AVDDH Result: Not short.
We checked for the signature problem of open line fault at PP1V8_LCM_CONN that often occurs in concert with backlight problems. Result: Not open.
We then did a voltage test using the iBridge with the device on and the screen active. I noticed that both the PP5V7_LCM_AVDDH and the PP5V7_MESON_AVDDH lines had 0v. Why would these lines not be active in this phone, when we already know that they are not short or open? On a line that is not short or open, but is missing its voltage we look to see if "the guy upstairs is doing his job" In this case, that would be the chestnut display driver. Do you think we should "change chestnut, bro?" Hold your horses! Wouldn't it be unusual for chestnut to simply stop working in a phone that was in for a screen replacement. One practical diagnostic strategy that we strongly believe in as we train students at Practical Board Repair School is to ask "How did it come to be this way?"
Something must have happened during the screen change that lead to either the death of chestnut, OR ONE OF THE SIGNALS that chestnut relies on when deciding whether or not to produce his outputs. Chestnut himself could be perfectly fine, but will not produce his outputs if the CPU doesn't tell him to, or if he can't detect the screen, etc. Aha, that reminds me--we still have to check for the good old AP_TO_LCM_RESET_L short within the CPU. We removed the battery and put the multimeter back in diode mode with the iBridge installed. A quick check showed that indeed, AP_TO_LCM_RESET_L was 0.00 DMU, aka short to ground.
With the image reset line short, it is held in the low position permanently. For those of you like me in your forties, this is like trying to play your old Atari 2600 with your friend's finger holding down the reset button. The CPU believes the screen is in reset, and therefore does not send signals to chestnut to compel it to generate image power outputs. Hence, PP5V7_LCM_AVDDH and PP5V7_MESON_AVDDH lines were 0v.
We have seen this same AP_TO_LCM_RESET_L short a few times before. In some of those cases, the history was "This happened right after I accidentally connected an iPhone 6 screen to my iPhone 6s connector" Aha--a clue. We know a little bit about how the screen signals "I am here" to the CPU in the iPhone 6. The iBridge has helped us learn even more by allowing us to study live voltages with the screen connected. Here is a recent video where we learned a little bit more than what we already knew. What we understand is that some voltages, like PHOTON_ALIVE are delivered to the screen connector from the logic board. This voltage asks the screen "Hey buddy, you here?" The screen itself routes this voltage through the screen and back out to the connector and board to signal "Yup, I'm here"
In an iPhone with a short on AP_TO_LCM_RESET_L, the actual location of the short is within the die of the CPU itself. We know that CPUs are brain-like chips made of wispy thin structures for transmitting ones and zeros. It is very easy to electrically damage a CPU. Routing even relatively low voltages into the CPU data lines is like connecting a fire hose to a balloon made of cotton candy. Disaster. So how did Phone B come to have a short on the AP_TO_LCM_RESET_L at a screen change? I think we will get a clue by looking at the actual screen that David was using on Phone B.
That'll do it. And now the entire story becomes clear. David's technician installed a screen on iPhone A without disconnecting the battery. There was a transient short created during the installation when ground momentarily brushed against screen connector pin 33 that always has live battery voltage. This short unleashed a burst of electricity that overheated and damaged FL4211--the backlight filter, creating an open circuit on the backlight line. David's technician tried to troubleshoot why the display was dark by multiple connect/disconnect cycles that ultimately lead to connector damage on both the logic board and the screen itself. After deciding that this 6s was not showing any display with multiple screens, David's tech wanted to know if the first screen was defective, so he installed it on Phone B. He pressed the damaged screen connector into Phone B and hit the power button. The reset line was normal, momentarily, which allowed for the normal activation of chestnut and consequently---chestnut boosted the PP5v7_MESON_AVDDH line to 5.7v. Due to the connector damage, this 5.7v was effectively bridged through the connector and injected into the CPU down the adjacent AP_TO_LCM_RESET_L line. And here is a rare case where we get to legitimately use everyone's favorite board repair term, FRIED. Yes, the likely case is that the 5.7 volts FRIED THE CPU in that spot, creating a permanent short to ground of the AP_TO_LCM_RESET_L line. Now, no matter what screen you install on iPhone B, it will not show any image because AP_TO_LCM_RESET_L is permanently in the low position. This is a situation that we can recover data from with a lot of effort, like in this video. But clearing an internal short within the die of the CPU is not a viable "for the sake of the phone" repair. He's dead, Jim.
What I like about this case is that even though we've seen this AP_TO_LCM_RESET_L short before, with the iBridge we can now easily investigate the consequences of some of our favorite signature problems on other aspects of the system. In this case, the iBridge showed us that when AP_TO_LCM_RESET_L is low, the chestnut PP5v7 and PN5v7 outputs are not generated. Before the iBridge, it would have been unwieldy and time consuming to make these voltage measurements. Have you discovered anything new since you've been using your iBridge flex sets? Comment below to tell us all about it. Still haven't ordered yours? What are you waiting for!