Beginner Troubleshooting

From LogiWiki - Your Wiki for all things Apple MacBook and iPhone related
Jump to navigation Jump to search

You have a laptop that does not power on, powers on but shows no image, crashes, etc… You want to fix it but you don't know where to start.

There are 2 situations:

  1. You just want your laptop fixed
  2. You want to learn about fixing laptops

Sorry, but if you're in situation number 1, we recommend you to use the services of a trustworthy professional repair shop.

Unfortunately, if you're not serious about learning, you will waste your time and money, and you will make us waste our time as well.

If you are in situation number 2, then please continue and read until the end. This serves both as directions on how to deal with your laptop problems and how to post on a forum, Discord, or this discussion page.

Remember that this is a place for mutual help. It's free as in not paid, so when you ask for something you have to realize people invest their time out of their day for you, without counterpart.

And when you help someone else, you do it for free without expecting anything in return.

It's also free as in freedom. You are free to ask questions, but people are also free to answer them or not. If you want to help this community, the best way is to learn and help people back when you're confident enough.

Tip: if you make effort in your communication and thinking process, people are more likely to want to help.

Link to Piernov's original on BadCaps

1. Where to begin

1.1 Identifying the hardware you are dealing with

It is very important for you and us to know exactly the machine you are working with and what is inside.

1.1.1 Machine model number

Find the full model number of the machine.

For example it may be "HP Pavilion 17-f166nf". "HP 17" or "HP Pavilion" is not enough, there are hundreds of laptop that fit this name.

If you find multiple model numbers, write them all.

1.1.2 Specifications

Find what CPU, GPU, RAM, storage device is actually in the machine. It is not always relevant to the problem, but it helps confirm the parts are compatible or recognize known unreliable parts.

  • CPU: specify full model number. Not "Core i7", but "Core i7-3615QM". If you can't find it, read the sSpec from the CPU package, it starts with SL or SR like "SR0MP".
  • GPU: specify full model number, like "Radeon HD 6490M". If you can't find it, read the part number on the GPU die. It starts with "216-" on mobile AMD GPU like "216-0809000". For NVidia it can be like "GF-GO7300-B-N-A3", "G84-602-A2", "N10M-GE2-S".
  • RAM: specify capacity and frequency like 1600MHz or bandwidth like 12800. For DDR3, confirm if it's PC3 or PC3L.
  • Storage device: capacity and type (IDE HDD, SATA HDD, SATA SSD, M.2 SATA SSD, M.2 NVMe SSD, eMMC…). Model number isn't necessarily required but in some cases it can help to know if there is a firmware upgrade available for exemple.

1.1.3 Motherboard model number

Motherboard model number is the most useful if you are troubleshooting a motherboard problem. It helps finding pictures, schematics, boardviews, known issues…
There is always a number written on the board itself. In some cases there's also a sticker that tells a variant.
There are several different board manufacturers, common ones are:

  • Quanta: found in HP, Dell, Acer, Toshiba, a few Lenovo… Model number on the board is "DA0xxxxMByyy" (first 0 is optional), "xxxxx" is the actual model and "yyy" contains the revision. For example DAY11AMB6E0 for Quanta Y11A rev E. Warning: 0 and O are hard to distinguish when reading from the board, for example DA0OP6MB6D0 for Quanta OP6.
  • Compal: found in HP, Dell, Acer, Toshiba, Lenovo… Model number is "LA-xxxxP" (P is optional), the older, the lower the number is. Like LA-4082P from 2007 or LA-E541P from 2017.
  • LCFC: found in Lenovo, it's actually Compal. Model number is "NM-xxxx"
  • Wistron: found in Acer (Wistron was part of Acer a long time ago), some Dell, HP, Lenovo… Model number is a "xxxxx-x" (x being a digit), often associated with a name. For example "Richie MB 11241-1".
  • Inventec: Toshiba, some Acer, HP… Model number is often "xxyyzz" with "x" being a letter, "y" a digit and "z" an optional letter. There's also a "6050xxxxxxxx-MB-xx" number that can be easier to spot. For example "SA10E" and "6050A2052401-MB-A04" for Inventec San Antonio 10E.
  • Foxconn: Sony, some HP. HP have a weird model number like "CHICAGO_HR_HPC MV_MB_V1" or "PM_I_HPC_S MV_MB_V3". Sony come with an "MBX-xxxx" number that's easy to identify, and another number, like "MBX-202" and "M790".
  • Pegatron: a few Acer and Toshiba. Acer in general have a "xxyyzz" with "x" being a letter, "y" a digit and "z" an optional letter. Toshiba have a weird model number like "PLF/PLR/CSF/CSR UMA"

The following manufacturers make their own boards (except maybe a few models here and there):

  • Apple: board model number is "820-xxxxx", like "820-3462". It's written black on black so it's hard to spot.
  • Asus: board model number follows the same format as the laptop model number, it can be the same or similar, and sometimes different.
  • MSI: model number is "MS-xxyxz", with x being digits, y a digit or a letter and z a optional digit for the revision. Some boards have 2 different model numbers like "MS-16J5" and "MS-1795" (with a 1 appended for the revision).
  • Samsung: model number is name associated with a "BA41-xxxxxx" number, like "Bremen-D" and "BA41-01197A".
  • Clevo: they manufacture laptops for a lot of "small local" brands. There are often 2 model numbers on the board, like "6-71-W25S0-D02" and "W251ESMB-D0".
  • Toshiba: a few business grade machines only. Something like "FLESY3" or "FHNSY1"

HannStar don't design boards, they manufacture the PCB. "J MV-4" and "94-V0" are not board model numbers, they are generic markings about PCB characteristics.

If unsure, send high-resolution pictures of the board. They are always welcome in any case.

1.2 History, symptoms, state of the machine…

Always collect the most information you can about what happened to the machine. What's not working right? When did the problem happen? What was being done with the machine at that moment? Did it start acting up months ago? Was it ever in contact with liquid? Was it ever dropped? Was it repaired in the past already? etc…

These give essential hints to where the problem can be. But of course, don't entirely trust what people tell you. Confirm the symptoms for yourself and write them down.
Once you identified what's not working as it should, do a visual inspection of the machine.

Look for any physical damage (bent, broken, missing stuff) and liquid damage, both on the chassis and the board, and take pictures if there is any.

Always make sure the correct AC adapter was used with the machine. Genuine original AC adapter of the correct brand and correct power rating. Even if the connector is correct, it doesn't mean it will work and it can even damage the machine.

1.3 Troubleshooting

As a first step you should always disconnect the battery. Also check RTC coin cell battery and perform a clear CMOS.

1.3.1 Parts

Always exclude the possibility of a part other than the motherboard being the problem before troubleshooting the board. Swap what you can, disconnect what you can and see if the behaviour changes. Write down your findings.

Battery-related issues are 99% of the time the battery itself. Even if the battery is new. There are so many garbage batteries on the market that you sometimes get 2 or even 3 bad batteries in a row. This is especially true for Apple machines.

1.3.2 Board

Here we will give just a few pointers to what kind of troubleshooting has to be done if you suspect a board issue.
Powers on but no display

  • Check if it's no backlight, no image or both
    • If no image:
      • Check if external display works, try with internal LCD disconnected as well.
      • Measure voltage on LCD power rail(s)
    • If no backlight but image is present
      • Try to locate the fuse (or resistor or inductor) to backlight rail and check if it's blown
      • Measure voltage on backlight rail, backlight enable signal, brightness adjustment signal
  • Check if laptop is POSTing: pressing caps lock key triggers caps lock LED, activity of USB drive blinks when turning on, audio plays, HDD is being read, etc… If it's not POSTing:
    • Check RAM compatibility
    • Reseat RAM and clean RAM slots
    • Measure exact voltage on CPU VCore
    • Check CPU Reset signal
    • Follow power sequence if it exists in the schematics
    • Follow list of power rails if it exists in the schematics

No power

  • Check if DC-in jack LED lights up
  • Check voltage on all the large inductors
  • Follow power sequence if it exists in the schematics
  • Follow list of power rails if it exists in the schematics

Note: don't forget to trigger power button when measuring for the rails that only come up when machine is turned on.

Powers on, display ok but something is wrong afterwards

  • Exclude software issue by booting from another drive with Windows 10/Linux/macOS


2. What not to do

2.1. Never reflow large BGA

It won't fix the problem since 99% of the time it's not an contact issue between the BGA and the logic board. There's a high risk of damage the board even further.

If you want to confirm if an AMD/NVidia chip is bad, you can heat it up to 200°C max for a minute and a half, no flux. It'll come back to life for a short while and then die again, because it's bad. In this case it has to be replaced.

2.2. Don't suspect that something is shorted with no reason

Shorted components are not the only failure mode of a laptop. Additionally, some places will measure low resistance to ground under normal operation. In general, high power low voltage (<2V) power rails will measure less than a few hundred ohms to ground. CPU VCore will measure a few ohms only. GPU VCore with newer NVidia GPU can even measure less than 1 ohm. It's normal. If you suspect a short to ground, always report your exact resistance to ground measurements.

2.3. Never randomly inject voltage

Voltage injection is used to find a short to ground. Only consider it after you found that there is a short to ground and after doing a visual inspection trying to find a broken component.

2.4. Never inject 19V

If the 19V main power rail is missing, there's a reason, often because a protection kicked in to avoid more damage. For example, the 19V power rail could be shorted to the CPU through a bad high-side MOSFET. In that case, the charging circuit protection kicks in and turns off the DC-in MOSFET. With enough luck, the CPU didn't take too much of a hit yet. If you inject 19V, you are sure to destroy it.

Always measure exact resistance to ground to see if there is a possibility of high-side MOSFET short. Always start at 1V and check if any of the large BGA (CPU/GPU/PCH) is warming up. Monitor the power consumption.

2.5. Never randomly replace components

If you don't follow a proper troubleshooting process, randomly replacing components will rarely solve the problem and it can create additional issues.

2.6. Never randomly bridge MOSFETs/fuse/components

Bridging components can easily lead to the destruction of the board without no way to repair it.

If it's a fuse, replace it with a matching fuse. If you don't, it will not be able to protect the board and if the source of the problem is still there, some other component will go up in flames.

If it's a MOSFET, its purpose is to stop current from flowing through it when turned off, and allow current to flow through it when turned on. It's not necessarily always turned on, otherwise it wouldn't be there. If it's turned off when it shouldn't be, then there is an issue elsewhere in the circuit, bridging the MOSFET won't solve it.

2.7. Don't apply power to the board with missing components

As a general rule, if a component is on the board, it's because it's useful for something. Engineers don't like adding a bunch of useless components on the board, it's harder to draw the schematics and route on the PCB.

Some of them can be removed without creating problem if power is applied, but others can kill the board with no possibility of repair.

For example, if you remove a buck controller or a MOSFET driver, the MOSFET gates will be floating. They'll charge up enough so that the MOSFET turns on. Of course, you don't want the 19V from the main power rail to go straight through the CPU VCore MOSFET to the CPU.