Visual inspection

From Repair Wiki
Revision as of 16:34, 2 May 2021 by Skolander (talk | contribs)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to navigation Jump to search

Visual inspection for logic board repairs[edit | edit source]

Trust your eyes![edit | edit source]

Most of the time, your eyes are more important than your brain.  Always inspect the edges of the logic board! 95% of the time the problem is somewhere in that area.

There is no greater time-sink for the new technician than over-intellectualizing a problem. Trying to decode what every single signal abbreviation means and trying to reverse engineer how an undocumented IC works, when the problem is staring you in the face. Do you need to know what a rusted, green, rotting probe point is for in order to fix it? NO!

Thinking before acting is a great thing, and one should always be thinking. However, when it comes to board repair, thinking almost always comes after seeing. 

What you see will tell a story, and that story will often teach you a lesson for the next board. If you have no WiFi, and have corrosion on a probe point - you now know that the signal going through that probe point is related to WiFi. Corrosion, rust, burned traces, and green dust are beautiful little markers that alert you to where your issue is. Without the corrosion there, you might have never found that this signal has anything to do with wifi. With the corrosion, you now know that this area is broken. Once you repair the damage and see that WiFi is back, you can add to your knowledge base that AP_RESET_L has something to do with WiFi. AP. "Ah - ACCESS POINT - THAT'S WHAT AP STANDS FOR!" 

Your eyes figured out what was broken before your brain knew that AP_RESET_L mattered. As a new technician, your eyes will be doing far more work than your brain, and will also serve as a teacher.

Probe point damage[edit | edit source]

Probe points are going to allow signals and power rails to pass from point to point, while giving you an exposed point that you can touch with your multimeter probe for measurement purposes. These exposed points can easily corrode and break away. It requires careful attention to detail and training to be able to scooby doo these probe points on a highly populated board - particularly when you need to figure out, out of 100 points, which are the ones that are causing your problem, and which are likely to cause future problems. 

Let's take the board below as an example. Knowledge of the one wire circuit and its role in creating a green light on the charger, would be necessary if using your brain for this board repair. Knowledge of how a logic gate works and the ability to see the data on a logic analyzer would help to see what is going on in the SYS_ONEWIRE line. To fix this board on a brains level, requires quite a bit of knowledge and experience. 

To fix this board on an eyes level, one must simply scrape the mask off the traces on each side of the probe point circled in yellow, tin with solder, and put a jumper over it. You need not know the voltages that must be present, or the purposes/identities of the many components. You don't even have to know what the circuit does. 

This doesn't mean you shouldn't go back after you're done to try and learn. See if you can infer how the signal/voltage that was missing as a result of the damage might be the reason that the machine is presenting the problem it has. This is a never-ending journey. 


Inspection1.jpg

Here is another example of probe points that work - but might stop working later.

Corrosion14.jpg Corrosion14 labeled.jpg

Here are some examples of corroded probe points (middle of the image) from a liquid damaged MacBook Air from a video with a lot of corrosion and several components need repair or replacement (https://www.youtube.com/watch?v=62lnfgMqmQU)

Screenshot Fixing a very badly water damaged Macbook Air.png Screenshot 2 Fixing a very badly water damaged Macbook Air.png

An example from a video where a corroded probe point prevented a Macbook Air from turning on (https://www.youtube.com/watch?v=uMNDHjo_vWQ) 

Screenshot Can spec of green dust kill a Macbook?.png

Short circuit by moisture combining with dust/hair/general debris[edit | edit source]

This may look like dust that does negligible damage, but it is partially short circuiting the pins. If the chip has not been damaged by short circuiting a high voltage pin to a low voltage pin, this can usually be fixed with some flux & solder, hitting each side with the T30-KN tip.

Corrosionqfn.png Corrosionqfnpointedout.jpg

Recognizing corrosion underneath a BGA package[edit | edit source]

Sometimes, corrosion may be underneath chips. Most modern technology makes use of BGA chipsets - ball grid array, where all of the conducting points are underneath the chip. The chip is sitting on a bunch of balls that conduct electricity to pads on the logic board. Sometimes it will be obvious, but in cases like below, it will require a trained eye to recognize the corrosion. Can you see it?

Inspection2.jpg

Working components & solder joints that won't work for long[edit | edit source]

Inspection3.jpg

This probe point and component have corrosion of the worst kind - it is just good enough to last for testing, but will fail when the customer starts using it. This solder joint will not stand the test of time, besides perhaps at L2 Computer with their 30-day warranty. This will come back for warranty service.

Since the corrosion was not addressed before ultrasonic cleaning, it is much more difficult to spot, as the green lump of dust has been washed away by the cleaner. It is imperative to address most corrosion before ultrasonically cleaning the board - or, to make a note of where it was, so you will know to come back and check the area later. 

Once the board is ultrasonically cleaned, and the evidence of the spill removed - the corroded areas are more difficult to spot. Out of sight, is out of mind!

The component & its solder pad are important - the probe point isn't. Why does this probe point not matter?[edit | edit source]

Look at the earlier picture. The bad probe point has a signal passing through it - in that case, from a connector, to the chip on the bottom right. Here, the probe point is for testing/measurement purposes only. That path goes to P3V42G3H_BOOST - which only goes to that one capacitor, and nothing else! The schematic and the board view below confirms that pin 3 of U7090 goes to pin 1 of C7094, not any other component/chip. The trace is already intact from the chip to the capacitor. The probe point that is under the capacitor serves no functional purpose. However, the capacitor & its solder pad do, and must be reworked. 

Inspection4.jpg
Inspection5.jpg

Another working component & solder joint that will not last[edit | edit source]

The liquid damage was repaired in one area, but the board was cleaned before it could be fixed in another. This joint technically conducts electricity, but if you smacked this component with the tweezer, it would immediately come off the board. This is less of a solder joint and more of a pile of sand waiting to be blown away by a breeze.

Inspection6.jpg

More bad areas post-ultrasonic cleaning[edit | edit source]

This board was working, but will soon have a dead CPU. Can you guess why? Note how the bottom joint of the capacitor, while having a lackluster solder joint, could likely be saved by applying some flux, and running back and forth with a T30-KN micropencil tip on a Hakko 2032, reinvigorating the joint. However, the top joint of that capacitor is too far gone. Removing it will showcase that the pad under it is destroyed, as a result of the corrosion surrounding it that has likely made its way under the solder joint. NEVER LEAVE THIS ON A BOARD! On the topmost capacitor, note the solder ball and the soot right above it - this is an excellent indication that this component should be replaced, and that something bad happened here. Keep in mind that tiny stray solder balls can be found on many perfectly working boards due to many reasons, and that in and of itself should not be attributed to a low quality factory assembly, poor rework, or a short circuit damage.

Corrosion10 unedited.jpg Corrosion10.jpg

Below you can see that in spite of the area being cleaned, a bad capacitor is still on the board - as well as several questionable vias. The vias with rust marks on the inside may still work, but should be checked diligently before calling the board repaired. Often you can run a piece of 46-50 awg wire through the hole, scrape the coating on each side, add solder on each side, make sure the wire is going where the via intended for it to go, and recreate the connection.

Corrosion13.jpg Corrosion13 spelled out.jpg

Bad legs on chips[edit | edit source]

Corrosion15.jpgCorrosion15labeled.jpg

Inspect at an angle as well as straight on. Tilting the board will reveal hidden corrosion[edit | edit source]

This QFN looks fine straight on. When you angle the board, you can inspect the joints and see little bits of green on the two on the right. This is visible at an angle, but not visible when viewing straight on.

Corrosion12.jpg Corrosion11.jpg

Examples of poor newbie soldering[edit | edit source]

You should have even joints that look like they naturally flowed into place - not clumps as pictured below on the left. On the right, you can see the same connector after the anchor pins have been soldered on better. You should not have voids where there is too little solder - it should be an even amount on each side, flowed into place naturally. Flux and heat will assist with these joints - add flux, hold the connector down with tweezers, and tap the joint with the iron so that it is touching both the pin AND the pad, and hold for half a second.

Solderclump.jpg

You should not have components that are barely held on the board by an incomplete solder joint. The one on the top is a mild example, the one on the bottom is an extreme example. It would be easy to flick this off with a fingernail tap.

Solderingbad02.jpg Solderingbad02edited.jpg