MacBook Pro mid-2012 frequent Kernal Panics

ketchup6 wrote:

In each case, I tried applying pressure to the RAM slot plate where HWTech circled. In one instance, this produced the aforementioned graphical errors, followed by the aforementioned boot loop involving a blank screen + 3 beeps.

That most likely indicates the memory slot nearest the Bottom Case should not be used. If the memory is not fully latched in place, then it may also cause this failure. The plastic arms holding the memory have very small triangular protrusions to hold the memory down in its latched state. These protrusions can easily be worn down to the point the memory is no longer fully held in place, plus this latch may loosen the latch for the memory in the slot nearest the Keyboard since the two arms are part of the same plastic arm.

If the module was fully latched in place and those protrusions are definitely holding the memory in place, then I would avoid using the memory slot nearest the Bottom Case.

FYI, pressing on the base of the memory slot will only cause an error when memory is installed into the slot nearest the Bottom Case. If memory is only installed into the other slot nearest the Keyboard, then it should not cause an issue when pressing on the base of the memory slot....at least I never encountered it.

Regarding the solder that sometimes breaks loose, would I be able to observe this fracture somewhere? Would I feel it while pressing? Overall the two RAM slots feel consistent, as does their mounting.

No, I've never actually seen the cracks in the solder joints, but there are a lot of pins which are not visible since there are about three rows of pins.

If both RAM sticks seemingly produce the same errors, could both of them have been corrupted at once? Is that even possible? Or does this (likely) mean something has gone awry with the RAM slots' mounting, the connection between RAM and the motherboard, etc?

Since it appears the memory slot nearest the Bottom Case is most likely bad, only rely on test results by testing memory in the other slot nearest the Keyboard. If you get a crash with memory installed in this slot, then that memory module is likely bad. If both modules have an issue in the slot nearest the keyboard, then I would suspect a Logic Board issue since it is extremely rare both memory modules fail. If they really have both failed, then it could mean the Logic Board damaged them both.

In the former case, I guess I can buy some RAM and hope to find different results. Which seems unlikely.

I would not spend any money on it. You can only use the slot nearest the Keyboard and you have two memory modules. One of those memory modules should be good. If both seem bad, then it is time to retire this laptop.

In the latter case, what can I do about this? Someone I know has a retired mid '12 MBP 2.5GHz (same model afaik). Could I somehow harvest parts off of theirs?

If it is definitely the same model series (2012), then yes. Just make sure they are the same model...they can have different CPU & GPU, but the model series must be the same. You cannot use a 2011 model Logic Board.

At what point does this become a hopeless endeavor?

If neither of those two memory modules work in the slot nearest the Keyboard without crashing the laptop (the other slot nearest the Bottom Case must remain empty), then it is time to call it.

Before I move to solve either of these hypotheticals, should I try reinstalling MacOS? Could a bad ram stick have somehow corrupted my boot drive? IE the consistent boot failure patterns may be the result of a now corrupted boot drive, which in turn hides a singularly bad RAM stick.

Yes, bad memory can easily corrupt macOS and the data written to the drive. I would first test the single memory module installed into the slot nearest the Keyboard by running Memtest86 using all of the CPU options (all cores, sequential, round robin). If it fails Memtest86 under these conditions, then there is no reason to perform a clean install since none of that is in use when using a bootable Memtest86 USB stick.

Btw another round of Apple Hardware Test resulted in zero detected issues.

Not surprising since its memory tests are much shorter and may not include all the variations that Memtest86 does.

I don't see any issue with the SSD's health other than the UDMA CRC Error Count you mentioned which may have occurred when using the bad SATA Cable. Now that you have a report of that RAW value, see if it increases after you encounter more Kernel Panics. If the RAW value remains at 639, then it is likely Ok.

The WD hard drive is worn out though. The Load Cycle Count has exceeded the manufacturer's expectations. This attribute is just a Lifetime health attribute which means exceeding it does not necessarily mean the drive is bad, but in my own personal experience....once the Load Cycle Count reaches this point, the drive begins to develop performance issues and may also begin to show other odd issues. I will no longer use a hard drive with an expired Load Cycle Count like this one even for a test drive. I wouldn't expect this WD drive to be causing your problem, although some drive failures can impact the boot drive. Just make sure that First Aid doesn't report any file system errors when scanning the WD hard drive (make sure to manually inspect the report for any unfixed errors since the First Aid summary can lie). File system issues can cause Kernel Panics.

You can remove the SSD and try booting it externally using a USB to SATA Adapter, drive dock, or enclosure. This can eliminate any possible issues involving the internal SATA Cable or compatibility issues between the laptop & this SSD. Samsung EVO SSDs don't have a very good track record with Macs when installed internally. You may also want to check Samsung's website to see if your SSD has any firmware updates available....I think there may be a firmware update for your SSD. I would wait to update the firmware until after you look at all other possible options to resolve the Kernel Panics though since I doubt the SSD is the cause, but the SSD could be bricked if the firmware update is interrupted.

https://semiconductor.samsung.com/us/consumer-storage/support/tools/

Check the "Retired" folder within that "/Library/Logs/DiagnosticReports" folder for any other Kernel Panic logs....sometimes they can be moved there fairly quickly. macOS has a terrible logging system these days. Look for any file names beginning with "kernel", or ending in ".panic", or even ".gpurestart".

I didn't really see any pattern between the two posted Kernel Panic logs. Usually if there is no pattern, then that usually indicates a memory issue. How long have you had the memory upgrade? What is the make & model of the memory? Macs can be very picky about the memory they use.

Also, the memory slot nearest the Bottom Case tends to develop cracked solder joints after a while. Pressing on the center of the Bottom Case can sometimes produce a Kernel Panic, or cause the laptop to freeze or to display graphical glitches. A much better test would be to actually press on the plastic base of the memory slot while the laptop is powered on (I like to do so while the laptop is sitting at the Option Boot screen). It can be a bit tricky to press on the plastic base of the memory slot (usually has a system serial number sticker or barcode on it). While pressing on the base of the memory slot, look at the screen to see if it shows any graphical glitches, and try moving the pointer to see if the laptop is frozen. Just make sure not to let any metallic item or surface touch any of the laptop's internals. I personally like to sit the laptop on its right side to perform this procedure.

You may also want to make sure both memory modules are securely latched in place. When inserting the second module located nearest the Bottom Case, it can sometimes cause the other module nearest the Keyboard to become unlatched. I like to gently press on the empty edge of the first module (make sure not to touch any components as they may break off) to make sure it is latched....use a non-metallic item like a wooden/plastic toothpick or a spudger (aka black stick).

You can also try running Memtest86 to check for bad memory. Use the downloaded Memtest86 .img file as a souce for Etcher (Mac/Windows/Linux) which will create a bootable Memtest86 USB stick. Option Boot the Memtest86 USB stick and select the orange icon labeled "EFI". Once the test begins to run you can disconnect the USB stick if you wish. I have recently discovered that I've been finding issues more easily when I modify the Memtest86 CPU options to "Sequential" or "Round Robin"....seems like I detect issues more quickly & more repeatable. I haven't received actual error messages with this mode, but the test has frozen the system which is just as useful.

Take look at all your reports. What you are trying to determine is whether there is a TREND.

Case A) There is a Trend: This same panic occurs in each case, with the same extensions present, in the same BSD process. A trend can indicate corrupted software or a Hardware problem that can be tracked down and solved.

Case B) There is a decided NON-Trend. The panics occur "all over the Map" for different panic reasons and in different BSD processes, with different extensions present.

A Non-Trend suggests you may have RAM memory problems. MacOS slightly randomizes the load point of key routines on each startup, as a hedge against fixed-address attacks. This causes a marginal memory cell to move into different routines each time your Mac starts up. This causes random and bizarre non-correlation in your panic reports from one run to the next.

Executive summary: From my reading of your reports, you appear to have a decided Non-Trend, suggesting ongoing RAM memory problems.

--------

There is one more 'cheap' Brute Force thing you can try, that MAY change your symptoms, not necessarily for the better. Swap the RAM modules with each other.

ketchup6 wrote:

I ran Memtest86, which resulted in no errors and PASS in big green letters. Sadly there was input delay, so I didn't understand how to navigate the post-test menu until the results' specifics were seemingly lost. Regardless I watched the process throughout the day, and nothing seemed out of order.

I never worry about the report screen, just whether any errors are listed on screen while the tests are running.

While on occasion I have seen the system freeze using the default Memtest86 settings, it is only when I modify the Memtest86's CPU settings when I see the computer freeze during testing fairly regularly & fairly quickly at least on the USB-C models. I haven't had occasion to try it recently on the older models.

My memory is two sticks manufactured by Crucial. DDR3, 1600 MHz, 8GB each (16 total). S/N 0xE0F7B277. I installed these mid 2020 if I recall correctly.

Crucial is one of the two recommended brands. As long as you purchased the exact part numbers recommended by using one of the two tools on the Crucial website to identify compatible memory, then you should be fine. If you only bought the memory based on technical specifications, then the memory may not be fully compatible although if you haven't had any issues in three years...it is probably compatible. Yes, Macs are that picky about their memory.

My memory sticks didn't appear to have any solder damage nor improper latching.

It is actually the solder joints of the memory slots connecting to the Logic Board which develop cracks from repeated pressure placed on the bottom case.

I tried pressing on their plastic green base while the machine was on, however no issues seemed to come of that.

I thought the plastic base of the memory slot on the Logic Board was usually a dark brown or black color, many times with an Apple sticker on the plastic. Here is a picture of a Logic Board (different model, but good enough for an example) where I circled the base of the memory slot to press on.

Thank you for pointing out my other drive's LCC, which I'd only seen for the first time after running DriveDX. I will try to replace it asap. My thoughts would be to simply replace it with another HDD with relatively high RPM, as it's simply a secondary drive I reserve for raw files. Do you have any recommendations here?

Not a lot of options for hard drives these days. I used to prefer the Hitachi or Toshiba drives....don't really trust WD these days (they now make the Hitachi branded drives). Not fond of Samsung hard drives, would rather have a WD drive.

My hope is that I can get my computer back to reliable working condition. If this naive goal is possible, I'd like to get it whatever drive would best suit it for a few+ years of service. I assume putting a second SSD in it, merely for storage, would be unnecessary.

While an SSD can work in the optical drive bay of a 2012 model, it may not always be 100% compatible. I know there have been issues in the past, but usually they involved the 2011 models I believe.

No luck with the "Retired" folder producing any lost logs. I remember noticing this issue previously -- what few crashes I had weren't producing logs where the internet said they should be -- and didn't know what to make of it. For clarity I've checked both user and system library/log/diagnostic reports folders, so I don't believe I was checking the wrong one. And they seem to be generating fine now, after having wiped and reinstalled the OS.

Sometimes the logs are just not saved, and other times macOS just deletes the logs quickly.

Here are the logs for the most recent three crashes: Two yesterday and one today.

I agree with @Grant's analysis of the Kernel Panic logs. It is what I originally suspect which is why I spent so much time focusing on memory in my last post. Either one of the memory slots is bad, or one of the memory modules.

I would remove the memory in the slot nearest the Bottom Case just because that slot has the highest probability of being a problem in my own personal experience. That is where I would start anyway, but I would also first try running Memtest86 with the other CPU options ("Sequential" and "Round Robin") to see if they report an error or cause a system freeze. Why? It is a lot easier if you can repeat a failure with the test when you begin removing modules or moving them around. Not definitive, but really helps.

the graphics goes crazy because on that model, the display RAM is the system RAM.

"an "integrated" Intel HD Graphics 4000 graphics processor that shares system memory"

In the US, these vendors:

http://MacSales.com (OWC) is still the most-recommended supplier here, because they have good return policies and a lab full of Macs. They pride themselves on their service, so they do not sell stuff that is so "difficult" it needs a lot of hand-holding to get working.

OWC sells their own private-label SSD drives, which are pretty good and reasonably priced, and you know they will work in your Mac.

Look for the "My Upgrades" link on their main page, and it will lead you to parts that will work in your specific model Mac.

http://DataMemorySystems.com is less often recommended, but very close in service and real-world testing, and has historically been a little cheaper for comparable products with the same terms.

http://crucial.com is often recommended because they have a great track record of supplying stuff that works first time, every time. Personally, I am not as fond of them because they use their database of what sales "stick" rather than a lab full of Macs to obtain their results. Their recommendations tend to be more conservative than others, and they sometimes discontinue parts for older Macs before other sites do.

--------

All of these vendors have a lookup program where you enter the specs of your computer, and they recommend the parts that will work under these terms:

"It works in your Mac,

or your money Back." (it costs them money to get it wrong, so they make certain they get it right.)

Also, a Lifetime Warranty is standard, and their added vendor stickers on the modules are all you need for proof of purchase if you ever need a return.