This raises the question of WHY does Apple solder their SSDs into the mother board?
Who gains? -- I see no material benefit to Apple -- I see no benefit to Apple's customers
But, as this points out: there is potential risk and harm
-- The drive can wear out
-- The laptop could die -- at which point the data on it is unrecoverable.
-- The user could have a need for more storage -- but is locked into what was originally purchased.
Returning to removeable, replaceable storage (even Apple has to do it) solves a lot of problems but (as far as I can see) causes none -- at least none that are material.
It's slightly cheaper for Apple as long as the NAND doesn't die and the logic board need replacing because of a bug like this. It also means people have to fork out for Apple's extortionate (4x at least) storage costs when people buy their Mac, which likely far offsets the cost of warranty replacements - and is almost certainly why Apple solders it on.
This raises the question of WHY does Apple solder their SSDs into the mother board?
Who gains? -- I see no material benefit to Apple -- I see no benefit to Apple's customers
But, as this points out: there is potential risk and harm
-- The drive can wear out
-- The laptop could die -- at which point the data on it is unrecoverable.
-- The user could have a need for more storage -- but is locked into what was originally purchased.
Returning to removeable, replaceable storage (even Apple has to do it) solves a lot of problems but (as far as I can see) causes none -- at least none that are material.
This has been debated ad nauseam, and yes there are benefits, including smaller packaging and improved cooling, increased power efficiency, better reliability, reduced costs, and higher transfer speeds. Apple is not soldering components simply to annoy you.
Of course, the downside of soldering is it stimulates unfounded paranoia in the uninformed.
Wow you've just reached David Icke levels of talking out your ass.
Smaller packaging is evidently unnecessary, since other, thinner laptops have removable SSDs. Oh and what about the Mini, that really needs to be compromised to save space does it? Many Intel NUCs are hugely more capable than the Mini with upgradable storage and RAM and similar footprint. Improved cooling is rubbish. Thermal pads between chassis and NAND are just as effective on M.2 as directly soldered. Power efficiency is not increased. The same components are just soldered on directly, that doesn't make it more efficient. Costs for Apple are decreased slightly, whilst costs for consumers are increased. How does this benefit customers? Higher transfer speeds is also rubbish. Whilst theoretically possible Apple doesn't take advantage of this. The T2 equipped iMacs with SSDs transfer at the same speed as the M1 Macs. The older non-T* Macs are capable of higher speeds too with faster SSDs. The interconnect is not the bottleneck.
Increased reliability is about the only customer-facing advantage, since the sockets are a possible (albeit absolutely minuscule) point of failure. Usually reseating is fine, and that's an easy fix compared to NAND that will eventually die and result in the whole computer being thrown away. Very eco-friendly. But Apple only really cares about being environmentally conscious that when it can increase profits, rather than increasing the longevity of a machine.
So in actual fact it is the informed in which paranoia is stimulated, rather than clueless know-alls with their heads up Apple's ass. Because the informed know that when bugs like this happen, and the SSD gets its lifetime eaten, it's impossible to replace it.
"At this stage, there is no outward indication that Apple is planning to widely discuss the matter. Given the potential for headlines declaring that the "New Mac mini will be dead in a year" or similar, it seems like a PR issue that Apple may want to intercept instead of letting it snowball."
I disagree with that opinion. This whole issue does not appear to be a valid concern. For one thing, who is to say that the rate of increase on usage will stay constant? Isn't it possible it may be stuck at 1% or 3% for a long time? The 1% might even be a placeholder figure. Perhaps it has to show some usage. Or, perhaps the figures are inaccurate as stated in the article. In the real world, is "wearing out" a Mac SSD really a concern? I've never experienced it, nor have I heard of anyone else experiencing it. I have experienced issues with non-OEM "upgrade" drives, but that was a particular brand/line that had known issues.
I really doubt we'll see those headlines unless there is much more data on the topic, and/or people's drives actually start failing.
This raises the question of WHY does Apple solder their SSDs into the mother board?
Who gains? -- I see no material benefit to Apple -- I see no benefit to Apple's customers
But, as this points out: there is potential risk and harm
-- The drive can wear out
-- The laptop could die -- at which point the data on it is unrecoverable.
-- The user could have a need for more storage -- but is locked into what was originally purchased.
Returning to removeable, replaceable storage (even Apple has to do it) solves a lot of problems but (as far as I can see) causes none -- at least none that are material.
This has been debated ad nauseam, and yes there are benefits, including smaller packaging and improved cooling, increased power efficiency, better reliability, reduced costs, and higher transfer speeds. Apple is not soldering components simply to annoy you.
Of course, the downside of soldering is it stimulates unfounded paranoia in the uninformed.
You cite a bunch of imaginative and, at best, weak reasons for soldering as SSD (except that its cheaper).
And the best argument you come up with against a replaceable SSD is an ad hominum.
This raises the question of WHY does Apple solder their SSDs into the mother board?
Who gains? -- I see no material benefit to Apple -- I see no benefit to Apple's customers
But, as this points out: there is potential risk and harm
-- The drive can wear out
-- The laptop could die -- at which point the data on it is unrecoverable.
-- The user could have a need for more storage -- but is locked into what was originally purchased.
Returning to removeable, replaceable storage (even Apple has to do it) solves a lot of problems but (as far as I can see) causes none -- at least none that are material.
Clearly no experience in electrical engineering.
Soldered SSD has literally 1/3 the number of possible failure points.
Why doesn’t Nvidia or AMD use sockets for their GPU chips or memory? Why solder them to the video card? If it quits working the entire video card is ruined.
Not seeing this issue. I suspect some people are using old SMARTMonTools that don't calculate it correctly. Here are my statistics for 2018 Mac mini 32GB 512GB, 2019 16" MacBook Pro 64GB 4TB, 2018 13" MacBook Pro 16GB 512GB, 2020 M1 Mac mini 8GB 512GB, 2020 M1 MacBook Air 16GB 512GB:
I always thought that S.M.A.R.T. (Self-Monitoring, Analysis, and Reporting Technology) was originally designed for original platter type of hard drives and has been tweaked over the years to accommodate SSD technology. Kinda like iTunes being used for music and then it evolved into the conduit into syncing and backing up your iOS device.
It’s time to retire SMART and come up with something written from the ground up that supports flash memory in a robust way.
This raises the question of WHY does Apple solder their SSDs into the mother board?
Who gains? -- I see no material benefit to Apple -- I see no benefit to Apple's customers
But, as this points out: there is potential risk and harm
-- The drive can wear out
-- The laptop could die -- at which point the data on it is unrecoverable.
-- The user could have a need for more storage -- but is locked into what was originally purchased.
Returning to removeable, replaceable storage (even Apple has to do it) solves a lot of problems but (as far as I can see) causes none -- at least none that are material.
Clearly no experience in electrical engineering.
Soldered SSD has literally 1/3 the number of possible failure points.
Why doesn’t Nvidia or AMD use sockets for their GPU chips or memory? Why solder them to the video card? If it quits working the entire video card is ruined.
You're comparing a system SSD to a Graphics card? And suggesting that there is high rate of failure of replaceable SSDs?
This raises the question of WHY does Apple solder their SSDs into the mother board?
Who gains? -- I see no material benefit to Apple -- I see no benefit to Apple's customers
But, as this points out: there is potential risk and harm
-- The drive can wear out
-- The laptop could die -- at which point the data on it is unrecoverable.
-- The user could have a need for more storage -- but is locked into what was originally purchased.
Returning to removeable, replaceable storage (even Apple has to do it) solves a lot of problems but (as far as I can see) causes none -- at least none that are material.
This has been debated ad nauseam, and yes there are benefits, including smaller packaging and improved cooling, increased power efficiency, better reliability, reduced costs, and higher transfer speeds. Apple is not soldering components simply to annoy you.
Of course, the downside of soldering is it stimulates unfounded paranoia in the uninformed.
Besides there is a very simple solution - boot from an external SSD. Heck a reasonable 1TB at US$300 is far cheaper then the US$600 apple wants.
This raises the question of WHY does Apple solder their SSDs into the mother board?
Who gains? -- I see no material benefit to Apple -- I see no benefit to Apple's customers
But, as this points out: there is potential risk and harm
-- The drive can wear out
-- The laptop could die -- at which point the data on it is unrecoverable.
-- The user could have a need for more storage -- but is locked into what was originally purchased.
Returning to removeable, replaceable storage (even Apple has to do it) solves a lot of problems but (as far as I can see) causes none -- at least none that are material.
This has been debated ad nauseam, and yes there are benefits, including smaller packaging and improved cooling, increased power efficiency, better reliability, reduced costs, and higher transfer speeds. Apple is not soldering components simply to annoy you.
Of course, the downside of soldering is it stimulates unfounded paranoia in the uninformed.
Besides there is a very simple solution - boot from an external SSD. Heck a reasonable 1TB at US$300 is far cheaper then the US$600 apple wants.
That won't save your data if you dumped your Starbuck's on the keyboard.
If Apple is going to solder the SSD they should at least initiate automatic backups to iCloud like on the iPhone and iPad.
I'm pretty certain that the latest release of smartmontools is accurately reporting the S.M.A.R.T. data from the NVMe SSDs in the M1 Macs. I've done some experiments to verify it.
I used:
smartctl -a disk0
to get a starting value of the "Data Units Written" and then wrote 400 GB to the internal SSD and then used the tool a second time to get final number. TLDR, it came out to almost exactly 400 GB using the tool. While I was writing the data, I watched the disk writes from the Activity Monitor Disk tab and also saw 400 GB being written.
Here are the numbers:
starting units: 12,724,656 [6.51 TB]
ending units: 13,506,170 [6.91 TB]
total: 781,514 [0.4 TB]
The math is simple arithmetic. subtract the starting units from the ending units and multiply by 512*1000. The result is 400,135 MB which is pretty much exactly 400 GB. This matches the 0.4TB of the result of subtracting 6.51 TB from 6.91 TB.
After my experimentation and research, I'm totally convinced that the latest version of smartmontools is working correctly and is returning the correct data on writes to the SSD and that offset to the Percentage Used value is the correct offset into the byte array. If Apple's NVMe SSD controller is not writing the correct value of the %used, that is on them but the actual number of bytes written to the SSD checks out and it is easy to repeat my experiment.
Edit: A point on the ambiguity of Power On Hours. The specification says this, "Power On Hours: Contains the number of power-on hours. This may not include time that the controller was powered and in a non-operational power state." So there is no mystery why the number for Power On Hours is low. The new M1 Macs are very low power machines that take every effort to reduce battery usage by turning off unused subsystems. This includes the SSD. And the SSD is very fast so the total on time is very short in general.
Comments
Wow you've just reached David Icke levels of talking out your ass.
Smaller packaging is evidently unnecessary, since other, thinner laptops have removable SSDs. Oh and what about the Mini, that really needs to be compromised to save space does it? Many Intel NUCs are hugely more capable than the Mini with upgradable storage and RAM and similar footprint. Improved cooling is rubbish. Thermal pads between chassis and NAND are just as effective on M.2 as directly soldered. Power efficiency is not increased. The same components are just soldered on directly, that doesn't make it more efficient. Costs for Apple are decreased slightly, whilst costs for consumers are increased. How does this benefit customers? Higher transfer speeds is also rubbish. Whilst theoretically possible Apple doesn't take advantage of this. The T2 equipped iMacs with SSDs transfer at the same speed as the M1 Macs. The older non-T* Macs are capable of higher speeds too with faster SSDs. The interconnect is not the bottleneck.
Increased reliability is about the only customer-facing advantage, since the sockets are a possible (albeit absolutely minuscule) point of failure. Usually reseating is fine, and that's an easy fix compared to NAND that will eventually die and result in the whole computer being thrown away. Very eco-friendly. But Apple only really cares about being environmentally conscious that when it can increase profits, rather than increasing the longevity of a machine.
So in actual fact it is the informed in which paranoia is stimulated, rather than clueless know-alls with their heads up Apple's ass. Because the informed know that when bugs like this happen, and the SSD gets its lifetime eaten, it's impossible to replace it.
"At this stage, there is no outward indication that Apple is planning to widely discuss the matter. Given the potential for headlines declaring that the "New Mac mini will be dead in a year" or similar, it seems like a PR issue that Apple may want to intercept instead of letting it snowball."
I disagree with that opinion. This whole issue does not appear to be a valid concern. For one thing, who is to say that the rate of increase on usage will stay constant? Isn't it possible it may be stuck at 1% or 3% for a long time? The 1% might even be a placeholder figure. Perhaps it has to show some usage. Or, perhaps the figures are inaccurate as stated in the article. In the real world, is "wearing out" a Mac SSD really a concern? I've never experienced it, nor have I heard of anyone else experiencing it. I have experienced issues with non-OEM "upgrade" drives, but that was a particular brand/line that had known issues.
I really doubt we'll see those headlines unless there is much more data on the topic, and/or people's drives actually start failing.
Soldered SSD has literally 1/3 the number of possible failure points.
Why doesn’t Nvidia or AMD use sockets for their GPU chips or memory? Why solder them to the video card? If it quits working the entire video card is ruined.
I used:
to get a starting value of the "Data Units Written" and then wrote 400 GB to the internal SSD and then used the tool a second time to get final number. TLDR, it came out to almost exactly 400 GB using the tool. While I was writing the data, I watched the disk writes from the Activity Monitor Disk tab and also saw 400 GB being written.
Here are the numbers:
starting units: 12,724,656 [6.51 TB]
ending units: 13,506,170 [6.91 TB]
total: 781,514 [0.4 TB]
The math is simple arithmetic. subtract the starting units from the ending units and multiply by 512*1000. The result is 400,135 MB which is pretty much exactly 400 GB. This matches the 0.4TB of the result of subtracting 6.51 TB from 6.91 TB.
I've also looked at the smartmontools code and verified that the offsets they are using within the byte array returned by the macOS IOKit api are correct according to the NVMe specification. https://nvmexpress.org/wp-content/uploads/NVM-Express-1_4-2019.06.10-Ratified.pdf (See section 5.14.1.2).
After my experimentation and research, I'm totally convinced that the latest version of smartmontools is working correctly and is returning the correct data on writes to the SSD and that offset to the Percentage Used value is the correct offset into the byte array. If Apple's NVMe SSD controller is not writing the correct value of the %used, that is on them but the actual number of bytes written to the SSD checks out and it is easy to repeat my experiment.
Edit: A point on the ambiguity of Power On Hours. The specification says this, "Power On Hours: Contains the number of power-on hours. This may not include time that the controller was powered and in a non-operational power state." So there is no mystery why the number for Power On Hours is low. The new M1 Macs are very low power machines that take every effort to reduce battery usage by turning off unused subsystems. This includes the SSD. And the SSD is very fast so the total on time is very short in general.