How is the SSD installed – M4 Mac Mini (2024)
ifixit.com
ifixit.com
Their install guide is interesting http://www.polysoft.fr/StudioDrive/MacStudio_SSD_test_note.p...
I had to buy an off the shelf M1 Studio due to a hardware failure, so I couldn’t wait for the lead time for one with more RAM and storage. It has been borderline unusable due to so many things requiring local storage — can’t even symlink to an external NVMe. (Many apps, but also Backblaze metadata and iMessage attachments)
In any case even with all the added complexity/cost from different providers you still come out 2 times cheaper than Apple, which is completely crazy considering they definitely don't pay the NAND modules as high and don't need to pay for a special reverse engineered board.
This is what leaves a very bad taste in mouth with Apple stuff. Even with all the absurd amounts of money they are making on base devices they still can't help themselves and price basic stuff in a predatory way.
Everyone would be completely fine with a little bit of markup, but it currently is at an insane level.
Not a user-friendly process, I think.
They are the same course, imo. That it's been done in this way is typical behaviour we see from them, but expected/unsurprising behaviour does not necessarily make it a justification.
My experience is only around trying to help a friend get the data off a fried board in a macbook, the desktops might be more standard.
Edit: not a 2230. Different pins.
I don't think the "U" notch and pins are the same.
But happy to be wrong :)
Standard has 4 this one has 11(?)
In order to upgrade the hard drive in a classic 68k or PPC Mac you had to either buy an Apple drive, or buy a third party disk utility capable of setting up a non-Apple drive.
Many Apple-specific retailers would bundle in the software needed to use a new hard drive with the purchase of one but you'd better believe they baked the cost in to the price of the drive.
If all you had was a Mac, your System disks, and a new non-Apple hard drive you 100% could not use it.
I used FWB Hard Disk Toolkit, but it was >$100.
Both Apple HD SC Setup and Drive Setup have been patched by the retro community to work with any drive, but in the 80s and 90s this was not the case.
For the most part (there were EXTREMELY rare exceptions) the only strings that would pass were specific custom models of widely-available drives that were only sold to apple.
FWB did not check.
You could use ResEdit to haxxor the list:
http://www.euronet.nl/users/ernstoud/drvsetup.html
I cannot believe that page is still online. I first referenced it ON a PPC Mac running netscape. My flabbers are gasted.
That's because there were vendor specific settings that it wanted to set and some vendor differences in getting some information. There was support for a generic drive but there was a compile time option that controlled whether or not it allowed it and I don't remember what that was set to on release builds.
Source: I worked at the contracting company that wrote FWB's early versions and am looking at the source right now to jog my memory. :-) (My current employer owns the IP of that contracting company and knows and is OK with me having these ancient files).
Here's the command dispatch table from mid 1990 to show what drives it recognized and which ones needed vendor specific routines.
struct command_switch
{
int (*read)(); /* read command */
int (*write)(); /* write command */
int (*format)(); /* format device */
int (*findsize)(); /* find limits of device */
int (*setup)(); /* setup device paramters */
int (*verblock)(); /* verify track */
int (*ready)(); /* test unit ready */
int (*rezero)(); /* rezero */
int (*reassign)(); /* reassign block */
} CmdSwitch[9] =
{
{
OMTIread, OMTIwrite, OMTIformat, OMTIfindsize,
OMTIsetup, OMTIvertrack, OMTIready, OMTIrezero, OMTIreassign
},
{
SEAGATEread, SEAGATEwrite, SEAGATEformat, SEAGATEfindsize,
SEAGATEsetup, SEAGATEvertrack, SEAGATEready, SEAGATErezero, SEAGATEreassign
},
{/* QUANTUM */
SEAGATEread, SEAGATEwrite, SEAGATEformat, QUANTUMfindsize,
QUANTUMsetup, SEAGATEvertrack, SEAGATEready, SEAGATErezero, SEAGATEreassign
},
{/* CONNER */
SEAGATEread, SEAGATEwrite, SEAGATEformat, CONNERfindsize,
nulldev, SEAGATEvertrack, SEAGATEready, SEAGATErezero, SEAGATEreassign
},
{/* WREN3 */
SEAGATEread, SEAGATEwrite, SEAGATEformat, QUANTUMfindsize,
CDCsetup, SEAGATEvertrack, SEAGATEready, SEAGATErezero, SEAGATEreassign
},
{/* DEFAULT */
SEAGATEread, SEAGATEwrite, SEAGATEformat, GENERICfindsize,
nulldev, SEAGATEvertrack, SEAGATEready, SEAGATErezero, SEAGATEreassign
},
{/* SyQuest */
SEAGATEread, SEAGATEwrite, SEAGATEformat, SYQUESTfindsize,
nulldev, SEAGATEvertrack, SEAGATEready, SEAGATErezero, SEAGATEreassign
},
{/* Sony optical */
SEAGATEread, SEAGATEwrite, SONYformat, SONYfindsize,
SONYsetup, SEAGATEvertrack, SEAGATEready, SEAGATErezero, SEAGATEreassign
},
{/* Ricoh optical */
SEAGATEread, SEAGATEwrite, RICOHformat, RICOHfindsize,
nulldev, SEAGATEvertrack, SEAGATEready, SEAGATErezero, RICOHreassign
},
};
If there was custom firmware, I think it was just the model ID or some other license key system.
Commercial storage systems (appliances, SANs) also have approved firmware versions (though not necessarily custom) as there are things like firmware bugs.
Supposedly some of the things that ZFS has found is SANs that should 'know better' doing dumb stuff like retrieving the wrong LBAs: the checksum according to the SAN is correct, because it is data+cksum in one sector, but it was the wrong sector that was read. ZFS caught this because its checksums are in the parent directory structure and not tied to the data blocks themselves. This also catches things like the two sides of a mirror being out of sync.
https://www.apple.com/shop/product/MR393AM/A/apple-2tb-ssd-u...
In practice if you have a Mac Pro you're better off just getting a PCIe SSD instead for a fraction of the price. Retail for a very fast 2TB SSD is ~$150 while Apple wants $1000.
I still buy Apple anyway, since the amortized cost over the life of owning the produce is lower in my experience. Plus it runs a lot cooler and quieter.
Words are not enough to eulogize how far Apple has fallen. Woz couldn't leave that dumpster fire fast enough.
Third party boards have already been designed and made, and they work. The main issue is that people don't really know much about NAND, so they assume it's like an SSD (it's not) or eMMC (it's not) where you plug it in and some hardware magic turns it into a disk (it doesn't).
What happens here is that the secure enclave, cryptographic accelerator and flash controllers are all packaged together. This gives you sick speeds and performance while also making it more secure than your average OPAL TCG trash that often isn't even implemented at all.
To make the embedded flash controller work with the NAND, you need two things:
- NAND chips that actually work with the controller. Not all chips do, especially low-end crappy bulk NAND chips won't do. People keep track of the NAND chips that apple uses and you can buy those and it will be fine
- The data in the NAND needs to make sense (so either have it empty or populate it ahead of time)
- Using USB-C you tell the embedded controller to revive and it will setup the NAND for you, this is available to anyone (so not locked behind some secret sauce).
In a way, this is similar to having a classic SSD, replacing the NAND chips on that SSD, and then telling the SSD controller that it has new NAND chips. Or to having a really old MFM/RLL hard drive before IDE and SCSI existed.
The storage device really is just dumb storage and the smarts are all in the controller. While we have moved this around back and forth a few times over the years, there is no conclusive benefit to one or the other. Having more smarts on the device means there is also more problems/variability on the now 'smarter' device. This is especially problematic during data recovery, or when you want your data storage to be trustworthy.
A lot of the hardware work has been done by dosdude1, and a table of NAND chips like this one: https://forums.macrumors.com/threads/apple-silicon-soldered-... can show you some options.
The reason Apple does this is the same reason as ever: if they feel like this is the best way to make some sort of experience work, they will do it. And if they make fat stacks of cash in the process, they aren't going to be sad about it. This is something that isn't exclusive to Apple, but most manufacturers don't have to luxury to design their own hardware, they have to integrate with a lot of partners, use reference designs or maybe even outsource their hardware to some white label manufacturer. This is also why you see more glue, foam pads, smaller components etc all over the industry: it gets the manufactured devices to do the thing they want it to do. If PCs became modular in the process, that was a side-effect, not much of a goal. (The goal was to upsell crap later down the line so your market is bigger)
As for pricing, that is just whatever the market will bear and not all that much related to the cost of raw material. This is of course not new and is default practice in most commercial businesses. So cheap components does not equal cheap products. (but cheap components might equal low quality products in some cases)
Apples SSD performance is nothing special, and if they wisely don't want to trust OPAL TCG they can encrypt the data in their own trusted silicon before handing it over to the untrusted SSD controller. That's pretty much what the Playstation 5 does, it supports standard NVMe drives but the disk encryption is done in custom Sony silicon so the third party SSD controller never sees the plaintext.
I'm not convinced there's any purpose for the way Apple does their storage aside from vendor lock-in.
As for that other methods may work (making the chain longer by introducing a separate bus, two transceivers, an additional controller from another vendor, extra firmware, extra power buses), they have done that in the past. In practically every shape:
- SCSI
- ATA/IDE
- SATA
- NVMe over custom physical port before M.2 was broadly available with the same specs, but probably also cheaper for them
So technology wise, it's not like they haven't gone back and forth with many, many combinations.
Business-wise:
> I'm not convinced there's any purpose for the way Apple does their storage aside from vendor lock-in.
First of all, I highly doubt Apple gives a shit. Their model of lock-in is making a better combined experience than the competition. If they could do that with some random western digital black SSD while making the same amount of money, they would do it. But more importantly: if Apple would do that, they would still charge you $1000 for that SSD, even if it's only 1TB. The concept of lock-in is making it so that people don't want to pay the cost of leaving. You can't lock people in with just the stick (an SSD-shaped stick with a $1000 price tag) if you don't have a carrot.
Secondly: I don't want or need to convince you, but trying to shoehorn a business in such a one-dimensional take is not exactly a good way to spend your energy.
Of course, if you started with the requirements that is has to be the exact same thing you cannot meet it, precisely because what Apple does is highly proprietary and dependent on their implementation of the controller and the encryption that is handled by their own silicon.
However, the fact is that there are SSDs that are faster than what Apple provides and there are also CPUs that are better at encryption or just competitive enough with what Apple does. You can't get a 1-to-1 solution because that a not something anyone actually desires but you can get a solution that is just as good if not better.
It is always weird to me that when it comes to Apple people want to go with the arbitrary requirements they dictated while dismissing any requirement of another competing solution, like modularity.
> First of all, I highly doubt Apple gives a shit. Their model of lock-in is making a better combined experience than the competition
What you say makes no sense and is highly contradictory in its nature. They charge the price they do for upgrades because there is literally no other way to do so. The reason they don't use some random WD SSD is precisely because if they would it would be trivial for anyone to bypass their predatory pricing and very few would actually use their upgrade options and they definitely could not realistically extract so much money from their power users. It's completely crazy that you believe otherwise even though they started soldering SSD/RAM on the Intel MacBooks, way before there was any technical rationalization about it (in marketing and by their fans).
The reason they do it is because of money; any other benefits are just a side effect. They get to save money from the vertical integration: suddenly they don't have to pay for a 3rd party integration of controller and software development (with the associated markup) and they go down the value chain, just buying NAND chips which are a lower value commodity with less supplier power. Since they already pay for their own silicon, the added cost of the controller on the silicon is minimal. And they get to dictate the price of storage/RAM upgrades since they become by definition their own supplier with no competition.
If they really wanted, they could have a solution that would allow their own custom NAND board and also support 3rd party SSDs for anyone who wanted (even if it would have some downsides). The added development cost would be minimal but of course they would never do that because realistically almost nobody would spend 3-4 times as much on storage no matter how much better their solution would be in theory.
When it comes to the lock-in, it is completely built into the software. They get to win because once people get used to and dependent on a particular software stack it is pretty costly to switch to another solution. And the cost is one of the things people hate the most: it is time and struggle to learn/adapt to new way of doing things. If you have been working a certain way for a very long time you also need to unlearn a good amount of it, which is a struggle (think about changing keyboard layout for example). Most people are not ready to pay this cost so they end up preferring to pay with cash because in the end they got the cash by selling their time and they find it is more efficient for them to go this route. And this is a fact Apple knows very well, in fact most of the Apple customers (at least when it comes to the Mac) are people who were wealthy enough to afford their hardware in the first place, those are people who generally have a large leverage on their time, meaning they have a great multiplier which is precisely why Apple feel that they can get away with this anticompetitive behavior and they are completely right.
That doesn't make any of it better, it is still highly abusive and extortionate behavior which is funny because it is a company that markets itself as highly virtuous. All the while they extract more money out of economies all other the place, in a way that is highly inefficient for society because they don't give back anything and the value they provide is highly out of whack with what they extract.
As for the carrot, there is some of it left but it has become very rotten. They make really nice hardware that's for sure, and the recent development of Apple Silicon has brough much needed value (if you don't care about compatibility outside of Apple focused software that is). But the reality is that outside of their base model that offer decent value, because of the cost of upgrades they wouldn't sell that much hardware if macOS could run on other hardware legally and without too much effort. Hackintosh would never have been a thing otherwise and we know that for a fact because historically when they licensed Mac OS to clone, their hardware sales tanked and then they later came with various technical mitigation to make it too costly to legally make a clone.
So, in a way I agree that there is good part of a carrot left but that's mostly if you can be happy with their base default configuration, if you have a requirement outside of that their pricing make it so that the carrot leaves a very bad taste in your mouth and you wish you could go another route even if it meant having to compromise on some other stuff.
Windows isn't the crown jewel - the real value is doc/xls/ppt that wrap people's data and bundling of everything into a MS license that marginalizes 3rd party entrants like Slack.
Microsoft has won because they were more competitive in bringing a desirable solution at a particular price, in other words they bring good value and even though some seem too ideologically tainted to understand it the whole world recognized it as such.
The open standard Office formats work fine and Microsoft supports them just fine. The real reason they still win is because their software is generally just better to use. Many have used the competing open-source solution and end up paying for the Microsoft stuff because it is just a much worse experience that is not worth wasting your time for the very competitive pricing Microsoft offer. They win because they provide more value than free stuff at the price they offer, it is just as simple as that.
You can see the bundling as predatory or you can see it as a competitive answer. Microsoft has seen that there was a market for some types of software, duplicated a competitive enough solution and sell it cheaper in order to bring more value, that's pretty much it. Even if they are legally prevented from bundling, I very much doubt it would change anything because they would just price the standalone solution in a competitive enough manner and would still win. They largely do that by the way, you can get plenty of their stuff in a standalone way. Even in the age of cloud/SASS subscription, Microsoft still sells standalone Office suite that you can very well use for years on end with zero need to upgrade (exactly what I do for my grandma, because she likes Publisher).
And if the bundling legal precedent is set, their competitors like Apple and Google are in way more trouble because they abuse the bundling way more.
It's hilarious the irrational hate Microsoft gets when they are objectively not that bad. Sure, they have some terrible behavior like any big corp, but by those standards, Apple is so much worse it is really stunning to see the difference in treatment.
There's no hate going on, irrational or otherwise. They have been really bad, and Teams vs Slack is a good example of how they are still bad in some instances. Not illegal bad, although maybe that, but just bad for competition, kept in place not by evergreen product quality (unlike Apple) but by 100 million IT departments all with mouse-only skills who couldn't choose anything but Microsoft without putting themselves out of a job.
https://support.apple.com/guide/security/boot-process-secac7...
Yeah dude. They only do this for greed. That said, a Micron 7450 Max is $270 for 800GB.
If they could save themselves the huge mountains of money to takes to create their own flash controllers, firmware, fabric and hardware design, and just buy a Micron 7450 Max and put that inside the computer, and still charge you $1000, they would do it. Because it means more profit for them, which is what they exist for.
But they didn't.
The same concept applies to say, the Red Mags, which are essentially mSATA SSDs in a metal box. Red takes on all the responsibility, claims, (reputational) damage for their product, but if they feel like they can make it happen with an mSATA SSD and still make lots of money, they will do it. They could use an SLC enterprise SSD, make a little less money and take a little less risk.
But they didn't.
This argument holds water only if the price were appropriate, not extortionate. Apple has shown the world for years they prioritize profit through control and not using industry standards.
The PCI/e, NVMe (+more) standards give you the best experience: a modularized peripheral connection system.
Appealing the capitalism, standards allow markets to exist and to compete in each component category. By not using the standard, Apple is saying "I don't want to compete on storage, I'd rather do it my own way and charge as much $$$ as possible."
Appealing to best practices programming, these standards are like interfaces. If you were to make a really useful *nix command line tool, but decided to not allow the output to be easily used by downstream programs through `|` pipelines by reason of "good experience" ("Why would you want to leave my good program? The flowers are so nice!", people would rightfully see you as stuck up and having the gall to think you know what's best for end users.
Appealing to psychology, Apple is not your friend. They are a powerful profit-driven company that dodges taxes and that fosters a cult-like belief system. I say this as an iPhone user and as a latest generation white-plastic macbook user. They make good devices. How much better could their devices be if they didn't spend so much effort in anti right to repair and anti compatibility?
> give you the best experience
Hardly. For practically every normal user, the best experience is when they don't have to think about hardware or specs at all, forever. Doesn't just apply to computers either. If anything, people doing the things they want with the hardware not existing at all, that would be even better since they didn't want to deal with it in the first place. Reality is of course that the things they want to do are implemented in software, and software runs on hardware. And they still would rather have it not exist, and not think about it. And anything you do that forces them to think about it is a detractor, even if it was "The Right Thing".
It's why Android beats Windows, and mobile devices like phones and tablets beat PCs; there is a whole lot less dealing with the details of the hardware and the software, and more "doing what you intended to do", whatever that might be. Even if the hardware and software is almost universally worse (at the very least somewhere until desktop-class ARM came around).
As far as all your other concepts, reality says the thing Apple does in the Apple way makes them more money than everyone else. So either they are already doing the things you want them to do and they work, or they are not doing the things you want them to do because they don't work to generate the same concentration of money (which is what they exist for, as I already wrote).
It really is cult-like behavior, there is no other way to put it. There is the famous piece of writing by Umberto Eco about Macs being catholic (https://www.simongrant.org/web/eco.html) and it really feels like that, their users tend to be highly religious, just not to a traditional religion.
You can use the stuff and like it, find it better in many ways but it is hard to actually accept the commercial practice with a straight face.
Apple isn't some magic company where if they have extreme margins the people that buy the stuff do mental gymnastics where the margins are not extreme. Of course they are. But people buying it have accepted that they can vote with their wallet, and they still vote to pay and get the product, rather than not pay and hope the signal is strong enough to lower the margins and maybe buy a product later if it is cheaper.
Most products you buy are not priced just a sliver higher than the cost of production, or even the cost of production plus R&D. They are priced at the maximum revenue possible. As long as Apple can sell NAND chips at 1000% markup, they will do it. And as long as it doesn't hurt people enough, they will keep buying it. Look at the clothing nonsense, bags, shoes etc. You can get a plastic bag made out of 10 cents of material and 2 cents of manufacturing line time but you'll have to pay $500 because it's special to some people. Or most of the plastic shoes that Nike makes, they aren't really made with $200 of plastic, that's going to be maybe $10 of material in total. There might be a much higher manufacturing cost, so maybe that's another $10. That means there's $180 going straight to Nike. But you don't see people coming to HN to complain about that, because somehow that's okay.
Where you assign some loud Apple users (let's call that the vocal minority) some cult or religious status for their behaviour, you're doing a disservice to yourself and the rest of the people that just want to stuff and will pay the price even if they don't really like it. Presumably because saving $2k in cost on a computer costs them more in UX, productivity or forced change, which they don't want, and the company extracts as much money as they can for it ($1k wheels anyone?).
Maybe the vocal minority thinks a computer can be a status symbol. But that's mostly just in their mind. The same high school mentality (and group behaviour) applies to the shoes or bags mentioned above.
Trying to make this seem like a special Apple case is pretty useless. It's capitalism at work, and you're part of it. Whether you can live with that is mostly just up to you and not much of a technology or startup discussion. We might even end up more in the area of philosophy than anything else.
This is absolutely huge news. I wonder if Apple will do something similar for the Studio, Pro, or -- dare I hope? -- even the Macbook Pros in the future? I can't imagine allowing this 'trapdoor' of money savings is a huge problem for profits since most businesses would never bother messing around with warranties for a spec upgrade. But this is absolutely MASSIVE for consumers. Just put in a little extra work and you can save hundreds or even thousands of dollars compared to Apple's upgrade pricing, for a good enough end result.
Not to mention the fact that this must also save _Apple itself_ an insane amount of money for repairs! Instead of throwing away the entire logic board, the CPU, the soldered-on RAM, and the soldered-on SSD whenever any of those components fail, you can just replace the malfunctioning part. Who'd have thunk (other than, y'know, every single computer company from 1980-2015)?
I would also of course love to see this upgradeability return to RAM. I'm curious if anyone more knowledgable than myself might know if the SoC/Apple Silicon Unified Memory system makes that more difficult, or if we've just accepted it because Apple Says So.
And while I'm on the subject of non-upgradeable RAM: does anyone know why no SBCs, from Raspberry Pi to Orange Pearl Jam Cake to Milk, allow for upgradeable RAM? Surely it's possible in the SBC form factor?
So Apple gets the best of both worlds. They can keep charging their high storage costs, but they themselves gets the flexibility of easily upgrading storage due to having this as a module.
Apple's was actually super late to the nvme storage party, and they're not even remotely at the top wrt maximum write/read bandwidth compared to pcie Gen 5 m.2 ssds you can buy for ~$200-300/TB.
So yeah, Apple's ssds cost at least 3x (and perform measurably worse). And I might add: gen 5 is already what, 1 1/2 yrs old now?
I'm not sure what you're talking about here, because Apple was one of the first PC OEMs to adopt PCIe storage, and very quickly followed that up with a transition to NVMe. This was circa 2015. They just didn't use the M.2 connector, and when they introduced the T2 chip they stopped using third-party SSDs in favor of their own built-in NVMe SSD controller.
Also, I don't think PCIe gen5 SSDs are being shipped in laptops yet (at least not in any significant volume), on account of the extra speed being completely not worth the power cost. Much like the transition from gen3 to gen4, availability of SSDs that are only suitable for desktops with large heatsinks comes long before availability of reasonably-efficient SSD controllers. Eg. Samsung's PM9E1 SSD for PC OEMs only started mass production a month ago: https://news.samsung.com/global/samsung-starts-mass-producti...
[0] from IDT, later sold to PMC-Sierra, then Microsemi, now Microchip's Flashtec product line
[1] Samsung XS1715 was at least the first to pass compliance testing at UNH-IOL
[2] Intel rebranded their enterprise NVMe drives as the Intel SSD 750 marketed for PC enthusiasts, but the real beginning of consumer NVMe was Samsung's SM951 when they started transitioning away from PCIe AHCI (used for compatibility with systems lacking NVMe-aware drivers and firmware).