Apple DTK teardown

[Kostask]

You can find out if there is a fan by listening to the system as it heats up. If you hear a fan, then it has a fan, if not, then it doesn't. Again, why would it matter if the DTK Mac Mini has user replaceable modules, its not like anybody is going to be upgrading it, or getting inside it. And even if it did, it is probably more of reflection on the A12Z using off SoC RAM than anything else. You will learn nothing about the AS Macs, because the DTK Mac Mini is only representative in the software sense, of the soon to come AS Macs.

You could see a smaller AS Mac Mini, for sure. That would be the regular/consumer level AS Mac Mini. I still contend that there will be a Server level AS Mac Mini, that will use the Mac Mini case in its current size, because of the infrastructure that has grown up around the current case size over the years.

 

[johngwheeler]

LOL they don't, they run on Azure and Google Cloud Platform.

Mostly Amazon Web Services actually, although that have also used (and probably still do) some MS Azure and GCP.

 

[Jorbanead]

You can find out if there is a fan by listening to the system as it heats up.

Except I don’t have a DTK, and so far nobody has told us if there is a fan or not. A picture would confirm one way or the other.

Again, why would it matter if the DTK Mac Mini has user replaceable modules, its not like anybody is going to be upgrading it, or getting inside it.

Because it could possibly indicate that they are still planning on using user replaceable ram vs soldered on ram. You would assume they would just solder it on if nobody plans to upgrade it, so if it was socketed it would just be interesting and fun to speculate. Yes it does not necessarily mean ASi macs will have user replaceable ram, I am aware.

We also don’t know what else may be inside the machine that could possibly offer other interesting changes to the current mini.

 

[ascender]

We’re not going to learn much if anything from a tear down as its just a DTK. They appear to have just thrown a SOC in to a Mac Mini chassis as that’s all you need for devs. Presumably they’ve taken the cheapest approach possible to these so it may be soldered RAM, but I’d be amazed if they took away upgradeable RAM from the Mini. There may be a fan because there’s a fan today. The purpose is for devs to start coding, its not a hardware test as such - that stuff will all be done in the usual Apple labs.

It would be interesting to see what’s inside it but doubtful if it will give much of a view as to their strategy.

 

[Cookie18]

You can find out if there is a fan by listening to the system as it heats up. If you hear a fan, then it has a fan, if not, then it doesn't. Again, why would it matter if the DTK Mac Mini has user replaceable modules, its not like anybody is going to be upgrading it, or getting inside it. And even if it did, it is probably more of reflection on the A12Z using off SoC RAM than anything else. You will learn nothing about the AS Macs, because the DTK Mac Mini is only representative in the software sense, of the soon to come AS Macs.

You could see a smaller AS Mac Mini, for sure. That would be the regular/consumer level AS Mac Mini. I still contend that there will be a Server level AS Mac Mini, that will use the Mac Mini case in its current size, because of the infrastructure that has grown up around the current case size over the years.

Actually, if the DTK Mac Mini doesn't have user upgradeable RAM that would tell us that ARM Macs are unlikely to have it since the current Mini already has upgradeable RAM. They'd have to go out of their way to remove it from the DTK. Although it isn't final there are a lot of questions that, while not being definitively answered, we could get a sense of which direction Apple is heading in with ARM Macs by knowing the insides of the DTK Mini.

 

[Kostask]

You make the assumption that the logic board in the DTK Mini is based on the 2018 and newer Minis. That isn't necessarily true. It could be, or it could very well be a new logic board specific to the DTK Mini.

The DTK Mini already deviates from the 2018 Mini in that it doesn't have a TB3 port, but does have 4 USB ports. which the A12Z cannot provide. To me, that indicates a new, DTK Mini logic board. It also means that due to the use of the A12Z SoC, it uses off SoC RAM.

I am of the belief that the AS Mac SoCs will come with 16GB RAM on the SoC, as well as on-board USB4/TB4 ports. This SoC, which I call the "entry level" SoC can be used to build a minimal complexity logic board for the basic MB/MBA, saving both space and power, and allowing for a lower price point (not huge, but $50-100), and very fast assembly.

So getting back to my points, don't look at the DTK as an indication of anything other than a very high level view of how the applications will run on the AS Macs. This is all it was designed to do, and will be disposed of when they are returned to Apple. It was the cheapest, fastest way for Apple to allow developers to do some verification work on a real, ARM based, Apple machine.

 

[Cookie18]

You make the assumption that the logic board in the DTK Mini is based on the 2018 and newer Minis. That isn't necessarily true. It could be, or it could very well be a new logic board specific to the DTK Mini.

The DTK Mini already deviates from the 2018 Mini in that it doesn't have a TB3 port, but does have 4 USB ports. which the A12Z cannot provide. To me, that indicates a new, DTK Mini logic board. It also means that due to the use of the A12Z SoC, it uses off SoC RAM.

I am of the belief that the AS Mac SoCs will come with 16GB RAM on the SoC, as well as on-board USB4/TB4 ports. This SoC, which I call the "entry level" SoC can be used to build a minimal complexity logic board for the basic MB/MBA, saving both space and power, and allowing for a lower price point (not huge, but $50-100), and very fast assembly.

So getting back to my points, don't look at the DTK as an indication of anything other than a very high level view of how the applications will run on the AS Macs. This is all it was designed to do, and will be disposed of when they are returned to Apple. It was the cheapest, fastest way for Apple to allow developers to do some verification work on a real, ARM based, Apple machine.

You are correct but you understand why people would want to know all of this for certain, right? If it has a different logic board that is new information. If it has RAM on the SoC that is also new information. The same for if it does/doesn't have a fan or anything else. What the DTK Mini does or does not have gives us an idea of where Macs are heading and that is something people want to know. It doesn't really matter if this is all a foolhardy effort, the hype train is rolling.

 

[Kostask]

I do understand, but I am trying to point out that they are trying to make the DTK Mini something it isn't, and was never intended to be. In other words, an indication of what will be in the AS Macs, and it isn't that. The DTK uses a 2 year old SoC from a tablet, and it is the best kludge that Apple could come up with to get developers going, and that is all it is. The AS Macs will use a current (A14 Generation), Mac specific (as per Apple statements) SoC, which has been optimized in a number of different ways compared to the A12Z.

 

[deconstruct60]

You are correct but you understand why people would want to know all of this for certain, right? If it has a different logic board that is new information.

New information about the DTK. It wouldn't necessarily be new information about a future Mac Mini.

The A12Z has relatively impoverished port support to what a Mini would require. There are two USB ports. Apple could have kludged the DTK ports with two USB hubs embedded internally. One doing the USB-C and Ethernet ports and the other doing the two USB-A ports. Is that going to be indicate of a future Apple Silicon SoC? No. Far more likely there is no Thunderbolt on the DTK because they can't possibly do it, rather than they were trying to control costs.

The DTK comes in one and only one configuration. A fixed set of RAM with no variation. A fixed storage capacity with no variation. A fixed Ethernet port with no variation. Is that going to be a constraint of a future Mini? Probably not. So why would be configuration requirements be reflected in a product with no variation at all?

Most likely to keep the DTK affordable, Apple mutated the iPad Pro board; not started from a Mini x86 board. Nor did they attempt to design something custom from scratch to "maximize" the synergy with the case.(it isn't a high volume product so "optimizing" the board buys what??? Nothing). It would be cheaper to take something that was already working and add some new adjustments . There is a decent chance that this could be a test harness board for the iPad Pro (something to throw the SoC into before the new case makes it out of Industrial design. Ports for debugging and connectivity before working with radios and Lightning. ).

the iPad Pro 3rd and 4th generations max out at 1TB SSD capacity. The DTK only has 512GB. Easily done with a iPad Pro board. The RAM would likely require a tweak, but since only a small double digit amount (16GB) not really a big leap at all. [ Not like have to get to 64 or 128GB levels. ]
Can decouple the Wifi to get rid of some of the "ruler" shape of the board. Attach the USB hubs to the USB connection point on the iPad Pro board.

Very good chance Apple made these DTK a substantively long time ago for internal use. ( or at least did a production run of a couple hundred. ). The capabilities here are all well within the era of the 3rd Gen iPad Pro. ( the A12X and A12Z are the exact same die. Apple could have gather some binned dies to do early A12Z in small numbers extremely easily. ) There is an implicit premise floating in these threads that this DTK is some kind of pre-production prototype of the Mini. Probably not. This is something they could have done even before starting work on the Mini ( much older). At which point it isn't particularly information of things going forward.

However, if put a high load on the Ethernet and USB ports at the same time and this system probably crumbles far more than an Apple Silicon system would.

If it has RAM on the SoC that is also new information.

The A12Z is known. There are iPad Pro teardowns. What going to see is the A12Z.

The A12Z isn't the Apple Silicon SoC. The AS SoC will have to deal with a wide variety of RAM capacities that span well past 16GB.

All Apple needs to coupled 16GB to the A12Z is two 64Gb DRAM chips and it is done. That doesn't scale very well past that though.

The same for if it does/doesn't have a fan or anything else. What the DTK Mini does or does not have gives us an idea of where Macs are heading

Not for Desktops. Apple's current desktop line up has DIMMs. (mostly so-DIMMs). Apple doesn't slavishly copy every single aspect of the iPhone design principles across the whole Mac line up.

and that is something people want to know. It doesn't really matter if this is all a foolhardy effort, the hype train is rolling.

I understand folks are curious. But the DTK is extremely likely to just be a large kludge that Apple whipped up to have something to test with. Something that was affordable to whip up just to have real hardware to test with. That's it.

Even Apple hints at that. In the system overview session of WWDC 2020 they make a reference to a "read me" file in the DTK distribution that highlights the differences in boot and some capabilities there are limitations of the DTK and not of Apple Siicon.

The DTK is highly lacking in being helpful for folks who need to write PCI-e card drivers and server different kinds of I/O peripherals. It is in no way the "end all , be all" of the Mac future capabilities.

 

[Cookie18]

New information about the DTK. It wouldn't necessarily be new information about a future Mac Mini.

The A12Z has relatively impoverished port support to what a Mini would require. There are two USB ports. Apple could have kludged the DTK ports with two USB hubs embedded internally. One doing the USB-C and Ethernet ports and the other doing the two USB-A ports. Is that going to be indicate of a future Apple Silicon SoC? No. Far more likely there is no Thunderbolt on the DTK because they can't possibly do it, rather than they were trying to control costs.

The DTK comes in one and only one configuration. A fixed set of RAM with no variation. A fixed storage capacity with no variation. A fixed Ethernet port with no variation. Is that going to be a constraint of a future Mini? Probably not. So why would be configuration requirements be reflected in a product with no variation at all?

Most likely to keep the DTK affordable, Apple mutated the iPad Pro board; not started from a Mini x86 board. Nor did they attempt to design something custom from scratch to "maximize" the synergy with the case.(it isn't a high volume product so "optimizing" the board buys what??? Nothing). It would be cheaper to take something that was already working and add some new adjustments . There is a decent chance that this could be a test harness board for the iPad Pro (something to throw the SoC into before the new case makes it out of Industrial design. Ports for debugging and connectivity before working with radios and Lightning. ).

the iPad Pro 3rd and 4th generations max out at 1TB SSD capacity. The DTK only has 512GB. Easily done with a iPad Pro board. The RAM would likely require a tweak, but since only a small double digit amount (16GB) not really a big leap at all. [ Not like have to get to 64 or 128GB levels. ]
Can decouple the Wifi to get rid of some of the "ruler" shape of the board. Attach the USB hubs to the USB connection point on the iPad Pro board.

Very good chance Apple made these DTK a substantively long time ago for internal use. ( or at least did a production run of a couple hundred. ). The capabilities here are all well within the era of the 3rd Gen iPad Pro. ( the A12X and A12Z are the exact same die. Apple could have gather some binned dies to do early A12Z in small numbers extremely easily. ) There is an implicit premise floating in these threads that this DTK is some kind of pre-production prototype of the Mini. Probably not. This is something they could have done even before starting work on the Mini ( much older). At which point it isn't particularly information of things going forward.

However, if put a high load on the Ethernet and USB ports at the same time and this system probably crumbles far more than an Apple Silicon system would.



The A12Z is known. There are iPad Pro teardowns. What going to see is the A12Z.

The A12Z isn't the Apple Silicon SoC. The AS SoC will have to deal with a wide variety of RAM capacities that span well past 16GB.

All Apple needs to coupled 16GB to the A12Z is two 64Gb DRAM chips and it is done. That doesn't scale very well past that though.




Not for Desktops. Apple's current desktop line up has DIMMs. (mostly so-DIMMs). Apple doesn't slavishly copy every single aspect of the iPhone design principles across the whole Mac line up.



I understand folks are curious. But the DTK is extremely likely to just be a large kludge that Apple whipped up to have something to test with. Something that was affordable to whip up just to have real hardware to test with. That's it.

Even Apple hints at that. In the system overview session of WWDC 2020 they make a reference to a "read me" file in the DTK distribution that highlights the differences in boot and some capabilities there are limitations of the DTK and not of Apple Siicon.

The DTK is highly lacking in being helpful for folks who need to write PCI-e card drivers and server different kinds of I/O peripherals. It is in no way the "end all , be all" of the Mac future capabilities.

That’s an awful long post to miss the point. People care because they’re excited, they’re curious. That’s it. End all be all. It doesn’t matter what is inside the DTK, people are curious about the first really new Mac since 2006.

You’re right about how none of it matters but it doesn’t matter that it is irrelevant. People want any crumb of info they can find.

You’re missing the point that people are basically speculating and guessing for fun. Who cares what actually comes to pass?

 

[projectle]

Apple is taking security very seriously this time around. Multiple parties have stated that Apple has returned to using steganography within screenshots to identify someone who takes screenshots from a DTK. Even the short lived AppleInsider article stated that the DTK does not have a removable bottom but instead felt like there were clips in the rear to seal it, and that it would be near impossible to open without leaving visible marks on the plastic back panel.

And, as always, Apple’s go to solution to breach of contract - just like they’re showing with Epic - is to delete their Developer account which destroys end user confidence. After all, if you lose the apps you paid for, you start seeking refund options as well as competitors.

These machines are built in China just like the Intel-based models, just at much lower quantity. My money is that if we ever see a board design, it will either come from a line worker in China snapping a pic and selling, or after the DTKs are to be returned from one that “was lost in a fire”.

 

[DeanL]

That’s an awful long post to miss the point. People care because they’re excited, they’re curious. That’s it. End all be all. It doesn’t matter what is inside the DTK, people are curious about the first really new Mac since 2006.

You’re right about how none of it matters but it doesn’t matter that it is irrelevant. People want any crumb of info they can find.

You’re missing the point that people are basically speculating and guessing for fun. Who cares what actually comes to pass?

But also everyone needs to stop presenting "the DTK won't reveal anything because of xyz" as factual information. It's not factual. We won't know for sure until it's torn down, which is why this thread was started in the first place.
Repeating over and over again and presenting 10,000 arguments for why the DTK supposedly would not reveal anything does not make that a fact, and people like me still want to see it torn down LOL.

 

[Wowfunhappy]

Frankly, I'm interested in seeing a tear-down regardless of whether it tells us anything about future Macs. The DTK is an interesting artifact in its own right.

 

[dmccloud]

Though I doubt we will ever see inside, it would tell us some stuff: Is there a fan? Is there user accessible RAM modules? How much unused space is there now? (apples SoC’s typically require smaller mobos which either means a smaller mini coming in the future, or maybe more storage options? Better cooling?).

The DTK won't really reveal anything on that regard though. Since we won't see a finalized design until the first AS Macs hit the shelves, opening the DTK is really about satisfying one's curiosity rather than gaining any insights into what the official products will look like inside.

 

[Kaida]

Frankly, I'm interested in seeing a tear-down regardless of whether it tells us anything about future Macs. The DTK is an interesting artifact in its own right.

me too. if there is a sticker or a seal or whatever, i want to see that or know there is, i don't want speculations of it..

 

[guzhogi]

Don't have to. Apple has some large data centers. Rack them up and they'd have a decent compute cluster for some 'embarrassingly parallel" workloads. Run them for a year or two until there is a surplus of Apple Silicon Minis and incrementally swap them out of the cluster(s).

Similarly I suspect more than a few of these were in use at Apple for internal developer build aids before handing them back out (maybe a new case ). The design is probably been around for more than a year or so.

Anyone remember System X?

 

[Manzanito]

But also everyone needs to stop presenting "the DTK won't reveal anything because of xyz" as factual information. It's not factual. We won't know for sure until it's torn down, which is why this thread was started in the first place.
Repeating over and over again and presenting 10,000 arguments for why the DTK supposedly would not reveal anything does not make that a fact, and people like me still want to see it torn down LOL.

We won’t know for sure until both the DTK AND the first AS mac mini are torn down.

If I recall correctly, the last time they made a transition computer it looked nothing like the actual intel mac pros (on the inside, I mean).

But I agree, I too would be eager to have a look inside the DTK. Everyone would be, even if there’s like zero chance that the final product has any resemblance.

 

[Seoras]

I'm sure someone will crack one open, anonymously post pictures/video, and then their DTK probably will take a nasty fall down some stairs or be the only thing stolen in a break in.

 

[Manzanito]

I'm sure someone will crack one open, anonymously post pictures/video, and then their DTK probably will take a nasty fall down some stairs or be the only thing stolen in a break in.

So little to gain, so much to loose, I wouldn’t bet on that. It’s possible we’ll get pictures after the actual as macs are released, though.

 

[Wowfunhappy]

We did get teardowns of the Intel DTK back in the day, right? Do we know how those happened?

 

[Ritsuka]

The Intel DTK was in a PowerMac G5 case, there wasn't even a screw to unscrew to open it.

 

[Manzanito]

The Intel DTK was in a PowerMac G5 case, there wasn't even a screw to unscrew to open it.

And there even wasn’t an app store apple could kick you out of. That’s before the iPhone, apple wasn’t in danger of disappearing anymore, but it wasn’t the juggernaut it is today either.

 

[t0mat0]

I really wish they could, I would love to see how it compares to the iPad logic board.

You seen the Nuvia lawsuit? Not going towards Apple Silicon on servers created some issues and a lawsuit... https://www.patentlyapple.com/paten...n-ip-agreement-to-not-compete-with-apple.html

 

[deconstruct60]

You seen the Nuvia lawsuit? Not going towards Apple Silicon on servers created some issues and a lawsuit... https://www.patentlyapple.com/paten...n-ip-agreement-to-not-compete-with-apple.html

Not much substantive there at all. California law prohibits broad, sweeping non compete clauses. IBM tried this with Papermaster after he left IBM and went to Apple. They lost. Apple is just as likely to loose. ( IBM has large research labs in California so they probably knew they had unenforceable clauses in their standard employee contract. It can 'pay' (at least short term ) to rattle the saber though. ) [ Nuvia opened up a Austin Tx office. TX law isn't quite so clear cut, but generally can't constraint ability to engage in a trade there either. Apple filed in CA so that isn't the primary driver. ]

Now if there is specific IP work of the Apple ARM implementation that is being copied here there is an issue. Likewise if was actively recruiting inside of the Apple during work hours and/or on Apple equipment . But neither of those would bar him (or the Nuvia) from producing something in an area overlapping with Apple's usages for their

That is basically what Apple's allegations are grounded on. Not some "you can't compete " clause.

"... alleging that he breached an intellectual property agreement and a duty of loyalty to the company by planning his new startup while on company time at Apple, spending hours on the phone with colleagues who eventually joined the venture. ... "

Apple lawsuit tests if an employee can plan rival startup while on payroll

Attorneys for a former Apple Inc executive on Tuesday will try to convince a skeptical judge of a core tenet of tech startup culture - that employees can plan a competing venture while still in a job.

www.reuters.com

"Servers" or "not servers" has nothing relevant to do with the suit. Not being able to take intellectual property (IP) that you don't clearly have sole ownership of to a new company is not a "non compete" issue. It is a don't steal issue. He was able to quit and start an overlapping business. But building a competitor which still collecting an Apple paycheck as full time job is a problem. That isn't "non compete" that is more "no backstabbing the company while on the payroll".


Apple's suit is going to do mostly what it did for IBM. It slows down the development at the new business. Time and effort devoted to handling the suit ( which isn't even to a larger company versus a smaller one (or smaller effective subsidiary that is being started in a larger company) ). It can also open up discovery so that do an technical inspection demonstrate that nothing was taken. ( e.g., where the ex-Google guy walked off with gobs and gobs data from the "self driving" car project and got caught. ) . The third goal probably is to slow the 'bleed' of folks being recruited away ( and 'encourage' Nuvia to recruits more broadly. If just spend the 'whole day' just combing through your old company's employees for hire candidates then there will probably be some blowback. )

 

[MacPoulet]

Maybe Epic will do a tear down of their developer kit since they aren’t afraid of pissing Apple off...