Hacker News Reader

Comment by MountDoom | Hacker News Reader

This is much less of a deal than it might appear. It also isn't an accidental victory for the right-to-repair.

The schematics are basically "here's a black-box SoC and here are the data lines connecting it to a black-box camera module". The magic of the iPhone are the black-box chips and their firmware, not the traces on the PCB.

About the most useful thing you're getting out of this are capacitor values, which are easy to measure either way.

Things used to be different 2-3 decades ago, but nowadays, most commodity PCBs are exceptionally boring - it's mostly just digital signal routing and some power-related stuff.

View on HN

It's still very useful for repair techs (and I don't mean parts-swappers, I mean actual techs). This data is pretty quickly reverse engineered and shared in China, and then made available to shops through services like ZXW, WuXinJi, JCID, etc. There's an entire shadow industry around creating and then sharing schematics and PCBs for common products which might need repair.

It's not really useful for most repairs. These PCBs are very dense. The traces and even vias are buried below the surface. At best it might give them some hints about where to probe if they've never worked on that phone model before, but in a practical sense they're not going to be doing many repairs based on the schematics.

The repair process for a modern phone involves swapping salvaged parts on to salvaged PCBs.

Most of the parts on the PCB are custom and not available for purchase separately. The passives aren't likely to be points of failure, but even when they are they're usually swapped from some other donor board rather than purchased new.

> The repair process for a modern phone involves swapping salvaged parts on to salvaged PCBs.

Right, but, knowing what the nominal in-line resistance is between any two pins on the board, will let you do some empirical tests to determine whether a given black-screen fault lies in the board itself, vs in one or more of the SMT components, no? (And so whether you should bother attempting to salvage the board, vs just scraping off the high-value parts and chucking it in the bin.)

>These PCBs are very dense. The traces and even vias are buried below the surface

You are so out of touch, PCB repair is very common.

This is just one example of via/pad repair: https://rewa.tech/iphone-12-pro-max-not-booting/

The whole repair industry in China is really insane.

>The repair process for a modern phone involves swapping salvaged parts on to salvaged PCBs.

This is always the case, but there is difference between swapping the whole camera module and just swapping the camera lens.

Hell they even use fiber laser to fix stripe line screen.

This reminds me of https://www.youtube.com/watch?v=reQq8fx4D0Q

That's assuming they can get their hands on the "black box" ASICs

I suppose you could transplant the chips from a damaged board.

Yes, that's what people like Louis Rossmann famously do. Apple, in turn, tries to serialise each chip and make it only boot if all the serial numbers match.

You don't need a schematic to do that.

Aftermarket board and passive components; transplant the Apple chips from a damaged board. Then you don't need to cannibalise the board, you use a new one.

I assume most physical damage and corrosion etc affects the board and passive components.

> This is much less of a deal than it might appear. It also isn't an accidental victory for the right-to-repair. The schematics are basically "here's a black-box SoC and here are the data lines connecting it to a black-box camera module".

I don't think right-to-repair folks would have much use for anything more than that hardware-wise. You can give a technician the HDL or lithography masks to Apple's A17 chip, but it's not like he has an EUV lithography machine in the back of his repair shop.

I mean, even the firmware binaries would be a game changer. But really, it'd be the appropriate tool chain to re-sign a replacement chip so that the overall device doesn't reject it. Provided there isn't some kind of mutual key exchange that gets fused out upon factory programming.

>The magic of the iPhone are the black-box chips and their firmware

But a technician won't care about those magic. All they care is whether they can get those FW and chips, and there is mechanism to flash/replace them.

A trace on PCB is much more important. Boardview is a must to do board repair.

I'm not a repair tech, but I do enjoy watching them work and I suspect they'd strongly disagree with you. It is always useful to know part numbers and how they are connected with each other even of the components are themselves "black boxes".

> It is always useful to know part numbers

In modern phones the key part numbers for ICs aren't going to correspond to anything you can buy.

At best, you can see the part numbers for the passives and their alternate suppliers. The part numbers for most of the actives aren't going to be useful to anyone outside of the manufacturing chain for these parts. A repair tech would simply source the donor parts from an identical phone model.

> aren't going to correspond to anything you can buy

...on the open market.

However, if your repair shop exists just down the street from one of Apple's component supplier's factories, then knowing the part number of one of the ICs or other actives, might be enough to let you walk in there one day and submit a hand-written PO for a few boxes of said part (...as a cash transaction, of course.)

Eh. YouTube nerd-bait aside, real-world phone repair techs replace screens, batteries, and charging ports. It's rare for anything else to fail, and when it does, it's generally not cost-effective to troubleshoot which of the BGA pads or microscopic inductors might be faulty. You just replace the PCB or tell the customer to buy a new phone.

> YouTube nerd-bait aside, real-world phone repair techs replace screens, batteries, and charging ports. It's rare for anything else to fail, and when it does, it's generally not cost-effective to troubleshoot which of the BGA pads or microscopic inductors might be faulty.

My mom's Samsung phone suddenly failed to charge a year ago, but in a weird way. Tried some tricks but it ended up getting stuck on a black screen during boot.

I handed it to a repair shop in town that had good reviews. Three days later they called back. I asked what was wrong, and it turned out they had to replace one charging chip, a power sequencing chip and a third chip I can't recall the purpose of, along with the USB port. The reason it took so long was they got a bit further for each chip they swapped.

I'm in Norway where prices are high, and we have a 25% VAT on top. I paid $180 including VAT, or $144 before VAT.

For that I might have gotten a very cheap Android phone on sale, however my mom's phone was a premium model from a few years earlier. Besides, less e-waste and saved me time reinstalling everything. Personally I think the repair was well worth it.

There's discussion here in Norway to drop VAT on repair jobs such as this, and in that case I think it would definitely make sense to repair such cases.

I'd argue that VAT is just a bad way to do taxes in general. The only thing it really accomplishes is slowing down the economy. Much better to target wealth/income instead IMO.

Much better to target resource extraction. That will have an immediate and very positive effect on the environment because it will make recycling so much cheaper than using virgin resources.

Your original comment talked about "most commodity PCBs," but it sounds like you really meant to limit your scope to phone PCBs.

Laptops, motherboards, microwave ovens and whatever else has "commodity PCBs" are much more repairable than the highly size-optimized phone market.

Not cost-effective for a repair shop in the US, but in other countries with lower cost of living, including China and many Asian countries, technicians go to great lengths to replace individual components on PCB boards.

They are doing some wild stuff in Asia like adding physical SIM card slots to eSIM only iPhones (see https://www.youtube.com/watch?v=-0Ja_x7JI0w). I imagine having schematics like this would be at least marginally helpful in developing those techniques.

You are the reason why we have so much e-waste

> it's mostly just digital signal routing and some power-related stuff.

It still saves competitors on R&D. As in "ah they routed it this way, nice!" or "They have this capacitor here, we don't. I wonder why it is there" etc.

Schematics are extremely valuable!

Sure thing if you copy the schematic, routing, you don't suddenly get a working iPhone, but you might use it to design your own device that will have robustness of a phone designed by multi trillion corporation.

> It still saves competitors on R&D. As in "ah they routed it this way, nice!" or "They have this capacitor here, we don't. I wonder why it is there" etc.

None of this is true.

Competitors save no R&D at all because these chips are custom to Apple.

The schematic doesn't show routing of the PCB. It shows logical connections between components. The routing is not represented.

Decoupling capacitor placement is also chip specific so it does not translate to other vendors. At this level of PCB design the board is simulated in very expensive software to determine how the power distribution network performs.

> but you might use it to design your own device that will have robustness of a phone designed by multi trillion corporation.

Not in the slightest. This shows logical connections between Apple specific parts. There are no portable industry secrets hidden in this document.

You're focusing on a technicality while missing the bigger picture. These leaks frequently include board layout files, so yes, competitors see the exact routing and component placement.

You claim this is worthless because of the custom Apple chips, but that’s precisely what makes it so valuable. Competitors aren't trying to drop an Apple A19 into their phone; they are learning how Apple solves the universal engineering problems that come with supporting a high-performance, custom SoC or their specific parts.

Seeing the complete power delivery network and high-speed routing for that specific chip is a masterclass. When they see 'a capacitor here,' they learn Apple's validated strategy for handling a transient load or suppressing a specific EMI source on a cutting-edge processor. That engineering principle is absolutely portable to their own custom silicon from Qualcomm or Samsung.

That is the very definition of saving R&D. It’s not about copying parts, it’s about learning the architectural patterns from a successful, finished product to avoid months of costly trial and error.

> You're focusing on a technicality while missing the bigger picture. These leaks frequently include board layout files, so yes, competitors see the exact routing and component placement.

The board files were not leaked.

I don’t know what you think you’re talking about, but it’s not relevant to the article and the leak we’re discussing.

> Seeing the complete power delivery network and high-speed routing for that specific chip is a masterclass. When they see 'a capacitor here,' they learn Apple's validated strategy for handling a transient load or suppressing a specific EMI source on a cutting-edge processor. That engineering principle is absolutely portable to their own custom silicon from Qualcomm or Samsung.

It is not portable nor priceless.

I’m telling you: I have experience doing this and working with these tools. Nothing you’re saying about this topic is accurate.

Fine, let's grant that only the schematic was leaked. The conclusion is the same. A schematic is not a mere netlist; it's an architectural document. It reveals the entire power delivery architecture - the choice of PMICs, the topology of every buck converter and LDO, and the meticulous partitioning of power rails for the CPU, GPU, and Neural Engine. That power strategy alone is a goldmine of portable R&D.

Furthermore, it details the complete bill of materials and the exact circuit design for every high-speed interface. To suggest this knowledge isn't portable is to claim that the principles of designing a stable power network or a clean USB4 interface are somehow unique to Apple. Any experienced engineer knows that learning from a competitor's complete, validated, and mass-produced system is an invaluable shortcut. An appeal to personal authority doesn't change these fundamental realities of hardware development.

I think varispeed is chatgpt.

I'm sorry but no. Unless you are producing a 1930s radio these sorts of schematics are next to worthless. There's nothing super clever about PCB routing or resistor placement and it is completely dependent on the chips used (which are increasingly including the actual resistors and capacitors internal to the chip).

Consider, for example, how useful this schematic would be vs an AMD motherboard schematic.

You imagine that because apple has a bunch of money to spend that they are spending more than a competitor would on PCB design. But that's just not how hardware works.

The money Apple is spending on R&D is almost entirely going into the design of their M series chips and their software/firmware.

This dismisses the entire field of modern high-speed digital design.

At GHz frequencies, the PCB isn't a passive substrate; it's an active RF component. Managing signal integrity, controlling impedance, and mitigating EMI in a dense layout is a monumental engineering challenge, not a solved problem from the "1930s radio" era.

The AMD comparison is backward. AMD provides public datasheets and reference designs. This schematic is valuable precisely because Apple's silicon is a black box. It's the only public blueprint for a successful, validated implementation.

The chip and board are a co-designed system. A brilliant SoC is useless without the board-level engineering to power it cleanly and cool it within a tiny physical envelope. That system integration is the R&D.

> This dismisses the entire field of modern high-speed digital design.

No it doesn't. That entire field operates primarily within the ASICs, not the PCB.

> At GHz frequencies, the PCB isn't a passive substrate;

Yes, it actually is. There is some noise and cross talk that's emitted from the signals but it's minimal and easily blocked. Very simple rules, like minimum trace distances, solve pretty much any problem that comes up from the frequency of the signals.

But further, those problems are peanuts compared to dealing with the exact same problem within the ASIC. Very simple rules are needed for the PCB.

> mitigating EMI in a dense layout is a monumental engineering challenge, not a solved problem from the "1930s radio" era.

You are vastly overblowing how big an issue or problem this is. The voltages and current on the lines are minimized to cut back on EMF and power consumption. Those are dictated by the chip design, not the PCB setup.

> A brilliant SoC is useless without the board-level engineering to power it cleanly and cool it within a tiny physical envelope. That system integration is the R&D.

There is some R&D that goes into the board, for sure. However you are completely overblowing how hard or costly it is for apple or others to develop.

This demonstrates a profound misunderstanding of the physics of high-speed electronics.

High-speed design doesn't stop at the ASIC boundary; that's often where the hardest part begins. The entire field of SerDes engineering exists to solve the "trivial" problem of getting multi-gigabit signals off the chip and across the board to RAM or a port without dissolving into noise. It's a system-level challenge.

At GHz frequencies, a PCB trace is not a passive wire; it is a transmission line. The idea that "simple rules" for trace distance solve this is fundamentally incorrect. Real-world design requires complex electromagnetic simulation to control impedance, manage signal reflections, and design via structures that don't destroy the signal. This is why Signal Integrity (SI) is a dedicated engineering discipline.

EMI is caused by the rate of change of current (dI/dt), not just the voltage level. Modern chips have incredibly fast signal rise/fall times to achieve high performance, which creates massive high-frequency energy that radiates easily. Passing FCC/CE certification is a famously difficult, multi-million-dollar R&D effort, not a trivial detail.

I don't think you realize it, but this completely refutes your initial point. That the schematics are valuable and something that someone could simply copy for their own board designs.

You can't hold both positions that there is a bunch of system dependent physics involved AND that someone can simply copy those design elements and pass CE/FCC licensing.

And no, I don't misunderstand it.

> The entire field of SerDes engineering exists to solve the "trivial" problem of getting multi-gigabit signals off the chip and across the board to RAM or a port without dissolving into noise. It's a system-level challenge.

Right, and most of the problem is inside the specification of the signal. It has (almost) nothing to do with the actual PCB design beyond talking about the voltages and acceptable trace lengths.

You are forgetting exactly what the initial discussion is about, whether or not the schematics are valuable to other hardware manufacturers.

> Real-world design requires complex electromagnetic simulation to control impedance, manage signal reflections, and design via structures that don't destroy the signal. This is why Signal Integrity (SI) is a dedicated engineering discipline.

Simulations that are completely missing from the schematics, correct? You don't even get to glean what those simulations would be and importantly minor changes could by your logic completely ruin a design. Correct?

This is the weakness of your argument.

Yes, these things matter. But you are frankly overblowing how complex or hard they are to create. Everyone is doing them and it's not a huge cost to the actual hardware manufacturing.

> Passing FCC/CE certification is a famously difficult, multi-million-dollar R&D effort, not a trivial detail.

There are hobbiests that have managed to pass the FCC/CE certification with their own custom PCBs. Look, I admit that it's something harder to do with high speed signals. However, it's not a "multi-million-dolar R&D effort" That is, unless you are wrapping the VP of engineering's salary into that calculation.

The biggest problem you actually face as a hardware company is that if there is an issue you have to redesign and resubmit to the FCC/CE which takes a time. Online, I'm finding people saying that 4 or 5 respins is common.

Apple spending more money here would be entirely around avoiding that respin. Buying the same testing equipment and shoring up the testing procedures to make sure they get the certification the first or second time rather than spending a bunch of time needing to retry.

Which, I can't stress enough, is not something that comes out of the schematics. You as a competitor don't get any real information about the testing that was done and you certainly get no information about what the high speed communications between chips looks like (though that is entirely going away).

Ok, so I saw varispeed's comments in this thread. They're really LLMish. I have a feeling you are just arguing with ChatGPT.

I hope not, my LLM detector is off if that's the case. But looking at how they've responded elsewhere it may be the case that they are pulling from LLMs to try and win arguments.

It definitely seems like they have a bunch of unconnected ways of arguing in this comment section. Like here they are arguing that the PCB design is the hardest act of engineering ever done. Elsewhere they are arguing that just seeing the schematics is all someone needs to build an iPhone.

Looking at their comment history outside this thread, I think they aren't always using LLMs. Perhaps just when something gets more technical than they have the experience for.

What ticked off my detector was the overuse of the "it's not [just] x, it's y" phrasing.

But all that information is in the engineering files that are neither the schematic or the layout. You might guess some information from the layout, but on some crazy HDI PCB like the iPhone uses, you'll be missing most of the information needed to take advantage of it.

Any signal that uses passive termination, and would therefore show up in the schematic, is basically a tacit admission that this line needs something, but basically anything remotely close to 50Ω will do. Which is something you could've gathered by inspection of the final product.

It is valid to say they do not have the majority of trade-secrets in the schematics.

However having high quality documentation that is easily accessible is always the first step, and if there are information there that are important it is _much_ more likely it will be known.

It is not "a much less of a deal" as you say, it is a pretty big deal for a global product that billions of people touch.

>The magic of the iPhone are the black-box chips and their firmware

But a technician won't care about those magic. All they care is whether they can get those FW and chips, and there is mechanism to flash/replace them.

Reply on news.ycombinator.com