A cable problem explained

In the past, I have complained bitterly about the difficulty of getting reliable cables for Apple computers, and also about how ridiculously overpriced they all are. I may have been wrong. There’s a good reason for that.

Adam Savage takes apart several USB-C cables, from a $3 cheapie to a $10 Amazon to a $130 (!!!!!) official Apple cable, and I learned something new.

Whoa…those expensive cables are packed full of complex circuitry, which I did not expect. Back in the day, I used to make custom RS-232 cable all the time — cut open the cable, splay out all the little wires, solder them to the appropriate pin in the connector, and you were done. That’s all it was, was a wire-to-wire connection between two adapters. I don’t think I could make a USB cable, and it’s not only the teeny-tiny wires that would exceed my soldering ability, but I wouldn’t be able to cope with all the miniature ICs in there.


  1. says

    This apple thunderbolt system seems to be purely UNNECESSARY complexity to create EXPENSIVE exclusivity in the quest for more PROFIT. Using almost microscopic conductors carrying ‘tiny voltages of the digital signal’ And, the miniscule size of the conductors makes them FRAGILE. I am certain (based on decades of experience by our organization in electronics) that this cable will not last as long as a well-made larger simpler cable. The UTP (unshielded twisted pair) system has been used in all CAT 5/6 cables for decades. All the added ‘computing power’ doesn’t seem to provide performance much greater than the lightning fast latest USB technology. This is the proliferation of the corporate computer tower of babel. Back in the 1970’s there was high-turn STP (shielded twisted pair) cable used in aircraft and high end audio which prevented any interference.
    And, all of this over-designed over-priced hardware is just so geeks can drool and go oooh! over it. And, we see it is mostly used by people so they can watch their nephew’s cat video or watch a streaming movie on a 7inch screen. I and my organization have no use for this needless extravagance. (o.k. now I’m going to climb off my soapbox and get back to work on setting up our mesh-network)

  2. says

    One more quick remark to summarize: PZ is correct, it is a ‘A cable problem explained’ but NOT justified to the satisfaction of our engineers.

  3. tacitus says

    You do get extra quality from more expensive Thunderbolt 4 USB-C cables but $130 is still a rip off. First of all that price is for a 1.8m cable. A shorter 1m cable from Apple with identical specs is almost half that price ($69) and even accounting for the possibility that you need to use better materials for the extra 0.8m length, it doesn’t justify the $60 upcharge.

    This is a perfect example of Apple offering a “reasonable” price for the lowest end spec device (like they do for their iMacs) but charging way more for the better (in this case longer) item which is likely the minimum spec that most people should buy.

    Either way, you can get a good quality (certified) 6ft Thunderbolt 4 cable for less than half what Apple wants to charge you, so you should save your money not pay the Apple tax.

  4. starblue says

    It’s a bit unfair to compare a Thunderbolt 4 cable to cheap cables that at most can do USB2 (they don’t even connect the wire pairs needed for USB3 and above). There are two or three levels in between, not to mention that there are also short Thunderbolt cables without amplifiers.
    And the Apple cable doesn’t seem much overpriced. I got a similar Thunderbolt 3 cable for my Lenovo Thinkpad in 2020, and that cost about 90€.

  5. says

    Beware, the comparisons in this cable are extremely inapt.

    While it’s correct that those Apple cables have a lot going on, there are a lot of unanswered questions in terms of comparing them to other brands of Thunderbolt cables, which tend to clock in a pretty pricy but half the price of Apple’s own.

    Comparing them to charging cables that don’t even implement the USB 2 communication standard, let alone 3, let alone Thunderbolt, makes the Apple ones look like a far bigger deal than they are.

    I found it particularly galling that they made a big thing about twisted pair construction, which isn’t an Apple innovation at all – that’s been standard in all data cables for decades and decades. Strip any Ethernet cable you like and you’ll find not only four twisted pairs, but four twisted pairs of different wavelengths to minimize further the ability of an interference source to do the same thing to multiple signals at once.

  6. says

    This comparison is meaningless—almost disingenuous. The video compares a Thunderbolt cable (which is vastly more complex and carries much more data) to a USB-C cable. Despite the fact that they use the same connector, these are NOT AT ALL the same thing. It’s like comparing An F1 race car to a Prius just because they both use four wheels.

  7. Jean says

    In addition to what has been mentioned in the comments above, using an Amazon basics cable as an example of a “good” cable is ridiculous. The quality of the Amazon basics product is all over the place and it’s just a product chosen to optimize Amazon’s profit.
    I like Adam Savage but his video is crap, at least from the point-of-view of cable comparison.

  8. wzrd1 says

    The techniques from the discrete component era aren’t quite so linearly applied to soldering components these days. Much more reflow soldering, via hot plates or even hot air gets used. Procedural microscopes are also commonly used, which is more than a godsend on the few times I’ve worked on such modern circuitry, as my 62 year old eyes just simply cannot resolve such components – even with my best reading glasses.
    Assembling from scratch, well, there is only one expression beyond “don’t do it!” that comes to mind, “Huge pain in the balls!”.
    Meanwhile, you can get cables at a fraction of what Apple sells by getting certified cables that aren’t resold by Apple (you do know that they don’t actually make their own cables, right?). Just dodge Amazon or ebay, as you really won’t know what counterfeit crap you’ll get and a very literal fire is a possibility. Almost had a fire once, thanks to a discount Amazon counterfeit, which shorted internally, overheating two other much pricier cables, destroying the lot of them.
    The local fire department learned by witnessing the conflagration that ensued due to my use of profanity, due to the spectacularly shitty timing of said failures. For in sooth, verily did fire spew from my lips, lightning from my eyes and thunderbolts from my arse.

    @shermanj, there are times one does want to use a thin conductor, such as with a high frequency sharp waveform signal, to avoid ringing issues. Alas, that’s simply not the issue with a computer or phone cable. There, it’s price and planned obsolescence. So fucking what if your house burns to the ground, as long as you’re in it and your estate lacks assets to litigate.
    Yeah, in a bad mood. Back is hurting, tired as hell from pacing to keep warm while waiting for the buses and the dentist kept me in the chair for over an hour while my BP climbed, just to defer all further treatment to an oral surgeon. Which could’ve been done last week.
    But, she did spend a bit more time talking to me after examination, asking some questions about my military service. I answered quite honestly and simply – after I got the referral.
    She learned, we preferred a peaceful solution first and foremost, but if fired upon, we swiftly and precisely ended the engagement successfully and I’m also a creative SOB, who literally made a terrorist’s sewage run in reverse to force everyone from their fortified villa, to see enough firepower to instantly suggest surrender over an unarmed charge upon that much noise. That particular specimen now resides in permanent exile in Qatar, under close guard, as the Afghans don’t want that one back.
    And I just realized that I ran out of naproxen. :/
    Smoke is now escaping my nostrils over my idiocy…
    Avoiding worse, as dinner will be ready in a half hour and it’s best experienced in a non-incinerated condition.

    Also learned, my BP is higher than it should be by a fairly alarming bit, gotta make an appointment with my doctor to get some hormone levels checked. Given this is a third consecutive monitored series of observations, I’ll score that one as a net gain.
    Largely, as I’ve found by experience that waking up dead really fucks up one’s weekend plans.

  9. Robert Webster says

    I got burned by Apple in the ’70’s. I had a project given to me that required me to write code on the Macintosh. However, they were in their Appliance phase, and the development system required was far, FAR too expensive. To the tune of thousands. In the ’70’s. So, I avoid Apple whenever possible. Never felt slighted in the least.

  10. says

    We have seen what the ridiculous miniaturization of connectors and cables does. We deal with and help people who don’t treat them as the fragile things they are. They jam the connectors and endlessly twist the cables and they fail in less than a year.
    @6 abbey St.Brendan confirms my assertion that UTP (unshielded twisted pair) cables have been in use for decades.
    @7 Pad Gallagher says: It’s like comparing An F1 race car to a Prius.
    And they are correct. But, how many people need an F1 race car for most of their driving needs; like going to the grocery store?

  11. says

    I admonish @9 wzrd1 Please, take care of yourself. We really do appreciate your contributions here. Get the naproxen sodium but also get a prescription for Fugitall pills. I understand all the BS frustration you are dealing with. But, you need help to relax and drop that blood pressure before it drops you.

  12. says

    The EU has ruled that Apple is liable for damages if third-party cables and accessories are used with Apple products, are wildly out of spec, and cause a fire, because at first they weren’t warning users when the things were plugged in. (About 10 years ago, some idiot over there ordered a $1 unbranded Lightning phone charger from China which was built to put out vastly more current than the phone was expecting, left it plugged in at night, and burned his house down, and sued Apple rather than the manufacturer of the charger or the reseller who was representing it as an iPhone accessory. And the courts sided with him, which is why I am always skeptical of EU decisions these days..)

    @#2 shermanj: That’s pretty amazing considering that Intel developed and owns Thunderbolt. Just because you losers on Windows have decided to pretend it doesn’t exist because it’s on Macs doesn’t mean it’s from Apple. (Reminds me of the days when USB was new and PC users were shunning it, and then got extra angry about it because Apple switched to it from Apple Desktop Bus. PC users did manage to kill FireWire, which was Apple, though, so I guess you can chalk up a win; instead, you got everybody to use USB 2 for years, which was noticeably inferior in every way. Hooray for your side.) (Also reminiscent of how PC users think the AAC file format is from Apple because iTunes uses it, when in fact it’s the MPEG group’s successor to the MP3 format.)

    @#11 Robert Webster: Well, of course it was going to be very expensive to write code for the Macintosh in the 1970s. The time machine you would need in order to get computers which weren’t sold until 1984 would, all by itself, cost a small fortune. Heck, even writing code for their first ready-made computer, the Apple II, might have required a time machine depending on when you did it, because the first Apple II didn’t ship until halfway through 1977.

  13. says

    Thunderbolt is just another flavor of USB-C of which there are far too many flavors, and annoyingly no easy way to tell them apart. The connector’s shape is no help, and there’s literally zero standard for marking connectors or ports.

    If the EU takes everyone involved to court I’ll be cheering from the sidelines.

    Oh, yea, what everyone else has said, that $130 price is absurd. The dudes with Adam (it’s sponsored content so 100% untrustworthy no matter how much you like Adam) are shills for something or other, probably Apple.

    Thumbs down.

  14. says

    As several people have mentioned, this is a terrible video. The Apple cable is a powered cable with active circuitry. They then compare that to a completely different spec cable (USB-C) that contains no active circuitry. Generally, you use an active cable if you need longer runs than is typical. And yes, you can get active USB cables. It would have been much more useful if they compared the same kind of cable (i.e., active USB-C to active USB-C). And their “amazement” over the differential twisted pairs, oye, that’s just embarrassing. That tech has been around for decades.

    Having said all of that, this is still not where the action is in terms of fleecing customers with cables. That would be in high end audio. Yes, you can drop $1k for a few meters of loudspeaker cable. My favorite, though, is replacement power cords for your amplifier and other components. You definitely need to spend $500 for a new AC power cord for your amp so that it can ensure that the electrons are nicely lined up and not inhibited in any way before entering the amplifier. I guess the maybe 50 feet of Romex that connects the AC outlet to the panel is not a problem*.

    *It’s not a problem because the sellers haven’t figured out a way to convince people that they need to rip out their house wiring, while swapping out a modular power cord is something anyone can do.

  15. Snarki, child of Loki says

    Give it another decade and USB-C will be held in the same regard as the old ethernet thickwire transceiver slide-lock connectors.

  16. says

    Forgot this. When it comes to Apple and connectors/cables, my favorite comes from the 1980s. Apple decided to use the established SCSI connector for their computers, but their marketing people just could not bring themselves to pronounce it the way everyone in the industry had been pronouncing it, namely “scuzzy”. Maybe too many bad memories from high school? Whatever. They tried to get people to pronounce it “sexy”. I laughed when I first read about it. So immature.

  17. hemidactylus says

    We’ve gone from the OS wars into the cabling wars which are OS (or manufacturer) adjacent. I do recall some Apple charging cables from older iPhones turning brownish toward the device end at the connector and not working properly. Maybe using the device and putting the cable under physical stress (bending) while charging is not a good idea. PEBKAC.

    My iPad has USB something or other for the charging cable but my iPhone 13 has Lightning. I try not to use the devices much while charging.

  18. says

    I do enjoy a good religious war.

    I think this was an informative video — I had no idea how much complexity in those connectors. At least know I know how the expense is justified.

    What the video is missing is any discussion of performance. If I go on Amazon to buy a USB cable, it’s nice to know why there are wide differences in price…but what I don’t know is whether I actually need the fancy expensive cable. It’s also annoying that there is nothing in the packaging or labeling that tells me exactly what I’m getting.

  19. hemidactylus says

    I’m kinda at a loss as to why cables are so important apart from charging in a highly mobile world with like wifi and Hedy Lamarr inspired Bluetooth. I was just listening to an audiobook on my iPhone using nice enough Bluetooth headphones because I live in 2023 amongst the desktop dinosaurs who hadn’t yet succumbed to the asteroid. I don’t use earbuds, so I guess I have my own traditions to uphold. Earbuds are gross.

    Full disclosure: I do have ethernet cable between my modem and router.

    And why waste paper and expensive toner on printing stuff (so Gutenberg era) when I can save as a pdf or bookmark to never read it again? I do have way too many stacks of printouts though, but have decelerated their accumulation along with no longer buying hardly any printed books.

  20. billseymour says

    abbeycadabra @6:

    I found it particularly galling that they made a big thing about twisted pair construction, which isn’t an Apple innovation at all – that’s been standard in all data cables for decades and decades.

    Indeed, for the better part of a century.  Remember the old five-line telephones with the pushbuttons to select which line you wanted?  Ma Bell used a cable with twenty-five twisted pairs, each twisted at a different pitch to avoid crosstalk.

  21. dontlikeusernames says

    Haha. I love the copium from the Apple fans in this thread. It’s absurdly overpriced (see also the 8GiB debacle). If you want to overpay, that’s on you. Given Apple’s lack of oversight wrt. labor practices in the countries they source from, I’d personally be wary of over paying.

    (Most if not all computer tech companies have very dubious practices in terms of sourcing, let’s be clear, but why are you rewarding Apple, specifically?)

  22. says

    I feel (as a left-over security guy) that it’s worth mentioning that complex cables full of stuff are a great place to put man-in-the-middle malware. Often when syncing your phone, you may have to tell it “trust this computer” and at that moment a key exchange can be faked. A good example is the “OMG Elite” cable:
    Cables like that can do keystroke injection and even mouse movements. They can also wait for periods of inactivity to do it. So, you plug in, go to the bathroom, and while you’re away the software in the cable starts dicking with your phone or your desktop, or both. They also can do keylogging, so entering passwords is actually a liability.
    [This is why I recommend that you carefully scope what devices and when they have access to what passwords, and never use the same password in two places.]

  23. jenorafeuer says

    The Vicar @ #14:
    Firewire was an Apple/Sony/(… checks …)Texas Instruments co-production. Sony obviously was interested from the video equipment end of things, and Firewire/i.Link/IEEE1394 was one of the standards for video equipment for years. Protocol-wise it was in many ways a serialized SCSI communications method.

    One of the reasons it ended up getting killed is that it was actually a security nightmare. One of the advantages of SCSI from a technical perspective (and disadvantages from a security perspective) is that any device can be the bus master. SCSI actually has a ‘copy’ command in the protocol where the CPU can say “Disk 1, read the data from sectors A-B and then send them all to disk 2 starting at sector C.” and then go off and do its own thing, meaning that the CPU and main memory doesn’t even need to be touched during copy operations.

    The problem with that was that you could plug in a device over Firewire which would become the bus master and actively read all the RAM on the computer you plugged it into by using the DMA controller on the main board. So you could literally just slurp off any data on the computer unless the computer properly handled access control on its own side, which of course most of them didn’t because that would have been expensive and complicated to do. You can literally use a Firewire port as a debug port on a computer, tracking application operations, and there are definitely valid uses for that… but it’s not necessarily something you want in consumer grade equipment.

  24. robro says

    Those $130 Thunderbolt 4 USB-C cables are the “Pro” version. There’s a 3 meter one that’s $160. But Apple has other Thunderbolt 4 USB-C cables that are considerably cheaper than that. I gather than the Pro version have high throughput so they are probably intended for people with high volumes of data like for audio and video.

    I’ve used Apple, DOS-PCs, Windows PCs, and some other types of gear. I’m happy with Apple but if you don’t like Apple you’re in luck…just like religion there are many options.

  25. nonlinear feedback says

    Helge @ #15: They’re not shilling for Apple. They’re shilling for Lumafield, the company which made the CT scanner used to see inside all the cables, used one of their scanners to generate all the imagery, sent someone to guest in Adam’s video, and wrote all this up as a corporate blog post even before paying Adam Savage to make a video about it.

    Things like their choices about which cables to compare and their commentary on the cables may seem odd until you realize that the real goal is to generate controversy. It’s a pretty tried and true technique: mention and praise Apple, especially by unjustly comparing an Apple product to something much cheaper which doesn’t even implement the same functionality, and you’re going to get hordes of angry people waving pitchforks and chanting about how much they hate Apple, as we’ve seen right here. It gets attention, and therefore gets more people noticing Lumafield, which is the goal. (One presumes.)

    And no, conspiracy-minded folks, Apple didn’t want this. They have a very particular way of presenting themselves to the public, almost Disneyesque in how tightly controlled it is. This ain’t it.

    @shermanj, you seem to have become a reactionary old curmudgeon at some point in your life. You should try pulling back from that a bit and learning more about the technology behind a 40 Gbps Thunderbolt 4 cable. No, they are not just pointlessly doing fancy things for marketing reasons. No, cable technology did not end when Cat-5 and Cat-6 were finalized. Those standards run out of steam once you need more than a few hundred MHz of analog bandwidth. TB4 needs to put 20 Gbps through a single twisted pair, which translates to several GHz of analog bandwidth. You can’t do that with Cat-5 and Cat-6. In fact, it turns out that getting copper to function as a transmission medium at all for signals well in excess of 10 Ghz is a very difficult engineering problem, which is why there’s lots of stuff going on here.

    @Marcus Ranum, these problems are the reason why there’s active silicon in these cables. To make 20 Gbps per pair work at medium to long cable lengths, they have to put advanced mixed-signal equalization chips in the cable heads. Your concerns about a cable MITM’ing the data stream are valid (we live in a disturbing future from an information security standpoint), but unfortunately, this active circuitry is absolutely required for the cable to function at all at its length and data rate. Thunderbolt 4 does permit passive cables, but this is only possible with lower combinations of data rate and cable length.

    To give an idea of how crazy this technology is, I’m pretty sure I recall hearing that merely bending a cable can cause changes to key parameters (parasitic inductance, capacitance, etc) significant enough to impact data transmission at these 20G line rates. The equalization chips in the cable heads monitor the signal passing through and adjust for such changes in real time.

  26. says

    @#10, Erlend Meyer: no, 40 Gbit/s (since I’m going to reference Wikipedia in a second, I’ll use their abbreviation… I prefer Gbps, myself) is entirely reasonable for any bus which is expected to handle a modern video signal, like Thunderbolt (or even non-Thunderbolt USB 4). If you’ll go look up HDMI in Wikipedia (tadaa!), you’ll find that transmitting a single video at 1080p at 60 Hz (which was the highest resolution of HDMI in version 1 of the spec, long since surpassed) uses 3.2 Gbit/s. A 4K display at 60 Hz uses 12.54 Gbit/s, and an 8K display at 60 Hz uses 49.65 Gbit/s (wow!), so actually a Thunderbolt 4 connection cannot do 8K video at 60 Hz without cheating by compressing the signal (in which it is not alone — actual HDMI cables can’t do it either!). Since Thunderbolt is supposed to be able to drive 2 monitors per port and video nuts are now pushing for 8K to be the new high-end for consumer-level gear, 40 Gbit/s is actually less speed than I would expect.

    @#15, Helge:

    Thunderbolt is just another flavor of USB-C of which there are far too many flavors

    “USB-C” refers to the connector, not the version. Thunderbolt just uses a plain USB-C connector. (USB-A is the rectangular 4-pin connector used at the computer end of cables in USB 1 and 2, USB-B proper was the weird irregular hexagon shape at the other end of early USB cables; mini-USB and micro-USB are technically variants of USB-B. The whole point was that USB cables were one-way from the computer to the peripheral. IIRC USB-C when proposed was supposed to replace only USB-A and maintain the one-way-ness, but because it is fairly obviously superior to all the other USB connector types, it has become a replacement for both A and B… so now everybody gets to deal with a mishmash of C-to-B, A-to-C, and C-to-C cables until all the old hardware gets replaced by C-only connectors.)

    @#26, jenorafeurer: you’re saying Windows users stuck with USB 2 because they were worried about security risks from FireWire? That’s like saying you’re going to vote for Joe Biden in 2024 because you’re worried that another Trump term might be bad for Palestinians. (It’s also inherently assuming that a machine can be secure when somebody has enough physical access to plug in a peripheral, which is false.)

    (But as a matter of fact, I know that’s not really true. I was around and paying attention, and PC users rejected FireWire because it was seen as being an Apple product and they didn’t want to use that even if it meant sticking with a slower, less-reliable bus which provided insufficient power to run any complex peripherals.)

  27. Silentbob says

    @ 27,28

    Sorry 違う. Team hyperliteral had a valid point, while your oblique pass-agg response was a bog-standard Trumpism.

    Morales wins this round, play on.

  28. EigenSprocketUK says

    My vague recollection was that USB2 was promised to be “coming soon” during the mid-late 90s when daisy-chains of FireWire 400 (plus Sony’s DV variant S100 i-Link) were already rock-solid in the video world (where reliable transfer is more important than pure speed).

    When 480Mbps of USB2 actually shipped, it was more widely adopted on peripherals, despite no-one ever actually seeing all of that labelled speed. It wasn’t long before it was leapfrogged by FireWire 800 which lasted for years (despite faster versions of IEEE1394 appearing during its reign) and didn’t have unwise connectors. Though that speed doubling wasn’t enough to make non-video people see beyond USB2. (To be fair to USB2, it is still quite handy for some protocols even when implemented over USB-C connectors.)

    During the mid 00s, USB was promising that USB3 would be coming soon. And then 3.1 would be coming soon. And 3.1 gen 2 or is it 3.2 gen 1 coming soon. And now we’ve got USB4 coming soon. (But if you want USB4 right now, just buy Thunderbolt 3 cable and kit, which even has USB-C connectors on it.)

    If nothing else, we have to applaud the USB consortium’s ability to sell its next generation bench demo promises, regardless of how short the cables have to be to sell for less than a hundred nicker to the end consumer. The bit they do badly (very very badly) is explaining the vast array of protocols and speeds; instead they hide it all behind one anonymous USB-C connector and pretend not to hear the confusion. This Adam Savage video owes its very existence to that confusion.

  29. says

    While there are reasonable people commenting here, I see many acting like a huge crowd of people shouting at each other, promoting only their own digital speedfreak deity on an ever growing chaotic tower of babel. Some are merely obsessed with determining how many electrons they can get dancing on the head of a pin. ROFLMAO

    @30 nonlinear feedback wrote about me: you seem to have become a reactionary old curmudgeon at some point in your life. You should try pulling back from that a bit and learning more about the technology behind a 40 Gbps Thunderbolt 4 cable.

    I reply: you seem to be hyperfocused on pimping for your tech fetish thunderbolt. (I love that you have to use $100,000 worth of CT scanner to understand it) I have continued learning my whole long life (to date), encompassing a wide range of technology and have learned that so many of the gimmicks developed were not carefully thought out or like monster cable, elitist expensive trinkets and ‘tech-religious quackery’ that were indeed like making people buy a formula one race car to go to the grocery store. It is simple to understand, there are times and places where you need an ‘F1’ and there are times when a prius-like vehicle is much more affordable, practical and satisfies the requirements of that context. The road of history is littered with fancy ‘state-of-the-art’ junk that wasn’t practical, affordable or widely usable. Eat an apple if you want, that’s your choice. But, I refuse to deify fruit!

    To echo PZ’s thought about a ‘good religious war’, that’s why I avoid deities. I have listened to him and seen his videos (I enjoy it all) But, low resolution lo-fi ‘Utube’ doesn’t require any of the ‘god-like’ capability of thunderbolt. Eat an apple if you want, that’s your choice. But, I refuse to deify fruit! I leave you zealots to your ‘high tech’ tower of babel. I’ve got work to do.

  30. says

    @#33, EigenSprocketUK:

    USB 2 did not actually have a practical data transfer rate of 480 Mbps — that was Intel’s published raw number, but it was a dishonest one (possibly deliberately). For various technical reasons including the overhead of the USB protocol itself, the actual maximum which you would get was something like 250 Mbps, which you only got if you had a single device which was the only thing on its chain (no hubs), while FireWire 400 in the same circumstances gave you a guaranteed 400.

    To make matters worse, USB 2, in order to maintain backwards compatibility with USB 1 without requiring much more complex circuitry, implemented hubs by giving every device an equal amount of time, which meant that as soon as you plugged a second device in, the maximum speed of any one device was more like 125, even if one of the devices was doing nothing. (And that was true even if the new device that was doing nothing was USB 1: equal time. Plug a USB 1 mouse and a USB 2 hard drive into the same USB 2 hub, and the hard drive’s transfer rate is cut in half even though the mouse is sending practically no data.) FireWire would only slow down when multiple devices were active on the same bus; inactive devices used a vanishingly small amount of data. (USB 3 fixes the “time slicing” problem, incidentally, as long as both the computer and the hub are USB 3 devices. It doesn’t matter what version the other devices are; USB 3 hubs are much smarter and can rearrange the time-slicing to avoid turning USB 1 devices into timewasters.) (Incidentally: I’m amazed that — as far as I can tell — nobody is making a USB 3 hub with both USB-A and USB-C ports. I’d love to get one, if a reasonably-priced one existed, but I haven’t seen a single one listed anywhere so far.)

    And since USB was built as a host-centric model, if your computer missed some data from a device, unless the device driver had a protocol to re-request it, that data was gone. That’s why USB 2 video capture quality is generally terrible — the maximum speed of USB 2 is just about enough for analog TV capture, but only just barely, so if your computer had a microsecond where it was doing something else instead of watching the USB bus for input (like waiting on a mechanical drive), you lost a frame of video. FireWire was built so that devices had queues in hardware, so the need to re-request data was rare anyway and video capture was smooth. As I said: a noticeably worse bus than FireWire in just about every way.

  31. Kagehi says

    Look, the basic rule here is 1) due in part to the idiocy of them recent changes to USB standards themselves, they don’t have to tell you what the cable or device “supports”, only if it meets minimum USB3 standards, and its uncertain if they will even bother to tell you if its a USB 4 cable, when those eventually show up, 2) if you are charging something all that matters is a) if they used all the pins needed, since not enough of them won’t charge as fast or effectively, and b) the quality/thickness of the wiring, because the cheap ones will fail “faster”, or not work well at all, 3) if you need speed, and yeah, some applications do need it, including cases where they now run bloody display data over it, etc., you need, at minimum, for it to be one with the electronics built in, and the capabilities of that electronics are going to have a serious bloody impact on how well it really works. This also includes getting the best results from an external storage device, which will be hampered by slower cabling, even slower versions of of the standards and in cable electronics.

    It, honestly, really annoys me that they don’t all have some basic stuff in them, I have had more than a few times where a cable, which should have been fine, has even been so bad that it won’t charge correctly, because it doesn’t even meet the basic spec needed for a flipping Amazon Kindle, never mind an external SSD, and the device whines about “charging using low power”. I don’t buy cheap any more, ever, but.. yeah, I don’t have a specific need for “super expensive” in the short term, so..

  32. cvoinescu says

    The video has a good number of factual errors.

    For instance, they say that the two cheap cables can’t transfer data. They can: they are USB 2.0 data cables. The middle two conductors are the single bidirectional data pair that USB 1 and 2 needs. It’s probably twisted less than the standard requires, but it should be fine over short distances. (USB 3 needs two more high-speed pairs, and a fully populated USB-C should have one USB 2 (slow) pair, and four high-speed pairs, plus at least one control wire.)

    They also say that the last cheap cable has nothing else on the PCB other than the connector and the cable. However, there are two or three SMD components on the back that they ignore.

    They also say that charging is a physics problem and determined by the thickness and length of the conductors. In reality, charging speed is limited by what the power supply can supply, what the device can sink, and whether they detect and negotiate each other’s capabilities correctly. Connectors in high-current USB-C cables (like the one labelled “6A”, which has wires too thin to safely carry 6 amps of current for any length of time) have a tiny IC that tells the device that the cable is physically capable of carrying that amount of current. That’s probably one of the components on the underside of the PCB.

    The IC in the Apple cable is not a processor. It’s an amplifier/signal conditioner that allows high-speed transmission over a long cable (well, long for the data rate it’s supposed to transfer: 80 Gbps).