The USB Type-C Cable That Will Break Your Computer

USB has been on our desktops and laptops since about 1997 or so, and since then it has been the mainstay of computer peripherals. No other connector is as useful for connecting mice, keyboards, webcams, microcontroller development boards, and everything else; it’s even the standard power connector for phones. The latest advance to come out of the USB Implementers Forum is the USB Type-C connector, a device with gigabits of bandwidth and can handle enough current to power a laptop. It’s the future, even if Apple’s one-port wonder isn’t.

Ground is red, V is Black. Photo: Benson Leung
Ground is red, V is black. Photo: Benson Leung

The cable of the future is, by default, new. This means manufacturers are still figuring out the port, and how to wire it up. You would think remembering ‘red = power, black = ground’ is easy, but some manufacturers get it so terribly wrong.

[Benson Leung] is a Google engineer who works on the Chromebook Pixel products, a huge proponent of the USB Type-C connector, and a very prolific reviewer of USB Type-C connectors on Amazon. The latest cable he tested destroyed his test equipment, including a $1500 Chromebook Pixel 2 (link dead, try Internet Archive). How did a cable manage to do this? The manufacturer switched black and red.

The cable in question was a SurjTech 3M cable that has thankfully been taken down from Amazon. Swapping GND and Vbus weren’t the only problem – the SuperSpeed wires were missing, meaning this was effectively only a USB 2 cable with a Type-C connector. The resistor required by USB spec was the wrong value, and was configured as a pull-down instead of a pull-up.

This isn’t an issue of a cable not meeting a design spec. Ethernet cables, specifically Cat6 cables, have been shown to work but fail to meet the specs for Cat6 cables. That’s shady manufacturing, but it won’t break a computer. This is a new low in the world of computer cables, but at least the cable has disappeared from Amazon.

125 thoughts on “The USB Type-C Cable That Will Break Your Computer

  1. The real WTF if you ask me, is that the Chromebook apparently had zero reverse polarity protection on it’s USB socket. I guess an argument can be made that if the assumption is made that the user’s cables are OK then physically they can’t reverse polarity, but that’s a pretty tenuous assumption – as has been shown.

        1. A diode is a problem considering the amount of current it may have to pass through, and because of the 0.6-0.7 Volt drop, which is significant in a 5 Volt system. One diode on both devices at the ends of the cable causes a 25% drop in voltage and consequently, a quarter of your power is lost.

          1. You could however use a schottky to reduce to about 0,4 or an MOSFET for near to 0,1V losses. Still, a power loss, a part cost and a PCB place you’ll need to invest.

          2. Besides, the real problem is that a USB-C cable is an active device with a vendor-ID chip inside that communicates with the host to report power handling capacity and other information, including which way the cable is plugged in.

            If the power leads aren’t soldered in right, the host will short-circuit anyways because it needs an active switch to select which pin it considers to be + and –

          3. Or you could use a Schottky, which of course would be even more expensive. Trust me, the electrical / electronic engineers that designed that chromebook had a diode in place and the bean counters took it out becuase it cost 7 tenths of a penny and multiply that over millions. I work in an industry that is that tight fisted. The issue is a wonky cable, not a bad design…at least according to the accountants.

          4. If you put the diode across the power so it conducts only if power is reversed, it will trip the PTC in the USB port and shut off power. When power is connected normally, the diode is reverse biased and invisible to the circuit.

          5. At the very least you can put the diode across the lines reverse biased, and a resettable fuse on the line. Reverse polarity (you could even spec the diode for over-voltage too) and the fuse trips.

        2. If he now has short circuit on the port maybe there is a diode in there. Cathode connected to +, anode to -, if you connect reverse polarized power diode burns and makes short, protecting other more expensive circuitry from reverse polarity. Sometimes zener is used, to provide reverse polarity protection and overvoltage protection. Problem can be fixed by removing diode (not recommended) or replacing it.

      1. Did you guys miss the “[Benson Leung] is a Google engineer who works on the Chromebook Pixel products” ?

        This wasn’t a premium product at the time, it was a prototype under test. It’s not unreasonable at all to have omitted full USB protection on the new-ish type C connector, when you’re busy trying to validate whether things like your motherboard work.

          1. To @Garbz and @mrz80 – I’ve replied to M who stated that the laptop itself was in the prototype stage and the motherboard was the thing being tested. Haven’t read his blog post myself though.

    1. “[Benson Leung] is a Google engineer” WTF? If I was a google engineer and I fried a chromebook I would just grab another one off the assembly line cost = $0.00 usd. That is a far better decision than redesigning the product. This also gives the consumer a reason to buy a second new chromebook in the same month.

      1. The connector is mechanically symmetrical too, so flipping it over still results in correct polarity. The laptop should have overcurrent protection on the VBus output.

        1. I was talking about the connector. You can turn it around and the polarity would be reversed, so the resistor or chip inside the cable is supposed to signal the host which way the cable is plugged in. If the cable is built wrong, the host has no way of knowing that it’s routing power the wrong way, and the cable is built to support between 3-5 Amps of current, which does damage very quickly before the host even realizes there’s anything wrong.

      2. The V+ and GND connections are symmetrical- there are 4 connections each that are consistently V+ and GND regardless of rotation. The end pins A1, B1, A12 and B12 are GND and the A4, B4, A9 and B9 pins are always V+.

        If the cable is wired backward they will be backward both orientations. There is no negotiation of which pin will be V+ and GND since they connect before the data pins.

  2. Well, this is the 21th century where a bucket of transistors cost peanuts.

    An external user connector in a consumer device not only should be ESD protected but also against shortcircuits and polarity reversal, and not just let fry the $1500 bloody thing.

    Blame also the faulty design of the Chromebook Pixel 2 and his test equipment.

    1. I recently switched from my old notebook to a new one. As a hardware engineer I shorted the USB lines in all the possible combination in the 5 years that I used that notebook (Fujitsu Amilo PI3560). The OS just said that there was a power faliure/communication error/etc at a specific port and life goes on. Usually I rebooted the machine after I accidentally “killed” all 3 of its USB ports.
      Now I bought a premium category ASUS machine (G551, cost 1200$, and is a complete piece of shit all together, but at the time I needed a dedicated Nvidia card and displayport and this was the only choice). Shorted the Vbus to the chassis and the whole thing shut down. Accidentally cut a usb cable while it was plugged in and guess what? It shut down again. Overloaded the port? Same thing. Im glad at this point that I did not short the +5V to the data lines… that will probably kill the machine for good…
      Seems like a 0.01$ current sensing resistor and a 0.05$ FET (or a PPTC) costs too much in a 1200$ product…

      1. Sometimes I wonder if too much competition is really a good idea. In you example, if the same machine is produced in a 1 million quantity (in a range of configurations) the saved cost is about 60’000. Still a lot of money for many peoples where this kind of devices are usually produced today, but certainly not a justification comparing the near half billion (because of the low range configurations) total value of the whole batch. So depending on the level where the decision was made, it look very different.

        An other point of view could be that USB standard require a open standard microelectronic implementation with all securities integrated. The USB standard is actually very complex because of the multiple backward compatibility it have to grant. Without a easy and cheap way for the industry to handle that, bad stories will continue.

        1. Even with no competition, the manufacturers have adopted a minimum effort strategy. The idea is to make a new model every year, just like the auto-manufacturers, so if there’s trouble with any particular model it won’t show up.

          It takes approximately 10% of your customer base to complain about a particular product for the whole group in general to take notice and identify that there’s something wrong, but since each new model is only bought by a limited number of people before it goes out of production, there’s only a handful of people per model complaining, and their voices are lost in the crowd.

          You can google for some particular model for user experiences, but that doesn’t give information about the other models, nor about the new model that they’re currently selling, for which nobody has any reliable opinion because it’s too new, and by the time there is enough experience to tell – it’s no longer being sold.

          That means the manufacturers can pass off significantly lower quality products without getting noticed and their reputation as a reliable brand remains despite. It only shows up in aggregate if you take all products from the brand over a number of years and count the returns/failures. Yet it tells you nothing about the latest model – you can still believe that the new model is better.

          Hence why the average laptop has a 20-25% probability of breaking after three years of use, yet everyone thinks their laptop is of better quality because they paid so much money for it and haven’t heard any complaints from other users of the same model.

          1. If there was only one or two models of laptop sold per brand for multiple years, perhaps only offering chip upgrades to keep them up to date, and they kept breaking down at a rate of 25% by year three, people would know plenty of other people who would say “I bought brand X model Y and it broke down on me” and they would simply switch to a better brand.

            But when every brand turns the product line around every two years, nobody can tell what’s happening, and the laptops keep breaking down.

          2. One aspect you neglected to mention might be unmarked revision changes on the same model or even entirely different parts. Such as one having RAM from company X and the other has RAM from company Y.

            I don’t collect enough of any one particular model but PCB revisions/same consumer model are well documented in long manufacturing runs.

          3. Dax, you don’t need manufacturers to produce the same model. The brand name itself should be a mark of quality, to tell you that a particular manufacturer does things properly, makes good stuff. The IBM Thinkpads, back when IBM made them, were the ne plus ultra of that.

            But the problem is EVERYBODY makes shit. Because people are happy to buy shit. And as you mention, nobody has access to the long-term large-scale failure results. Except maybe large retailers, or companies that buy a lot of stuff.

            Reviews in magazines and websites don’t make a difference, they’re only testing one item at a time. I don’t think there’s much you can do. It’s only large-scale fuckups that the community ever hear about.

      2. I had 12V (with enough amps to burn things available) applied to Samsung ultrabook (NP900X4D) USB VBUS, which also killed USB-FTDI adapter on its way (MEGA16U2-based). It shut down after a couple of seconds, but I’ve just unplugged the cord and pressed the power button, and it booted back up, and the port’s as well as before. I was very pleasantly surprised – heck, the 5V is probably the same 5V line that’s all across the motherboard powering things, and the board’s tiny – yet seems to have the proper protection. I’m surprised to hear about this one

      3. Had the same problem with one of those mini laptops from asus some years ago, but even worse: some usb devices plugged in would reset the thing. My guess….. there was no protection between the internal 5V supply and the usb.

      1. For those diving down the archives: Type-C is a standard for the connector. It is used for USB, audio, Display Port, thunderbolt, charging…

        Maybe there is still a need for these resistors for when the peripheral is not using a Type-C cable, but the host has a Type-C port instead of these large rectangle standard Type-A port.

        It is 2022, and everywhere, it is still Type-A ports on hosts.

  3. I thought that notebook designers always included some kind of of overcurrent protection in their PSU design. Most of those buck converters work in current mode anyway, so why Google f***d up the design, while those controller chips for notebooks include in their datasheets a reference design with all the protection circuits? Worst case scenario would be a blown P-MOSFET and/or controller chip. Best case: Vusb line would switch off until short is removed.
    I’d expect that [Benson Leung] knows, how to fix this. That might be a good read for people who never fixed any broken notebook themselves. But from Google’s perspective teaching people to fix their notebooks would be bad for business. Western economy is based on the concept that if something breaks, owner replaces it. Here we try to fix things first…

    1. The idea behind this is that today equipements are made to last 1 year before replacement by the last version. Manufacturers try to save any penny on manufacturing cost. If it brake after the warranty they don’t care at all.

      This spring my Samsung TV failed 2 months after the 1 year warranty expired. I send it to a shop for repair but they told me it would cost the price of a new one…So I bought a new one from another brand.

      1. You might try to fix old one, it would cost you probably some time and less than a dollar in parts.
        “Repair” in most cases is just another word for replacing the entire module. I fixed my LCD monitor once, while being drunk, by replacing faulty capacitor with spare from broken computer PSU. Other time my washing machine started to spill dirty water all over my bathroom. Replacement part costed me less than shipping, and that was 3 USD. Fixing it took me one hour. New machine would cost me at least 150USD…

  4. Well, atleast the one port not wonder MacBook has proper input protection. I once shorted a 12V rail to a data pin on a macbook and it instantly switched off. I thought that was it, but it had only shut down.

        1. Have not heard about Cutlery, but I have heard about radioactive cast iron furniture from Mexico. What usually happens in Cobalt-60 form old medical radiation equipment gets mixed into the scrap heap accidentally and becomes a part of new iron/steel products. I think they screen shipping containers for radiation now.

    1. china is a ‘sell and run!’ country. they could care less if people die due to their neglegance. its just westerners; they don’t matter. more will buy our crap and when they die, more will still buy from us.

      this whole thing about china really need to be addressed. its criminal what that country gets away with, and we LET THEM because nothing much is made in the west anymore (for consumer tech, at least). we let all our manufacturing leave our shores and this is what we got.

      I wonder if this can be fixed. things don’t look good for this generation, but maybe the next one will find a way to fix it. the way we’re going, though, china will end up killing us all.

  5. Planned Obsolescence is the deliberate shortening of product life spans to guarantee consumer demand.

    Go google the light bulb conspiracy. I design LED lighting and I am not against LED technology or anything like that. It is, however, a sad fact that incandescents that could last much longer were made to burn out…

    Washer, Dryer, and thank God we pushed back on cars crapping out at 50,000 miles. Don’t remember those days? Tires that lasted 20k? China makes products that are chinese quality because they are happy to have anything and they have a short life span (the people, not the products).

      1. +1 I live in a VERY wet (IE Rainforest), I for one will happy buy new tires any day as long as I can stop and maneuver when the drunk/drug drivers do their thing. Solid metal tires I am sure would last a LONG time but wouldn’t help you stop. In fact my dad went to a family reunion one year and someone had a Model-T he said the tires were only like 2 inches wide as big as wagon wheels and had “NO-SLIP” or something like that embossed where you would expect to see tire sipping or tread. The owner told my father it was his favorite winter vehicle because the sheer mass and narrow tire profile basically meant that you couldn’t get the tires to slip even on most ice and slick roads.

        1. Narrow tires have other problems. They tend to buckle under the car, they’re easier to slip off the rim, they have very little sideways traction, which they lose quickly. But they do bite the road well because of the small contact patch.

          So it’s basically easy to get going and go fast, but don’t take a hard turn on the ice or you’ll be in the ditch.

      2. Your average passenger car tire should last 30K to 35K before it needs to be replaced. BFGoodrich Z rated tires come with a 35K warranty and their Advantage T/A tires come with a 75K warranty. You should replace your tires when the tread depth is less than 1/8th inch. The only tires I’ve owned that needed to be replace before hitting 35k were the cheap Korean tires that came with my Hyundai accent.

    1. >”Go google the light bulb conspiracy.”

      The lightbulb conspiracy is a myth/hoax/propaganda. It’s not possible to make an incandecent bulb last longer without sacrificing the already bad efficiency even further.

      The point of the “conspiracy” was to punish those manufacturers who would sell sub-standard products with the promise of longer bulb life without telling the customers about the trade-off of dimmer bulbs.

    2. Maybe *you* should Google the “Light Bulb Conspiracy”, and read carefully. I did, and here’s what I found, edited a bit for better clarity:

      “Incandescent bulbs all use a tungsten filament. A hotter filament is more efficient but burns up more quickly. It is very simple to make a bulb last forever: use a longer and thinner filament, which does not get as hot, and glows more red than white… A standard 100 watt bulb costs 50 cents, lasts 1500 hours, and uses $18 in electricity over that time (at 12 cents per kWh). The new everlasting bulb [if designed to produce the same amount of visible light but with a longer and cooler filament] will use about 3 times as much electricity over its first 1500 hours, costing an extra $36 to save a half dollar. And another $36 for the next 1500 hours. This is about $200 per year more than the standard bulb which is designed to burn out quickly and save money. The money saved represents a large quantity of coal or natural gas that would be burned to save a few little bulbs.”

      You can also take a standard 100W incandescent, and dim it down to 50W with either a dimmer, or a diode. And it will last basically forever. But it’s still inefficient, producing only about as much light as an undimmed, standard 25W bulb. If you needed the original amount of light to adequately light the room, you’ll have to use four 100W bulbs; which at 50W each, will now consume 200W instead of 100W. Again, you pay much more than you save.

      There are enough examples of REAL planned obsolescence in the world, those due to nothing more than corporate greed. But not the light bulbs, or the tires; they’re merely examples of people misunderstanding valid compromises.

      1. The misunderstanding comes from the fact that there did exist a Phoebus cartel which standardized the lightbulb to 1,000 hours and -then- did what cartels do which is agree to non-compete and sell the bulbs at higher margins.

        But before the cartel there were a number of bulb manufacturers who each made wild promises of bulbs lasting any number of hours, and since the consumers had no reliable means to measure light output (no photodiodes/transistors etc. because this was the 1930’s) they could pass off substandard products without a care in the world.

        1. Sorry to burst your bubbles guys, but we have an incandescent bulb that puts out great light, and lasts a minimum of 20,000 hours. They go into places where replacing it is a hard thing to do. My point is more about the planned obsolesce of consumer goods that don’t need to be. I have seen 100 year old electric fans running just fine. Go look around at all the SubZero products that last for generations. You can get a replacement wood door panel for a rolls royce. Somewhere there has to be a balance between affordable and junk. That is my point.

  6. I actually truly benefitted from this post. I’ve had a usb port on my desktop that was killing webcams, etc. So I just took a meter to the ports, and… sure enough. Instead of +5v, I was registering -10v. Any idea why this might be the case? I haven’t checked the wiring yet, but.

  7. What made me laugh is that the engineer reset and then re-flashed the test equipment expecting that it would some how come back to life.

    Makes me imagine a airliner pilot that has just crashed into a mountain and miraculously surviving and then reaching for the switch to re-start the engines.

    1. yeah, if it blew the test equipment, i wouldn’t be turning around and just plugging it into my premium laptop… He said he checked with a multimeter afterwards – maybe that should be his first step from now on….

    1. due to corporate ‘diversity quotas’, each week a new pair of wires get to be this week’s “red and black”. this week it was yellow and green; and next week, I hear purple and brown are the ones that will carry power. how exciting!

      black and red don’t deserve the monopoly on carrying power. not if management has any say in it!

      (lol)

  8. 1. “Engineer” Benson Leung deserves to have his equipment damaged. The idiot plugged Chinese hardware of unknown quality in without doing even minimal tests to see if damage would ensue.

    2. Ether the USB-C specification and/or the hardware that was damaged by the USB cable does not have even basic electrical protection.

  9. If the USB guys want to innovate and insist on changing the connector shape, why not change it so that it’s obvious which way the connector is plugged in? I see people trying to force usb plugs in the wrong way all the time, I can only imagine what’s going to happen once we are working with such high voltages.

          1. Unless Howard thinks the average user is dumb enough to try to force an oblong into the jack perpendicular to its intended orientation. I wouldn’t put it past some of my customers.

    1. I wonder, why they never got the idea to use something in form of 4-pole audio jack. It can be plugged only one way, it’s rugged and was actually designed to last. Not that USB micro A or B rubbish that is in everything now…

      1. The problem with a four pole jack is that there are three possible ways it can connect, and all three happen every time it’s inserted/removed. Not to mention the contacts are exposed to the environment when not inserted. Also the USB C connector is reversible, so the problem is already solved.

      2. A jack-plug shorts its contacts for a moment when inserted. No big deal for an amplifier, a really big deal for a data bus. Add that to the wonky contact, most older plugs experience, and the fact that a jack plug alread has its purpose for connecting audio, there is a whole mess waiting to happen.

  10. Hi, I had a near miss due to a knock-off micro USB cable supplied WITH THE CHARGER!!!! this was a name brand fast one too.woke up to the smell of burning, fortunately it hadn’t ignited anything but the cable was a black and charred mess.
    Pretty sure moisture got into the end and caused enough of a resistance to induce a thermal runaway event.
    Took pictures but the charger seems undamaged and still using it with a new cable.

  11. This cable can’t have been a mistake. I mean, really… How do you get not only the polarity of the power portion of the cable wrong, but also manage to use the wrong value resistor AND swapped pull-ups and pull-downs on the data lines? This makes it sound like they were TRYING to burn down someone’s PC.

  12. Maybe I’m missing something fundamental in this, but am I correct in thinking the engineer killed his computer by plugging in an open ended cable? Even if the power leads were swapped, where’s the current path?

    Go the next route, assuming the non computer end of the cable was plugged into some other device, why didn’t it die, or take on some kind of damage just as the chromebook did? In the latter case, one could make the argument that the chromebook is under designed on a good day, and that it can’t properly support the power capacity inherent in it’s own design.

    How else would you explain the computer killing itself by being plugged into a peripheral, while the peripheral walks away from the same conditions unscathed?

  13. Chinese quality control = ship 130% of your product order – omit the inspection and let the customer determine if product is acceptable.

    Thats why I think a google or apple car will fail. It’s one thing to accept a $10 piece of throw away electronic junk. You need a whole lot of “affluenza” to throw away $30,000 cars because they suddenly stop working or after a “software update” you can’t make right hand turns anymore.

  14. My kid told me how they “hack” the PCs at school (he’s 10) to reboot the machine and get out of whatever walled garden they put up. He showed me at home. Yup, he was putting a paperclip in the USB port to trigger an overcurrent and reboot the thing. What happened to PTCs on the Vbus line and an overcurrent message?

    Mark my words, this’ll be the least of your worries. We’re heading toward active cables with power negotiation through the CC channels for USB power delivery. Some of those solutions are microcontroller based. What’s preventing a malicious firmware load that spoofs that CC traffic to send bursts of +20V occasionally? It’ll work fine for a while and every now and then throw a slug of voltage at the poor little webcam at the other end. Ffft.

  15. lenovo laptops have pretty good usb port protection. I bought some refurbs from newegg, and every single one had a damaged usb port (the plastic insulator block was broken off) I plugged a mouse into one of those on accident, and the mouse is dead forever, but the laptop told me ‘USB short- circuit detected’ or ‘usb over-current’ or something. so i pulled the battery and unplugged the AC adapter, and isolated the pins from one another. and upon booting and testing all the ports with a chargerdoctor (aka USB ammeter/voltmeter) they all tested good, and devices would connect to them again. but i still returned the laptop to newegg because that wasn’t very ‘refurbished’, blocking off the port would have been an acceptable solution, not a proper fix, but would have prevented the end-user from inadvertently plugging something into a usb port that was missing its insulator block. plus it only came with 3 out of 4 pins. i kept the others and just blocked the ports off for now to keep people from shorting them out.

  16. I think what we really need is a revision of USB that is wired symmetrically, like a palindrome, “stop pots” then you could plug it in either way. Yes, I am aware of the existence of the Apple Lightning Connector,(how could I not recognize the only good idea that’s come out of a company in 20 years?) but there needs to be a new kind of USB port to really do it right.

    1. kinda, except they used too many wires. 8 is a much safer number. they got too complicated with it. just mirror the same 4 wires on both sides. how can they expect the child laborers in the sweatshops to properly place 24 tiny wires?. I never blew up my devices when i ripped the male ground shield off of a usb adapter, so i could use it as a male or female. I think the real problem here is chromebooks and tablets and all these newfangled devices are designed to fail. you buy them, they break, repair is often not cost-effective, eventually they sell you another. and it’ll either break within the year or fail just after the warranty expires.

  17. Who would trust a cable named “surj” anyway?
    My usb 2 and 3 port are protected against reverse plug (their plastic is not tight ) I tried. It says surge and works great after reboot.

Leave a Reply to DroneCancel reply

Please be kind and respectful to help make the comments section excellent. (Comment Policy)

This site uses Akismet to reduce spam. Learn how your comment data is processed.