Slashdot Mirror
Amazon.com Now Bans USB Type-C Cables That Aren't Up To Spec (google.com)
Google engineer, Benson Leung has been on a mission to get rid of USB Type-C cables that aren't compliant with Type-C 1.1 spec. He reminds us that these cables could potentially lead to damage. Over the past few months, he has reviewed over a dozen of USB Type-C cables on Amazon.com and concluded that the vast majority of them aren't compliant with the aforementioned standard. Now he reports: Amazon.com has just made a change to their "Prohibited listings" for Electronics. They've added the following line: Any USB-C (or USB Type-C) cable or adapter product that is not compliant with standard specifications issued by "USB Implementers Forum Inc." What does this mean? It means that cable manufacturers who sell poorly made or intentionally deceptive USB Type-C cables and adapters are banned from Amazon, officially. Really great news, but we all have to continue to be vigilant and call out any bad products we find on Amazon and other stores (both online and brick and mortar) as we find them.
Great news would be Amazon white-listing compliant cables, I have a hard time imaging El Cheapo Cables Inc. being overly concerned about a bullet point in the amazon ToS.
Slashdot bans postings that aren't up to spec. Like this one.
Put it in the devices where it belongs.
So back in the day when I did motherboard design, the biggest headache we had during our automated testing was USB keys and USB hard disks that had bad FW such that they would randomly disconnect, or otherwise hang up host-side code. MS Windows is least tolerant of this, and would often blue-screen. It seems every generation during our testing we'd get either blue-screens or BIOS lockups with some of these devices, have to go on a 2-3 week crusade of signal integrity analysis and measurements to prove that electrically nothing was wrong. Then inevitably we'd hook up a protocol analyzer and see things that just plain didn't make sense: the disconnects happened for NO reason. They happened with some vendors and not others, or certain devices from one vendor but not others.
Lots of money spent, lots of time wasted, but it turns out that that cheap overseas shit we all love so much doesn't always work so great. The bottom line is if you are going to have a standard you have to have some way of keeping people from sticking your logo on it if they cannot meet the requirements. It's great this Google engineer took up the mantle of shaming bad products, but the problem is more widespread than mere cables.
Now Amazon needs to deal with their entire counterfeit sex toy problem. If you're not aware, never buy sex toys off Amazon. Most of their products are low quality, counterfeits of more respectable brands. Often they're unsafe or made to low standards. Most manufactures will stop selling to any store that uses Amazon.
Amazon has gotten better about such things. You no longer have to go through the foreign support people with the forms and scripts. They now have a direct contact for unsafe product issues:
I would add UL (underwriters laboratories) and several others. UL moves a bit slow and reactive instead of proactive, but they certainly are zealous about protecting their brand. Products with their mark, that test out unsafe, will be quickly dropped from Amazon and elsewhere.
Unfortunately, this doesn't help with all those 2GB USB flash drives from China, which are labeled and firmware hacked to appear to have 64+ gigabytes of usable space.
Slashdot gets worse every day... Pipedot: News for nerds, without the corporate slant
USB is still a serial protocol with the change between USB 2 and 3 was 1) splitting a common pair used for sending and receiving (simplex) to two pairs to allow for duplex transfer 2) specifying the higher speed pairs to have tighter ratings to enable higher bit rates.
The idea that Ethernet (assuming you mean common Cat 5/6 cables) works better than USB cables is ludicrous! They aren't specified to tolerate the plug/unplug cycles of even a cheap USB cable and the plug itself is fragile.
I wonder what you are comparing to? The USB is a good standard and all common devices need no specific driver. Keyboards, mice, joysticks, printers, Ethernet adapters etc. just simply work with a common USB stack. So why do you call it shitty?
The complication we are talking about here is that the _cable_ between a host and a device surprisingly need to be specified to tolerate the currents it conducts (yes it's a strange thing). Enumeration is between the host and the device, not between the cable, the host and the device.
This isn't a problem with the USB standard, it is a problem of manufacturers making crap and not caring of potential hazards.
This is an important issue to me because I have devices that need good USB-C cables. If anyone else is in the same boat, here's a direct link to Benson Leung's reviews. Focus in on the 5-star ones and look for the value buys (if the product is still available):
https://www.amazon.com/gp/pdp/...
One of these cheap cables wasn't wired correctly and when Benson connected it to a Pixel C to charge some device (Nexus Phone?), the Pixel was destroyed. Apparently during manufacture two wires in one of the connectors were switched.
One difference between older USB cables is that the Type C cables contain a 56 k ohm pull up resistor for current control purpose. Some of the out of spec USB C cables with at least one USB Type C plug - probably a USB Type A plug at the other end - have a lower valued resistor and can cause problems. The problem is that if a lower resistance is used with a power supply that can only provide 1 Amp instead of 3 Amps at 5 Volts, the power supply can be fried as it tries to deliver 3 Amps. This could be the case for powered USB ports on computers. I've read that Apple laptops with a Type A compatible connector cannot deliver 3 Amps (1 Amp?) and might be at risk of damage when using an out of spec USB Type C connector cable with the wrong resistor. Further more, these out of spec cables may not be cheap. For more information, check the linked page and scroll down a bit:
http://www.androidauthority.co...
In a time of universal deceit, telling the truth is a revolutionary act. George Orwell
so that's why they call their tablet fire...
---
It's the port design on shitty (Intel, mainly) mobos. They're not individually fused (or fused at all).
The cable in question simply had the wrong pinout, and threw voltage onto lines that shouldn't have had that voltage.
You can't physically stop someone from applying potential to your exposed pins, but you can reasonably guard against it. Intel mobos typically don't (or didn't). All the brands people use for building their own (ASUS, GIGABYTE, ASRock, EVGA, MSI, Biostar, etc.) advertised USB (and other) short/spike/etc. protection as a feature years ago when it was becoming a frequent problem.
USB-C is designed to be able to deliver 100W DC (20A) - brains or not, swapped wires in such a cable is quite likely to be able to fry something. Lots of electronics aren't going to be able to survive up to 20A of current in a reversed polarity, or delivered on a pin that was supposed to be an outgoing signal or voltage.
In this case it sounds much simpler - and is a problem that could affect standard USB 2 and 3 cables as well: The wrong identifying resistor was included in a C-to-A adapter, making the device think it was plugged into a high-current power source, when the reality was that the USB C port was only able to deliver 2 amps. The resulting current draw then fried the USB port's power source, destroying the port and possibly the connected device.
A related problem is commonly responsible for slow charging with old-fashioned USB ports: The spec defines a 0.1A maximum current draw unless the device has negotiated for more. But having to talk to electronics makes for expensive wall-warts, so an auxiliary standard was created whereby the port could identify itself as a "dumb charger capable of delivering X amps" by including a ingle resistor, whose resistance was used to specify X within a few tiers, including tiers far in excess of what a "real" USB port can deliver (As I recall USB 2 ports are specced up to ~2A, assuming the connected device successfully negotiates for more than it's default 0.1A. Dumb chargers can be specced up to 5A with the right resistor) Some cables can interfere with that, generally resulting in well-behaved electronics "failing gracefully" and charging at a much slower pace
Not being versed in the intricacies of type--C lore, it sounds like what probably happened is that adapter cable *should* have identified itself as something like a normal low-current type-A port to connected devices, but instead delivered a garbage resistance that got interpreted as "take all the power you want", and the type-C port just couldn't handle the resulting load.
--- Most topics have many sides worth arguing, allow me to take one opposite you.
The max current at any power level is 5A. The 100W spec is 5A at 20V.