[HN Gopher] All About USB-C: Illegal Adapters
___________________________________________________________________
All About USB-C: Illegal Adapters
Author : zdw
Score : 183 points
Date : 2022-12-27 18:14 UTC (4 hours ago)
(HTM) web link (hackaday.com)
(TXT) w3m dump (hackaday.com)
| zzo38computer wrote:
| USB is really terrible for many reasons. This article describes
| some of them, but it is worse than that. (I knew USB was bad; I
| did not know it was quite that bad!!)
|
| I think for charging, there should be a charging only port, and
| for data, having separate ports which you can tell which physical
| port is in use, and perhaps RS-232 or MIDI would be good, than
| USB which is no good.
| RockRobotRock wrote:
| USB is the worst connection standard, except for all the other
| ones.
| ec109685 wrote:
| I get what the EU is aiming for by banning lightening connectors,
| but it would be nice if it was coupled with improvements to
| USB-C.
| curiousgal wrote:
| This is why I don't think that would change anything. I am more
| that certain Apple would abuse the standard to force the users
| to only buy its cables.
|
| Edit: I know complaining about downvotes only solicits more but
| I am genuinely curious why two people thought this comment was
| so irrelevant to the discussion that they downvoted it. Only
| explanation is that they believe that Apple could do no wrong.
| Lol
| elzbardico wrote:
| Really, the day the iphone gets a USB-C connector will be a
| sad day to me. I hate that stupid male connector disguised as
| a female smd mounted in a board.
| scarface74 wrote:
| Yes because it's not like Apple already has been shipping iOS
| based devices that support USB C that can be used with any
| USB C cable and supports all the relevant standards including
| video over USB C.
|
| Also, Apple will be forcing all Mac apps to go through the
| Mac App Store any day now. Since it introduced the Mac App
| Store a decade ago.
| jasonhansel wrote:
| For a while, I've been using a USB-C female to USB-A male adapter
| to connect my MagSafe charger to various old-school USB wall
| chargers. Somewhat surprisingly, this works perfectly; the
| charging speed is probably slower but I haven't noticed.
|
| That said, I'm guessing this violates some specification or
| other. Ah, well, no explosions yet.
| puzzlingcaptcha wrote:
| About USB-C to barrel plug connectors: I think people generally
| use a "PD decoy" board with that setup, which does Power Delivery
| negotiation (usually based on the cheap IP2721 chip but you can
| get a fancier one from ST) with a PD-capable power brick. This
| way you can easily get 5/9/12/15/20V out of a single decent
| quality GaN charger.
|
| Now obviously you want to be careful with where you plug it in
| but it's still quite convenient.
|
| A more advanced way to do this is use an arduino board that can
| negotiate arbitrary voltage in 20mV steps with a PPS-compliant
| charger, effectively giving you a tiny lab power supply (although
| you can't get under 3.3V and the chargers are still a bit pricy).
| matthewfcarlson wrote:
| I know I'm biased towards hackaday but I think this series is one
| of the better ones they've ever done. The writing is tight and
| clear. You can tell Arya is an expert in the area.
| _Microft wrote:
| Have a look at this tool to help find out what 'capabilities'
| your cables have:
|
| https://github.com/alvarop/usb_c_cable_tester
|
| It has a number of connectors, the cable gets plugged into the
| board with both ends and labelled LEDs will light up for wires
| that are connected. It runs off a coin cell, no microcontroller,
| no single-board computer, no nothing. Just a power source, LEDs
| and connectors.
|
| From what I remember, the creator said that the repo should
| contain everything required to have some manufactured by JLCPCB
| (that's a large Chinese PCB prototyping service that also
| populates boards with components if you like).
| sschueller wrote:
| Yes, you can get 5 boards assembled for around USD 44.-
| https://github.com/alvarop/usb_c_cable_tester/blob/main/ORDE...
| kps wrote:
| I still have my RS232 version, but I thought those days were
| over.
| carabiner wrote:
| I guarantee that within 3 years, a "USB-lite" will be announced
| with minimal features in an attempt to solve this mess. Yay, now
| you have n standards.
| whatever1 wrote:
| Is there any certification body that also has a list of certified
| usb products ?
|
| I literally cannot predict if a cable can charge, transfer data
| slowly, fast or if it will fry my device.
|
| I mostly want a *hard requirement for a specs label ON THE
| CABLE*. How else am I supposed to know what they are for? Even
| Apple is not labeling their cables.
|
| Same physical appearance and no labeling is recipe for disaster
| randy408 wrote:
| https://www.usb.org/products
| whatever1 wrote:
| That is very helpful thanks!
| googlryas wrote:
| What ever happened to "Benson from Google", who was probably
| the closest thing to a certifying body that the public could
| trust.
| theshrike79 wrote:
| If it's the same person I'm thinking of, most of the content
| disappeared with Google Wave.
| cactusplant7374 wrote:
| Holy shit. I do have to flip my Type C extension cable sometimes.
| This sort explains what is happening.
|
| I would like to have really long cables (12ft+). That way I don't
| need to bring a bulky extension cord with me when traveling. Has
| anyone found a way to make this work?
| layer8 wrote:
| At this point, they should rename it to MSB - Multiversal Serial
| Bus.
| hoosieree wrote:
| Ah, this takes me back to that time I had to add a USB-2.0
| adapter inline with an all USB-C signal chain in order for the
| operating system to recognize the device.
|
| And that charger I have which is USB-C but only works in one
| orientation.
|
| And those expensive OpenFlow/P4 compatible switches in the
| datacenter whose management interface is serial (as in RS-232),
| but the form factor of the connector is USB-mini.
| LarryMullins wrote:
| I think there must be something fundamentally wrong with many
| of the USB-C / USB 2.0 adapters on the market. I've burned out
| 10 of them already, using them to connect my old trusty mouse
| (5V, 100mA). The mouse works fine (20 years after I bought it!)
| but the adapters only last a few months at most before they
| start intermittently disconnecting, eventually failing
| completely. I'm at my wits end with this because there is no
| mouse on the market today with the same form and button layout.
| TeMPOraL wrote:
| > _before they start intermittently disconnecting, eventually
| failing completely_
|
| My phone had this problem. I went through several cheap USB
| cables over the course of a year, before finally realizing
| it's not the cables that were the issue (other than being
| cheap and easy to deform): the issue was caused by lint
| accumulated in the USB port. After scraping it out with a
| needle the issue went away.
|
| The lint and dirt will accumulate over time, and get
| compacted in the port by the charging cable, making it very
| hard to spot, as it looks just like the back wall of the port
| would. Cheap USB cables will easily deform after being
| plugged a couple times, making them much easier to pull out -
| but this alone will _not_ make the connection intermittent,
| even though it _feels_ that way if you 're not aware of the
| lint problem.
| devwastaken wrote:
| On top of bad usb-C, the chips used for Ethernet and HDMI are
| also overwhelmingly poor. Every adapter overheats, so hot you
| cannot touch it, even with a full metal case. We need a standards
| body to ban junk products designed to be unsafe and/or destroy
| your devices.
| Robotbeat wrote:
| Are there any competing connectors under development now? I
| personally wonder of the mechanical robustness of a device-side
| port that isn't a pure female port.
| onewheeltom wrote:
| The best thing about USB-C is the connector. Everything else is a
| mess. I have a powered speaker with a USB-C port for "charging"
| that does not work with a USB-C to USB-C cable. Only USB-A to
| USB-C cables will charge the speaker.
| Foobar8568 wrote:
| Ducky keyboards are the same...
| wlesieutre wrote:
| I have a USB recharged LED lantern with the same problem
| zlsa wrote:
| This is a super common f*ckup in a truly ridiculous number of
| devices. It's so common that somebody published a PCB design to
| retrofit noncompliant devices to make them work:
| https://github.com/ide/usb-c-to-c-power-mod
| clumsysmurf wrote:
| I hate USB-C connectors. All my phones have had the same issue:
| USB-C connector busts and you can't insert the cable any more.
| I use high quality Anker cables, and I keep having to buy new
| Google Pixels over this. Replacing the USB-C port is around
| $160.
| jacquesm wrote:
| Username checks out ;) But seriously, yes, they are
| ridiculously fragile and usually only SMD soldered on, no
| hole through part to give it some mechanical strength. Very
| frustrating, I am _super_ careful inserting USB-C connectors
| and have broken a couple already, fortunately not on anything
| expensive.
| clumsysmurf wrote:
| Do you recommend any kind of port-saver magnetic cable ...
| something like https://www.amazon.com/Magnetic-Charging-
| Charger-Transfer-20... but with high speed data?
| jacquesm wrote:
| I haven't used anything like that, for me the problem
| isn't tripping over cables but basically the force
| required to attach the cable in the first place. I guess
| a magnetic pigtail would reduce the number of times I
| need to insert the cable but a regular USB-C extension
| cable would do just as well for that purpose (and that's
| what I use right now).
|
| That's still not perfect because the 2 cm of plug
| sticking out of the side of the device is long enough to
| put substantial leverage on the soldered part inside the
| device but that seems to be a given. I do try to find the
| shortest possible plug to minimize that force, but can't
| find them much shorter than that.
| sephamorr wrote:
| This probably isn't your issue, but I've had issues in the
| past, usually with a cell phone that sits in my pocket, that
| show up as cables no longer maintaining detent and falling
| out. The issue was actually accumulated dirt in the bottom of
| the female connector. Scraping it out with a toothpick fixed
| it for me. Just a FYI.
| genewitch wrote:
| i feel as though the port in the phone is male, and the
| cord side is female, since the part that actually "mates"
| goes into the cord, not the other way around.
|
| when i started typing this i had a coaxial cable in mind as
| an example of "looks like female but isn't", although UHF/F
| come close. the "male threaded" part is not the male
| connector.
| jacquesm wrote:
| There are little rubber covers that you can use to close
| the connector hole when it isn't in use.
| pigsty wrote:
| If you keep buying the same cables and they keep breaking
| things, they're probably not high quality.
|
| In my experience, Anker products are nothing short of
| garbage. Keyboards and adapters all break within weeks. I
| feel like all the "anker is the best!" stuff I see online is
| stealth marketing because I've had better results buying
| absolutely anything else.
| deathanatos wrote:
| The connector on every phone I have ceases to work eventually
| not because the connector is broken, but because the socket
| has (sufficiently) filled with pocket lint such that the plug
| isn't able to reliably connect anymore. One just needs to
| remove the lint from the port. (And note that you won't see
| the lint by gazing into the port, either; it has usually been
| compacted into the bottom of the socket by the plug squishing
| it down.)
|
| I've fixed ~4 phones by doing that.
|
| This was a problem with USB micro, too.
| axiolite wrote:
| Magnetic cables are a good solution to this and other
| problems. The magnetic tip stays inserted at all times,
| preventing lint infiltration.
| Certhas wrote:
| Any suggestions for how to clear the socket out?
| hallway_monitor wrote:
| I know the sibling suggested a toothpick, but everything
| besides a needle has been too large and too week in my
| experience. There's nothing important at the bottom of
| the port so just jam it in there and scrape around until
| you can get a bunch of lint out. You will be surprised
| what you find.
| Dma54rhs wrote:
| Plastic sharp object to pick. Maybe toothpick and loads
| of lighting.
| delecti wrote:
| I'll second the "relevant username" sentiment, but without
| the winky face. I've had phones break in a variety of ways,
| but the USB-C ports have always kept working.
| eternityforest wrote:
| I have never once seen a USB-C connector fail device side.
| It's so far the most reliable connector I've ever used long
| enough to have an opinion, perhaps with the exception of US
| 120v plugs which suck for their own reasons.
| gambiting wrote:
| I was going to say the same thing. I don't understand how
| that's possible - it's incredibly frustrating.
| elzbardico wrote:
| Really, compared to lightning USB-C connector is pure trash.
| But compared to USB-A I concede it is an improvement.
| genewitch wrote:
| i always worry about the lightning connector shorting out
| when it's not plugged in to something.
| gumby wrote:
| The way USB C is designed you have to go out of your way to
| make that happen. I believe you (!) but I'm not sure how I
| could build a device to behave that way. Could it be due to a
| weirdly noncompliant power adaptor?
| wiredfool wrote:
| I've got a pair of cheap Bluetooth headphones that do that
| too -- they only work with a usb a to c cable, not with a
| proper usb c power supply and cable.
| fragmede wrote:
| That device, the speaker in this case, is missing the pull-
| down (Rd) resistors on CC1 and CC2 to say that it's happy to
| get 5v@500ma when connected using a USB-C to USB-C cable.
| Those resistors aren't needed/used for signaling when using a
| USB-C to USB-A cable, which is why that works.
|
| https://community.silabs.com/s/article/what-s-the-role-of-
| cc...
| RobotToaster wrote:
| So they saved a fraction of a cent per unit on two
| resistors?
| fragmede wrote:
| To GP's point, that's the problem with Postel's Law. It's
| possible that was intentionally skipped to save
| fractional cents, but I think it's also possible they
| made the device and just never tested with a usb-c to
| usb-c cable. "It works on my machine (which is too old to
| have usb-c ports because we don't buy macbooks for
| hardware designers)!"
| numpad0 wrote:
| No, it's the peripheral device. There were initial confusions
| among bootleg/hacky manufacturers such as:
|
| - whether CC pins are to be handled by the cable or by the
| device,
|
| - whether added pins are for insertion detection or for
| protocol use by the host,
|
| - which pins are to be pulled down or left NC for charging,
|
| - whether USB-C is meant to be a "mobile" connector, or a
| replacement for the USB-A connector, or how soon the
| replacement is to occur.
|
| As the result, there were some devices that has a USB-C input
| but only works with an A to C cable.
| TulliusCicero wrote:
| I have a retro games portable that behaves the same way.
| yjftsjthsd-h wrote:
| Always, or did it just break the ability to flip it upside-
| down?
| zlsa wrote:
| If the 5.1k pull-down resistors are missing on the device
| side, it will not charge from a USB-C wall brick, laptop,
| etc. in either orientation, but it will work fine when
| charging from a USB-A outlet with a USB-A to USB-C cable.
| This is because USB-C outlets are not permitted to provide
| any power until they detect the presence of a "legacy" device
| that just wants 5v (as signaled by the pull-down resistors on
| the device).
| Animats wrote:
| The USB crowd has managed to replicate the decades old problem
| with DB-25 RS-232 cables.
|
| I once worked in a data center where there was a rule that all
| cables over 1 foot long must have all wires and be 1:1. Any
| "adapters" were very short, labelled cables. USB-C is now there.
| karmakaze wrote:
| This was far more interesting than expected. Title more
| appropriately "illegal and/or useful".
|
| This talks about power. I'd also like to read one about all the
| possible data and video usages and gotchas.
| blacksmith_tb wrote:
| And likely not "illegal" (housefires notwithstanding...) but
| just in violation of spec (which is great when it scratches
| your itch, but obnoxious when some combo of device and cable
| won't do what you expect).
| yjftsjthsd-h wrote:
| Eh, that's within the way the word is used; giving your CPU
| an illegal instruction won't call the cops, but it will crash
| your software.
| ChuckMcM wrote:
| This is a great discussion of all the ambiguity that has been
| added to the USB spec along the way and how that ambiguity can
| bite you. I'm fairly convinced at this late stage that the IETF
| motto of "precise in what you send, flexible in what you receive"
| is basically evil. When standards work around the non-compliance
| of clients in creative ways (like HTTP browsers did, A LOT) they
| become "well it isn't in the standard but if you don't do this
| you'll break a lot of things out in the field" problems. That
| resistance to breaking existing clients is understandable, but
| ultimately it is a trap (in my opinion) that results in both
| security issues and safety issues.
|
| It would be interesting to have a USB 4.x spec that was both
| explicit in all the behaviors AND had explicit tests for and
| responses to non-compliant stuff. I don't think the IEEE or
| anyone else for that matter has the stones to try the experiment
| though.
| brigade wrote:
| Of these, I think the only one that can work solely because of
| an overly "flexible in what you receive" is the USB condoms,
| which shouldn't be encouraged in the first place. Specifically,
| that people wanted to charge at more than 100mA without doing
| real negotiation, so they ignored official USB specs came up
| with a variety of static or simpler negotiations that could be
| mimicked.
|
| The un-negotiated 20V power input is a WTF and I've never seen
| that before, but the typical behavior from a port not expecting
| that is getting fried so it's not like anyone else is being
| overly liberal.
|
| The extension cable works because basically all physical
| connections from the last 30 years or so are designed to still
| try to work at a degraded level with subpar signal integrity,
| because the world is analogue and you can't have a 100% perfect
| signal. The sometimes missing pins can't be required without
| breaking compatibility with pre-C ports that don't have them.
| Likewise, the male-A to female-C adapter can't really be
| distinguished from a male-A to male-C cable, so you can't
| prevent it from working without also blocking those cables.
|
| IMO the real issue is the USB-IF stubbornly refusing to allow
| any legit versions of an extension cable (or I guess male-A to
| female-C) with active circuitry to make them reliable/safe. A
| hub allows male-A to male-A connections and length extension
| after all...
| jzwinck wrote:
| If IEEE does that, some manufacturers will just ignore it. Sure
| they may have to write weasel words like "USB compatible but
| not USB authorized" on their products, but they'll dominate the
| market by allowing people to use the largest variety of
| devices.
| zdragnar wrote:
| The simplest practical way around that (imho) is to license
| the name for the 4.x standard and be proactive in taking
| action against anyone who uses the name or logo if their
| cables or ports aren't compliant.
|
| Of course, depending on the jurisdiction _who_ can actually
| bring claims, and how damages can be calculated, vary
| greatly.
|
| I, for one, wouldn't buy "USB compatible but not USB
| authorized" cables if it meant being secure in knowing that
| the damned cable will actually work. Less tech savvy people
| might not, but there's also people who can be tricked into
| buying blinker fluid for their cars. There's gotta be a happy
| medium somewhere.
| solardev wrote:
| Is that basically what Thunderbolt does?
| marcosdumay wrote:
| > is to license the name for the 4.x standard
|
| Oh, while you are at it, they will also need names that
| don't keep changing and don't have synonyms.
|
| The 4.x names are ok for reference in a standard, but they
| are useless for consumer communication.
| gumby wrote:
| The negative impact of Postel's suggestion ("be conservative in
| what you send, be liberal in what you accept", AKA the
| "Robustness principle") is worse on end users than developers
| in cases like USB.
|
| USB is silently super backwards compatible. This is really good
| in a pinch: I need to get the data off this drive and even if
| it takes six hours because I only have a USB 2 cable I'm better
| off than if I couldn't do it at all.
|
| The problem is the opposite case for users: They have a fast
| USB-4 drive (potentially 40 Gbps) and a USB 4 laptop but pick
| up a USB-2 only cable. Hard for most people to diagnise,
| especially as such a cable won't have an e-marker so the laptop
| can't even query it. Or a slow charging cable for their device,
| so they fancy laptop doesn't visibly charge, or only charges
| overnight.
|
| Even for me these cases are hard to diagnose. And harder to
| understand than, say, using an inferior cable with your monitor
| so you et no signal at all.
| eternityforest wrote:
| Most people who want 40Gbps can diagnose the issues, the rest
| will assume "Oh, it takes 6 hours, better plan for it, that's
| how it is". But it will work every. single. time. at those
| very slow 2.0 speeds.
|
| What they should do is give a popup saying "You have a slow
| cable, it could be faster, go rummage till this message goes
| away and mark the good one with a sharpie", and that wouldn't
| need any changes to the spec
|
| They've chosen reliability over quality. Cheap reliability is
| a great thing. We did fine before fast hard drives, we can
| often live without them. But we'll want to scream if we can't
| get it working at all.
| solardev wrote:
| > go rummage till this message goes away and mark the good
| one with a sharpie
|
| I dunno if I even have the right kind, or how to even shop
| for the right kind, because there are like 500 wrong kinds
| that are all wrong for a different reason =/
|
| Basically I just look for Thunderbolt 3/4 support and if it
| doesn't have that, I give up, because who the heck has time
| to understand 4000 kinds of incompatible USB?
| eternityforest wrote:
| Maybe they could have a link in the popup?
|
| I wonder if you could start an online cables boutique
| that just sold tested and reviewed cables and cable
| management with a 50% markup for people who just don't
| want to deal with it.
| 2muchcoffeeman wrote:
| Then you have to wonder about how to make sure the link
| doesn't 404.
|
| And your store idea invalidates having cheaper devices
| and cables and a ubiquitous protocol.
|
| The only good thing about USBC is that if you moderately
| keep up to date with all your devices, eventually,
| recharging becomes relatively easy for all your devices.
| I don't think I've ever had a problem charging a USBC
| device. Generally everything charges everything. The high
| power devices like laptops are probably the exception
| since you need a brick that can deliver 65-80W[1].
|
| 1. One reason why Apple Silicon is so appealing. The M2
| Air is plenty fast and basically consumes the power of an
| iPad. You don't need a 30W supply to charge it.
| TeMPOraL wrote:
| > _What they should do is give a popup saying "You have a
| slow cable, it could be faster, go rummage till this
| message goes away and mark the good one with a sharpie",
| and that wouldn't need any changes to the spec_
|
| I think something like this exists in Windows. I'm not sure
| under what conditions it triggers, but I remember
| occasionally seeing a tooltip saying something like "this
| device could work faster", directing me to plug it into a
| different USB port.
| xg15 wrote:
| I remember this as well, but wasn't this way back if a
| usb 2.0 device was on a 1.1 (!) port?
|
| Not sure if they continued this for USB 3.0 and USB-C.
| brookst wrote:
| +1
|
| I recently threw out all of my USB-C cables that are longer
| than a foot but can't charge at 100w, because I decided USB-C
| is a primarily a charging protocol, and I'll use wireless,
| short cables, or Ethernet for data transfer and HDMI for
| display.
|
| The matrix of charging + data speeds in USB-C is a nightmare
| and always resulted in having the wrong cable for the desired
| devices and usage.
| scarface74 wrote:
| Until I need a USB C cable that can carry power and video for
| my portable external display and almost no random cable that
| you pick up supports it at all.
|
| So I have to be extra sure that I have the right cable before
| I leave for a trip.
| slaymaker1907 wrote:
| I'm sure the Quest 2 link cable is theoretically capable of
| that. However, given that the standard is such a mess,
| there's a decent probability that the display, computer, or
| cable won't trust another component and thus fall back to
| something much slower. It's also kind of hilarious that the
| Quest 2 needs a separate USB-C cable from the included
| USB-C charger to actually talk with a computer.
| triska wrote:
| Regarding the negative impact of Postel's suggestion, see
| also _The Harmful Consequences of Postel 's Maxim_:
|
| https://datatracker.ietf.org/doc/html/draft-thomson-
| postel-w...
|
| In Section 4, they define an alternative design principle
| based on the lessons of protocol deployment:
|
| _Protocol designs and implementations should be maximally
| strict._
| nullish_signal wrote:
| I tried using a power-only USB cable for data once, then
| tossed it back in the bin. Later, tried it again, back in the
| bin. I think on the 5th try, months later, I finally threw it
| away.
|
| Also, Dial-Up/Ethernet pairs seem to be too complex? Each
| house I bring my Ethernet/wire Tester+Detector to, somewhere
| I find a rat's nest of copper cable. Just today I found an
| Ethernet cable that connects 4/8 to nearby the router, and
| the other 2 Pairs of wire terminate in a Dial Up cable in
| another room!?
| scarface74 wrote:
| I threw all of my random USB C cables and Lightning cables
| away and standardized on these:
|
| https://www.amazon.com/dp/B093YVRHMB?
|
| - USB A and USB C on one end
|
| - USB C, micro USB and Lightning on the other end
|
| - 100W PD
|
| - 20GB data
|
| - video over USB C for my portable second display
| (https://www.amazon.com/dp/B095GG31KX)
|
| I never have to worry about the "right" cord. I did keep my
| MagSafe cord for my Mac.
| lostlogin wrote:
| Does that standardisation have 6 cables (including the
| MagSafe)?
| baq wrote:
| Fast Ethernet (the 100Mbps kind) doesn't use those pairs
| and works just fine with low frequency noise of a landline
| phone.
|
| Any cable ran in the previous decade should be dedicated to
| Ethernet, though...
| asveikau wrote:
| > "precise in what you send, flexible in what you receive" is
| basically evil
|
| I think it's a matter of where you apply it.
|
| For a purely software construct such as a network protocol,
| probably ok.
|
| For something like negotiating current and voltage where making
| the wrong decision can melt a cable or fry a circuit board, as
| stated in TFA, obviously it doesn't fly.
| Taywee wrote:
| Even a purely software construct, it's evil. If you flexibly
| receive, then popularly wrong data becomes the new standard.
| HPsquared wrote:
| Evil, but strong. The two often go together.
| asveikau wrote:
| I feel like you are making an assumption that your flexible
| implementation, or a single "wrong" implementation, becomes
| dominant.
|
| If there are _many_ implementations with differing amounts
| of perceived flexibility or wrongness in different areas,
| the average is going to be OK.
|
| If misunderstanding a spec is so common that many
| implementers are universally confused and "wrong" in the
| same way, maybe it's the spec that was wrong, and the
| misinterpretation should have been the spec.
|
| Ps. Phrasing this as "evil" contrasting with good reads to
| me as paranoid and melodramatic.
| layer8 wrote:
| The problem is that those implementations that do end up
| being dominant then factually constitute an ill-defined
| unwritten standard. There are just too many examples of
| this. So it's good advice to not start being lenient,
| because you never know what ends up being dominant. As a
| sibling comment points out, being lenient in fact
| increases your chances of becoming dominant.
| speed_spread wrote:
| Unfortunately the flexible implementation has a better
| chance at becoming dominant as it will work in more cases
| hence user are more likely to adopt it. Users don't care
| about standards compliance in the short term (if ever),
| they just want a solution to their use case, which is
| understandable. Same reason Microsoft worked so hard to
| keep deviant apps running on newer versions of Windows,
| there's a quality to "it just keeps working no matter
| what I do".
| Taywee wrote:
| I agree with "evil" being melodromatic of a term. I was
| just using it for consistency.
|
| Bad behavior being eventually codified into a standard
| because implementations were lax and it's impossible to
| go back on what everybody is now already doing is the
| story behind much of the modern web as it exists. HTML
| requiring closing tags for some elements, requiring a
| lack of closing tags for others, and leaving the closing
| tag still optional for yet others is a side effect of
| common browser behaviors.
|
| You're right that many implementations with different
| wrongness smooths out the average in many cases, except
| when all those implementations are expected to behave the
| same way on the same bad data. Then they all converge to
| the same bad behavior, which then becomes the new
| standard. Perhaps that's organic and natural, but I still
| wish we had a world where HTML was strictly compatible
| with XML, or with any other structured and predictable
| format.
| duped wrote:
| The problem isn't the philosophy but applying the philosophy
| out of context. I don't want my device to be "flexible" in what
| it accepts when we're talking about dozens of watts fed by
| amperes of current into my hand or pocket.
|
| Like there's a reason that Speak-On (1) was invented. It turns
| out sharing a common connector for carrying power and signal
| can be fucking dangerous.
|
| (1) https://en.wikipedia.org/wiki/Speakon_connector
| jacquesm wrote:
| I've seen more than one XLR based cable go up in a puff of
| magic smoke, this is a nice industry response to a very real
| problem. Better yet if the problem isn't detected during a
| soundcheck. Whoever thought that using the same connector for
| microphones as for speakers was a good idea wasn't thinking
| clearly that day. Unfortunately there is still a lot of old
| gear floating around so you get this:
|
| https://www.bax-shop.nl/speakerkabels/klotz-
| scasf030-neutrik...
| duped wrote:
| I've seen older stage monitors use powered TRS cables. I
| still see it in amp heads and cabs. Powered signals on
| exposed conductors. Better not touch it.
| jacquesm wrote:
| Oh, that must work great when you insert them and short
| the connectors against ground...
| duped wrote:
| It's actually not that bad since the only thing on the
| other end is a transducer that's already rated for a lot
| of current. It's just rather dangerous to have powered
| audio signals on exposed copper so you generally don't
| hot swap the things. It's not so bad for amps since the
| cables are very short and inaccessible, but for stage
| monitors you can accidentally unplug it on stage.
| jacquesm wrote:
| I've been zapped by microphone stands more than once and
| seen one musician take a serious hit just after plugging
| in his guitar. Stage gear is handled quite rough by non-
| experts and there is a ton of wear on plugs and cables.
| Especially outdoor concerts can be pretty nasty from an
| electrical point of view.
| wkat4242 wrote:
| Haha I use speakon for charging + data of computer backpacks.
| Because they can have more pins than xlr, can carry decent
| load, have good strain relief and locking mechanism.
| eternityforest wrote:
| That's a little less horrible than XLR, because the common
| use of speakon is already a designated "You're on your own"
| protocol, since you have to verify yourself that the amp
| isn't too much for the speaker and you're not feeding it
| 80hz square waves.
| joebergeron wrote:
| Do standards organizations tend to release/sell/provide any sort
| of "test harness" that vendors can use to validate their
| implementations against? Some coworkers and I recently authored
| an open API standard to be implemented by external vendors, and
| the test harness we released alongside it was the single most
| valuable tool we could possibly have provided to speed partner
| integrations/catch problems early.
| kennywinker wrote:
| I had a recent brush with the chaos of USB-C. A device I built
| was using what I assumed was a "standard" 6-pin usb port
| (https://www.lcsc.com/product-detail/USB-Connectors_Korean-Hr...)
| - but it wouldn't work with the cheap usb-c cables I had bought
| for it. After a bunch of time with a continuity tester and
| ultimately ripping one of the cables apart, it turns out the it
| was a 6pin cable, but it had a different 6 pins connected. One of
| them was "left handed" while the other was "right handed".
| Maddening.
|
| To help with this in the future, I recently ordered a run of
| these usb cable testers from jlcpcb with some friends -
| https://github.com/alvarop/usb_c_cable_tester - totally worth it
| to never have to guess about a cable again.
|
| (for the curious, a technical description of what was happening:
| the cable had pins A1, A4, A5, B1, B4, and B5 all connected. The
| port had pins A12, A9, A8, B12, B9, and B8 connected - If you
| look at this in a pin diagram like the one on this page
| https://arstechnica.com/gadgets/2015/01/usb-3-1-and-type-c-t...
| you can see that no matter which way around you put it in the
| port you'll never get contact)
| gumby wrote:
| This is part of a hackaday series on USB C. Other posts:
| https://hackaday.com/series_of_posts/all-about-usb-c/
___________________________________________________________________
(page generated 2022-12-27 23:00 UTC)