For a "senior editor" of PC World admitting to using multiple cables, not just the TB5, that rate less than he needs, to complain things didn't seem to work right, feels like the real headline is "first, use the right cables".
The biggest problem with all this is that it's very nearly impossible to tell what kind of cable you have most of the time. If you're lucky (or purchased wisely), the cable will have it's type and class printed on it, but most don't -- certainly not OEM cables that ship with devices. And by now, most of us probably have a dozen or more USB-C cables of various sorts. Some of these may be TB cables. Some may even be TB4 cables. Some may even be data-only cables. But it's hard to keep track, and it's stupidly expensive to just replace everything every time a new standard is available [on one specific device].
The most annoying part is that there is actually a perfectly fine standard for labeling USB-C cables.[0] If a cable can transfer 80Gbps of data and 240W of power, it's supposed to simply have "80Gbps / 240W" written right on the connector. Some rare cables actually do this[1], but the vast majority of manufacturers just can't be bothered with it.
Is there a robust library for good quality cables somewhere? I remember a few blog posts that did larger tear downs, but I’d love to just be able to go to a site that has validated cables and know with some confidence “this is a real TB5 cable”
Not an app, but some USB testers (e.g. POWER-Z KM003) can do this by reading an e-Marker chip on the cable. I'm not sure if there's a way to access eMarker purely through software without some sort of driver support, though.
The problem is that a cable saying that it's something does not mean that it actually conforms to the standard and can deliver that speed/power. And proper testing requires complex hardware.
That's true of every cable that has ever existed, including basic electrical wiring, and so feels like a separate problem than merely identifying what kind of cable a cable even intends/claims to be in a world where they all have the same shape connector: don't let the perfect be the enemy of the good.
The ones I hate most are those cursed only-charging/no-data usb-c cables. Otherwise I just assume everything is USB2 unless it looks expensive. And all my thunderbolt cables right the thunderbolt version, not sure if that is universal though.
He says that he used a Kensington Thunderbolt 5 cable, so I think that just reinforces the story that the vendors don't have things implemented properly yet.
That teardown (and the similar Lumafield scan) is quite misleading, though. Nothing about that is Apple-specific, or makes the cable in any way "better".
A signal will degrade as it travels through a cable. The longer the cable, the more it degrades. At USB4/TB4/TB5 speeds you can make a cable up to about 80cm before it starts to become a problem. Any cable longer than ~80cm needs active electronics to boost the signal's quality.
Apple's cable is 180cm, so it has those electronics. Any 180cm long USB4 cable made by any manufacturer is going to have them.
The issues encountered here is exactly why we need good reviews for new technologies that try a bunch of stuff and find what works and what doesn't. It is a pity that the author did not get into the trouble of trying another cable at least.
Re: shuttering etc display issues "despite having a 4090 nvidia gpu", I am a bit confused whether the displays connected through a thunderbolt dock are supposed to be controlled by the discrete gpu or only by the igpu. I would assume that it can only go through the on board one but I may be wrong. This could also explain why the power delivered capped at 87W while gaming instead of the claimed 140W capacity of the dock, because if the igpu was exclusively used maybe the consumption may not have reached that high.
Also he says that once he encountered some ssd transfer issues when both transferring data and using the dock's ethernet connection, he "allowed windows to use the wifi" but if both ethernet and wifi are available windows (and most OSs I know) tend to prefer the ethernet by default, so I doubt this would change anything anyway. Still the bandwidth should be more than enough for TB5, which is puzzling. One thing that the author could have checked is whether the connection was actually TB5 or it fell back to a lesser version. The way it is all described it would not actually surprise me even if the connection fell back to USB3gen2 or so.
It would have been a better review imo if the author checked and reported on such very probable causes of the issues he encountered.
Yeah, it really seems like the author doesn't really understand the technology. Which is fine for an average user, but you'd expect better from a professional tech reviewer.
First, the power thing. He brings 240W charging as a "benefit of TB5". It isn't, it's a basic USB PD thing. Any USB-C device could use it - and of course it's only going to work if all the components actually support 240W charging. What does he mean that he "later discovered" his laptop was charging from one port, and at 87W, did he not read the specifications or the manual beforehand? It was never going to support more than 100W (the laptop's limit, dock can do 140W), so considering the laptop's own power brick delivers 330W it is guaranteed to be running in some power-saving mode.
The display issue could just as well be caused by the GPU only supporting DisplayPort 1.4, so the link to the GPU-USB4 controller is limited to an effective data rate of 26Gbps. A single uncompressed 4K 144Hz stream takes 31.35Gbps, even with a 3x DSC compression ratio that's not going to work! This is a well-known limitation of the RTX 4000 series, and people have been complaining about a lack of DP 2.0 for years. The part about a cable being "rated for 1440p165, not 4K" is nonsense: the only thing that matters is bandwidth, 4k60 requires 12.54Gbps, 1440p165 requires 16.30Gbps, and both match what you'd expect from a cable designed for DisplayPort 1.2's HBR2 (17.28Gbps).
Same story with the SSD performance. Giving a benchmark result is useless without actually explaining what you are benchmarking, and "folder copy time" is entirely meaningless. He gives a number of 252.3 MB/s for the benchmark - but the SSD is capable of 6000MB/s! Sure, the SSD's spec is for best-case sequential access, but if you want to test TB5 that's what you should be going for. 252.3 MB/s sounds more like random read/write, which is far less relevant here. And the stuff about "streaming a video via the internet": how about you tell us whether you were outputting video via the dock in either case? You know, the part which actually matters?
"Windows 11 reports the controller as USB4 v2": yes, because that's what it is. TB4 and TB5 are nothing more than USB4 with some optional parts made mandatory.
Look, I completely agree that the USB / Thunderbolt ecosystem has gotten quite complicated for the average user and has developed way too many gotcha's. But if you're a tech journalist you're supposed to know this! Don't tell us that your monitors mysteriously didn't work, do some basic research and tell us that they don't work because Nvidia screwed up. Don't tell us that it mysteriously didn't charge at maximum speed, tell us that it doesn't do so because the laptop manufacturer screwed up, and you got a dock which couldn't supply more than 140W anyways. Why bother writing an article if you're not going to provide any more value than some random Reddit rant?
> But if you're a tech journalist you're supposed to know this!
Agreed, but the sad fact is, every "journalist" makes very obvious mistakes in every article they write. It is just that we (the readers) don't notice the obvious mistakes in the stories for which we lack the insider or specialized knowledge to see the mistakes jump out of the page.
This is the effect that "Gell-Mann Amnesia" describes:
My Lenovo P1 Gen 5 needs to reboot once after booting from power off, otherwise the Thunderbolt dock won't connect and I only have display, and no network or USB. Super annoying. I bought a docker station to save time.
That sounds bad, have any more details on what happened? You plugged in something external and ‘fried’ your ssd ?
I don’t know how likely that is here, but there was a class of attacks called Thunderspy which could allow an external malicious device full access to your computer via thunderbolt.
I actually think it could also have been something that resetted the T2 chip during the process..
I bought a new MacBook and wanter to transfer data using the cable / migration assistant. I think it froze. After that it wouldn't boot, and there was no partition table. The SSD worked after formatting though.
The cable was a cable directly from Apple, but yeah product return fraud on Amazon could be scary with these kinds of attacks
Because https://appleinsider.com/articles/22/04/23/teardown-of-apple...
For a "senior editor" of PC World admitting to using multiple cables, not just the TB5, that rate less than he needs, to complain things didn't seem to work right, feels like the real headline is "first, use the right cables".
[0]: https://www.usb.org/sites/default/files/usb_type-c_cable_log...
[1]: https://www.amazon.com/Anker-Supports-Display-Transfer-Charg...
A signal will degrade as it travels through a cable. The longer the cable, the more it degrades. At USB4/TB4/TB5 speeds you can make a cable up to about 80cm before it starts to become a problem. Any cable longer than ~80cm needs active electronics to boost the signal's quality.
Apple's cable is 180cm, so it has those electronics. Any 180cm long USB4 cable made by any manufacturer is going to have them.
Re: shuttering etc display issues "despite having a 4090 nvidia gpu", I am a bit confused whether the displays connected through a thunderbolt dock are supposed to be controlled by the discrete gpu or only by the igpu. I would assume that it can only go through the on board one but I may be wrong. This could also explain why the power delivered capped at 87W while gaming instead of the claimed 140W capacity of the dock, because if the igpu was exclusively used maybe the consumption may not have reached that high.
Also he says that once he encountered some ssd transfer issues when both transferring data and using the dock's ethernet connection, he "allowed windows to use the wifi" but if both ethernet and wifi are available windows (and most OSs I know) tend to prefer the ethernet by default, so I doubt this would change anything anyway. Still the bandwidth should be more than enough for TB5, which is puzzling. One thing that the author could have checked is whether the connection was actually TB5 or it fell back to a lesser version. The way it is all described it would not actually surprise me even if the connection fell back to USB3gen2 or so.
It would have been a better review imo if the author checked and reported on such very probable causes of the issues he encountered.
First, the power thing. He brings 240W charging as a "benefit of TB5". It isn't, it's a basic USB PD thing. Any USB-C device could use it - and of course it's only going to work if all the components actually support 240W charging. What does he mean that he "later discovered" his laptop was charging from one port, and at 87W, did he not read the specifications or the manual beforehand? It was never going to support more than 100W (the laptop's limit, dock can do 140W), so considering the laptop's own power brick delivers 330W it is guaranteed to be running in some power-saving mode.
The display issue could just as well be caused by the GPU only supporting DisplayPort 1.4, so the link to the GPU-USB4 controller is limited to an effective data rate of 26Gbps. A single uncompressed 4K 144Hz stream takes 31.35Gbps, even with a 3x DSC compression ratio that's not going to work! This is a well-known limitation of the RTX 4000 series, and people have been complaining about a lack of DP 2.0 for years. The part about a cable being "rated for 1440p165, not 4K" is nonsense: the only thing that matters is bandwidth, 4k60 requires 12.54Gbps, 1440p165 requires 16.30Gbps, and both match what you'd expect from a cable designed for DisplayPort 1.2's HBR2 (17.28Gbps).
Same story with the SSD performance. Giving a benchmark result is useless without actually explaining what you are benchmarking, and "folder copy time" is entirely meaningless. He gives a number of 252.3 MB/s for the benchmark - but the SSD is capable of 6000MB/s! Sure, the SSD's spec is for best-case sequential access, but if you want to test TB5 that's what you should be going for. 252.3 MB/s sounds more like random read/write, which is far less relevant here. And the stuff about "streaming a video via the internet": how about you tell us whether you were outputting video via the dock in either case? You know, the part which actually matters?
"Windows 11 reports the controller as USB4 v2": yes, because that's what it is. TB4 and TB5 are nothing more than USB4 with some optional parts made mandatory.
Look, I completely agree that the USB / Thunderbolt ecosystem has gotten quite complicated for the average user and has developed way too many gotcha's. But if you're a tech journalist you're supposed to know this! Don't tell us that your monitors mysteriously didn't work, do some basic research and tell us that they don't work because Nvidia screwed up. Don't tell us that it mysteriously didn't charge at maximum speed, tell us that it doesn't do so because the laptop manufacturer screwed up, and you got a dock which couldn't supply more than 140W anyways. Why bother writing an article if you're not going to provide any more value than some random Reddit rant?
Agreed, but the sad fact is, every "journalist" makes very obvious mistakes in every article they write. It is just that we (the readers) don't notice the obvious mistakes in the stories for which we lack the insider or specialized knowledge to see the mistakes jump out of the page.
This is the effect that "Gell-Mann Amnesia" describes:
https://www.epsilontheory.com/gell-mann-amnesia/
If the cable alone needs to be $100+ to maybe be reliable… that’s not a great sign for widespread consumer adoption.
I guess he proves his point, it's hard to build a cutting edge system without much more research if you assembly it based on your own insights.
And most importantly: Connector labels don't mean what you should be able to expect them to mean. That point is no news sadly.
I would love to read from someone who actually knows what's happening and who goes on til the reason is found.
Any other articles?
I don’t think plugging a cable directly into the pci bus is a great idea
I don’t know how likely that is here, but there was a class of attacks called Thunderspy which could allow an external malicious device full access to your computer via thunderbolt.
https://thunderspy.io/
I bought a new MacBook and wanter to transfer data using the cable / migration assistant. I think it froze. After that it wouldn't boot, and there was no partition table. The SSD worked after formatting though.
The cable was a cable directly from Apple, but yeah product return fraud on Amazon could be scary with these kinds of attacks