Regarding macOS lying about the connection speed, I'd be curious to see detailed output for such a device connected using such a cable. XHCI controllers build entirely separate device hierarchies for devices connected via the high speed versus the superspeed data pairs - each USB port effectively becomes two virtual ports, one superspeed, one high speed. So the device should turn up in the device tree as connected to a high speed port, even if it's tricked the driver into reporting
On macOS, every USB device turns up in the I/O registry as an IOUSBHostDevice service; if system profiler is claiming reporting a connection speed of 10Gbps, I assume its "Device Speed" property is coming back as 0x4. (Though I'd be curious to verify that.) What should however give it away as maxing out as high speed is that the (eventual) parent service would be of type AppleUSB20XHCITypeCPort/AppleUSB20XHCIPort because that's what devices show up under when they're connected via the legacy data pair. (note the "USB20")
I've long been toying with the idea of building a macOS utility that surfaces all this information about connected devices in a much more accessible and human-readable way. Sounds like maybe there might be some interest in such a tool out there?
This product should not have to exist. And yet my subconscious frustration with USB-C has apparently grown to the point where I didn't even question immediately spending £35 on the thing.
If anyone else feels similarly, heads up that the official store won't ship orders until the end of the month (related to Chinese New Year?) so my cable drawer got itself a brief reprieve...
Cable diagnostics ought to be built in to the PHY of any USB-C port that uses the fast differential pairs. (Slow USB-2 ports can be exempt.) Surely they already have most of the necessary logic?
When I need a cable, I usually don't want to wait for shipping or waste time trying to verify it's a good cable, so I'll just wander to the closest computer/electronics store and buy an Apple branded one.
There are other good brands, but you sometimes have to be careful, you can pretty much never find them in physical stores, etc.
That said, very neat that someone built a good USB cable tester. Especially one that isn't multiple hundreds of dollars like lots of decent testing equipment.
My family eats USB cables and chargers, so I've had to buy (no exaggeration) hundreds over the years. The problem arises when you need a specific cable, e.g. a Thunderbolt 5 certified cable at least 2m long, and you have a pile of 20 almost-identically-looking cables (from 20w to 240w, from no data to 80Gbps) and exactly one of them is what you're looking for ...
you have a pile of 20 almost-identically-looking cables
Surely there must be some add-on to the popular labelmakers for cable labelling? The addressable market -- nerds with hundreds of cables and OCD -- is huge.
My guess is it ships with a battery already installed, but there's a plastic film between the battery's connector and the device's contact. Common way to prevent unnecessary discharge until the customer is ready to use the device. The odd phrasing is probably down do machine or amateur translation.
they put a better translated sticker on the packaging that said "For first-time use peel off the battery's plastic film", the battery was installed on arrival but shrunk wrapped to prevent it powering the device in transit.
pmdj | 16 hours ago
Regarding macOS lying about the connection speed, I'd be curious to see detailed output for such a device connected using such a cable. XHCI controllers build entirely separate device hierarchies for devices connected via the high speed versus the superspeed data pairs - each USB port effectively becomes two virtual ports, one superspeed, one high speed. So the device should turn up in the device tree as connected to a high speed port, even if it's tricked the driver into reporting
On macOS, every USB device turns up in the I/O registry as an
IOUSBHostDeviceservice; if system profiler is claiming reporting a connection speed of 10Gbps, I assume its "Device Speed" property is coming back as 0x4. (Though I'd be curious to verify that.) What should however give it away as maxing out as high speed is that the (eventual) parent service would be of typeAppleUSB20XHCITypeCPort/AppleUSB20XHCIPortbecause that's what devices show up under when they're connected via the legacy data pair. (note the "USB20")I've long been toying with the idea of building a macOS utility that surfaces all this information about connected devices in a much more accessible and human-readable way. Sounds like maybe there might be some interest in such a tool out there?
callahad | 14 hours ago
This product should not have to exist. And yet my subconscious frustration with USB-C has apparently grown to the point where I didn't even question immediately spending £35 on the thing.
If anyone else feels similarly, heads up that the official store won't ship orders until the end of the month (related to Chinese New Year?) so my cable drawer got itself a brief reprieve...
fanf | 13 hours ago
Cable diagnostics ought to be built in to the PHY of any USB-C port that uses the fast differential pairs. (Slow USB-2 ports can be exempt.) Surely they already have most of the necessary logic?
janiczek | 13 hours ago
Would be great to see examples of cables (with store links) that actually are decent according to the tester.
zie | 15 hours ago
When I need a cable, I usually don't want to wait for shipping or waste time trying to verify it's a good cable, so I'll just wander to the closest computer/electronics store and buy an Apple branded one.
There are other good brands, but you sometimes have to be careful, you can pretty much never find them in physical stores, etc.
That said, very neat that someone built a good USB cable tester. Especially one that isn't multiple hundreds of dollars like lots of decent testing equipment.
cpurdy | 11 hours ago
My family eats USB cables and chargers, so I've had to buy (no exaggeration) hundreds over the years. The problem arises when you need a specific cable, e.g. a Thunderbolt 5 certified cable at least 2m long, and you have a pile of 20 almost-identically-looking cables (from 20w to 240w, from no data to 80Gbps) and exactly one of them is what you're looking for ...
gerikson | 11 hours ago
The basis of a healthy diet
Surely there must be some add-on to the popular labelmakers for cable labelling? The addressable market -- nerds with hundreds of cables and OCD -- is huge.
rpc | 13 hours ago
I also got one of this gizmos from japan when I had to check cable quality for video transmission.
Related, just watched this from Linus Tech Tips the other day, regarding producing their own USB cables (video)
NB: no endorsement implied, just interesting how hard is to build them.
fcbsd | 9 hours ago
the Treedix TRX5-0816 has been part of my debugging kit for the last year, like the author I would recommend it.
hoistbypetard | 16 hours ago
The tester that the blogger landed on looks nice, but on the vendor page, I found this graphic puzzling:
https://treedix.com/cdn/shop/files/TwoPowerSupplyModule_2_3750x.jpg
Why would you need to peel off the film on the surface of the battery, and isn't that hazardous?
pmdj | 15 hours ago
My guess is it ships with a battery already installed, but there's a plastic film between the battery's connector and the device's contact. Common way to prevent unnecessary discharge until the customer is ready to use the device. The odd phrasing is probably down do machine or amateur translation.
Garbi | 15 hours ago
Or the supplied battery has shrink wrap and someone once put it in without removing the wrap and the complained.
hoistbypetard | 15 hours ago
Ah. That makes sense. I was imagining peeling the outer cover off a commercial AAA battery, and was quite puzzled.
fcbsd | 9 hours ago
they put a better translated sticker on the packaging that said "For first-time use peel off the battery's plastic film", the battery was installed on arrival but shrunk wrapped to prevent it powering the device in transit.
james | 12 hours ago
I've needed wanted anything more but needed anything less