Comment by traceroute66
3 days ago
> I'm using the hardware tester from $INSERT_NAME_HERE
Oh god not again... before everyone starts naming their favourite brand, here's a 265 comment HN thread[1] (from last week !) on why these cheap USB-cable "testers" are not worth anything.
TL;DR You need a six-figure wallet to seriously call anything a "hardware tester" for USB cables. User ChrisMarshallNY has a one-liner summary buried in that thread:
Think about it. You need to generate a precise, calibrated, 40GHz signal, transmit it, then precisely measure the returned signal for phase, amplitude, and waveform integrity -at that speed. Oscilloscopes and logic analyzers that run at that speed, are damn expensive.
If you only paid $40 for a black-box on your favourite online marketplace, then what you've got is definitively NOT a USB-cable tester.
OK sure your 6 figures thing or even that 20k thing are way better testers than the Treedix. But for my use case, i.e. verify what's actually in the cable (not blindly trusting the marker) the latter is sufficient and very ergonomic, and the "real" testers are overkill.
> verify what's actually in the cable (not blindly trusting the marker) the latter is sufficient
But as was repeated 101 times on that thread ....
All that those cheap $40 boxes are doing IS reading the marker.
Sadly you are kidding yourself if you think a $40 box off Amazon is doing any sort of "verification". Let alone in any sort of calibrated, reproducible sense free of random unknown variables.
Then I am kidding myself, because I'm convinced it does a lot more.
1 reply →
Are you suggesting there is no middle ground and that simply knowing if a USB-C cable is a USB C cable or a fake one can't be done with a cheap tester?
> Are you suggesting there is no middle ground
Sadly yes.
Like many people on that thread concluded:
The best "middle ground" thing you can do is pay a premium to buy your cables from reputable manufacturers.
Buying my cables from decent manufacturers does not tell me what specific version of USB a cable purports to support, what voltage of USB-PD it can transmit, whether or not it can do Thunderbolt or DisplayPort, or a dozen other random capabilities inherent in these cables.
4 replies →
Am I missing something? For practical use you only care if it carries a certain amount of packets per second without excessive retries, not whether it preserves the waveform exactly.
> For practical use you only care if it carries a certain amount of packets per second without excessive retries
The $40 black-boxes do two things:
There is zero chance that a $40 black-box, even if it did a naïve packet pushing test, would deliver you reproducible results and introduces all sorts of other random variables.
"Excessive retries" sounds like exactly the sort of intermittent, hard-to-diagnose cable issues that you will not find unless you test the USB cable on a 6-figure rig. :)
Man you're missing the mark.
If I want to tell a cable that can do only USB 2 from one that can do 3 from one that is actually thunderbolt, the treedix is a godsend. So. Many. Devices. Come with some unknown usb cable. Can it only charge? Transfer data? How fast? Stuff like signal integrity, dispersion, differential timing, phase shift etc are totally out of scope for this simple thing - telling apart these cables.
> If I want to tell a cable that can do only USB 2 from one that can do 3 from one that is actually thunderbolt
And if the marker chips in your "some unknown usb cable" are fake / counterfeit ? Then what ?
Your $40 toy won't help with that scenario and you will be back at square one. That's what.
You're arguing that the ability to read a label is absolutely worthless, because they might occasionally lie to you. It's not so absolute.
Maybe you need absolutely 100% confidence your labels don't lie, but I and presumably others on this thread merely want to check for the presence of the label and the declarations on it. That's enough for me to sort my cables.
Maybe occasionally I'll come across a counterfeit cable (or more likely, a damaged cable, because I generally buy reputable ones) and have one underperform its label, but I can get rid of it then.
Huh? Decent quality cable testers can be had for like $20,000. There's a difference between a general-purpose oscilloscope or logic analyzer and a purpose-built domain-specific appliance. I probably could've saved up for one in just a few months back when I had a job earlier this year.
In the world of specialized test equipment, $20k isn't too bad at all. It's a certainly still a lot of money in absolutes, but the barrier to entry is simple very high for this class of kit. The sales volume is low and the engineering to do precise work with RF signals at these kinds of frequencies is complex. The people doing the work of designing, building, and getting this stuff into the hands of end-users still need enough money to live healthy lives.
And $20,000 is five figures, so an off-the-cuff six figure estimate is an error of an order of magnitude. That's quite a large error.
But to bring it all back 'round again: The oft-repeated suggestion that a $40 device can do the same job represents an error of 3 or more orders of magnitude. :)
Some of the stuff in this cheaper class of kit can observe what happens in terms of voltage and/or current on the wire between two devices, some of them can have a chat with a cable and/or a device on their own to see what they state of their capabilities (as this software does), and some of them can also test things like DC continuity. That's useful, but it doesn't help to grade the actual signal characteristics of the cable.
There's a galaxy of difference between DC and a 40 GHz symbol rate.
> Decent quality cable testers can be had for like $20,000.
See the discusison linked to.
Trust me, my interest was piqued by that thread and I went down the rabbit warren with Mr Google and spent far too much time on the subject. I looked at the various manufacturers and the setups required.
TL;DR: To start with, you need fancy harness assemblies. Then you need to plug those harnesses into very, very expensive equipment.
You're not going to find anything for $20k. High-throughput USB-cable testing simply can't be done the same way copper CAT6 certification can with a 20k Fluke.
For those who want to play with Mr Google, some manufacturers from my notes: Wilder Technologies, Allion, Keysight, Tektronix, Teledyne LeCroy
Here's one blog post I uncovered during my researches: https://www.keysight.com/blogs/en/tech/bench/2023/02/17/powe...
Aren't you both agreeing.. that it is way beyond $40? :)
You need a six figure wallet to want to spend $20k on a tester :)
1 reply →