Cutting the Ends Off USB Cords.
Patrick, Mon Sept 30 2019, 03:13AMis it possible or wise to cut off the mini end of a USB cable then solder it to the USB board end directly. Making for a permanent attachment ?
I would keep +,-,D+,D- the way the are then keep the strands of shield soldered to the metal shroud of the original board side connector.
I fear the connector as used normally will wiggle to the point of damage of the course of a year then become intermittent. The only sketchy part is that its all high speed data until the FTDI chip, then it slows down to 115kbps. But at 480Mbps is the soldered connection sensitive to modification ?
Re: Cutting the Ends Off USB Cords.
klugesmith, Tue Oct 01 2019, 06:28PM
Sounds like a good idea, if your cable strain relief is good near the soldered connection.
It's not hard to achieve signal integrity good enough for 480 Mbps (fundamental frequency 240 MHz) at the scale of USB connectors.
klugesmith, Tue Oct 01 2019, 06:28PM
Sounds like a good idea, if your cable strain relief is good near the soldered connection.
It's not hard to achieve signal integrity good enough for 480 Mbps (fundamental frequency 240 MHz) at the scale of USB connectors.
Re: Cutting the Ends Off USB Cords.
Patrick, Tue Oct 01 2019, 06:40PM
ok ill post a pic when i get around to that mod.
Patrick, Tue Oct 01 2019, 06:40PM
ok ill post a pic when i get around to that mod.
Re: Cutting the Ends Off USB Cords.
2Spoons, Thu Oct 03 2019, 02:20AM
Just keep the wire sticking out of the cable as short as you can - you want to maintain the correct impedance as much as possible.
@Klugesmith: where do you get 240MHz fundamental? USB bit rate is the same as the modulation rate. Which is why we just bought a 6GHz scope for doing compliance testing -so we can see the 3rd and 5th harmonics when checking eye patterns.
2Spoons, Thu Oct 03 2019, 02:20AM
Just keep the wire sticking out of the cable as short as you can - you want to maintain the correct impedance as much as possible.
@Klugesmith: where do you get 240MHz fundamental? USB bit rate is the same as the modulation rate. Which is why we just bought a 6GHz scope for doing compliance testing -so we can see the 3rd and 5th harmonics when checking eye patterns.
Re: Cutting the Ends Off USB Cords.
klugesmith, Thu Oct 03 2019, 03:50AM
Good question, that sent me to the Internet for learning about USB for the first time. Found out about High Speed Mode, whose physical layer seems to use the familiar balanced-pair line with termination R on both ends of both wires, switched-current driver, roughly half-volt swings, etc. Here's one app note about compliance testing that you might already be familiar with:
We aren't talking about Super Speed, are we?
Correct me if this is wrong, but HS mode seems to use plain binary data stream at 480 Mbps. No 8b10b coding, multi-level or multi-phase modulation, etc.
So if you transmit JKJKJKJK at 480 Mbps, that's a 240 MHz square wave. A spectrum analyzer would display spikes at 240 MHz and its odd multiples.
That Tektronix paper shows why, as you said, you need a multi-GHz scope to detect aberrations like non-monotonic transitions (figure 7). Given the cheapness of knock-off USB cables, it's a wonder that they work at 12 Mbps! 480 Mbps mode might be harder to break, because it's purely differential, with pretty good impedance control at both ends no matter who's driving and who's receiving.
klugesmith, Thu Oct 03 2019, 03:50AM
Good question, that sent me to the Internet for learning about USB for the first time. Found out about High Speed Mode, whose physical layer seems to use the familiar balanced-pair line with termination R on both ends of both wires, switched-current driver, roughly half-volt swings, etc. Here's one app note about compliance testing that you might already be familiar with:
We aren't talking about Super Speed, are we?
Correct me if this is wrong, but HS mode seems to use plain binary data stream at 480 Mbps. No 8b10b coding, multi-level or multi-phase modulation, etc.
So if you transmit JKJKJKJK at 480 Mbps, that's a 240 MHz square wave. A spectrum analyzer would display spikes at 240 MHz and its odd multiples.
That Tektronix paper shows why, as you said, you need a multi-GHz scope to detect aberrations like non-monotonic transitions (figure 7). Given the cheapness of knock-off USB cables, it's a wonder that they work at 12 Mbps! 480 Mbps mode might be harder to break, because it's purely differential, with pretty good impedance control at both ends no matter who's driving and who's receiving.
Re: Cutting the Ends Off USB Cords.
2Spoons, Thu Oct 03 2019, 04:54AM
Ah yes I forgot the JKJK is effectively half frequency. 12Mbps still uses differential signalling , but the levels are higher. 480Mbps is only +/-300mV. Its actually quite easy to screw up 480Mbps if you mess up your board layout ( which we have), You have to run controlled impedance diff pairs over a good ground for anything over a few cm long. 12Mbps is easier, being a longer wavelength.
2Spoons, Thu Oct 03 2019, 04:54AM
Ah yes I forgot the JKJK is effectively half frequency. 12Mbps still uses differential signalling , but the levels are higher. 480Mbps is only +/-300mV. Its actually quite easy to screw up 480Mbps if you mess up your board layout ( which we have), You have to run controlled impedance diff pairs over a good ground for anything over a few cm long. 12Mbps is easier, being a longer wavelength.
Re: Cutting the Ends Off USB Cords.
Plasma, Thu Oct 03 2019, 08:44AM
I did a network cabling course in the past, and we had to keep the unwinding of the cables to were it was connected to the plug less than 20mm that was for cat5 ,you have a link budget each plug join etc had a value if it exceed the budget it failed the fluke test, became of errors.
Not directly related but...
Plasma, Thu Oct 03 2019, 08:44AM
I did a network cabling course in the past, and we had to keep the unwinding of the cables to were it was connected to the plug less than 20mm that was for cat5 ,you have a link budget each plug join etc had a value if it exceed the budget it failed the fluke test, became of errors.
Not directly related but...
Re: Cutting the Ends Off USB Cords.
Patrick, Fri Oct 04 2019, 05:31PM
yes its just 2.0 if it were 3.0 id be scared.
Patrick, Fri Oct 04 2019, 05:31PM
yes its just 2.0 if it were 3.0 id be scared.
Re: Cutting the Ends Off USB Cords.
Conundrum, Thu Oct 10 2019, 07:45PM
Have a nice armoured cable from an old iPad lead, same problem (ie ripped off Lightning end)
Conundrum, Thu Oct 10 2019, 07:45PM
Have a nice armoured cable from an old iPad lead, same problem (ie ripped off Lightning end)
Print this page