What Cable is This? Wanna make interconnects.

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Jump to Latest
qusp

qusp

Member
Joined 2009
udailey said:
Well, I think it would be fine for USB. Get out your DMM and measure resistance from one end of a usb cable to the other. I bet its not even ONE OHM.
When manufacturers talk about 75Ohm impedance they dont mean a 3 ft piece of copper has 75Ohms. For a 1 meter piece of copper to have 75 ohms of DC resistance, not AC, it needs to be .017mm in diameter. Kind of tiny. AC would be a bit larger but not a lot. The manufacturers are talking about the impedance of 1000ft of the cable. Thats the standard. Your USB sockets dont care about the resistance of the one meter cable as long as its low low low. If you want to build your own USB cable just do it. I did it and it works fine. Used silver for no good reason other than it was on the table.
Click to expand...

its impedance (at high frequency) and USB is balanced mode impedance at that. that impedance needs to take into account the twist rate, frequency, the dielectric thickness, the dielectric type and the wire thickness (ratio between these last 3). that and its not 75Ω impedance, its ~90Ω even the 75Ω spec for coax isnt resistance either, its impedance too, but single ended.

the above advice is misleading and straight up incorrect. if its just for a standard USB 1 cable it will probably work with most transceivers as its low speed and the spec is pretty rugged, if its USB 2, not a chance

probably best to have read the USB cable specifications at some point before offering advice here.
 
Last edited:
qusp

qusp

Member
Joined 2009
sorry for being uptight above, but I made a point of mentioning much of this, then you contradicted me with psuedo-technical 'information'

its related to the propagation delay and signal attenuation caused by the lumped differential mode impedance of the cable and the matching drivers. although there is impedance specified for the USB1 spec, its pretty lax, but for USB2 its very important and reasonably tightly specified

7.1.1.3 High-speed (480 Mb/s) Driver Characteristics
A high-speed USB connection is made through a shielded, twisted pair cable with a differential characteristic impedance (Z0) of 90 Ω ±15%, a common mode impedance (ZCM) of 30 Ω ±30%, and a maximum one-way delay of 26 ns (TFSCBL). The D+ and D- circuit board traces which run between a transceiver and its associated connector should also have a nominal differential impedance of 90 Ω, and together they may add an additional
4 ns of delay between the transceivers. (See Section 7.1.6 for details on impedance specifications of boards and transceivers.) The differential output impedance of a high-speed capable driver is required to be 90 Ω ±10%. When either the D+ or D- lines are driven high, VHSOH (the high-speed mode high-level output voltage driven on a data line with a precision 45 Ω load to GND) must be 400 mV ±10%. On a line which is not driven, either because the transceiver is not transmitting or because the opposite line is being driven high, VHSOL (the high- speed mode low-level output voltage driven on a data line with a 45 Ω load to GND) must be 0 V ± 10 mV.
Note: Unless indicated otherwise, all voltage measurements are to be made with respect to the local circuit ground.
Note: This specification requires that a high-speed capable transceiver operating in full-speed or low-speed mode must have a driver impedance (ZHSDRV) of 45 Ω ±10%. It is recommended that the driver impedances be matched to within 5 Ω within a transceiver. For upstream facing transceivers which do not support high-speed mode, the driver output impedance (ZDRV) must fall within the range of 28 Ω to 44 Ω.
On downstream facing ports, RPD resistors (15 kΩ ±5%) must be connected from D+ and D- to ground.
When a high-speed capable transceiver transitions to high-speed mode, the high-speed idle state is achieved by driving SE0 with the low-/full-speed drivers at each end of the link (so as to provide the required terminations), and by disconnecting the D+ pull-up resistor in the upstream facing transceiver.
In the preferred embodiment, a transceiver activates its high-speed current driver only when transmitting high- speed signals. This is a potential design challenge, however, since the signal amplitude and timing specifications must be met even on the first symbol within a packet. As a less efficient alternative, a transceiver may cause its high-speed current source to be continually active while in high-speed mode. When the transceiver is not transmitting, the current may be directed into the device ground rather than through the current steering switch which is used for data signaling. In the example circuit, steering the current to ground is accomplished by setting HS_Drive_Enable low.
In CMOS implementations, the driver impedance will typically be realized by the combination of the driver’s intrinsic output impedance and RS. To optimally control ZHSDRV and to minimize parasitics, it is preferred the driver impedance be minimized (under 5 Ω) and the balance of the 45 Ω should be contributed by the RS component.
When a transceiver operating in high-speed mode transmits, the transmit current is directed into either the D+ or D- data line. A J is asserted by directing the current to the D+ line, a K by directing it to the D- line.
When each of the data lines is terminated with a 45 Ω resistor to the device ground, the effective load resistance on each side is 22.5 Ω. Therefore, the line into which the drive current is being directed rises to 17.78 ma * 22.5 Ω or 400 mV (nominal). The other line remains at the device ground voltage. When the current is directed to the opposite line, these voltages are reversed.
Click to expand...
 
Last edited:
qusp

qusp

Member
Joined 2009
the detail was not just for you, udailey posted a response that used technical enough terms to make it seem there was something behind it, but it was based on incorrect assumptions rather than factual information, so I posted some of the relevant spec from usb.org (there is a lot more where that came from, its a very extensive document).

Its just a reminder of what specifically we are dealing with with USB2 and above, USB 3 is even more harsh. its an excerpt from the USB 2 high speed specification. its point 7.1.1.2 page 129 of the usb_20.pdf at usb.org

re cat6, yes, some of the higher quality well qualified network cable as I mentioned its OK, but only for reasonably short lengths, but OK only, it will work, but why? just get a belkin gold or something. the ethernet spec is 100Ω +/-10% differential (+ to -) under different conditions, but the usb spec is 90Ω +/-10%. so its only just within range if the ethernet cable is bang on its 100Ω or below.

why are you doing it? what makes you think it will improve the audio quality you hear? is there actually a specific problem you have identified and are trying to solve, or is it more of a case of you reading about somebody elses opinion/anecdotal results based on different gear? where does your USB DAC get its power from? are you sure you are even able to use external power and ground with your DAC?
 
Last edited:
qusp

qusp

Member
Joined 2009
alternatively you could of course buy something slightly OTT like the belkin gold, which is fairly reasonably priced and has good quality, but non wanky construction. Take that cable and modify it, cut the source/PC end off, split it into the 2 twisted pairs for say 100mm, re-sleeve it in something of your choosing and re-terminate it with 2 x USB (+/- signal and power/ground) or 1 x USB and 1 x DC connector
 
4

4PLaY

Member
Joined 2013
The problem to begin with is that I want to use the DAC on a raspberry pi, and I am worried that the supplied power from the USB 5v will be inadequate or at best mediocre. :)

I'm not against modifying a belkin gold, sounds like a good idea. Maybe even just modify one that I already have on hand. I really just want to run an external 5v on the power line and while im at it, to just isolate it from the data lines.

I have this guide to help me out. now the hard part for me would be the LED and the Resistor part. Let's not start another foodfight about the type of wires he used. :p

Check it out:

2 Channel Audio : DIY Solid core USB cable with linear power injection
 
D

DF96

R.I.P.
Joined 2007
myweb.tiscali.co.uk/g8hqp/
udailey said:
Well, I think it would be fine for USB. Get out your DMM and measure resistance from one end of a usb cable to the other. I bet its not even ONE OHM.
When manufacturers talk about 75Ohm impedance they dont mean a 3 ft piece of copper has 75Ohms. For a 1 meter piece of copper to have 75 ohms of DC resistance, not AC, it needs to be .017mm in diameter. Kind of tiny. AC would be a bit larger but not a lot. The manufacturers are talking about the impedance of 1000ft of the cable. Thats the standard. Your USB sockets dont care about the resistance of the one meter cable as long as its low low low. If you want to build your own USB cable just do it. I did it and it works fine. Used silver for no good reason other than it was on the table.
Click to expand...
Do I detect a tiny smidgeon of confusion between transmission line characteristic impedance and conductor resistance? They are both measured in ohms, but that is about it. Reading a good book might help.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Top Bottom