What may be behind belief in the "90Ω specification" for USB

[bluesman]

There is no "90 ohm specification" as far as I can determine, Chris. The actual impedance spec for USB 1 and 2 from the 2007 USB.org document is a minimum of 76.5Ω and a maximum of 103.5Ω for the cable. The USB 3 specs only seem to document the need to normalize the resistance across the connection between mated cable assemblies to 90Ω.

I don't know if this is the reason many believe there's a standard 90Ω impedance for USB cables, but it makes sense. USB uses a 45-ohm system in which the driver has a source impedance of 45 ohms and the receiver has a termination of 45 ohms to ground. All USB cables and tracks should have a single-ended impedance of 45 ohms - here's a circuit diagram:

 

 

As both D+ and D- lines terminate into 45 ohm resistors to ground, one could conclude (wrongly) that the cables have a 90Ω impedance. In reality, these resistors are in the devices, not the cable.

 

David

[Harpy]

Wouldn't that be that termination spec of the device? 44 ohms to 28 ohms with the (+15%). "The 90ohms +/- 15% impedance has to do with the internal components, the high frequency signals USB transmits to match the other USB devices, and to reduce noise. Each data line uses a 1.5K and a 15K ohms pull down and up resistors respectively."

[ar-t]

ARGH!!!!!!!!!!!!!!!!!!!

 

Yes, the resistors are in the equipment. But, the characteristic impedance, of the cable, has to match the source and load impedance.

 

No, it is not a physical resistance. Oy vey. It is the cable impedance. 90R + /- 15%.

 

And not all stuff has those 1.5k pull-up/-down resistors. Only the really slow stuff. And it has NO termination resistors.

 

The higher speed stuff does not have those pull-up/-down ones, or it would not work worth diddly squat. Which is why it is terminated.