Continue to Site

Welcome to

Welcome to our site! is an international Electronics Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.

USB D+ and D- signal frequency

Not open for further replies.


Advanced Member level 2
Jul 4, 2005
Reaction score
Trophy points
Activity points
The USB Full-Speed interface runs at a data rate of 12 MHz. The interface runs at four times this frequency and must be clocked from an accurate 48 MHz clock source

Due to this the USB D+ and D- lines needs to be 90ohm from at least 12MHz to 48MHz?


USB specification is free to download.

[Moderator action: moved from "RF, Microwave, Antennas and Optics"}



then I misunderstood your question.

If you are asking about characteristic impedance: Then it´s specified.

If you are asking about frequency range: The data rate is 12MHz. But since the waveform is not sine but square .. you theoretically have infinite frequency overtones. So you can´t limit at 48MHz.

When designing a trace with characteristic impedance.. it´s theoretically for unlimited frequency. It is not designed for a dedicated range like 12MHz .. 48MHz. There are physical limits like trace length and lso limits like dielctric loss.


i know the trace needs to be 90ohm up to infinite frequency.

However, it's impossible to design a 90ohm trace up infinite frequency range due to PCB parasitics. how do solve this problem?

i dont think we can practically design 90ohm D+ and D- up to infinite frequency.
--- Updated ---

The only way is to assume the D+ D- frequency range—that's why I asked the question.


So if we assume 12MHz ... and you don´t want to get much impedance problems, then going to 100MHz should be well enough.

So if the 5th harmonic (60MHz) which has only 20% of the fundamental amplitude on asquare wave... then the "mismatch problem" shouldn´t harm, even with FR4 material.

But 12MHz is rather low frequency. ... and means signal wavelength of maybe 40m.
So when we say (rule of thumb) impedance matters maybe down to 1/10 to 1/20 of wavelength, then we still ar at least at 2m. I guess your traces are much shorter. Thus the whole impedance problem is not that critical.

So my recommendation:
Design your traces according specified impedance and you are safe.
Relax. Don´t overcomplicate a 12MHz interface.

12 MBPS ( full speed) USB waveform has 12 MHz maximal toggle rate, more like a 6 MHz square wave.

The cable properties are specified in the USB spec in term of TDR measurements and frequency dependent signal attenuation, no other requirements. You may additionally refer to the rise time spec of 4 - 20 ns to evaluate the effect of discontinuities, although it's not strictly suggested by the spec. 4 ns corresponds to about 90 MHz signal bandwidth.

In practice, it's not easy to miss the full speed requirements with your design. High speed is a different thing.
No, the commonly used relation of rise time and bandwidth is based on first or second order low pass filters. 2 ns rise time corresponds to about 175 MHz in this relation. Full speed USB drivers are specified to have minimal 4 ns rise time.
Not open for further replies.

Part and Inventory Search

Welcome to