Do I need to design a LPF for 30GHz Diff signal?

[tony_lth]

Hi, Gurus,
Now I am studying some 60GHz circuits. And 60GHz is double freq of 30Ghz VCO.
The 30GHz VCO outputs differetial signals.
Normally I should insert a BPF after the 30GHz VCO, but now it is diff-signal.
I am afraid that insert two BPF on the two diff traces may cause delay time difference more than the requirements.
So I think LPF may reduce the possibility of such problem.
But I think maybe no filter should be better.
Could you comment this issue?
Best,
Tony Liu

 

[biff44]

First: How are you getting to 60 GHz from a 30 Ghz VCO? Are you using a frequency doubler? Are you trying to pull off a 2nd harmonic from the 30 GHz VCO?

is it a truly differential signal, i.e. two conductors close together that have a differential mode impedance of 100 or 50 ohms? (where the ground plane has no currents flowing?)

you would then use a differential highpass or bandpass filter, i.e. ONE filter with differential input and output, and therefore the signal will not end up with any time skew.

If instead you have two independent 50 ohm outputs from this vco, and want to install two separate single ended filters, yes you have to worry about time skew in the filter, and time skew from the two independent transmission lines. I personally would use a highpass filter, over a bandpass filter (which has group delay as a funciton of the bandwidth of the filter). it would be easier to match two highpass filters for group delay.

 

[BigBoss]

What is your purpose ?? Cut-Off the main Carrier ? ( 30GHz)

 

[tony_lth]

Hi, @BigBoss & @biff44 ,
Thanks.
It's two independent 50R traces. My purpose is to filter the major spurs out of the 30G VCO.
Even for two same designed HPFs, is the fabricated error of the two traces critical for time skew? Or can I ignore the error? Due to 30GHz, T is 33ps.
Best,
Tony Liu