Post subject: VCO,Packaging model and ESD [help] Posted:
Wed Jul 27, 2005 6:13 am
Joined: Mon May 24,
2004 11:30 pm
In above figure, Lg means the packaing's parasitic inductor(e.g.,~2nH
in 128pin QFP) and Cp means ESD and bond pad's parasitic capacitor(e.g.,~3pF).
It means that this LgCp tank's resonance frequency is about 2GHz.
Unfortunately, the frequency of my signal from Port1 is just
about 2GHz(it is a VCO signal and vary from 1.8GHz to 2.2GHz). Even
when my signal's peak-to-peak voltage value is 0.4v, the simulation
shows that the peak-to-peak voltage value at Port2(it is the gate of
a NMOS) is about 10v(my vdd is 2.5v). It will make the NMOS breakdown!
In theory,I found that if the LgCp's res. freq.(i.e.,1/sqrt(Lg*Cp))
equal the signal's frequency, the signal at Port2 will be large very
How to solve this problem? Would you please give me any
advice or references?
Thanks very much.
Post subject: Posted: Wed Jul 27, 2005 6:25 am
Joined: Mon Jun 27, 2005 2:02 pm
Just a comment: The resonance frequency
It will make the resonance
frequency even lower.
Please provide more details about your
Post subject: Posted: Wed Jul 27, 2005
Joined: Mon May 24, 2004 11:30 pm
Nice to see you again.
I want to introduce
an external VCO signal(about 2GHz) into my circuit(a divider) in order
to realize a frequency synthesizer.(i.e.,the vco in this frequency synthesizer
is external,not in my die).
So I add the packaing model(it is
a 4-port model for vco differential siganl) and sweep the Cp(acted as
the 100um*100um bond pad's capacitor and ESD capacitor) in the schematic
of my first post. Then connect with my divider(i.e.,Port2 is connected
with my divider).
When the input siganl(at Port1 in the schematic
in my first post) peak-to-peak voltage is 0.4v,I found the Port2's voltage
increase from about 0.5v to 15v with the increasing Cp from 0p to 2p.
Then Port1 voltage will decrease to 0.1v with continually increasing
Cp from 2p to 10p. In other words, the peak-to-peak voltage value at
Port1 have a peak value(about ~15v) and this peak value exists when
input signal's frequency close to the LgCp tank's resonance frequency.
I am afraid that such a large voltage will demage the MOS in divider
practically. Seriously it will make my divider out of work!
there any similar phenomenon in RFIC circuit? for example, when use
a off-chip RF PLL to provide LO signal for mixer(because in this situation,
a RF LO signal has to be introduced into mixer and this signal must
be affected by mixer's packaging and ESD). How to handle this situation?
Thank you very much.
Posted: Wed Jul 27, 2005 10:29 am
Mon Jun 27, 2005 2:02 pm
I don't have experience in RFIC design, however I can
tell you that in any PLL synthesizer that I saw or designed myself,
I used to connect a pad (Resisitive element) between the divider's input
and the VCO output in order to provide match to the divider's input
and to adjust the level of the power from the VCO into the divider.
I don't know if this is possible in your design. This pad has no capacitive
effect, since it is composed of resistors, why don't you give it a go?
Post subject: Posted: Wed Jul
27, 2005 8:57 pm
Joined: Mon May 24, 2004 11:30
I will use an attenuator between
the divider and vco in order to do impedance match,power match and bias
voltage. I am not sure whether such a resistive netwoek will help me.
Thank you again.