regards,
Larry Smith
Sun Microsystems
> Date: Wed, 17 Mar 1999 16:18:25 -0800
> From: Vinu Arumugham <[email protected]>
> MIME-Version: 1.0
> To: Larry Smith <[email protected]>
> CC: [email protected], [email protected]
> Subject: Re: [SI-LIST] : A Question About Power Noise.
> Content-Transfer-Encoding: 7bit
>
> Larry,
>
> Looking at the impedance/frequency plot of say a 0.01uF and a 0.1 uF capacitor in
> the same package/layout, one finds that the impedance of the 0.01 uF capacitor only
> briefly dips below the inductive region of the 0.1uF line. If the goal is broadband
> low impedance for the power distribution system, then the larger value is a better
> choice. On the other hand, if you are trying to cover a narrow band of frequencies,
> then a carefully chosen smaller value may be better.
>
> Thanks,
> Vinu
>
> Larry Smith wrote:
>
> > Yes, I believe a low value of capacitance is important. A good
> > model for a decoupling capacitor is a series RLC circuit. The
> > circuit resonates (becomes a low impedance) at the frequency
> >
> > f0 = 1/(2pi*sqrt(LC))
> >
> > At that frequency, the capacitive reactance and the inductive reactance
> > cancel each other and you are left with the ESR of the capacitor as the
> > impedance.
> >
> > The total inductance (reference Ray Anderson's note on this thread) of
> > an 805 or 603 capacitor mounted on carefully designed pads can be 1 nH
> > or less. Working backwards in the above formula, you would want a 25
> > pF capacitor to decouple a 1 GHz frequency. Our studies show that an
> > NPO 27 pF capacitor has an ESR of about 0.46 Ohms
> >
> > The inductive reactance jwL = j6.28 Ohms and the capacitive reactance
> > 1/jwC = -j6.28 Ohms at 1 GHz. In the series RLC, those reactances
> > cancel at resonance and we are left with the 0.46 Ohm ESR, far less
> > than the reactances.
> >
> > This resonance has a pretty sharp Q and you have to dial in the decap
> > value pretty carefully if you want to hit an EMI problem frequency. A
> > 25 pF capcitor mounted on a padset that gives a total of 1 nH
> > inductance provides a far smaller decoupling impdeance than a larger
> > valued capacitor.
> >
> > regards,
> > Larry Smith
> > Sun Microsystems
> >
> > > Date: Wed, 17 Mar 1999 09:39:23 -0800
> > > From: Vinu Arumugham <[email protected]>
> > > MIME-Version: 1.0
> > > To: Mark Randol <[email protected]>
> > > CC: "[email protected]" <[email protected]>
> > > Subject: Re: [SI-LIST] : A Question About Power Noise.
> > > Content-Transfer-Encoding: 7bit
> > >
> > > Mark,
> > >
> > > Is a low value of capacitance really important? At 1GHz+, the package
> > inductance dominates the characteristics and so the impedance of a 0.1uF and a
> > 100pF capacitor in an 0603 package will not be very different. As you point out,
> > > for better performance at higher frequencies one would need a lower inductance
> > 0402.
> > >
> > > Thanks,
> > > Vinu
> > >
> > > Mark Randol wrote:
> > >
> > > > SMT 0603 ceramic caps are still pretty good at 1GHz+ as long as you keep
> > > > the value under a couple 100pF. 0402's go up through a couple GHz.
> > > >
> > > > --
> > > > ---------------------------------------------------------------
> > > > Mark Randol, RF Systems Engineer | Motorola SPS, Inc.
> > > > (602)413-8052 Voice | M/S EL379
> > > > (602)413-4150 FAX | 2100 E. Elliot Road
> > > > [email protected] | Tempe, AZ 85284
> > > > ---------------------------------------------------------------
> > > >
> > > > **** To unsubscribe from si-list: send e-mail to
> > [email protected] In the BODY of message put: UNSUBSCRIBE si-list,
> > for more help, put HELP. si-list archives are accessible at
> > http://www.qsl.net/wb6tpu/si-list ****
> > >
> > >
> > >
> > >
> > > **** To unsubscribe from si-list: send e-mail to [email protected]
> > In the BODY of message put: UNSUBSCRIBE si-list, for more help, put HELP.
> > si-list archives are accessible at http://www.qsl.net/wb6tpu/si-list ****
> > >
>
>
>
**** To unsubscribe from si-list: send e-mail to [email protected] In the BODY of message put: UNSUBSCRIBE si-list, for more help, put HELP. si-list archives are accessible at http://www.qsl.net/wb6tpu/si-list ****