[SI-LIST] Re: Decoupling for PLL

Ray, these days, you can get a decent inductor from muRata - 10uH 15mA Isat 
in an 0805.  Add a 0.5 to 1 ohm resistor and for the 10KHz case you are 
good to go.  You can get better than 40dB rejection at common SMPS 
operating frequencies and more than 60 through most of the signal frequency 
range.   The inductor gets larger as the demand for a higher impedance goes 
up, or the cut-off frequency goes down.  The insertion loss is ultimately 
limited by the capacitor ESR and ESL.  So a little bead with a very low 
inductance will result in a low Z filter with limited insertion loss at 
HF.  But if it gets the job done, that is all that matters.

I have funny attitudes about things like power and clocks.  Without them 
working properly nothing else does.  So I give them a lot of priority.  It 
doesn't always make me popular.

Regards,


Steve.

At 07:49 AM 8/27/2004 -0700, Ray Anderson wrote:
>Steve-
>
>Thanks for the details of your synthesis procedure. Makes perfect sense 
>technically, but I now see where our "disconnect" is.
>
>In the majority of the designs that I've worked on real estate is at a 
>dear premium. The system designers grouse about having to put in a couple 
>0603 parts (L and C) [and even louder if they are 0805] to implement a 
>power filter. When we advise them to include an extra R to tame the 
>peaking invariably we get to listen to 15 minutes of their PCB real estate 
>woes. Usually the L's that we had to work with were relatively low 
>inductance ferrite beads or ocassionaly actual inductors. In any case 0805 
>was about the largest physical size that was allowed for the filter 
>components(sometimes a 3215 bead if we were real lucky). B case tantalum 
>caps and inductors in the mH range were pretty much out of the question. 
>Hence we didn't have the luxury synthesizing a proper filter as you've 
>suggested. Also, when we did apply the series resistor fix to filters made 
>of relatively small  LC parts, the ohmic value usually turned out to be < 
>5 ohms (more like 1-2) in order to keep the DC levels within narrow 
>ranges. So in the end we usually ended up having to devise suboptimal 
>filters in order to comply with board area constraints. Although they may 
>have been suboptimal, they did indeed provide the requisite filtering at 
>the necessary frequencies without peaking or excessive DC drop.
>
>Thanks again!
>
>-Ray
>
>
>steve weir wrote:
>
>>Ray, being an old stick in the mud, the synthesis procedure I use is 
>>pretty trivial for the series R-L-C.
>>
>>Since Butterworth is our friend, the ratio of the characteristic Z of 
>>sqrt( L/C ) will be 1.4X our target Z, which will equal R.  So the 
>>procedure is:
>>
>>1) Set R from the acceptable IR drop.  Let's take your example of 100 ohms.
>>2) Determine L and C from Zchar and Wn.  For Butterworth Zchar = .707 * 
>>Ztarget
>>a. C = 1/( Zchar * Wn ) = 1/( 71 * 6280 ) = 2.26uF
>>b. L = Zchar / Wn. = 71 / 6280 = 11.3mH
>>
>>A drum inductor at that high of an inductance already has about 40 ohms 
>>resistance, so we just plop a 62 ohm resistor in series and we are 
>>done.  We have a filter that is well behaved from DC up for any load from 
>>open circuit down to the saturation current of the inductor.  Where I/R 
>>drop is an issue we scale the impedance, but the inductor physical size 
>>is essentially fixed by the voltage swing and the cut-off frequency.  For 
>>example if we drop down by 10:1 in inductance, the winding resistance 
>>similarly falls in the same package and the saturation current rises as 
>>the square root.  The same inductor size and cost will apply whether we 
>>synthesize a 100 ohm, 10 ohm or 1 ohm circuit.  All that is going to 
>>change is the size of C, and the size of the series R that we add.
>>
>>If we drop this to a 1 ohm circuit then L drops to 113uH, and C increases 
>>to 226uF ( B box tantalum at low voltage ).  For a bit more reasonable 
>>cut-off frequency of 10KHz we get to drop the energy storage, and the 1 
>>ohm circuit becomes 11.3uH and 22.6uF, both very tolerable values.
>>
>>If we go to the shunt load form then we can keep moving L down and C up, 
>>but that 1KHz cut-off frequency is going to keep the LC product fixed, 
>>and we will need to invert Zchar to 1.4 Ztarget, meaning that the ratio 
>>of L/C will be 4X that of the comparable R-L-C.  So for any given 
>>characteristic impedance, 3A results in a higher L than the R-L-C.  That 
>>was a blessing to voltage mode converters, because the problem was 
>>getting L down.
>>
>>Regards,
>>
>>
>>Steve.
>
>.


------------------------------------------------------------------
To unsubscribe from si-list:
si-list-request@xxxxxxxxxxxxx with 'unsubscribe' in the Subject field

or to administer your membership from a web page, go to:
http://www.freelists.org/webpage/si-list

For help:
si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field

List FAQ wiki page is located at:
                http://si-list.org/wiki/wiki.pl?Si-List_FAQ

List technical documents are available at:
                http://www.si-list.org

List archives are viewable at:     
                http://www.freelists.org/archives/si-list
or at our remote archives:
                http://groups.yahoo.com/group/si-list/messages
Old (prior to June 6, 2001) list archives are viewable at:
                http://www.qsl.net/wb6tpu

Other related posts: