DESIGNERCOUNCIL Archives

1996

DesignerCouncil@IPC.ORG

Options: Use Monospaced Font
Show Text Part by Default
Show All Mail Headers

Message: [<< First] [< Prev] [Next >] [Last >>]
Topic: [<< First] [< Prev] [Next >] [Last >>]
Author: [<< First] [< Prev] [Next >] [Last >>]

Print Reply
Subject:
From:
Date:
Thu, 09 May 96 17:07:05 CST
Content-Type:
text/plain
Parts/Attachments:
text/plain (270 lines)
     
     Ralph,
     
     As usual you've come through.
     
     Although, I understand your points well and can't disagree with 
     you, I'm still not sure I can totally agree.
     
     This "won't disagree but not really agree" stand comes from a 
     couple of observations:
     
     1. That the telegraph equations get split up into "high
        frequency" and "low frequency" equations depending upon 
        assumptions that retain or remove different parameters from 
        the complete equations, i.e. removing some of these parameters 
        from these equations allow one to study/predict results in 
        coax cables as I'm sure you're well aware.  And, one can 
        "lump" parameters successfully for coax.  But, I'm using the 
        same procedure for the case of a trace.
     
     2. In my model, the units for resistance, inductance, and 
        capacitance are "per length".  To extend the length of the 
        trace, I would not "cascade" the parameters as in adding more 
        sections of a filter network. (I don't know, but does Spice do 
        this in modeling?) The individual parameters only become 
        larger.
     
     3. This is a 1 inch trace.  Not feet of wire.
     
     4. Finally, I can't believe that the trace DOES NOT act as a 
        low pass filter no matter what school or books have showed me 
        about transmission lines.
     
     As stated by someone off line, frequency problems above 800 MHz 
     are well documented with FR-4.  This forces me to believe I'm on 
     to "something".  Yes, I agree with what you say about lumped 
     parameters versus distributed parameters, but even worst case 
     rough estimating tells me I'm near "something".
     
     I guess if I were to be tacked against the wall, I'd have to say 
     OK.  But, that still won't help me.
     
     Doug McKean, ADC Video, [log in to unmask]
     
     
     ______________________________ Reply Separator 
     _________________________________ Subject: DES-PC Board Frequency 
     Limi
     Author:  [log in to unmask] at internet-mail Date:    5/9/96 2:33 PM
     
     
     Subject:                              Time:  7:45 AM
     OFFICE MEMO         DES:PC Board Frequency Limits         Date:  
     5/9/96
     
     Doug:
     
     The following your messages and are combined and reproduced here 
     to complete the query and information, my comments follow at the 
     end.  Ralph
     
     ----------------- initial query ------
     
     >I would like any responses to be off-line for the following 
     discussion.
     
     >Please use the following email address:  [log in to unmask]
     
     >I made a simple mathematical model of a stripline, 1 inch trace, 
     on     FR-4, nothing out of the ordinary as far as dimensions and 
     found to my     surprise a resonant freq at about 800 MHz with a 
     20 to 40 dB roll-off     above 800 MHz
     
     >Could someone in the RF/Microwave world of nuts and bolts pc 
     board     design tell me what I'm supposed to do at say 10 GHz 
     with regular pc     board materials/construction/fabrication???
     
     ---end of initial query---
     ---start of second message with tech. info. follows ---- #004# >I 
     don't have any problem with going on the Net with this.
     
     >This is more of a theoretical issue than anything else.  Not 
     really     related to IPC.  With all the talk about too much 
     mail, I decided to     throw the issue on the net with the option 
     to email direct.
     
     >So without any further talk, here goes:
     
     >I am trying to write a technical paper for my company on what I 
     have     found.
     
     >My basic model was based on a lumped parameter model of a one 
     inch     copper trace with one end terminated with a signal 
     source, the trace     lumped with trace resistance and trace 
     inductance in series, to a     termination of a resistor (load) 
     and capacitor (dielectric) in parallel.
     
     >I drop into the frequency domain with Laplace and write my 
     system   response.  
     
     >Up to this point, I am still in line with any cross check of a 
     similar     circuit in any circuit analysis book.  If anyone is 
     into it, I'm following the basic derivation of 
     telegraph/telephone equations for transmission line signal 
     analysis.  OK, so big deal.
     
     >Now, I plug in "standard" values for the lumped parameters of 
     copper     trace resistance per inch, copper trace inductance per 
     inch...
     
     >I DERIVE the impedance of the trace using IPC impedance relation 
     for a     stripline trace.  From this, I derive capacitance per 
     inch.  This is     where I'm a little skittish about what I'm 
     doing.
     
     >If I set the load resistance to infinity, I get a large resonant 
     spike at about 800 MHz.  If I adjust the load resistance to below 
     150 ohms, the spike goes away.  To anyone interested, I'm playing 
     around with the system damping factor from my second order system 
     response     equation.  Alright, so what???
     
     >No matter what I do with the load resistance, there is a large 
     roll-off above 800 MHz.  Like 40 dB!!!
     
     >I'm using MathCad for printable output.  I have Word, 
     Wordperfect,     Excel, and MathCad software on an IBM compatible 
     486/66 MHz machine.
     
     >So far, I have had two people say that they are interested, a 
     third has had no problem with boards (ECL based circuitry - I was 
     sort of expecting that), a fourth has seen designers wrestle with 
     running simulations - looking at results - do a little adjusting 
     - run simulation again and so on for a couple of weeks (Spice 
     based).
     
     -------  then then the third provided the details ------
     
     I used the wrong word!
     I apologize  MICROSRIP is the correct word. NOT stripline.  
     
     I am doing Microstrip calculations on a Microstrip construction 
     Sorry about that.
     
     Here's some specifics:
     
     Dielectric Constant = 4.5
     
     Dielectric Height = 0.029 inches
     
     Trace Width = 0.008 inches
     
     Trace Thickness = 0.0024 inches
     
     Trace Length = 1.00 inch
     
     Inductance per inch (assumed value) = 20 nH per inch
     
     Calculated Impedance Value for Trace =  107 ohms
     
     Derived Capacitance from Impedance =  2pF per inch
     
     Signal Delay along trace = 187 psec
     
     Resonant Frequency of trace without a load = 850 MHz
     
     Thanks in advance, Doug McKean
     From: [log in to unmask]
     
     ------------------------------------------------------
     
     Doug-
     
     You got different results because of different design and 
     modeling techniques/metholologies.  Transmission lines are 
     modeled differently than a lumped parameter spice-like network 
     analysis model.  What you analyzed using you lumped parameter 
     model is called a "series resonant circuit".
     
     First, the transmission line (in theory) are frequency 
     independant, and therefore, the formulas do not take frequency 
     into account.  Most of the transmission line modeling formulas 
     are based on the variables of  dielectric constant and 
     capacitance.  They extract these out of the more complex 
     transmission line formulas.  The general formula for transmission 
     line impedance is:
     
     Z(o) = sqrt ((R length + i omega (L length)) / (G length + i 
     omega (C length)))
     
     where Z(o) is the impedance of the transmission line
     R is the (series) resistance per unit length (including skin 
     effects) of the conductors
     G is the (shunt) conductance per unit length between conductors L 
     is the inductance per unit length of the conductors
     C is the capacitance per unit length between conductors i omega 
     are the complex quadratic operators.
     
     ----
     Frequency domain network analysis is different because you are 
     now performing an electrical characterization of a resonant 
     circuit and low-pass filter.
     
     When you do a "network analysis" using lumped parameter models, 
     the electrical characteristics of the network is very sensitive 
     to frequency and you are therefore performing an frequency domain 
     analysis.
     
     In using a simple lumped parameter model, the output of your 
     source is connected to your lumped parameter model, which 
     consists of a series inductor in series with the distributed 
     capacitance to ground.
     
     The resonant frequency of a series resonant circuit is when the 
     inductive and capacitive reactances are equal, and when solved 
     for frequency is equal to:
     
     f(o) = 1 / (2 * pi * sqrt ( LC))
     Which using your provided values of L (20 nanoH)  and C  (2 
     picoF)
     
     Plug the numbers in and solve, your resonant frequency f(o) is 
     795.+ MHz.
     
     In theory, at resonance, the impedance of a series resonant goes 
     to 0 (zero) and current is liminted only by the source impedance 
     and series resistances. The current waveform through the series 
     components is common to both the inductor and capacitor.  The 
     derrived voltages across the capacitor and inductor are 90 
     degrees (lead/lag in quadature) out of phase with the current. 
     Because resonance, the series current is infinite, you will 
     develop very high
     voltages (180 degrees out of phase) across the capacitor and 
     inductance (due to low intrinsic impedances).  And because the 
     induced voltages are in quadrature and 180 degrees out of phase 
     they essentially cancel each other across the lumped parameter 
     network.  This is why at resonance, you obtain the very high 
     voltage across the output capacitance of you lumped parameter 
     model.
     Then as you load the capacitor with resistance you will change 
     the phase
     relationships of the electrical network and the two voltages 
     across the inductance and capacitance are no longer 180 degrees 
     out of phase, rather some vector quantity.
     
     As I mentioned in my previous post, the simple series inductace 
     an capacitance is a low-pass filter.  In the frequency domain, as 
     the frequency increases the low-pass filter will do just that, 
     pass low frequencies based on the electrical characteristics at 
     that frequency(s).  At resonance, you get what you observed, a 
     peak.  Then depending on the number of lumped parameter models 
     you have in you model, the higher frequencies will be attenuated, 
     and using a multi-stage model, you will obtain the roll-off or 
     attenuation you observed (calculated).
     
     And yes Doug -- as the more traditional printed board designs 
     make the transition into the world high frequency/speed we get 
     into the world of "serious" printed circuit design, 
     manufacturing, assembly and test.
     
     And yes also to "Groovy" Dave Hoover, those individuals/companies 
     who are currently successfully working in this area have spent a 
     lot of resources to develop their CAE modeling tools in order to 
     obtain correlaton between their CAE tools and reality (what they 
     get as an end product).
     
     Ralph Hersey
     e-mail:  [log in to unmask]



ATOM RSS1 RSS2