Radio Frequency Design Project 4
Agilent ADS and Cadence 6 Software Tutorials, continued
Overview
Remain in same project groups for the
semester.
The objective of this project is to work with S-parameters, Smith
chart, and Cadence layouts
NOTE: Use the Project
Report
Template and keep answers to
questions on consecutive sheets of paper with all plots at the
end.
IN NO CASE may code or files be exchanged between students, and
each student must answer the questions themselves and do their own
plots, NO COPYING of any sort! Nevertheless, students are
encouraged to collaborate in the lab session.
Only turn in requested plots ( Pxx )
and requested answers to questions ( Qxx ).
Part 1
Select the plot options to plot S11 and S21 from -20 to 0 dB
in 1 dB steps.
Print the S-parameter plot and turn it in. You should have a
"rippled" plot of both S11 and S21. S21 should ripple between -6
and -8 dB. ( P2 )
Make sure that your
plots, legends, axes, and fonts are legible in your report!
What is the line impedance? (from formula or RFMD tables or
from linecalc) Hint: see the substrate block on the
schematic. ( Q1 )
What is the effective source impedance to the left of the
line (TERM1 and the resistor R3, as below)? ( Q2 )
What is the effective load impedance to the right of the line
(TERM2 and the resistor R2, as below)? (
Q3 )

Note that the S-parameters are plotted only from the
perspective of the TERM devices, so the power transmitted to the
right TERM device is only the power delivered to the TERM
resistor, and the reflected power is only that reflected back to
the left TERM.
Draw a schematic on a sheet of paper with a 50 ohm 2 volt
rms source (2 volts rms open-circuit voltage) and a load
comprised of three 50 ohm resistors in parallel. At extremely low frequencies, this is a
good model of the circuit since the transmission is
electrically short (i.e., much less than a quarter wave).
This is a good model at low frequency since the line is not
very long in wavelengths at 1 MHz! Questions Q4 - Q11 relate
to hand calculations using this hand-drawn schematic.
Calculate the voltage across the three 50 ohm resistors
that are in parallel at 1 MHz in your hand-drawn schematic?
( Q4 )
Calculate the power in mW delivered to one of the three
parallel resistors.? (This would correspond to the power
delivered to one of the three 50 ohm load resistors in
parallel in your hand-drawn schematic.) ( Q5 )
Calculate the maximum power in mW available available
from the source at 1 MHz (i.e., replace the three 50 ohm
resistors with a single 50 ohm load)? (
Q6 )
Calculate how many dB down is the power in Q6 relative to
Q5 at 1 MHz? (10 log10( Q5/Q6 ) ) ( Q7
)
How does your answer to Q7 relate to the plot P2 at 1
MHz? ( Q8 )
For Q9- Q11, assume that the maximum available power from
the source is 1 watt:
(The answers to the questions below
should agree with the foregoing calculations in Q9-Q11 AND
with plot P2 for the above ADS schematic at low
frequency(1MHz), otherwise you have an error somewhere)
Calculate the power in mW delivered to the right-hand
TERM at 1 MHz in the above ADS schematic? ( Q9 )
Calculate the power in mW delivered to the right hand
50 ohm resistor R2 at 1 MHz in the above ADS schematic?
( Q10 )
Calculate the power in mW delivered from the source
at 1 MHz in the above ADS schematic? ( Q11 )
What is the length of the line in wavelengths at the bottom
of the S11 dip near 14 MHz?
(you should be able to figure this out
from the schematic and plot P2, but you can also confirm this
from the transmission line geometry using linecalc to
double-check your answer) ( Q12
)
Looking into the left side of the line at the bottom of the
S11 dip near 14 MHz, use a Smith chart to figure out the
impedance. What is the impedance looking into the left side of
the line (as shown below)? ( Q13 )

Note that the S-parameters are plotted only from the
perspective of the TERM devices, so the reflected power that is
plotted is only that reflected back to the left TERM.
Draw a schematic on a sheet of paper with a 50 ohm 2 volt
rms source and a load comprised of the left-hand 50 ohm
resistor in parallel with the answer to Q13. At the S11 dip near 14 MHz, this is a good
model of the circuit since the transmission has transformed
the impedance (we cant use the previous low frequency model
anymore!). Questions Q14 - Q20 relate to hand
calculations using this hand-drawn schematic.
Calculate the voltage across the 50 ohm resistor in
parallel with the Q13 calculated resistance near 14 MHz in
your hand-drawn schematic? ( Q14 )
Calculate the power in mW delivered to the Q13 resistance
near 14 MHz in your hand-drawn schematic.? ( Q15 )
Calculate the maximum power available available from the
source (i.e., replace the 50 ohm resistor and Q13 resistance
with a single 50 ohm load)? ( Q16 )
Calculate how many dB down the power in Q15 is relative
to Q16 near 14 MHz? (10 log10( Q15/Q16 ) ) ( Q17 )
Consider that all the power in the Q13 resistance near 14
MHz must be dissipated in the right-hand TERM and right-hand
50 ohm resistor (the line is lossless). And since the
voltage across these two resistors is the same, and they
have equal resistance, the total power must be equally split
between them.
Considering this, and the answer to Q17, calculate how
many dB down is the power delivered to the right-hand TERM
relative to the maximum power available from the source;
hint: your answer should properly relate to the S21 plot
P2 at 14 MHz ? ( Q18 )
Using the Q13 impedance in parallel with the left-hand 50
ohm resistor, calculate the effective load resistance seen
by the left-hand TERM at 14 MHz (as shown below)? ( Q19 )

From the foregoing answer, calculate the return loss of
the impedance as seen by the left-hand TERM at 14 MHz (this
should correspond to S11 at 14 MHz)? (
Q20 )
Change the S-parameter box in the ADS
schematic to a stop frequency of 20 MHz by double
clicking the S parameter box on the schematic. Rerun the
simulation.
Click the "smith plot" icon in the data plotting window.
Drop the plotting box in the visible area, and in the pop-up
window:
Select DataSet -> S(1,1) -> Add
(only plot S11 on the Smith chart)
Click on the plot options, click Coordinate->both, select
Grid, and select admittance line type as long dash, and pick a
green/blue shade color and red/yellow shade for impedance.
Click OK and the Smith Chart should appear. To figure out
which end of the curve is at what frequency, Select
Marker->New from the Dataplot menu bar and click on one end
of the plotted curve. Then Edit->Undo to delete the marker.
The Smith chart should be similar to below:

Print the smith chart plot (without the marker) and turn it
in. Please write by hand (or markup) on the plot the
start and stop frequencies on the curve (mark the end-points).
You should have a "semi-circle" in the left side of the chart
(Hint: similar to above). ( P3 )
From the Smith Chart (using a compass, compare the ADS Smith
Chart plot to the Smith Charts used in class) what is the
impedance (unnormalized) at 1 MHz? ( Q21 )
From the Smith Chart (using a compass, compare the ADS Smith
Chart plot to the Smith Charts used in class) what is the
impedance (unnormalized) at 14 MHz? ( Q22
)
From the Smith Chart (using a compass, compare the ADS Smith
Chart plot to the Smith Charts used in class) what is the Susceptance (unnormalized) at 20
MHz? ( Q23 )
Again use the marker and drag it with the mouse to observe
the read-out and find the impedance (unnormalized) at 20 MHz? ( Q24 )
( NOTE: the above answers should agree
with all of your hand calculations above. Remember, the
impedance you observe consists of everything except the
left-hand TERM. )
Change the S-parameter box in the ADS schematic to a stop
frequency of 100 MHz by double clicking the S parameter box on
the schematic. Rerun the simulation.
Replot the Smith chart from 1 to 100 MHz and plot as before,
Print the Smith chart plot (without markers) and turn it in.
You should have a "circle" in the left side of the chart. ( P4 )
Why do you have a "circle" now? ( Q25 )
Part 2
In this part, we will experimentally determine the wavelength
and impedance of a microstrip line of width 0.025 inches (25
mils) on a 0.05 inch (50 mils) thick epoxy-fiberglas FR4 board
(apologies to the metric system ...). This is a good way to
check your microstrip line impedance for future projects.
Go down through the directory tree to p2txline3, and double
click that design file, and the following schematic should
appear.
Double-click the "gear" icon in the upper right of the window
to simulate.
Click the "rectangular plot" icon in the pop-up data plotting
tool.
Select the plot options to plot S11 and S21 from -30 to 0 dB
in 3 dB steps.
Print the S-parameter plot and turn it in. You should have a
plot of both S11 and S21. S11 should have a minimum (or null)
near 30 MHz. ( P5 )
What is the electrical length in wavelengths of the line at
the minimum of S11 near 30 MHz??
(you should be able to figure this out
from the schematic and P5 alone, but you can also confirm this
from the transmission line geometry using linecalc to
double-check your answer) ( Q26
)
What is the peak value of S11 near 15 MHz? ( Q27 )
What is input impedance looking into the left end of the line
at the peak value of S11 near 15 MHz (as shown below)?
(Hint: is the impedance looking into the
line at its largest or smallest value there?) ( Q28 )

Change the S-parameter box in the schematic to a stop
frequency of 25 MHz by double clicking the S parameter box on
the schematic. Rerun the simulation.
Click the "smith plot" icon in the data plotting window.
Click on the plot options, click Coordinate->both, select
Grid, and select admittance line type as long dash, and pick a
green color. Click OK and the Smith Chart should appear. To
figure out which end of the curve is at what frequency, Select
Marker->New from the Dataplot menu bar and click on one end
of the plotted curve. Then Edit->Undo to delete the marker.
Print the S11 smith chart plot (without the marker) and turn
it in. Please write by hand on the plot the start and stop
frequencies on the curve (mark the end-points). Also mark the
frequency where the impedance crosses the purely resistive axis.
You should have a "semi-circle" in the right side of the chart.
( P6 )
From the Smith Chart what is the impedance ("un-normalized")
looking into the left end of the line at 15 MHz? ( Q29 )
What is the line impedance?
(DONT USE A FORMULA to calculte this
answer! figure it out from the experimental data.)?
(Hint: experiment with quarter wave line sections on the smith
chart, and see that an impedance of Zo/k is transformed into
an impedance of kZo; a transformation ratio of k^2.)
( Q30 )
What is the effective dielectric constant of the line?
(DONT USE A FORMULA to calculte this
answer! figure it out from the experimental data.)
(Hint: lambda = lambda0 / sqrt(relative dielectric constant).
( Q31 )
Part 3
In this part, we design matching networks.
Design a 2-element matching network to match a load of 250
ohms in series with a 2 nH inductor into 50 ohms at 4 GHz.
Proceeding from the load, use a shunt inductor then a series
capacitor for the matching network. (use file p2match1 as a
starting point and find the proper values for the 1000 pF
capacitor and 1000 nH inductor).
For the load alone (as shown below with no matching
network), Plot the the Immitance Smith chart from 3 to
5 GHz., and make sure to add a Marker
at 4 GHz ( P7 )
Design the matching network as shown below.
What are the final design values for the matching network
inductor (nH) and capacitor (pF)?. (
Q32 )
Printout the schematic of your final matching network
along with the load. ( P8 )
Plot the the Immitance Smith chart for the matched load
from 3 to 5 GHZ (label, by hand, start and stop frequency),
and make sure to add a Marker at 4 GHz
. ( P9 )
Plot S11 (-30 to 0 dB) for the matched load from 3 to 5
GHZ. ( P10 )
Over what frequency range (from what frequency to what
frequency) is your return loss better than 10 dB? ( Q33 )
Design a 2-element matching network to match a load of 20
ohms in series with a 5 nH inductor into 50 ohms at 4 GHz. (use
file p2match2 as a starting point and add a 2-element input
matching network).
NOTE: you might not be able to use a
shunt element as the first element of the 2-element matching
network! So redraw the schematic as required.
Plot the the Immitance Smith chart for the load alone (no
matching network) from 3 to 5 GHz. (
P11 )
What are the values for the matching network inductor
(nH) and capacitor (pF)?. ( Q34 )
Printout the schematic of your final matching network
along with the load. ( P12 )
Plot the the Immitance Smith chart for the matched load
from 3 to 5 GHZ (label start and stop frequency). ( P13 )
Plot S11 (-30 to 0 dB) for the matched load from 3 to 5
GHZ. ( P14 )
Over what frequency range is your return loss better than
10 dB?. ( Q35 )
Part 4
For the past 4 months, you have been designing the world's
greatest 4 GHz amplifier chip with 300 ohm output impedance and
only 0.05 pF stray capacitance ... a perfect impedance match to
the 300 ohm antennas you bought last month for the prototype
radios!
To your horror, your manager comes into your office and says
that the president of the company insists that you package
your chips into the packages he bought from a friend on a
fishing trip. The packages are not very good for RF. After the
obligatory bond-wire inductance of 2 nH connecting your chip
to the package bond area, they have a stray capacitance of 1
pF to ground and another 1.5 nH series inductance because the
pins are quite long ... And if that wasnt bad enough, the
president traded all of your 300 ohm antennas for a box of
fancy 50 ohm antennas ... so you now need a 50 ohm output!
To make matters worse, your manager says the program is over
budget so you cant use any inductors or capacitors because
they will cost too much. He insists that you can only
use microstrip transmission line sections laid out as foil
patterns on the system board.
Based on the great stuff you learned in RF class, you save
the day by matching the antenna with a single little piece of
microstrip ... and the boss gives you 10,000 Enron stock
options as a bonus.
A schematic of the whole situation is shown below. You
need to make the impedance look like 50 ohms looking into the
transmission line from the left. (also see the really handy
file p2match3)
Plot the the Immitance Smith chart for the chip alone
(the 300 ohm and 0.05 pf) from 3 to 5 GHz. ( P15 )
Plot the the Immitance Smith chart for the chip in the
package (the chip plus the two inductors and 1 pF, with no
transmission line matching network) from 3 to 5 GHz. ( P16 )
Notice what the package did to the chip!
The package is so bad, it almost took the impedance from an
open circuit to a short circuit ...
What are the values for the microstrip matching network
length (meters) and width (mils)?. (
Q36 )
Printout the schematic of your final circuit. (including
matching network along with the load, ie., the chip plus the
two inductors and 1 pF). ( P17 )
Plot the the Immitance Smith chart for the matched load
from 3 to 5 GHZ (label start and stop frequency). ( P18 )
Plot S11 (-30 to 0 dB) for the matched load from 3 to 5
GHZ. ( P19 )
Over what frequency range is your return loss better than
10 dB?. ( Q37 )
Sketch your transmission line to 1:1 scale. ( Q38 )
Part 5
- Run the Cadence6 software
- Open your 2-resistor voltage divider schematic from the
previous project in the Cadence 6 software, by double-clicking
the schematic view of resDiv_cell in the Library Manager tool.
- Change both resistors in the
schematic to 100 ohm resistors from the
"ami05_c6_customlib"
- Make sure that none of your pins have names that could
interfere with global names/variables. NEVER name a pin "gnd" or "vdd" or "vcc" or
"vss"
- Check the schematic in Schematic-XL tool using
MenuBar::Check::CurrentCellView (yellow arrow below) and make
sure no errors are reported in the CIF window
- As shown below, launch the Layout XL tool (red arrow below)
- In the popup window, select create new (blue arrow below)
- Print out your new schematic. ( P20 )
- In the Layout XL window, create a new layout using
MenuBar::Connectivity::Generate::AllFromSource as below
- In the popup options windows, check the tabs for the options
(blue arrows below) and make sure PR
boundary (place and route boundary) is turned off as
below (red arrow below)
- Note that all of the pins from your schematic will be created
as metal1 0.9x0.9 micron pins (yellow arrows below)
- The devices from the schematic should now appear in the layout
- As before in Layout XL, use MenuBar::Options::Display to set
the "Display Levels" to 0 to 30, and to set the check buttons
"on" for "Pin Names" and "Instance Pins."
- If you click on a pin in the layout or schematic, the
corresponding item should also be highlighted
- Selecting a pin in the layout highlights a pin in the
schematic (blue arrow below)
- Selecting a resistor in the layout highlights a resistor in
the schematic (yellow arrow below)
- Activate the annotation bowser pane using
MenuBar::Window::Assistants::AnnotationBrowser
- Then, use MenuBar::Connectivity::Nets::ShowAllIncompleteNets
to see the nets that must be connected, as illustrated
below
- You can move devices by typing "m" in the layout area
- Then, move your pins and devices into positions approximately
as in the schematic, where the nets will be easy to connect and
wire, as shown below
- Print out your preliminary layout with "fly lines" of nets
visible as above . ( P21 )
- Select metal1 on the LayersPalette in the left side of the
Layout XL window (red arrow below), then type "p" when your
mouse is in the layout area to create a metal1 path.
(Alternatively use MeuBar::Create::Shape::Path)
- Connect your circuit as illustrated below using metal1 paths
- Experiment with typing "r" in the layout area to draw a metal1
rectangle
- Print out your final layout as above. (
P22 )
- When you finish wiring the layout, run DRC (design rule check)
using MenuBar::Verify::DRC as shown below
- Make sure no errors are present by inspecting the message in
the CIW window (red arrow below)
- For a list of rules see MOSIS.com
ON Semiconductor C5N SCMOS_SUBM
rules
- Take a snapshot of your CIW output for DRC showing "Total
errors found: 0" as above. ( P23 )
- When you finish DRC, run extraction (extract schematic from
layout) using MenuBar::Verify::Extract as shown below
- Make sure no errors are present again by inspecting the
message in the CIW window (red arrow below)
- When you finish extraction, run LVS (layout versus schematic)
using MenuBar::Verify::LVS as shown below
- Use the "form contents" option (green arrow below)
- Use the browse buttons (yellow arrows below) to navigate to
the schematic view and extracted views of your design
- Check that the proper files and views are loaded into the form
(red arrows below)
- Make sure no errors are present again by inspecting the
message in the CIW window to see the message "netlists match"
- If your design layout matches your schematic, you should get a
netlist match popup as follows
- UNTIL YOU SEE THE FOLLOWING MESSAGE,
YOUR DESIGN IS INCORRECT!
- Repeat the DRC, extract, and LVS steps on revisions of your
design until you get the above message
- The LVS output file often has good information for debugging
errors
- You should always check the output file,
as below, even if you get the "netlists match" popup
- To access the LVS output file, click the output button (red
arrow below)
- Take a snapshot of your LVS output window as shown above
showing "the netlists match" as above. MAKE SURE that you
include the full output up through the "netlists match" line
(blue arrow above), and include the full file name of your
design (green arrow above) ( P24 )
Report
NOTE: Use the
Project
Report
Template and
keep answers to
questions on consecutive sheets of paper with all plots at the
end.
Do not add extraneous pages or put
explanations on separate pages unless specifically directed to do
so. The instructor will not read extraneous pages!
Only turn in requested plots ( Pxx ) and requested answers to
questions ( Qxx ). All plots must be
labeled P1, P2, etc. and all questions must be numbered Q1, Q2,
etc. YOU MUST ADD CAPTIONS AND FIGURE NUMBERS TO ALL
FIGURES!!
Copyright © 2010-2012 T. Weldon
Cadence, Spectre and Virtuoso are registered trademarks of
Cadence Design Systems, Inc., 2655 Seely Avenue, San Jose, CA
95134. Agilent and ADS are registered trademarks of Agilent
Technologies, Inc.