18-729 Fall 2004 Announcements

Date Announcements
8/30
  • Welcome to 18-729!
9/1
  • The username and password given in the class for the course materials are correct.
  • You can now find 18-729 in CMU Blackboard with course ID "F04-18729", though there is only a link to this real class website under "Course Information".
  • TA's office hour will be on Friday, from 11:00 to 13:00. The location is HH 2138.
9/9
  • Slides for the next two lectures have been posted.
  • TA's office hours have been changed! New time slots: Mon. & Fri., 11:30AM ~ 12:30PM. The location is still in HH 2138.
  • Rev.2 slides for Lecture 1-4 has been posted.
9/14
  • Rev.3 slides for Lecture 1-4 has been posted.
9/16
  • Homework 1 is due today. Solutions will be posted at 4 PM.
  • Homework 2 has been posted. Due date is Sept. 28 (Tuesday).
  • Homework 1 solutions has been posted.
9/20
  • Rev.1 slides for Lecture 6 has been posted.
  • Slides for Lecture 7 on radio architecture has been posted.
9/21
  • Rev.1 slides for Lecture 7 on radio architecture has been posted.
  • Two papers on sub-sampling receiver architecture have been posted.
9/22
  • HW#1 has been graded. It will be returned before tomorrow's lecture.
  • One comment on homework: please make your answers clear and neat! This will help TA make correct decisions. Thank you!
  • A change in HW#2: please neglect the last sentence in the description of Problem 2. Basically, it want to tell you that all L's and C's are idel.
  • A hint about Problem 2 in HW#2: There is an equivalent input impedance for BJT but we want to study the effect of Rb.
9/25
  • More hints and clarification about HW#2:
    • Problem 2: Focused on Rb's effect on the NF of the circuit! You can assume BJT is noiseless (why?).
    • Problem 4: KT/C noise is for each R-C.
    • Problem 6: "Ae" and "PHIe" are feedback quantities. Your goal is to show if they can reduce the gain and phase error caused by polyphase filter.
  • There is a mistake in Problem 6 of HW#2, the phase shift in Q-path should be +90. Same mistake occurs in your lecture notes ("Lec4_RadioArch_rev1.pdf", p29).
9/26
  • There is a typo in Problem 1. Part (d) is actually part (c). This won't bother you anyway.
9/27
  • For Problem 4, part(b): in the I-path, the capacitor is supposed to be laid out as the bottom plate is connected to the substrate (ground); and in the Q-path, the capacitor is laid out as the bottom plate is connected to the resistor.
9/30
  • Lecture Slides II for CMOS RF Device Modeling have been posted.
  • HW #2 Solutions has been posted.
  • HW #3 has been posted. Due at 4 PM on Oct. 7 (Thursday).
10/4
  • In HW#3, Problem3, please use Vth,nmos=0.49 V.
10/5
  • In HW#3, Problem4, "Vod=Vgs-Vth" is constant at any temperature.
  • There is no extension for HW#3. It is due on this Thursday (Oct. 07).
10/6
  • HW#3, Problem 1: I(t) and Q(t) are baseband signal, so there are no carriers associated with these signals. In other words, they are just "I(t)" and "Q(t)", not something like "Icos(wt)" and "Qsin(wt)". The function of a GOOD dual-image-reject transmitter in this case is to give you a correct quadrature modulation output. You need to justify if this is true for the given structure based on what you really get at the output.
  • HW#3, Problem 3: plot Ids ~ Vgs, and assume Vds = Vdd.
10/7
  • HW#3 due date has been extended to 10/8 Friday at noon.
  • Lecture slides on BJT RF model have been posted.
10/12
  • For "number of fingers", please use "multiplier" option for MOSFET devices in Cadence.
  • For those plots, you need to use these functions given in the "Direct Plot Form" for S-parameter analysis:
    • YP -- Y parameter
    • HP -- H parameter
    • Gmax -- Power gain
  • The derivative function is in the Calculator: Special Functions --> deriv.
  • If you are curious and you want to learn more about Cadence, please use help documentation of Cadence, as discussed in Cadence tutorial. Have fun!
10/14
  • HW#4, Problem 1, the total width should be "60 um", not "90 um".
10/19
  • For midterm, you can bring all of lecture notes and/or books. GOOD LUCK!
11/2
  • Rev. 2 of lecture slides on LNA design has been posted.
  • HW#5 has been posted. Due date is 11/18 (Thurs.) by 4 PM. This homework requires a lot of Cadence simulation, so start early.
11/5
  • Update of HW#5, Problem 2: the specification has been changed to:
    • Impedance matching: |S11| & |S22| < -12 dB from DC to 2.4 GHz.
    • Voltage gain: |vout(jw)/vin(jw)| > 10 dB.
    • Bandwidth: BW > 2.4 GHz.
    • Additionally, for this problem, you may set R1 = 0.
11/9
  • Rev. 2 of HW#5 has been posted.
11/16
  • There is one additional specification for the Problem 4 in HW#5: gain > 20 dB at 2.4 GHz.
  • A little about noise simulation: there are at least 3 ways to get noise figure (NF):
    1. S-parameter analysis
    2. AC analysis + Noise analyais
    3. PSS analysis + PNoise analysis
    So, if you cannot use one of them then try others. However, for phase noise and mixer simulation, you may have to use the last one.
  • To get 1-dB compression point, as stated in the handout, you should use Transient analyis, not PSS analysis.
  • Please always check the operation points of the transistors in your circuits, and make sure they are properly biased -- i.e., use "gm vs. Iden" and "ft vs. Iden" to check the operation region.
11/17
  • For simulation issues, for example, if you forget how to perform a PSS analysis to get IIP3, please take a look at "SpectreRF User's Guide" in Cadence help documentation. You can open the help docs by choose "Help" menu in icfb window.
11/29
  • Please read this additional note, and use those numbers to get IIP3 of your mixer!