Chapter 5: Design

Questions:

1. Which MMIC transistor technology has an operating frequency down to dc?

2. Which MMIC transistor technology is most suited to low-noise mm-wave applications?

3. Which MMIC transistor technology can operate over 100GHz?

4. What is a good figure for the gain per stage in a multistage amplifier MMIC?

5. What range of input match is reasonable to expect for a 10GHz MMIC?

6. What are the dc bias pads usually placed, and what characteristics of their placement are generally a foundry standard?

7. In a chain of amplifiers, which amplifier has the most impact on the overall noise figure, and what characteristic of this amplifier can reduce the contributions from the other amplifiers in the chain?

8. What technique is employed to ensure good input match and low noise figure simultaneously?

9. Which is the other important amplifier parameter, as well as noise figure, gain, and matches versus frequency?

10. How are oscillations prevented without degrading the noise figure?

11. In the distributed amplifier, where is the wave traveling from and to?

12. If a distributed amplifier is to be constructed with transistors with an input capacitance of 0.10pF, what element is needed to connect their inputs together, and what is its value to match to 50-Ohm characteristic system impedance?

13. What three advantages come with using cascode FETs?

14. Which circuit technique trades less current for move voltage?

15. Which circuit technique reduces the number of different supply voltages required by FET MMICs?

16. A 4x100 FET can be biased at 50% Idss with Vgs=-0.5V. If Idss is 400mA per millimeter of gate, what self-bias resistor value is required?

17. What is the main drawback for both the self- and stack-biasing technique?

18. What is the definition of a linear device?

19. What assumption is made for large-signal operation?

20. Power amplifier design is concerned with handling the effects of mainly strong or weak nonlinearities?

21. What are the four main effects of increasing the input power level into an FET with a nonlinear transconductance?

22. What are the three steps in the power amplifier design methodology described in this chapter?

23. Which MMIC transistor technology is best for an efficient 2.5GHz power amplifier, and how is the process capability rated?

24. As FET device size increases, which parameters trade off against each other, and what determines the ideal device size?

25. In a multistage power amplifier, which stage has most influence over the overall efficiency?

26. Which stages should operate linearly, and how far should the output power from them be backed off to ensure this?

27. Why do we need power splitters and not just one large device?

28. Which are the most common power-splitting or power-combining techniques?

29. Complete the power budget in the following figure. The requirement specification is that Pout=34dBm at P1dB. FET A has 11-dB small-signal gain and P1dB of 29dBm. (a) What input power is needed? (b) Will this architecture meet the specification?



30. Are the matching circuits in a power amplifier designed for the frequency response using small- or large-signal simulation?

31. As well as the RF response, what other aspects of the MMIC design must the matching circuits cater to?

32. What parameter is traded off for better efficiency in bias modes other than Class A?

33. For large-signal optimization, which technique is based on the dc I/V characteristics of the FETs?

34. Which large-signal optimization technique is based on real measurements?

35. Which large-signal optimization techniques predict the final output power?

36. Are soft or hard compression characteristics more desirable for a power amplifier?

37. Which MMIC transistor technology allows design of oscillators with the lowest phase noise?

38. Which MMIC transistor technology allows design of oscillators with the lowest cost?

39. For a general oscillator, what are the three sections that the circuit can be broken down into?

40. For high-Q or low-phase-noise applications, which part of the oscillator could be off chip?

41. Which component can produce negative resistance?

42. What is the main function of mixers?

43. What do LO, RF, and IF stand for?

44. In addition to the harmonics of the LO and RF frequencies, what other frequency components are present in a general mixer?

45. Which devices are used in an active mixer, and which are used in a passive mixer?

46. Of active or passive mixers, which can be reciprocal, which can have conversion gain, and which are easier to bias?

47. What are the advantages of using a balanced mixer?

48. How many nonlinear devices are used in a doubly balanced mixer?

49. When might you use and antiparallel diode pair?

50. What is the image frequency, and how does its suppression affect the mixer noise figure?

51. What are the two modes of operation for FET mixers?

52. Which of the two FET mixer modes gives the best linearity, and which gives conversion gain?

53. Which of the following is a balanced mixer, a dual-gate FET mixer, or a Gilbert-cell mixer?

54. What is the function of the balun?

55. What bias voltages are applied to an FET acting as a switch?

56. What are the advantages of FET swtiches over PIN diode switches?

57. What are four main issues that start to become more important for mm-wave MMICs?

58. What limits the number of turns of track in a spiral inductor that a designer can use in an MMIC design?

59. What is the effective dielectric constant for an 80um wide line on a 100um substrate at 30GHz?

60. What is the wavelength in 50-Ohm GaAs microstrip at 40GHz?

61. What is the impedance of a via through a 100um substrate at 60GHz, and what can be used instead of a through-substrate via for a ground at 60GHz?

62. What are the two common transmission-line types used at mm-wave frequencies?

63. Which mm-wave transmission-line type has the lowest cost?

64. Which mm-wave transmission-line type is better for power amplifiers, and what are the two main reasons why?

65. Which mm-wave transmission-line type has shunt parasitics, and what is the advantage of thi?

66. Which mm-wave transmission-line type is easily characterized by s-parameter blocks and does not require 3D simulation?

67. Which mm-wave transmission-line type is limited in the range of transmission-line characteristic impedance?

68. What fraction of a wavelength along a lossless transmission-line transforms an open circuit to a short circuit?

69. What are the two main reasons for needing 3D simulations at mm-wave frequencies?

Questions:

1. HBT and bipolar transistors.

2. HEMT.

3. InP HEMT and InP HBT.

4. 10dB.

5. Between -10 and -15dB.

6. At the side of the chip, and their size and pitch.

7. The first and its gain.

8. Series inductive feedback.

9. Stability.

10. Placing stabilizing resistors at the ground end of shunt inductive matching elements.

11. From the input to the input-artificial-transmission-line terminating resistor, and from the output-artificial-transmission-line terminating resistor to the output.

12. An inductor of value 250pH.

13. Higher gain, wider bandwidth and higher output impedance.

14. Stack bias.

15. Self-bias.

16. 6.25-Ohm.

17. The performance will vary with the resistor process variations.

18. A linear device is defined as a device with properties (e.g., resistance, transconductance) that are independent of the voltage or current applied to the device.

19. Large-signal operation assumes that the voltage and current signals applied to the circuit are large enough that the devices may become nonlinear.

20. Weak nonlinearities.

21. Less power in the fundamental as the input power increases corresponds to compression of the device gain. The dc component produced can change the dc bias point of the device. Increasing power at the harmonics frequencies distorts the output waveform. The generation of power at other frequencies can have serious consequences on the system perfomance.

22. Architecture design, small-signal design, and large-signal optimization.

23. GaAs HBT, and power processes are rated in terms of watts per millimeter of gate width or emitter-finger length.

24. As FET device size increases, output power increases, and the gain decreases. The optimum device is one that gives the maximum amount of power and still has useable signal gain over the specified frequency range. The rule of thumb is to select the larger device that still exhibits a Gmax of 10dB at the high end of the specified frequency range.

25. The output (last) stage.

26. All the stages before the last stage should be backed off by 3db below their P1dB output power level.

27. Large power devices will have too little gain or get too hog if just one large device is used.

28. Wilkinson splitters and bus-bar and parallel-matching networks.

29. The completed power budget is shown below. The input power needs to be +16.5dBm. The output power from FET A in the first stage is +27dBm, but the P1dB of FET A is +29dBm; therefore, linear behavior (3dB back-off from P1dB) is limited up to +26dBm, so this does not quite meet the required specification.



30. Small-signal simulation.

31. dc bias paths.

32. Gain.

33. The Cripps load line technique.

34. The load-pull technique.

35. Large-signal simulation using nonlinear models.

36. Hard.

37. HBT.

38. MESFET.

39. Resistor, negative-resistance circuit, and load.

40. The resonator.

41. The active devices (transistors).

42. Frequency conversion (to higher or lower frequencies) while retaining any modulation data on the signal.

43. Local oscillator, radio frequency, and intermediate frequency.

44. All the sum and difference products.

45. Biased transistors in an active mixer, and unbiased transistors and diodes in a passive mixer.

46. Passive mixers are reciprocal, active mixers can have conversion gain, and diode mixers are easiest to bias.

47. Additional suppression of the even harmonics of the LO frequency.

48. Four.

49. In a subharmonic mixer.

50. The image frequency is the unwanted sideband in a single-sideband mixer, which if the mixer is designed to reject it, reduces the mixer noise frequency by 3dB.

51. Active and passive modes.

52. Passive FET mixers give the best linearity, and active FET mixers can give gain.

53. A Gilbert-cell mixer is a balanced mixer.

54. A balun is used to convert a balanced to an unbalanced signal.

55. 0V and -5V.

56. Near-zero power consumption and easy control bias via the gate contact.

57. Higher frequency, shorter wavelength, parasitics have a larger effect, and 3D simulation may be required.

58. The self-resonant frequency.

59. 8.4 to 8.5.

60. 0.25mm.

61. 10-Ohm and a radial stub.

62. Microstrip and CPW.

63. CPW.

64. Microstrip because the thermal dissipation is better through a thinner substrate and the high dc bias currents are easy to return to ground using vias.

65. CPW has fewer shunt parasitics, so the active devices have more gain.

66. Microstrip.

67. Microstrip.

68. One-quarter of the wavelength.

69. The short wavelength means that radiations is more likely, and coupling between close components is also possible.