Ask Harlan, May 3, 2006

From: Hasan Ajam, Faramoj

I have a question regarding quality factor of varactor diodes at different frequencies. In the data sheet of varactor diodes the quality factor is determined only at one frequency. What is their quality factor at other frequencies?

Dear Hasan,
You have touched on a practice that has bothered me for years. Most manufacturers specify Q at around 50 MHz. I once asked why they did this and was jokingly told that, "nobody works at 50 MHz, so it is fair for everyone." There is no universal direct correlation to Q at other frequencies. It is a function of the resistance and construction of the individual diode. The only valid solution is to make some measurements.

From: David Wang, AFEU

The program I am involved in now needs a delay circuit that can achieve 20 to 40 nanoseconds delay for transient signals (impulse trains) up to several GHz (subnanosecond pulse width). Are there solutions that can meet these needs while remaining compact and small in waveform distortion? The area for the delay circuit in the program is very limited.

Dear David,
I do not know of any circuit that will give you that much delay in a small package with the bandwidth that you need. Cables will do it but they take a lot of space. SAW delay lines have an upper frequency limit of about 1600 MHz and BAW lines are limited to about 800 MHz. There are filter circuits but they have limited bandwidth and take space too. I have heard that there are digital techniques but I am not personally familiar with them.

From: Bharatkumar Mehta, Space Applications Centre

What are the different ways to measure AM to PM transfer co-efficient in TWT amplifiers?

Dear Bharatkumar,
AM-PM conversion in any active circuit can be measured either with a spectrum analyzer or with a vector network analyzer. There are block diagrams and descriptions for several methods in Chapter 2 of Intermodulation Distortion in Microwave and Wireless Circuits, Pedro & Carvalho, Artech House Inc., 2003, ISBN#1-58053-356-6.

From: Robert Badzey, Boston University

I have a question about RF MEMS and their impact on the wireless market. Literature I have read implies that two of the big advantages of RF MEMS are size and power savings. Size/integration is obvious. It is the second point, however, that has me confused. How will the replacement of a passive SAW filter with a MEMS component affect the power consumption? Again, the literature I have read implies that because of the high-Q of MEMS filters, it is possible to reduce the dynamic range of the LNA (and possibly the transmit PA) and therefore its power consumption. Just how much of a savings would that yield? What is the relationship between the LNA's power draw and its dynamic range? I am not in the RF engineering field (I have a degree in physics), but am curious as to what concrete savings such advancements might bring.

Dear Robert,
The reference to power savings for MEMS relates to active components such as switches. Replacing one passive component with another passive component will not provide any savings. With regard to the power draw of LNAs, there is no significant savings as a function of dynamic range. However, it is a significant factor for power amplifiers where an increase is generally at the high end, which does require more power draw. How much is dependent on the design.

From: Mohammad Reza Modarresi, Sharif University of Technology

I want to design and implement a low phase noise microwave oscillator. I would like to know if there is an on-line article that has done a complete literature review on this subject.

Dear Mohammad,
I am not aware of any on-line literature review. I can recommend two books on the subject, however, both of which contain hundreds of references. They are: Design of RF and Microwave Amplifiers and Oscillators, Pieter Abrie, Artech House Inc., 1999, ISBN#0-89006-797-X and RF and Microwave Oscillator Design, Michal Odyniec, Artech House Inc., 2002, ISBN#1-58053-320-5.

From: Justin May, Stanford Linear Accelerator Center

I have been trying to develop a 20 dB coupler, either microstrip or stripline, for beam position monitoring equipment at ~6.5 GHz. I need as directive a coupler as possible (10 to 20 dB), as the coupled signal is part of our calibration scheme. So far I have been trying to do an all FR4 board (budget-constrained). I have checked designs with ADS and calculation by hand. However, none of them have worked. In most cases they were non-directive, in some the directionality was in the other direction. Is this difficulty (with something that should be simple) intrinsic to the material (FR4) being used at C-band? Can I even expect to find a solution without going to a hybrid board?

Dear Justin,
There are two possible problems. FR-4 is not an appropriate material for use at 6.5 GHz. It is lossy and the dielectric constant is not a controlled parameter at that frequency. The difference in cost for a few inches of Teflon-glass is rather small and you will get much better results. The directivity cannot be any better than the return loss of the termination on the fourth port of the coupler. This is probably the principal reason for your directivity problem, assuming as you say, the coupler has been properly designed.

From: Jeffrey Chuan, Indra Sistemas

I am trying to design filters with a suspended stripline substrate. Do you have any idea of where I can find information relating to this topic?

Dear Jeffrey,
While there have been a few articles describing suspended stripline filters, I have never seen any papers or books that provide design information. If your filters are direct-coupled, double registered, the standard equations for impedances in air are reasonably accurate. If they are coupled-line filters, you may have to determine the coupling factors empirically by building and measuring several test fixtures. Once you know the coupling factors for your specific construction, standard design procedures should be applicable.