Published March 20, 2007

From: Luis Abril, Alcatel Alenia Space

Dear Harlan,
For an RF application, which SPDT is better: an absorptive one or a reflective one?

Dear Luis,

In general it is always better if everything in a system is matched in all conditions. However, an absorbtive switch frequently costs more and may have more insertion loss. If the reflection does not cause a problem, I would use the simpler design.

From: Satyajit Chakrabarti, SAMEER Kolkata Centre

Dear Harlan,
How can a single antenna be made to operate at three or four frequencies? Is there any bandwidth limitation?

Dear Satyajit,
Some antennas like parabolic reflectors have no bandwidth limitation. Others like dipoles and printed arrays are generally limited, but can be adapted to cover several frequency bands by the addition of tuning structures. A good reference is Antenna Theory, C. Balanis, Wiley, 2005, ISBN#0-471-66782-X.

From: Andreas Koroulis, Intracom

Dear Harlan,
I could use a little bit of help in a problem I am facing. I am on a project involving the design of a filter bank. The filters are already built, so I have to integrate them with the switches (in/out). My first approach was to design a microstrip PIN diode switch, but I came up with a bulky design. So I did a little search and discovered that the approach that many companies follow is a little different. The signal is driven through the input connector to the center of "star like" metal ribons/bars construction, and each of these "bars" is "travelling" to a specific filter housing through milled "channels" in the housing of the filter bank. What I am trying to find out is how this thing operates (principles of operation, is this star of bars a standard part and how it is described, where are the PIN diodes located, etc.)? I have also seen this topology in many microwave control assemblies, power splitters, etc. I would appreciate it if you could give me some help. For reference I am posting this url, where you can find some pictures that will help you to understand what I have previously described: http://www.klmicrowave.com/ecommerce/files/4_channel_sfb.pdf.

Dear Andreas,
The switch that you describe is a fairly common SPMT switch built in channelled, low dielectric microstrip with beam lead PIN diodes for each arm in series at the common junction. There are sometimes additional shunt chip diodes spaced a quarter-wavelength away to increase the isolation. The end of the common arm generally looks like a lollipop, which provides more area for attaching the diodes and also serves to tune the junction.

From: Joppet Karlo Quinones, Teradyne Philippines Ltd.

Dear Harlan,
What is the best way to calibrate an S-parameter measuring instrument?

Dear Joppet,
The best and simplest way is by using the "short - open - through line" technique. Test kits with these calibration standards are available from a number of manufacturers of test equipment.

From: Kwang Seah, Imperial College

Dear Harlan,
I am a student working on power amplifiers and power combining. I would like to sum the power of the outputs of several power amplifiers (high efficient power amplifiers like class E or F, for example) with different power outputs. Can I simply sum the voltages/currents at the output? Or should I sum the powers using power combiners like Wilkinson dividers/combiners? Sincere apologies for asking such a trivial question.

Dear Kwang,
It is not a trivial question. Power amplifiers are generally combined using matched combiners like Wilkinsons in order to avoid interaction between the output circuits. The problem occurs in that the amplifier outputs must be of equal amplitude and phase or part of the power will be lost in the isolation load. If your amplifiers do not meet that requirement, you cannot use matched combiners.

From: Gilbert Brunette, Consultant

Dear Harlan,
I would like to find references that clearly show how one would derive the number of bits needed in an ADC used to demodulate everything from basic modulations like BPSK all the way up to 64 QAM, at various bit rates. I understand the link between S/N and clock jitter but I am still not sure about the number of bits with respect to something like 64 QAM demodulation.

Dear Gilbert,
I think the best reference for your question can be found in Modulation and Coding for Wireless Communications, A. Burr, Prentice-Hall, 2001, ISBN#0-201-39857-5.

From: Afshin Safari, IUT

Dear Harlan,
We can see some RF ICs with many not connected (NC) pins like HMC462LP5 here. Do you know the reason?

Dear Afshin,
I cannot speak for the specific Hittite product that you reference. However, many RF IC manufacturers use standard patterns and packages for different products and not all of them need the same number of pins, thus the extra NC pins on some products.

From: Richard Ollins

Dear Harlan,
Do you know where I can find a table comparing the spectral efficiency of various modulation techniques?

Dear Richard,
I do not know of any simple comparison table. However, there are several references that may be helpful to you, including Modulation and Coding for Wireless Communications, A. Burr, Prentice Hall, 2001, ISBN#0-201-39857-5 and Radio System Design for Telecommunications, R. Freeman, Wiley, 1997, ISBN#0-471-16260-4.

Harlan Howe, Jr. received his BS degree in optics from the University of Rochester in 1957. He has been actively engaged in the microwave industry for 48 years, first as a design engineer and then as an engineering manager. In 1990 he became the publisher/editor of Microwave Journal. He retired as publisher in 2001, but remains the editor. He is a Life Fellow of IEEE, past president of MTT-S and the recipient of an IEEE Third Millennium Medal in 2000 and the MTT-S Distinguished Service Award in 2005.

Do you have a question for Harlan?