Ask Harlan, February 20, 2007

From: Stephen Morley, Independent Communications Consultant

Dear Harlan,
I am just joining a wireless company and would like to have your view on the impact of millimeter-wave links in the marketplace, having previously experienced FSO links. Do they have more of a chance of gaining some market?

Dear Stephen,

I think that millimeter-wave links have the potential for substantial market growth for short haul applications. They are smaller, have great bandwidth capability, can be highly directional and are therefore easier to secure and cause less interference to other links.

From: Tom Stokowski, NWS Corp.

Dear Harlan,
How do I convert HDMI to an 8VSB signal?

Dear Tom,

HDMI (High Definition Multi-media Interface) and 8VSB (8 Level Vestigal Sideband Modulation) are techniques related to high definition digital television. They are outside the microwave field and my area of expertise. I am sorry that I cannot help you. I'm assuming there are texts on television broadcasting, but I am not aware of them.

From: SM Hasan, Virginia Tech

Dear Harlan,
I am building a PCB board that contains several high frequency diplexers and filters. The cut-off frequency of my diplexers is 1250 and 1550 MHz and the cut-off frequency of the LPF is 1000 MHz. The simulation results are perfect. However, in my PCB the frequency responses of these filters are very bad - the cut-off frequency goes down around 300 MHz than my design and the passband has lot of ripples and losses. I tested these circuits by using some high frequency capacitors, but it did not improve the results. Do you have any idea where the problem is? All of my components are SMT in a 0603 package size.

Dear SM,

Filters at these frequencies are usually made using distributed transmission lines rather than lumped elements. Your problems are due to inadequate Q, internal losses, physical dimensions (particularly in the inductors) and package parasitics. Since you are already using a PCB, you should try printing the filters. By folding the elements you can keep the size reasonable.

From: Beenamole Sreedharan, Research Organization

Dear Harlan,
What is the power handling capacity (average and peak) of a stripline and a microstrip line?

Dear Beenamole,

Peak power is related to voltage breakdown, which is a function of dimensions and dielectric strength of the substrate material. Average power is related to line losses, heat sinking, ambient temperature and the allowable rise in temperature. There is a good discussion of peak power breakdown in Microwave Engineering - Passive Circuits, P. Rizzi, Prentice Hall, 1988, ISBN#0-13-586702-9. Average power in stripline (which can be extrapolated to microstrip) is covered in Stripline Circuit Design, H. Howe, Artech House Inc., 1974, ISBN#0-89006-020-7.

From: Thomas N.C. Wang

Dear Harlan,
For a "MIMO antenna" to be used for a sectorial BTS application, will it be workable for a TDD system (i.e., the downlink and uplink have the same carrier frequencies)?

Dear Thomas,

If the Multiple Input Multiple Output (MIMO) antenna has sufficient bandwidth to support the BTS application, it should be OK if both signals are at the same carrier frequency. There is a good discussion of MIMO techniques in Wireless Communications over MIMO Channels, V. Kuhn, Wiley, 2006, ISBN#0-470-02716-9.

From: Tom Wang

Dear Harlan,
In the famous Shannov's channel capacity formula: C=B log (1+S/N), if the channel noise "n"=2, what is the limiting value of the equation? Can it be the same as Nyquist capacity formulae C=2B for the noiseless channel? Also, if S/N=1 then C=B so if B=20 MHz then C=20 mbps of information then another half of C=B... i.e., 20 mbps should be the noise brts!

Dear Tom,

Your question is a communications systems question that is outside my area of expertise, so I cannot answer you directly. However, I can recommend several texts that may be helpful:
Digital Techniques for Wideband Receivers, J. Tsui, Artech House Inc., 2001, ISBN#1-58053-299-3; Advanced Techniques for Digital Receivers, P. Page, Artech House Inc., 2000, ISBN#1-58053-053-2; and Modulation and Coding for Wireless Communications, A. Burr, Prentice Hall, 2001, ISBN#0-201-39857-5.

From: Boon Wei Kang, Agilent Technologies

Dear Harlan,
How do we determine the R_bias for a particular low noise amplifier? Let us say it is a voltage-biased device, and operates from a +3 V power supply, with current drain of 84 mA. Does that mean the R_bias should be 3/0.084?

Dear Boon Wei Kang,
Yes. Ohm's Law still works.

From: Madhusudhan, PhD Student

Dear Harlan,
I would like to introduce myself as a PhD student in microwave and MEMS doing research in India. I have just started my research in modeling of MEMS devices and I have encountered a problem. I am very much confused by different things like field solvers and SPICE models. Could you provide any clarity between the two and also suggest field solvers or a low cost field solver that I can work with for my degree?

Dear Madhusudhan,
SPICE programs are designed to solve problems in one dimension only, while most field solvers work in two or three dimensions. Most field solving software is quite expensive; however, Sonnet Software does offer some free software at www.sonnetsoftware.com/lite.

From: Fahd Ahmed

Dear Harlan,
I am dealing with an emp testing procedure. I have calculated the shielding effectiveness using the formula: A=kt(fug)0.5. This shows that k is constant for thickness t in inches or mils for sheet and the square root of permeability, conductivity and frequency for this is sheet, but for a rebar structure having center to center distance is 6 inches length and width and bar diameter is 0.25 inches. I do not have the appropriate formula, but the calculation can be carried out using graphs by comparing spacing/wavelength x axis with diameter of bar/spacing y axis. Please help me in giving the formula for rebar calculation having e field of 100 Kv/m outside the rebar structure and steel having u=1000 at 10 kHz and g=0.17.

Dear Fahd,
I do not have any idea where to start in trying to calculate the effect of re-bars in a dielectric. I am sure that it cannot be done with a simple formula. In order to get any meaningful answer, I think you will have to model it and use a field solver software package to analyze it. There are several major software companies that have programs that can address this.