Ask Harlan

From: Liu Shumin, MW Center CHINA

I have designed a 1.5 to 1.8 GHz low noise amplifier. When I put a lid on its cavity, the input VSWR changes to 1.3 to 1.5 GHz. Do you know why this happens?

Dear Liu,
The change is occurring because your circuit is radiating from the board or components and the cover is reflecting it back to the circuit. The normal solution, if you cannot prevent the radiation, is to place some lossy absorbing material on the inside of the cover.

From: Steven Chan, Northrop Grumman Space Technology

I am using Artech House's 4th printing of Stripline Circuit Design (1982) and was wondering if equation 4-8 has a typo? I would like to know if there is an errata sheet for this book?

For equation 4-7 there is a term in the numerator w/b; in equation 4-8 this term becomes w/(2*e) where e=2.718. When I typed these equations to compute the even and odd mode impedence, I got a strange answer and only by changing the 2*e term to b did I get an answer that seems more like what I was expecting. For one thing the 2*e term caused Zoo>Zoe, which is strange.

Dear Steven,
Yes, there is a typo in equation 4-8. The first term in the denominator should be w/b. I am sorry, but there is no errata sheet for the book. I apologize for the confusion.

From: Yashwant Wanmode, Centre for Advanced Technology

I fully agree with your answer/suggestion to the question asked by Mark Smith of Airnet Communication. I have also made some loop type directional couplers at S-band frequencies of WR-284 waveguide. I did not find the right literature to fix the loop dimensions and hole size. Finally, I used the "Cut and Try" method. Now my question is do I have to use the loop type directional coupler for forward and reflected waves monitoring/measuring? If one loop is measuring the forward waves what should be the distance of another loop that can measure the reflected wave? Also, how do I define the directivity in case of loop type directional coupler?

Dear Yashwant,
I too have used the "Cut and Try" method when I built loop couplers. However, there is a more definitive paper on the subject. It is called "Criteria for the Design of Loop-type Directional Couplers for the L-Band" by Lombardini, Schartz & Kelly, MTT-S Transactions, October 1956, pp. 234-239. In it, the authors report directivities of 30 dB or greater. As for spacing, I would try to keep them from interfering with each other (perhaps a half wavelength). Directivity is defined as the difference in output for the same signal fed from opposite directions. Hope this helps.

From: Toks

How can I design a microwave horn antenna?

Dear Toks,
There are many types of horn antennas. In general, however, the design is based on the needed gain, beamwidth and bandwidth. There is a good discussion of horn design for several types of horns in Microwave Antenna Theory and Design, S. Silver, Peregrinus, 1986, ISBN#0-86341-017-0.

From: Yaron

What is the mathematical conversion (formula) between EVM to BER?

Dear Yaron,
There is no simple conversion between Bit Error Rate (BER) and Error Vector Magnitude (EVM) since BER is dependent on many factors and is affected by the type of modulation used. There is an extensive discussion of BER and its relationship to various system elements in Radio System Design for Telecommunications, R. Freeman, Wiley, 1997, ISBN#0-471-16260-4.

From: Aizu Dean, Kisarazu

I have a few questions regarding milimeter-wave circuits:
1. What is a radial stub and how do we calculate its impedance?
2. Does the slotlines' impedance only depend on its width and height or are there other calculations that one can make? If there are, what does the formula look like and what are the parameters?

Dear Aizu,
A radial stub is a fan shaped stub, which has a greater effective bandwidth than a single impedance stub. There are equations for its reactance in Chapter 23 of Planar Microwave Engineering, T. Lee, Cambridge University Press, 2004, ISBN#0-521-835267-7.
The impedance of a slotline is dependent on the width, height and dielectric constant of the substrate. Because of the non-TEM nature of the mode in a slotline, the characteristic impedance cannot be defined uniquely. There are several definitions and methods of calculation. These are discussed in detail in Chapter 5 of Microstriplines and Slotlines, Garg, et al., Artech House Inc., 1996, ISBN#0-89006-766-X.

From: Tri Chau, TQT

I am interested to know why a stantion transmitter video signal does not exist in 250 to 350 MHz diapason (a transmit video signal of public safety cameras to police officers 10 km distance, for example)? Has it only been viewed by microwave bands?

Dear Tri,
The 250 to 350 MHZ band is not licensed for that purpose. There are public service (police, fire, etc.) bands on either side of it as well as microwave bands.

From: Jason Kneier, Pratt & Whitney

I am designing a balanced, air-filled stripline fixture to be used for material characterization. It will be a rather large-scale fixture compared to what one might typically associate with stripline (the "b" dimension may be close to 3"). I am seeking advice on whether to use a broadwall or end launch at the coaxial junction. In your book, Stripline Circuit Design, I see you recommend that a broadwall launch not be used over 4 GHz, which should not be a problem for my design. Could you tell me if one of these launch methods has inherently lower return loss, or lower loss over a wider bandwidth, than the other?

Dear Jason,
With a "b" dimension of 3" your stripwidth will be in the order of 4.5" depending on the dielectric. You are going to launch higher order modes at frequencies just over 1 GHz, which will impact your measurements. As for launchers, unless you are using very large connectors, like 1-7/8, you are going to have a major mechanical discontinuity that will be hard to compensate. It will be easier to work with an end-launch rather than a broadwall launch. However, it may be better to characterize your dielectrics with one of the test cavity systems developed at NIST. If you contact them they will give you appropriate references at no charge.