Ask Harlan

From: Adeline Rebeiro, Twin Antennas

Could you provide details of the various divider networks used in microstrip patch antennas? I am trying to design a microstrip patch array for 2.4 GHz and 19 dBi gain. From a single patch, I get a gain of around 9 dBi. Do I need to go to an array configuration? What method do you suggest should be used to combine the patch elements?

Dear Adeline,
Most patch arrays are fed by corporate feeds consisting of simple reactive power dividers. These dividers are typically three-port tees with appropriate transformers to match the junction. There is a discussion of corporate feeds in Chapter 12 of Microstrip Antenna Design Handbook, Garg, et al., Artech House Inc., 2001, ISBN#0-89006-513-6.

From: Tapas Bhuiya, SAMEER

How is insertion loss related to the return loss of a filter? Is it true that good return loss of an antenna is always efficient (efficiency is more, for example)?

Dear Tapas,
Return loss is the reflected power, expressed in dB, from any discontinuity on a transmission line. Insertion loss is the total loss of the circuit, again expressed in dB, and includes resistive losses as well as the reflected loss. The efficiency of an antenna includes reflective losses, resistive losses and power lost to unwanted radiation such as sidelobes.

From: Subhash Janakiraman, Icon Systems

I have a question on a RF board. The system works at 902-928 MHz ISM band. I am connecting two boards using a header. The PCB dielectric material is FR4 and they are six-layer boards. The distance between the two boards is 0.5 inches. I am concerned about the antenna traces getting connected through this header. Is this arrangement fine or am I supposed to have some effects due to the connector metal acting as a radiating source? Please explain.

Dear Subhash,
Although it is probably not matched to 50 Ohms, the header will act as a transmission line provided the outer shell is grounded. If it is not grounded, then the line will couple to it and it will be a source of spurious radiation. At your frequency, 0.5 inches is a small fraction of a wavelength, but it will still represent a mismatch.

From: Syed Shafqat Hussain, University of Leeds

I am a novice trying to research loss compensation in microwave filters, but could not find any relevant material on the Internet or in books. Could you please help me understand this topic and also guide me to where can I find reading material relevant to the topic?

Dear Syed,
The long established text on microwave filters is Microwave Filters, Impedance-matching Networks and Coupling Structures, Matthaei, Young & Jones, Artech House Inc., 1980, ISBN#0-89006-999-1. A newer book that may be helpful is Analog and Digital Filter Design, Winder, Newnes/Elsevier, 2002, ISBN#0-7506-7547-0.

From: Murali Ellanti, LNT Electronics

What are the CW and pulse power handling capabilities of improved SMA connectors at 18 GHz? I want to handle 100 W CW and 3 kW pulse. Can I use SMA connectors?

Dear Murali,
Standard SMAs are rated at 100 W CW up to 12 GHz at 20 degrees C. There are “Super SMAs” available that are rated at 150 W CW at 18 GHz and 125 degrees C. I have not seen any published data on peak power ratings. That capability is related to many conditions, including VSWR, humidity, dirt and the back interface. My own “rule of thumb” for high power is to allow a minimum of a 2:1 safety factor. I would not risk using an SMA.

From: Ashim Jolly

Could you let me know where I can acquire the circuits for mobile phones that I can develop in my lab? I am an engineering student and require some designs that I can simulate in software like PSPice, etc. Could you also recommend some other ways of learning how to build your own mobile phone?

Dear Ashim,
A reference with some circuit examples is RF/Microwave Circuit Design for Wireless Applications, Rohde & Newkirk, Wiley, 2000, ISBN#0-471-29818-2. However, since all mobile phones are built with complex, multifunction IC chips, it is unlikely that you can build your own phone.

From: Moazzam Loynmoon, SIGNET

In a loop antenna a 1 Amp/Volt current transformer is required to terminate both ends of the loop. Output is single-ended. Frequency required is 19 kHz to 30 MHz minimum. Could you tell me how to fabricate it? I know that a toroid is required for the same.

Dear Moazzam,
For your frequency and bandwidth, a ferrite is definitely needed. Specific design equations for such an antenna are given in Chapter 5 of Modern Antenna Design, Milligan, Wiley, 2005, ISBN#10- 0-471-45776-0.

From: Arshed

Our customer would like to establish, on turn-key basis, a general-purpose RF/microwave workshop/lab to test/repair their radar assemblies.
The lab should include the following:
Lab environmental pre-requisites, such as temp, humidity, electromagnetic/electrostatic shielding (screen room), earthing, TEM cell, etc.
Soldering/de soldering technique/eqpt
EMI/EMC measurements eqpt
Test equipment and software required to perform test/repair of assemblies
Please note that this will be a turn-key project, to include all peripherals infrastructure eqpt and also training of working staff. The customer does not have any list tools/test equipment to be procured to establish such facility. Could you provide a comprehensive proposal on turn-key basis?

Dear Arshed,
I do not believe that any company provides turn-key labs of the complexity that you describe. Despite the fact that any test consultant capable of providing you with a design plan and implementing it is likely to charge a significant amount of money, that cost will be small compared to the cost of the equipment needed. Your customer may be better served by using existing commercial testing labs and simply building a repair shop with minimal test equipment.

From: John Bradin, Assa Abloy ITG

I am researching compact antennas for a 2.45 GHz WPAN. The radiation efficiency is (radiation resistance)/(radiation resistance + loss resistance). I believe that the correct calculation of the loss resistance for the case where the conductor thickness is > 5X (skin depth) is given by (loss resistance) = (conductivity/(skin depth))*(# of squares in the antenna conductor). There are other equivalent expressions that reduce to this. Is this correct? Given the amount of work in very small 2.45 GHz wireless devices, I am surprised not to see this concisely discussed. I have found a few cases where the skin depth issue is ignored. What am I missing?

Dear John,
I do not think that you are missing anything. It is just not mentioned. While some of the antenna references refer to ohmic losses, they do not say how to calculate them. Skin depth should not be ignored. There is a discussion of skin depth and its effect on surface resistivity in Chapter 2 of Microwave Engineering: Passive Circuits, P. Rizzi, Prentice-Hall, 1988, ISBN#0-13-586702-9. Based on his discussion, I think your calculation is correct. There is also a discussion of compact patches in Modern Antenna Design, Milligan, Wiley, 2005, ISBN#0-471-45776-0.