Ask Harlan: October 29, 2003
YOUR RF & MICROWAVE TECH Q&A RESOURCE
You may know Harlan Howe from his twelve years as publisher and editor of Microwave Journal ®, or from his 34 years as a Microwave design engineer and engineering manager, or from his service as an IEEE fellow and past president of MTT-S.
Now, although semi-retired, Harlan is available to answer your questions about RF and Microwave engineering. If he doesn't have the answer, he will find an industry expert who does.
FROM: Reza Esfandiari, Dynalinear Technologies, Inc.
My question is about phase noise in CDMA power amlpifiers. Does reducing the BW of the power amplifier help to reduce the phase noise? Is there a theoretical formula between the minimum noise figure and the phase noise? Also please give some further advise of how to minimize the phase noise in CDMA PA.
Phase noise results from noise processes in the semi-conductors and is particularly sensitive to 1/f noise. Narrowing the bandwidth will improve phase noise but the choice of semiconductors is more important. I don't know of any specific equation that relates noise figure to phase noise since the effect is a function of the type of noise, however, there is a good discussion of phase noise in: Nonlinear Microwave and RF Circuits, Stephen Maas, Artech House, 2003, ISBN# 1-58053-484-8, pp.566-572. It specifically addresses oscillators, however, the theory is applicable to amplifiers. I would also recommend: RF Power Amplifiers for Wireless Communications, Steve Cripps, Artech House, 1999, ISBN# 0-89006-989-1.
FROM: Olivier Guillemant, CSR
I've just received a couple of Orthogonal Mode Transducers (frequency range: 12.75-13.25GHz) that I am trying to optimize, before sending to customers. I've got both horizontal and vertical polarisations available. For example, to improve the return losses for horizontal polarisation, I can insert/move tiny copper rods horizontally in a circular cavity. I would like to see what kind of spacing I should ideally use between those rods (quarter of a wavelength, half a wavelength,...), but I can't find any interesting publication about that kind of OMT.
Do you have some interesting documents about OMT theory that I could refer to?
Orthoganal transducers are a "black art" practiced by a few old time waveguide engineers. The wires that you describe add a reactance to the circuit for matching or mode suppression and are typically placed much closer than a quarter wavelength. They actually act as a vane but are easier to fabricate. My own experience (many, many moons ago) was entirely empirical and I know of no references, however, there are commercially available finite element and EM field solving software programs which can analyze this structure.
FROM: Linda Bui-Le, NAWCWD
I need a notch filter which rejects all signals between 429.65MHz-436.15MHz (peak power level of 49dBm is at 433.1MHz).
Is this product available on martket? If not, would you please point me to some resources on how to build it?
Thank you very much.
Since you have a very specific requirement, it is probably not available off-the-shelf. There are a number of filter manufacturers listed in our on-line buyers guide, who would design it for you.
If you want to build it yourself, I suggest that you look at chapter 12 of Microwave Filters, Impedance-Matching Networks and Coupling Structures by Matthaei, Young and Jones, Artech House, 1980, ISBN#0-89006-099-1.
FROM: Athanasios Tsitouras, Skatatron
I'm a student at the Patras University in Greece, department of EE engineering and it's the 1ST time I'm working on an AGC loop for a tranceiver 5ghz. I use for this 3 components: an LNA, a directional coupler and a power detector, all at 5ghz. I must design the coupler and the power detector. I need some detail-directions on how to design a 15dB Microstrip Backward Coupler for 5ghz and what cautions should I take in order to have improved results. For the power detector I'm thinking to use a 4 diode-bridge and a LPfilter in order to rectifiy the output of my coupler. Do you have any suggestions on the power detector's architecture?
Thanks for your time and best regards,
The normalized even mode impedance (Zoe) for a 15 dB quarter wave coupler is 1.2. If you plug that into any of the commercial design software programs available, you can get the line width and gap for the substrate that you are using. I would not use a 4 diode bridge. A single diode will do the job followed by a low-pass filter. (A simple quarter-wave open circuit stub will do the job). I would put a small resistive pad in front of the diode to improve the match to the coupler, since I don't think you have a sensitivity problem. The pad will also provide the DC Return for the diode.
FROM: Alberto Carbajo, Analog Devices
I want to measure the RF output power from an IC pin. This pin is internally terminated to ground with a 50ohms resistor.
When I connect the Network Analalyzer (50ohms input impedance) to this pin, the signal will see a parallel equivalent resistance of the two resistors, the internal one in my DUT and the input impedance of the analyzer. It means a mismatch and thus, reflections in the line and stuff.
Is the above assumption correct?
Thank you very much in advance.
Yes your assumption is correct. You will have a 2:1 mismatch. If your analyzer will accept an input from a high impedance probe, that would solve the problem.
FROM: Ken Santefort , W. H. Leary Co. Inc.
We are a company that makes inspection systems for the packaging industry. One item we inspect is glue.
I am looking for a solution to detect a 3mm wide film of glue on a carton. I believe a sensor could be developed using microwaves to measure the absorption of water in the glue.
Where can I find someone to develop this sensor?
A microwave interferometer is sensitive enough to do the job. I developed an experimental system of this sort years ago when I was at M/A-COM. Unfortunately, the market did not develop and it was never produced. Since the application is limited, I doubt that anyone will fund the development themselves. However, you could contact them or any of the instrumentation companies in our on-line buyers guide. The chances are that a smaller company would be more interested than the giants.
FROM: Halid Mustacoglu, Syracuse University
I'm comparing the unloaded Q factor of different types of resonators like waveguide resonators, helical resonators, microstrip resonators, CPW resonators...
My question is for Coplanar waveguide and Coplanar Stripline and Hairpin Resonators unloaded Q factors. What is the Q values for these resonators? I found for CPW around 150, for Coplanar stripline 110, for hairpin resonators 150. Are they ok or are they wrong?
And also I have another question. The bandwidths that we can achieve for different types of filters: How much BW we can get with hairpin filters, edge-coupled filters, waveguide filters... And if we can not achieve wider bandwidths with these filters why can not we achieve? What are the limitations for BW?
Thank you very much.
The unloaded Q of any resonator is a function of the construction and materials as well as the type of resonator. Your figures are in the correct range but the actual unloaded Q can vary by a factor of two or more.
The filters that you mention are all coupled line filters. The maximum bandwidth is a function of how tight the resonator coupling can be made in the construction that is being used. Coupled line filters can be made with bandwidths up to an octave. However, if you really need broad bandwidth, stub filters are easier to make and can go as high as 3:1 bandwidths.
FROM: Yat Sen Lie, L & L Design Works
I am preparing a technical paper. What is the procedure to submit this paper to the Microwave Journal for publication?
If you have not written the paper yet, you should submit an abstract and outline to us. We can help shorten the process. If it is already written, simply send it by post (685 Canton St., Norwood, MA 02062 USA) or e-mail to email@example.com.
Please do not embed figures in the text. They should be separate and need not be in final form, since we will re-draw them for publication.
Note that our papers are peer reviewed prior to acceptance.
FROM: NGUYEN HANG NGA, VIETNAM POST AND TELECOMMUNICATION
Dear Mr. Harlan!
I'm interested in the nonlinear effect evaluation, especially, the nonlinear effects of HPA of M-QAM radio relay systems, satellite and mobile system.
Can you help me with some deep information on this field of study, such as the way to determine these effects and studying tends to linearise.
I'm very pleased if you can help me.
I'm very sorry for my bad english.
I'm looking forward to hearing from you.
Nguyen Hang Nga
The best source of information that I know is: Nonlinear Microwave and RF Circuits, Second Edition, Steve Mass, Artech House, 2003, ISBN# 1-58053-484-8.
FROM: Mi Zhou, UESTC (China)
Dear Mr. Harlan:
I'm confused about the behavior of diode employed multiple junction. For example, an ISIS (Integrated Series IMPATT structure) diode is used as multiplier, but is it working in IMPATT mode or simplely as a series or a few PN junction? Some one told me that such structure can amplify harmonic signal through parametric effect in doubler use, but in theory, parametric amplifiers need a 2fo driven, and in doubler, 2fo is the frequency we abstract. So my head has become a clump of mess. Can you help me, Thanks!
The ISIS diode, which was developed by M/A-COM in the '80s, is a monlithic construction of the series connection of varactors previously sold as a "Stack-Pac". It is not an IMPATT, it is a varactor. It can be used in any circuit as a varactor, with the advantage of higher power capability with lower capacitance.
FROM: Madhu Sudan, CENTRE FOR DEVELOPMENT OF TELEMATICS,C-DOT
I'm Madhu from India. I 'm working on WCDMA Low noise amplifier &RSSI. To meet the Dynamic Range requirements of WCDMA Rx signal, RSSI ckt should be able to detect -70dBm RFsignal. For this, an RF Variable Gain amplifier (VGA)(RFMD2377) followed by Anaog's RF detector chip solution (AD8362) was used. It is working fine delivering an output voltage of 0.7v to 3.4v correponding to -70dBm to +10dBm.
When the same ckt is connected in the following arrangement, the output is shrinking to 2.4v to 2.8v.
When RFsignal 0f -70dBm to +10 dBm is applied at the RFVGA input ,DeTECT OUT is 0.7v to 3.4v, which is desirable. When the complete circuit is operated as shown above, the DETECT OUT is only 2.4v to 2.8v (not linear also). pls suggest me the solutions to make it work independent whether I connect LNA or not.
It sounds to me like some element in your chain is saturating. There are a number of books in the Artech House library that may be helpful to you. Go to www.artechhouse.com.
FROM: Clifford Vaughan, Motorola SPS
How does a UWB receiver work?
How do you specify its sensitivity/Noise Figure/Bit-Error-Rate?
There have many articles about creating the UWB signal, its spectrum and short range applications.
I have not senn a detailed explination of the complete Trans/Rec system.
Motorola - Tempe, AZ
UWB is a signal that occupies more than 500 MHz in the 3.1 to 10.6 GHz bands. The Microwave portion of a UWB receiver is no different than any other reasonable bandwidth receiver. The difference comes in the chosen modulation/de-modulation scheme used. Thus the definitions you seek are dependent on the chosen scheme.
I have not seen any textbook on UWB, although Artech House has one in the works for release next year. There was a recent IEEE conference on UWB in San Francisco on July 25, 2003. It is possible that the proceedings from that conference may be helpful.
FROM: Steven Grubor, Cinergy
Mr. Howe, Have you heard of any documented attempts to jam or hack licensed terrestrial microwave? Is digital microwave less susceptible? What are the ways to reduce the risks?
I have not seen any documented examples, however, there are companies that make very sophisticated intercept equipment, which is used by Government security agencies. This equipment is designed to break codes as well. There is no security advantage to digital vs. analog transmission.
FROM: Arthur Greenman, Stabro Laboratories
Is there an accepted definition for sub-harmonics relating to the output of a synthesized signal generator?
I am not aware of any universal definition. Most manufacturers specify the specific frequencies at power levels expressed in dB below the desired output.
FROM: Ho Kenneth, In-tech Electronic Ltd
Currently I need to design a microstrip balun for a DECT transceiver.
I have some questions belows:
1)How to measure the impedance of differential devices using network analyser?
2)What are the considerations when design a balun and how it works ?
Could you kindly suggest some articles and books about the microstrip balun?
Thank you so much.
There are test accessory kits availabe to permit the measurement of a balanced line with a network analyzer. I suggest that you contact the applications engineers at the manufacturer of your network analyzer.
The choice of balun type is a function of bandwidth needed and available construction materials. There is a good discussion of Baluns in: RF Design Guide, Peter Vizmuller, Artech House, 1995, ISBN#0-89006-754-6, pp. 69-76.
FROM: Alberto Giraldez
I would greately appreciate to receive information on extra-stable frequency TR modules in the band 1.2 to 1.4 GHz.
Virtually all TR modules are custom designed for the specific application rather than off-the-shelf standard products. We do not recommend specific suppliers in this column, however, if you read the new product releases and the ads in Microwave Journal , you will find a number of companies capable of producing a custom module for you.
FROM: Roland Teoh, Dynamic Engineers, Inc.
We are looking for a Ku-Band Duplexer < Without filter
Can you please recommend a manufacturer that can help us?
The Ku-Band Duplexer will be use in commercial earth station.
We don't recommend specific companies, however, there are a number of possible sources in the microwave component and tube categories of our On-line Buyers Guide.
FROM: Young Yang, Unique
I am an RF Designer. Some special Terms and detailed testing of RF circuits I do not know for sure. Who can tell me where can I learn them? Thanks.
Young Yang, Unique, Beijing, PR China
Your question covers a lot of territory, so there is no one single answer. There are a number of books on RF circuits in the Artech House library. I suggest that you go to www.artechhouse.com.
FROM: Robert Lacoste, Alciom
I'm wondering how low cost RF field level meters actually works. I mean devices that display field levels in V/m or mW/cm2, over a very wide band (MHz to GHz).
Of course it is easy to build a wideband "field level" receiver, giving the level of a received signal from a dipole antenna expressed in µV or dBm, but the relationship between received level and field strength is frequency dependant, right? Something like "field level=received level-antenna gain-20log(F)+constant "? So how do they get the field level without knowing the frequency? Or do they measure it?
I'm sure there is something I'm missing, but what?
Moreover what are the best antennas for a flat wideband frequency response? Very short dipoles?
Many thanks for your help...
Field level meters use a large number of detectors in different orientations to eliminate directional and polarization effects. The "antennas" are very short with respect to frequency for the range of the instrument, so that frequency is not measured nor is it a factor.
FROM: Binh Ng, Westcan Wireless
How much fresnel zone or high need for 23 ghz compare to 2.4 ghz?
The rquirements are related to the topography of the transmit path as well as the frequency. An excellent reference is: Fresnel Zones in Wireless Links, Zone Plate Lenses and Antennas, H. Hristov, Artech House, 2000, ISBN# 0-89006-849-6.
FROM: John Drelling, Microtel
I am a new analyst in my company, and I got a mission to map the microwave industry: who are the main players in the market, market shares, annual sales, future directions of the industry....
If you could help me Iwould be very grateful.
The major players are all exhibitors at the annual MTT-S exhibition. A list can be found in the May issue of Microwave Journal. Market share is very difficult to determine as are salaries. You might contact Allied Business Systems or Frost & Sullivan. Be aware that they charge for their services.
FROM: Walter Chew, Minimatics, Inc.
We are a jobbing machine shop located in Mountain View CA that has served the local cluster of microwave power tube manufactures with machining services for thirty years. The cluster includes Varian Associates (now CPI), Litton (now L3), Teledyne and Watkins Johnson (now defunct). Our customer list also includes non locals Northrop, Triton and the sorely missed Raytheon. All those names provided as reference to our level of service to the industry for these thirty years.
For planning purposes we never located a marketing resource for providing industry metrics on projected levels of units to be purchased by tube users.
Our question: can you suggest an industry expert that we could contact?
Thank you. W. Chew, President, Minimatics, Inc. 433 Clyde Avenue, Mountain View CA 94043 650-969-5630
I suggest that you contact Allied Business Systems or Frost & Sullivan.