PowerSmith ©™
PowerSmith by David Tomanek VA3DIW and Martin Kratoska OK1RR is program designed for impedance matching. Use the program for broad range of applications, from discrete circuits to antennas. Generally, it is a modeling tool and simulator program.
The Smith chart employs series impedance values. e.g. values measured with an impedance bridge or vector network analyzer. PowerSmith handles conversion between parallel impedance of Zp and series impedance of Zs. Model of lumped elements and lossy transmission line is included. You may load data file with complex reflection coefficients. ELNEC, EZNEC, PowerSmith as well as microwave network analyzers generate these files. You may stop your design at any time and save your project using the proprietary data format. Features include s-parameters and scattering matrices. The program has a passive frequency sweeper and an active component tuner. Special cursor works as a magnifying glass, providing the means of exact impedance reading. The cursor is adjustable and there is no need for mouse. Of course, the PowerSmith has detailed SWR chart. The graphical interface reveals behavior of a complex load with a complex matching network. You may manually design your matching network, or use the features of AutoMatch. Choose frequency and impedance of your interest. The AutoMatch will search for matching combinations. The final draft is a schematic diagram of suggested matching networks. Added features are advanced LNA and amplifier analysis from the s-parameters, stability, gain circles and noise figure circles. The package is shareware. The software never expires. Please register at the authors.
System:
Try the Caldera OpenDOS, IBM DOS, DrDOS, FreeDOS, PharLap, gentoo Linux with prompt emulation, Ubuntu, and Lynx. More interesting is Andrew Tanenbaum's Minix OS. Try VMware for multiple OS platforms. It can switch over to different OS within seconds. Windoze is not designed for time-sensitive fast mathematical routines. Windoze is an obscure toy. MS will take some time to reinvent UNIX. Stable system is an OPEN SYSTEM with free source code, where everybody can contribute. We have tried to do something for small companies and hobbyist.
The non-profit users are encouraged to use it. The profit based entities are encouraged to support further development. The software was written during winter in Wintepeg MB (-40C outside, -80C windshield factor). During the next winter, the graphical user interface was added.
FAT32 is supported, NTFS is not. The best way to aviod MS system compatibility BS is to run it from USB flash disk formatted with FAT. Have fun.
Download PowerSmith.zip ver1.81b (118k) 30 day evaluation per request
Any use of the software by any government agency is prohibited. Non-profit organizations can use the software after registration. The MOM simulation packages only upon signed NDA. no exceptions.
mail for location from where to download
The HOWTO example:
There were few questions how to use Smith Chart and PowerSmith. Here is a short example. We want to match an antenna. To match an antenna is not a simple task. Measure the antenna impedance at the feedpoint. The antenna works as impedance transformer. If you take generic antenna and beam it to the next barn, you will read different impedance compared with the same antenna beaming up to the sky. That's all OK and right. The next anomaly is the matching sequence. If you compensate the imaginary part of the impedance, the real part will change! Yes, the antenna works as an impedance transformer. Do not get discouraged. The following steps may help.
1/ Measure the impedance. Make a few frequency points and mark the impedance. Estimate the impedance profile in the band. The game is to hit the 50-ohm center of the Smith Chart. You can get there using as many ways as you want. You may try to match the whole band, or make a narrow band match. Every component has loss. Sometimes at the microwave frequencies, a perfect match has extensive loss. Sometimes, the simplest match is the best match.
Point number Frequency MHz measured impedance SWR 1 14.015 12 + j15.45 - 2 14.075 36 + j56.47 - 3 14.150 67 - j37.19 -
2/ Simulate different matching solutions. Save each solution into particular match file. Alternatively, make a note on paper. Check bandwidth with the software. Check few pre-compiled solutions in the software. Check the loss.
3/ Match the real antenna. Compensate the imaginary part of the impedance. You get then only the real part. It is easier to handle. First, compensate the jX.
4/ Measure the real part now. Make a note. It might be different from the real part measured before. Check the feasibility with the software.
5/ Make the final step. Use high pass or low pass match to hit 50 ohms. Consult the "solution" button. Verify the solution in real life conditions.
At this point, you will find out the calculated match fits, but there is an error. A small tweak is required. Where is the problem? The Q of the components and measurement error makes the difference. Then component tolerances. The antenna finally started radiating. Most of the energy is radiated. The impedance is not exactly what you measured the first time. That's all right. Everything fits. You can backtrack the present antenna impedance by using simulation software, loading the components with negative values, and stepping towards the feedpoint.
This step created many discussions about antenna simulation programs. By my opinion, whatever antenna simulation software lies in terms of antenna impedance. No matter what objects and loads are in the vicinity. The free-space modelling makes nice pictures. (The Lulu land). The radiation pattern is obviously close to "right". The gain is very questionable. (translated to plain English: it is completely wrong). Watch single item. The antenna simulator gurus never give away the antenna Gain related to a dipole, and antenna impedance. The pointers are gain numbers like -47dBm etc. Change the simulation frequency in NEC codes to 1296MHz. Suddenly the segments are larger than wavelength. Within few years somebody builds the extra super special wire antenna. He finds out, it was not worth of the effort. The simulation software was wrong.
6/ If you get lost, or the weather conditions are complicated >> Make a RC or RL load on you work desk similar to the antenna. Check the match performance with SWR meter, and 1Watt TX. Check how it works. Sometimes the whole problem is bad connector or water in the waveguide.
Note: There is match and conjugate match (the maximum power transfer).
You will measure a fictive load E.g. Z= 56+j13. You can match it.
The conjugate match will use impedance Z=56-j13 and will take it to 50 Ohms. The last combination provides maximum power transfer. We have not talked about component Q, and the circuit bandwidth. That is for a book to read.
A circuit with pure real impedance and zero jX has infinite bandwidth. However, this circuit never existed, and will never exist. Therefore, use the time rather more productive way. This may answer the question of British exorcist.
Questions and Answers
Q: What is better? To match the antenna at the feedpoint, or match it at the Transmitter output with automatic transmatch?
A: Well, depends what you want. The best solution is to match the antenna at the feedpoint. Then there is 50-Ohms. Then there is minimum loss on the 50-Ohm cable. The TX automatic tuner will compensate for small SWR differencies and the band edges.
Note: When you use automatic tuner by the transmitter, and you run 1kW into mismatched antenna, it is the worst solution. There is additional loss on the cable due to cable impedance mismatch. There is high SWR on the cable, resulting in current nodes and spots with very high voltage. Touch the cable. There are hot spots. The heat means you loose energy. The reflected power overheats the final amplifier. The connectors suffer. Sooner or later, you will burn the transmitter or cook the power supply. Maybe you will cook both. The best SWR I have seen on batch of devices was around 1:1.05. SWR under 1.2 is not bad for 100 Watts shortwave TX. Properly matched microwave antenna can have reflection coefficient s11 of -20dB to -35dB. If you walk close-by, you might see the s11 dips moving up and down. Sometimes, the match is impossible. Depends, what you want..
Q: Connectors & cables
A: HF. For low power, any type of connector works. The PL259 is quite unreliable, and suffers from outdoor use. I don't know who came with that. The N-type is the best for powers up to few kWatts. The N-connector is designed for 50 Ohms, it is weather resistant. The coaxial audio-connectors (the 3.5mm and Cinch) are not bad for QRP and pocket-size devices.
A: VHF to microwaves. Use N-type, BNC, SMA , SMC, and similar ones. They are all 50 Ohms. The gold plated are the best. The SMA is 26GHz, SMB snap-on is 4GHz, SMC with thread is 10GHz, N is 11GHz. Good teflon BNC is 4GHz, it can take 100W@1GHz. There is slight difference in quality from manufacturer to manufacturer. Get ONLY connectors with teflon dielectric. There are some metric-based connectors, worth of nothing. The metric doesn't fit anywhere, there are compatibility issues. Only European nuts can manufacture 60 Ohm connectors. The problem with metric SMA flanges is simple, nothing fits to the imperial units, boards, boxes, PCB boards, and holes.
From time to time, clean the grease on the gold plated thread and teflon center. You might feel the difference during measurements. The silver-plated connectors are totally useless after few years. (oxidated silver). Good clean BNC connetor will take 250Watts on 400MHz without any problem. They are rated up to 4GHz. But it is the indoor-type of connector.
A: Cables. The teflon types are best. Any type of cable that can get soaked with water is worth of nothing. The so-called low loss PVC with helicoil inside, gosh, a marketing joke.
The silicon grease and weather-proof gasket helps. For outdoor perm. mount, just paint it. It will last for years. Cables for HF are not quite critical up to few watts. Watch the cable loss. E.g. the tiny RG174 is not the best choice. Sometimes, it is good to read the spec from cable manufacturer.
Q: What is the best antenna?
A: Any antenna. There are resonant and non-resonant antennas. Any beer can shorter than lambda/4 can be tuned to resonance.
Q: What is the difference?
A: The shorter, the less efficient antenna. Most of the software simulations never deliver realistic gain and front-back ratio.
A short juice-can antenna with perfect matching is not efficient, because it is short/to lambda. Larger antennas with proportions appropriate to lambda, verified designs, need matching.
Q: Vertical or horizontal mount?
A: Lower bands tend to propagate with vertical polarization. The polarization plane probably spins from horizontal to vertical plane. Low height horizontally mounted radiator will need transmatch outside.
Twenty meters band and higher frequency bands like horizontal polarization. In most of the cases, the antenna mount will be dictated by the space you got.
Q: Is the EM antenna better?
A: I don't think so. It might be possible to manage the EM field in different way than we are used to. Presently we know Greek. Any shortwave beam beats any super-duper wire antennas.
---
VA3DIW
Meetings: Starbucks at the corner of California Blvd and Lake Ave.
One block south of Del Mar Blvd, the old Pasadena,CA.
Members of W6KA welcome new members.
Powersmith Smith Chart is not associated with powersmith drilling tools and electrical screw drivers, or powersmith engineering inc., nor steam power generators. Please drill your own hole for yourself.
mirrors
< The EM Routines and Codes of JPL Pasadena (web disposed) >
< ( an exiled US web library finds home in Greece ) >
Circuit Sage - The software library, good stuff or propaganda
RFCafe - notes and corporate advertising.
The figure explains where you can match. The forbidden areas are in black.
Remember when it starts to oscillate : -"when you see the negative resistance".-
PowerSmith Users:
Small business Amateur radio fans Professional R&D institutions University teaching Marine services
Policy change. No more software. Radio switched off.
