A sample text widget

Etiam pulvinar consectetur dolor sed malesuada. Ut convallis euismod dolor nec pretium. Nunc ut tristique massa.

Nam sodales mi vitae dolor ullamcorper et vulputate enim accumsan. Morbi orci magna, tincidunt vitae molestie nec, molestie at mi. Nulla nulla lorem, suscipit in posuere in, interdum non magna.





Antenna Tunner





Manpack Homebrew



MFJ259 ช็อต

IMG_25570913_081937 IMG_25570913_082407 IMG_25570913_082417 IMG_25570913_082435 IMG_25570913_082447 IMG_25570913_082456 IMG_25570913_082535 IMG_25570913_082559 IMG_25570913_082602 IMG_25570913_082630 IMG_25570913_082705 IMG_25570913_082924

Coaxial Resonators

Cavity filters, or more correctly coaxial resonators, are the cornerstone of good repeater performance. However, there appears to be some sort of black magic in the way they work or how they achieve their performance.

A coaxial resonator is much like a quarter wave stub but open-ended. It can also be likened to a quarter wave antenna enclosed within a conducting chamber and a resonant L/C tuned circuit, encompassing many of the features of each.

In most coaxial resonators the “antenna” is excited by a magnetic coupling loop at the resonant frequency and energy is removed by another loop placed next to the base of the centre resonator. The centre resonator is surrounded by air and there is in fact an optimum ratio of inner case diameter to outer diameter of the resonator of 3.25, which works out to give a characteristic impedance of around 70Ω

Tek 1
Tek 1 – Construction drawing cavity.

Factors affecting coaxial resonator performance

1. Diameter of centre resonators

This is chosen to achieve an inner resonator to outer casing diameter of 3.25, arriving at the optimum 70Ω impedance. This gives the highest available Q.

2. Length of resonator

This determines the operating frequency of the formed tuned circuit. Length is inversely proportional to frequency, much like any antenna.

3. Coupling loops

These are the means by which energy is input or extracted from the tuned circuit. Loop size determines the amount of coupling. Two major factors have a pronounced effect on coupling: (a) the length of the coupling loop wire adjacentto the centre resonator. The longer this is the more of the energy field around the resonator is intercepted hence a greater influence this has on the coaxial resonator’s performance. The longer this wire or loop is next to the centre resonator the higher impedence the input loop becomes; and (b) the spacing of the coupling loop to the centre resonator – ie, the closer the above mentioned loop, the more stronger field it intercepts around the centre resonator, the greater influence it has on tuned circuit performance. The reverse is also true with respect to loop-to-resonator spacing.

It is the manipulation of these two factors that plays a major part in obtaining desired performance of coaxial resonators and generates the biggest headache and level of frustration. For it is with these factors that filter loss, bandwidth and notch depth are juggled.

4. Materials of construction

Most commercial coaxial resonators are constructed of copper but there is an increasing use of aluminium as a casing material. The actual resonator itself, and quite often the inside of the casing, will be silver-plated to increase electrical conductivity, reducing “skin loss” and hence assisting in the maintenance of high Q. The coupling loops are of silver-plated copper and the centre resonator may even be gold-plated at the far end, in the region of high RF impedance.

The centre resonator is usually silver-soldered to the top-cap and the far end of the casing closed off with a copper endcap. The top-cap and end-cap are usually secured to the casing by machine screws or even soldered once all adjustments have been made.

5. Frequency of operation

This is governed by the electrical length of the centre resonator and is usually very close to a quarter wavelength at the frequency of interest. This central portion has within it a screw mechanism by which a telescoping metal tube at the free end of the coaxial resonator is able to effect its mechanical and electrical length, and hence its operating frequency. Most high electrically conductive metals have large coefficients of thermal expansion hence it is possible for the length of the centre resonator to change with extremes of temperature and more importantly change its resonant frequency. To combat this effect commercial designs incorporate a threaded rod of invar, an alloy with an extremely low expansion coefficient which holds the length change to a minimum over extremes of temperature.

Modifications and performance

The set of coaxial resonators used in this article were ex-government (now ofcourse corporatized) made by TCA of Australia, 100 MHz versions type 1564/101/1NZ.

They had been cut down in length externally and internally to bring the operating frequency up to 146 MHz band. The outer casing had a finished length of 610 mm, the centre resonator length ended at 510 mm at 144.350 MHz, but allowed some adjustment length for other frequencies. The resonator itself and its screw thread mechanism were somewhat difficult mechanically to modify. The resonator was removed from the top-cap to achieve this.

The original coupling loops of 100 MHz coaxial resonator are designed for 50 ft in and out, hence to use them for 146 MHz scale their dimensions by 100/146 in overall length parallel to the resonator. The coupling loops were re-connected to the BNC sockets and top-cap. The spacing between the pick-up loop and centre resonator controls the degree of coupling energy into and out of the tuned circuit at its resonant frequency.

Experimental section

Tek 2
Tek 2 – Set up for bandpass/inductive short/capacitive short/notch.

With the aid of a spectrum analyser in storage mode and a wide range signal generator, with a constant output over the swept frequency, the characteristics of coaxial resonators on the 2 m band can be examined. The frequencies of interest were the linear repeater input and output frequencies for which these resonators were modified. In the test set up the 50 Ω generator feeds the coaxial resonator input and output feeding the spectrum analyser.

1. Bandpass performance

Fig 1
Fig 1

Sweeping the signal over 10 MHz produces the response of Figure 1. The 3 dB points are about 200 kHz and insertion loss under 0.5 dB. When you look at the response of this tuned circuit you see the attenuation is very poor 600 kHz away from centre frequency. Clearly this set up would be very inadequate for repeater filtering purposes to keep the transmitter out of the receiver.

2. The anti-resonant notch

Fig 2
Fig 2

There is a simple modification to this bandpass configuration that will produce the required frequency response. A notch at the unwanted frequency of the transmitter to keep it out of the receiver is obtained by placing a small value capacitor (2-20 pF variable trimmer) across the input and output loops (Figure 2). This produces an amazing effect. Figure 3 shows a notch in the swept response below the passband peak. The notch depth depends on the Q of the tuned circuit. Compare this to the simple band-pass response. On first examination the capacitor looks like a short-circuit across the loops, but the overall summation of phase shift across the capacitor and phase shift across the tuned circuit, both frequency dependant, produces this curious but very useful effect. The amount of capacitance controls the peak-to-notch frequency difference; more capacitance brings the notch and peak closer together while less spreads them further apart. How close one can get them depends entirely on the overall Q of the system. Circuit Q depends directly on the amount of coupling hence, in an overcoupled resonator with its inherent very low insertion loss, it will be very difficult to achieve a close peak-to-notch with good rejection notch depth.

Fig 3
Fig 3

Fig 4
Fig 4

The same phenomena occurs with a “shorting conductor” or a “shorting coax” (Figures 4 and 5). However, the passband peak and rejection notch are reversed in frequency position (Figure 6). The peak to notch separation is once again controlled by the amount of inductance; the more inductance the wider the separation between notch and peak, while less inductance allows them closer together. Once again the depth of the rejection notch and separation are controlled by the circuit Q. This effect can be obtained by using a piece of 50Ω coax between the input and output. Its length governs the peak to notch separation in the same way as does the shorting inductor (Figure 5).

Fig 5
Fig 5

Fig 6
Fig 6

3. Suckout Notch

Tek 3
Tek 3 – Notch configuration.

This is the opposite of the bandpass response and the coaxial resonator is placed in the feedline as is shown. At the frequency of resonance the tuned circuit begins to accept energy from the coax feed reducing the amount of power reaching the spectrum analyser, hence the dip in the swept response. However, as you can see, the overall depth of the notch is only 18 dB. This depth is governed by the amount of coupling; the greater the coupling the deeper and broader the notch becomes, the less the coupling the shallower and the narrower it becomes (Figure 7).

Fig 7
Fig 7

Coaxial resonators designed with the same input and output impedance can be coupled together to perform various bandpass configurations. Placing them in series will add losses and add notch depth if they are on identical frequencies.

By placing several coaxial resonators in series with passbands and rejection notches in the correct places one can get enough isolation between a transmitter and receiver to run them both off the one antenna. This is called duplexing and the group of connected coaxial resonators called a duplexer. Transmitter to receiver isolation figures are measured in decibels and one usually requires at least 80 dB to achieve duplexing.

In a repeater set up (eg, transmitting on 144.350 MHz and a receiver on 144.950 MHz) the connection of coaxial resonators would be as follows: for receive coaxial resonators are tuned to pass 144.950 MHz and have “capacitive shorts” between input and output connectors to place a rejection notch at the transmitter frequency of 144.350 MHz. This keeps the transmitter from overloading the receiver and requires several resonators connected in series to achieve this.

On the transmit side the coaxial resonators are required to pass 144.350 MHz and reject 144.950 MHz. It may seem silly rejecting the repeater input frequency from the transmitter but a transistor or even (dare I say) valve transmitter has a broad noise pedestal centred symetrically around the carrier and, at 600 kHz awayfrom the carrier, the noise level generated by the transmitter may still be in the millivolt level. The rejection notch at the receive frequency attempts to keep noise off the input frequency and let the receiver respond to the wanted microvolt signals that would normally be buried in the transmitter noise.

A novel feature of coaxial resonators is the phenomenon of harmonic resonance. A resonator tuned for 144.350 MHz functions well at the 3rd harmonic at 433.050 MHz as it is a three-quarter wavelength structure. However, it is over-coupled at 70 cm using 2 m coupling loops. At 288.7 MHz, or 2nd harmonic, the resonator exhibits a shallow notch a few decibels deep due to the resonant half wave structure.

The same occurs for the inductive and capacitive “loaded” coaxial resonators. The message is that although the circuits are very high Q they will not eliminate harmonics and you still need a low pass filter on the transmitter. Running these three-quarter wave resonators at 70 cm yields another problem – thermal drift. Because the resonators are three times as long thermal expansion will be three times as great, hence they are going todrift much more in frequency with temperature changes.

Well, that’s a very simplistic look at the coaxial resonator and its usefulness. There are other variations of it, in the form of loaded versions and helical versions, but they all are very closely related in the resonant family. For greater understanding of them you should have a look at the references, especially the FM and Repeaters by ARRL. If I can be of any assistance to anyone contemplating building a set of coaxial resonators I would be more than happy to help in any way I can. The first time round they can be a great source of frustration and loss of hair by manual means.

Tek 4
Tek 4 – Repeater duplexer configuration.


  1. RSGB, VHF UHF Manual, 4th Edition
  2. “A Lo Cost PC Board Duplexer”, QST, April 1979.
  3. ARRL, FM and Repeater Handbook
  4. K. Wiener DJ9HO, The UHF Compendium, Parts 1 and 2.
  5. Bitter Experience

The photographs don’t show the full depth of the notches in the responses due to the wide sweep width and phosphor blur on the screen. On narrow sweeps the depths of the notches are apparent but the overall shape of the tuned circuit response is not seen.

ZL1BTB, Mike Penfold


ทำ Filer จากสาย Heliax

This filter is designed to remove interference caused by pager transmitters operating between 148MHz and 150MHz.

They are in effect cavity resonators constructed from LCF158 coaxial cable.

Specifications for this cable can be found here. This is some serious coaxial cable. It has a 50mm outer diameter and very low losses (0.82db/100m @ 150MHz). Both the outer and inner conductors are mode from solid corrugated copper.

The first three images show a test model. The bottom is shorted by soldering either a piece of PCB or 16 copper wires between the inner and outer conductors. The section of LCF158 is 41cm long. The bolt in the top is 60mm long (threaded section, 65mm total including the head). The actual size is not critical. There is a nut soldered to a piece of PCB that sits across the top of the LCF158 supported by two smaller pieces of PCB soldered to the outer copper shield. Copper is removed from the PCB so there is no connection between the bolt and the LCF158 shield.

Click on the pictures to view hi-res images.

T_IMGP2625 T_IMGP2623 T_IMGP2624

The following images show the frequency response of the test model over various frequency ranges. The vertical scale is 10db per division.

Image 1    Image 2    Image 3    Image 4 

Image 1 shows the response between 145MHz and 150MHz (markers at 1MHz intervals). The notch depth is almost 30db. The attenuation starts at 147MHz, reaching 10db at 148MHz with the maximum attenuation occuring at about 148.25MHz.

Image 2 shows the response between 125MHz and 175MHz. Aside from the notch, there is almost zero insertion loss.

Image 3 shows the response between 400MHz and 500MHz (markers at 10MHz intervals). The notch here is at the 3rd harmonic frequency. The depth of this notch is about 46db. Note that the attenuation starts at about 430MHz and reaches about 12db at 440MHz.

Image 4 shows the response between 100MHz and 600MHz. Note there is about 3-4db of ripple in the pass band.

The next set of images show the final version which is made up of three filters in series. In my case, I’m using the filter on a Yaesu FT-8900 with a single antenna. This meant that the notch in the 70cm band should not attenuate frequencies up to 440MHz. To do this, I reduced the length of the LCF158 and tuned it using a 5pf trimmer capacitor soldered between the inner and outer of the LCF158 at the top. This has the effect of moving the 70cm notch higher in frequency (the attenuation starts at about 440MHz). Adjustment is also simpler with this design.

For this version, the length of the LCF158 inner conductor is 39cm. The two sides of the out conductor were left 20mm longer to provide a support for the PCB.

Adjusting either the trimmer or the bolt changes the notch frequency. The length of the bolt also affects the depth and width of the notch. The bolt should be adjusted to set the size of the notch and the trimmer used to move it on to the desired frequency.

The cables connecting the filter sections are 1/4 wavelength (500mm x Velocity Factor). This reduces the influence of one filter on the next when making adjustments.



ซ่อม MFJ-269

Page 10 of 61« First...89101112...203040...Last »