|
A
single helix notch filter for the 2m-band. |
This filter is designed to suppress signals between 155 and 165 MHz, but to pass signals in the 2mtrs amateur band nearly without attenuation. They can prevent noises, caused by IMD in your receiver, due to strong out-band signals.
The filters are made from cheap, easy to get materials. For tuning, measuring-instruments and experience using them is needed. So this is NOT a beginners project.
What is a helix
resonator and what can it do. A helix resonator circuit comprises of a coil inside a screening. It has no discrete tuning condenser. The parasitic capacitance between the coil windings and the screening is used like this. This "condenser" has a very small value, and low loss, because the dielectric is made of air. Consequently, the circuit-Q is large. The mechanical and electrical properties of a helix circuits are between "conventional" parallel-tuned-circuits and coaxial resonators. Helix circuits mostly have larger dimensions than parallel-tuned circuits using a discrete tuning condenser. In modern HAM-equipment helix circuits are not found often, presumably because they are bigger, and prevent miniaturization. In professional communication equipment helix band pass filters were used more often, in order to get good selectivity. Compared with 1/4 wavelength long coaxial resonators, helix circuits are smaller, and have therefore lower Q. Duplex filters of (amateur-) repeaters with small shift (0,6MHz) therefore need coaxial resonators. In professional repeaters, using wider shift (4 to 10 MHz), duplex filters often contain helix resonators. They are smaller, and cost less.
|
|
As notch filter.
When experimenting with home-made helix resonators,
i discovered:
- If a tap on the helix is connected to a transmission line, it becomes a narrow-band absorbing circuit ("notch filter"). The notch can reach a depth of 30-40 dB.
- On a certain lower frequency, a low SWR and insertion loss is measurable. The frequency-difference between the notch and pass is called "shift". The shift can be adjusted by adjusting the coil-tap nearer to, or further away from the mass-connection of the helix.
- If the helix notch filter is connected to the transmission line through a 1/4 wave cable piece, the notch will be below the pass frequency. This behavior is comparable to that of coaxial resonators.
Needed equipment.
For adjustment of this type of filters you need measuring equipment, and
experience using it. Adjustment is most easy using spectrum analyzer, tracking
generator and directional coupler.
Adjustment is also possible using less sophisticated equipment. For measuring the frequency characteristics, a 100-200 MHz signal generator with calibrated attenuator, and a receiver for 100-200 MHz with a good (analog) signal strength meter, can be used.
To measure the insertion loss, and the correct match to the transmission line at 146 MHz, a good SWR meter, power meter, dummy load, and a transmitter for 130-146 MHz can be used.
Dimensioning the filter.
During experiments i found, that the way of coupling between transmission line
and the helix resonator had a decisive influence on the properties of the
filter. Studying different ways of coupling and assembling, and the measuring
and comparison of it all, took a lot of time. If you stick to the dimensions i
state here, you will save a lot of time getting good results.
A tuning screw was not used. The filters were accurately adjusted to a fixed frequency. I you whish, a tuning screw can be added.
The diameter of the screening, the thickness of the coil wire, and the amount of stray capacitance between the coil end and the screening, determine the circuit-Q. The bigger they are, and the smaller the stray capacitance, the higher the Q, the lower the insertion loss, and the deeper the notch will be.
According to a nomogram i found in an old book,
the coil has to have a diameter half of that of the screening.
The length of the coil should be twice the diameter of the coil.
The wire diameter should be half of the pitch.
These sizes give good, predictable results.
The filters described here sometimes differ from these recommendations. This because cheap materials readily available were used.
Above that, the notch filters proved to give better results using shortest connections possible, and therefore the helix sometimes had to be positioned out of the center of the can.
Materials:
1. | Clean new paint can 3/4 liters. Diameter x height = 100 x 115 mm. Higher cans can be used too. |
2. | Tinned copper wire 3mm diameter. |
3. | BNC chassis receptacle, one-hole mounting (PTFE insulation!). |
4. | Transparent glue cartridge for thermal glue pistol. |
Construction: See fig.
1 to 4 for details.
Wind a coil from the 3mm copper wire. Diameter 48 mm, pitch 10 mm, 5 turns. Bend
the mass-side sharply at 90 degr. in the direction of the side of the screening.
Both coil-ends will be trimmed later to fit, so leave them long enough.
Drill a hole of 9.5 mm diameter at 25 mm above the bottom in the side of the can.
Solder the BNC receptacle into the hole. If the insulation melts, it was not made of PTFE (HI).
Important
construction details: The shape and length of the mass-end of the coil, and the way the BNC receptacle is connected to the coil, have big influence on the filter properties! Best results are obtained, if the coil end which will be connected to the screen, is bent 900 with respect to the coil, and if this end is soldered to the SIDE of the screen in a short way. The connection between BNC and coil MUST be AS SHORT AS POSSIBLE. Solder the coil therefore DIRECTLY onto the upper side of the pin of the BNC, close against the BNC insulation. The consequence is, that the coil is positioned not in the center of the can. This proved to have no serious consequences. Wrong: Should you connect through a short piece of wire, the notch will be about 10 dB less deep! Keep enough free space between the top of the coil, and the side and top lid of the can. This will reduce stray-capacitance, and detuning when the lid is placed.
|
|
Tap.
The shift can be adjusted, by changing the distance between the tap-point and
the the point where the coil end is bended towards the screen. For details see
fig. 1-4. The table below shows the approximate taps i found:
|
|
Assembling:
- | Cut the mass-connection of the coil so, that the coil will fit onto the BNC at the correct point, the mass connection of the coil to the screen, while keeping enough distance between the coil and the screen. |
- | Temporarily put an connector onto the BNC receptacle. It will cool the insulation, en keeps the pin centered. |
- | Tin the coil, BNC-pin and can at the appropriate places. |
- | Place the coil onto the BNC-pin,
with the mass end of the coil resting to the side of the can.
Temporarily solder the coil onto the pin. |
- | Solder the mass-end of the coil to the can. |
- | Solder the coil to the BNC-pin. |
Another configuration:
*** If the tap has to be 1/2 turn,
a short connection between the BNC and the coil is impossible. You now have two
possibilities:
or
|
|
Adjustment of the shift (the frequency difference between pass- and notch frequencies).
- | Connect the transmission cable between the measuring equipment to the BNC, using a BNC-T-piece. |
- | Measure the notch frequency. It will be much
to low.
Course-tune the notch frequency just below the wanted frequency. This is done by repeatedly clipping off a short piece of wire from the free open end of the coil. Tip: 3 mm equals approx. 1 MHz. Note this "notch frequency". |
- | While measuring the SWR, lower the measuring
frequency, until an SWR of less than 1:1,2 is measured (return loss 15 dB
or so).
Note this "pass frequency". |
- | Calculate the shift = notch frequency - pass
frequency.
For a 155 MHz notch filter it should be 10 MHz. |
- | If necessary, readjust distance between tap and mass connection, by desoldering and resoldering the coil. Larger distance means larger shift and vice versa. |
The total number of turns should be between 3.5 and 4.25 .
Fine tuning: After the shift is adjusted correctly, fine tuning of the notch frequency can be done. First, put a bit of solder on the sharp tip of the coil. It should be round to prevent flash-over. (See foto 1). By bending the last two centimeters of the coil in or out, the tuned frequency will become higher or lower. If you want, now you can drill a 4 mm hole in the side of the can, beneath the position of the top of the coil. Solder a 4mm nut on it. A 4mm screw will act as fine tuning condenser. When the lid is placed on the can, the resonator will lower in frequency. The more "headroom" the coil has, the less detuning. Always place the lid FIRMLY on the can. After the filter is tuned correctly, check the SWR (<= 1:1,2), insertion loss (0,25 dB), and notch deepness (abt. 38 dB).
|
|
Coil support.
If the filter works satisfactory, you can put a coil support to it. It will make
the construction more stable (foto 1). The support material MUST have
good VHF-properties. The coil-Q can be spoiled by it. I used a thermal glue
cartridge. When molten, it stinks exactly like coaxial cable inner-insulation.
The coil is now mechanically supported, free of mechanical strain.
|
Foto 1. Observe: as the tap is at 1/2 turn, the shortest mass connection at a right angle was here to the bottom of the can. Note also the slightly bended coil-end in order to tune the resonance frequency exactly. |
Fine tuning:
- | Press the lid FIRMLY onto the can. Measure
the notch frequency.
Note how much it is lower than the wanted frequency. |
- | Remove the lid.
Again measure the notch frequency. Calculate the difference due to closing the lid. Bend the coil tip. Measure the notch frequency. Repeat this, until the notch frequency is so much "to high", as will be lowered, when closing the lid. |
- | Close the lid firmly. Measure the notch frequency. It must be within 100 kHz from the wanted notch frequency. |
At last, check the insertion loss at 146 MHz, SWR and notch depth.