25 July 2010

Wi-Fi Antenna & Biquad Antenna Construction 25-07-2010

                                         Wi-Fi Antenna Construction 

Many of you have probably already seen this on The Screensavers, but we wanted to test this out for ourselves.
The whole concept came from this website here: Freeantennas.com

The idea is to boost the output from your wireless access point. The design is based on a Deep Dish Cylindrical Parabolic.

This is what you'll need: some construction paper, scissors, a glue stick, aluminum foil and the antenna template. You can get the template HERE.

Now the template was a bit big for my needs so i resized the two pictures down to 75 % in paintshop and then printed them out.

First i trimmed off the excess paper of the template to fit the construction paper i happened to have. I then glued the template on to the construction paper using the glue stick. This is to give the antenna some strength.

Using the scissors i cut out the template now stuck to the construction paper and followed the pattern's lines that needed to be cut.

Next i cut the six slits out of the rectangle piece and the glued the aluminum foil onto it. After gluing and wrapping the rectangle part in foil, i used the slits i cut before to cut through the aluminum.

Next up is to bend the other part nearly in half and slide the tabs into the slots on the rectangle piece. I stapled the tabs down for good measure.

And finally you have your antenna! Wow that thing is sure butt ugly, but it didn't cost me more that 10 cents to make.

I placed the antenna on the wifi antenna as shown in the picture and then set off to test it.

I used net stumbler to record the results. First with the home made antenna (left) and then without (right).

Low and behold the damn thing works! Now it's not a big gain, but it might be good enough to reach that particular spot in your home you couldn't get any wifi access. Before buying a store bought antenna try this and see if it works for you first, besides it'll only cost you a dime.

                                              Biquad Antenna Construction    

I've done quite a bit of experimentation and testing with various home made dipoles for 24dBi Conifer dishes, and have managed to increase the performance of the dish.

Trevor Marshall has a webpage with information about using a biquad as a feed on a Primestar satellite dish, with very good results. I decided to try using a biquad as a feed on a 24dBi Conifer dish, to see if I could improve the performance of it of the dish.

Note that the photos on Trevor Marshall's webpage do not clearly show the construction of the biquad - particularly the way in which the quad is attached to the coax. Numerous people (including myself) have constructed biquads incorrectly, based on his photos, and found that they perform very poorly.
Use the photos of my biquad below, and refer to the websites listed in the references section at the bottom of this page for more information on the correct construction of the biquad.

Parts Required
I used the following bits and pieces:
  • 123x123mm square section of blank PCB
  • 50mm length of 1/2" copper pipe
  • short length of CNT-400 or LMR-400 low loss coax (~300mm long)
  • 250mm of 2.5mm2 copper wire (approx 1.5mm diameter)
  • N connector
Note that you don't have to use blank PCB for the reflector. You can use any material that's electrically conductive, can be electrically connected to the coax braid, and will reflect microwaves (ie, any metal plate will do fine).
I've also heard of people using CDROM as the reflector, as the foil on it will certainly reflect microwaves.

Cut a square piece of blank printed circuit board, 123x123mm.

Note that Trevor Marshall recommends a size of 123x123mm if using the biquad as a stand-alone antenna, while 110x110 is optimal if using it as a feed for a large dish.
He also recommends attaching some lips to two sides of the reflector, to reduce radiation from the rear lobes.

Use some steel wool to remove any tarnish and polish it up. Cleaning the copper in this way will make it easier to solder.

blank printed circuit board

Cut a 50mm section of copper pipe, and file both ends smooth. Using some sandpaper and/or some files, polish up the copper pipe (including the inside of the copper pipe, to ensure a good connection with the coax braid).

the dimensions of the copper pipe

Cut a notch into one end of the copper pipe, removing approx 2mm from half the circumference.

a short secion of copper pipe, notched at one end

Drill a hole in the centre of the blank PCB so that the copper pipe is a tight fit in the hole. I found a reamer to be very useful for enlarging the hole to the correct size.

making a hole in the centre

Insert the copper pipe into the hole, with the notched end on the copper side of the blank PCB. The copper pipe should be protruding approx 16mm through the hole, measured on the copper side of the PCB.

insert the copper pipe into the reflector

Solder the copper pipe to the PCB, to ensure a good physical and electrical connection.

solder the copper pipe to the PCB

Quite a bit of heat is needed, due to the thickness of the copper pipe, and an electrical soldering iron probably won't be able to deliver sufficent heat. I found a small gas torch works quite well.

Making the Element
The element is made from a length of copper wire, bent into the appropriate shape.

Note that the length of each "side" should be as close to 30.5mm as possible (measured from the centre of the copper wire to the centre of the copper wire), which is a quarter of a wavelength at 2.4GHz

the shape and dimensions of the element

I had some offcuts of electrical power cable lying around, and found that 2.5mm2 power cable had a diameter of approx 1.6mm - a little bigger than the 1.2mm that Trevor Marshall specifies, but didn't think it would make a significant difference to the performance of the biquad.

recycling power cable offcuts

Remove the insulation, measure and cut a 244mm length the copper wire, and straighten it as best as you can.

straighten the wire

Measure the mid-point of the wire, and make a 90 degree bend. The bend should be quite sharp and pronounced.

90 degree bend

Measure the midpoints of each half, and make two more 90 degree bends in the wire, so that it looks like that shown in the photo below.

another two bends

Once again, measure the midpoints of each section, and make some more 90 degree bends, resulting in what is shown below.

bend it some more...

Do the same to the other side, resulting in the biquad shape.

make it symetrical...

Clean up all your bends, and ensure each side of the element is as straight as possible, and as close to 30.5mm as possible.
Note that you may need to trim a small amount off each end of the wire to achieve this.

The element must now be attached to the reflector. Note that only the two "ends" of the copper wire are to be attached to the copper pipe - the centre of the copper wire must not touch the copper pipe (hence the notch which was cut into the end of the copper pipe.

The copper wire element should be approximately 15mm away from the reflector. Testing antenna performance while varying the spacing between the copper wire element and the rear reflector indicates that a spacing of approx 15mm provides the lowest SWR (test results available here).

the element soldered onto the copper pipe

Strip approx 30mm of the outer sheath from the end of the coax.

strip the outer sheath

Fold the braid back over the outer sheath, and trim the centre conductor, so that about 4mm is protruding.

fold the braid back, trim the centre conductor

Insert the braid into the copper pipe, so that the end of the centre conductor lines up with the extreme end of the copper pipe, and solder the centre of the element to it, ensuring the centre of the element is not in contact with the copper pipe. Refer to some of the additional photos below for details.

solder the centre conductor to the element

another view

Note that the feed between the rear reflector and the biquad element needs to be shielded. Using coax to feed the biquad element directly, and positioning the coax inside the copper tube achieves this.
Use of bare conductors as a feed between the reflector and biquad element results in a radiating feed (such as this one), which will have a detrimental effect on the biquad's performance.

I used a coax crimper to crimp the end of the copper pipe onto the coax. This ensures that the coax would not move inside the copper pipe.

the copper pipe crimped onto the coax

the completed biquad

Now terminate the other end of the coax with an N connector.

If desired, you can add spacers at each end of the element, to ensure the element doesn't move in relation to the reflector. Refer to my double biquad page for more details on making spacers to support the element.

If you intend to mount the biquad outside, I'd recommend you place it into a weather-proof enclosure, to prevent corrosion, and to prevent water ingress into the coax.
Numerous people have used small tuppaware containers successfully.

This can be achieved by drilling a hole in one side of the container, and pass the coax tail through the hole, leaving the biquad itself inside the container. Seal up the hole for the coax with some silicone, and your biquad should be protected against the elements.

another view of the completed biquad

Some very rough initial testing using the biquad as a feed on a 24dBi Conifer dish looks very promising, with the signal strength being at least as as good as my home made Conifer dipole (I was holding the biquad at approximately the focal point of the dish, and hadn't even removed the Conifer dipole).

I also managed to get a marginal link to a 180 degree waveguide on an access point 10km away, using only the biquad by itself, connected to a 30mW RoamAbout wireless card.

Some more detailed testing with multiple antennas, including the biquad shown above, indicates the biquad has a gain of approx 11-12dBi.

A friend has access to some antenna test equipment, and performed some tests on the biquad featured on this page.
The azimuth plot (ie, radiation pattern) of the biquad is shown below, and shows a 3dB beamwidth of about 50 degrees.

azimuth plot of the biquad

A number of people have suggested the spacing between the element and the rear reflector should be a 1/4 wavelength (ie, 30.5mm) instead of 15mm. However, test results (such as these) indicate the SWR of the biquad is minimised when the spacing is about 15-17mm. Increasing the spacing to 30.5mm increases the SWR significantly, thus reducing the efficiency of the biquad.

For a higher-gain variation of the biquad that's virtually just as easy to build, have a look at the double biquad.

When using a biquad to establish a link to another wireless device, you should ensure the polarisation of the biquad is the same as the antenna you are connecting to. Similarily, if establishing a link with two biquads, ensure they are both oriented for the same polarisation.
Failing to match the polarisation will result in significant signal loss.

vertically polarised

horizontally polarised

Changing the polarisation is just a matter of rotating the entire biquad antenna by 90 degrees.

The biquad antenna is not particularly directional, but has a fairly wide beamwidth.
The 3dB beamwidth for a biquad (without side lips) is typically about 40-50 degrees, thus making it ideal for any applications where you want fairly wide coverage.

The relatively wide beamwidth also makes a biquad very suitable for war-driving and stumbling, allowing you to pick up signals without having to align the antenna directly with the signal source.

While a directional antenna, such as a Conifer dish (3dB beamwidth of a 24dBi Conifer dish is approx 7 degrees), is better suited for point-to-point links, the narrow beamwidth of a Conifer dish requires more precision when aligning the antennas (the narrower the beamwidth, the less susceptible it will be to interferance from other sources). An antenna with a wider beamwidth, such as a biquad, doesn't require the same precision for alignment, thus making it easier to get a link working.

If you're one of those people who may not have all the tools required for building a biquad antenna from scratch, or you don't want to shop around for all the parts required, you can buy a DIY kit containing all components from WarDrivingWorld.

In November 2006, WarDrivingWorld sent me one of their DIY biquad kits to review. The kit contains all the pre-cut and pre-drilled parts required to build a biquad antenna.
For more infomation on this kit, including antenna comparision test results, read my review of the WarDrivingWorld DIY Biquad Kit.

Biquad and double biquad antennas, as well as DIY kits for building your own antennas are also available from Biquad Supply.



LED flashers are great fun for the novice or seasoned builder!.
In modern times, analog circuits such as the astable BJT multivibrator or a 555 timer IC driving 4017 decade counter(s) are being replaced with small microprocessor-based circuits such as the PIC Microcontroller. The QRP/SWL HomeBuilder web site has featured a number of multivibrator and LED flasher circuits over the past decade and each year at least 40 - 50 emails are received requesting help, specific designs or calling for more projects. Thus, another page of LED flasher circuits was added and building these circuits was quite enjoyable.
Lower millicandela LEDS were used for all the experiments. The 2N3904 transistor was used exclusively and the schematics do not indicate this part. By request, this web page will avoid using a lot of math

Basic Multivibrator Flasher

To the left, in schematic 1 is the basic 2 LED astable multivibrator flasher built by Mary. She chose to use 2 different colored LEDs and the red LED is clear when unlit. It is quite bright when lit compared to the yellow LED despite the fact that it only draws 0.5 mA more. The bread boarded test circuit was powered by a new 9 volt battery and was regulated by a L78L05 (in a TO-92 package as shown in the schematic). The 5 volt regulator was used to avoid exceeding the reverse breakdown voltage of the 2N3904. This topic will be discussed a little later on.

470 ohm current dropping resistors were chosen to keep the collector current draw less than 10 mA. The LEDS were bright enough to see well in dim lighting. You may change this resistor "R" value (lower R = brighter), but do not exceed the maximum current rating for the LED or transistor (this is more applicable to higher voltage multivibrators). You may also place 2 or more LEDs in series on each half of a multivibrator, however, the current dropping resistor may need to be reduced to maintain brightness. Consider using a power supply as opposed to battery power for your flashers.

To change the pulse (oscillation) frequency, you can change the base resistor or the timing capacitor values. For example, increasing the capacitor or the base resistor values will increase the time OFF per cycle and thus reduce the oscillation frequency. The oscillation frequency is 60 divided by the sum of the time OFF for each half of the multivibrator. This is discussed a little more here . Do not feel you have to use the same timing capacitor for each 1/2 of the multivibrator. Multivibrators with different timing components on each 1/2 are termed asymmetrical.

Over time, some builders sent me emails that they could not get their multivibrator to run. I problem-solved with them and discovered many problems including bad parts, bread boarding errors, the oscillation frequency was too fast to observe, transistors were not saturated during their ON time and failure of the transistors due to excessive current or perhaps even reverse emitter-base breakdown.
Above photograph: The breadboard of the 5 volt flasher. The yellow LED is switched ON, the red LED is switched OFF.

Basic Astable 555 Timer IC Flasher

For a lower parts count than the 2 transistor multivibrators, 2 LEDs can be alternately flashed with a 555 integrated circuit configured as shown in Schematic 2.  I chose the combination of a 2K2 and a 47K resistor to determine the oscillation frequency along with the 10 uF capacitor connected to pins 2 and 6. You can practically change the (R Speed) 47K value to between 10K and 100K or more. Greater resistance = lower speed. You may also wish to connect up a 100K or so potentiometer instead of the 47K resistor for a variable speed version. Additionally, the 10 uF capacitor value can be changed. Feel free to experiment.

Although, alternately flashing LEDs  is great for the beginner to electronics, the basic one ON, one OFF circuit gets boring quickly. In the next section, we will try to improve the look and try to approximate a flash like a police car (within limits).
Above photograph: The breadboard of the 555 IC flasher. The R Speed resistor was 100K at the time.

A Better Looking Multivibrator Flasher?

To the right is a multivibrator circuit suitable for 6 VDC or greater. Protective diodes have been added to prevent emitter-base reverse breakdown which may occur with higher power supply voltages. Reverse DC voltage spikes approaching that of the power supply voltage may occur each time the transistor switches OFF and in the worst case scenario, they can damage the transistor or in the best case scenario will have no effect or possibly just decrease the time OFF of the transistor. The minimum emitter-base breakdown voltage of a 2N3904 is 6 VDC. The reverse breakdown voltage of a small-signal diode such as the 1N4148 or 1N914 is probably 100 VDC. Protective diodes may be added to our basic circuit 2 different ways and both of these methods are shown in Schematic 3 as circled 1 or 2. The emitter-base break down protective diodes are colored blue.

I prefer Method 2. With Method 1, the protective diode slightly increases the collector-emitter saturation voltage and decreases the collector current by the small voltage drop across the diode (0.6 VDC for a silicon diode). For Method 2, during the part of the cycle when the transistor base normally goes negative, just the anode of the protective diode will go negative and not the transistor base. Other than the protective diodes, the circuits of A and B are identical.
To make a better looking flasher, I now use a multivibrator as the timer and leave the work of switching the LEDs on and off by auxiliary circuits called half-shots, which are in essence, 1/2 of a regular multivibrator. The take off points of the basic timer multivibrator are on each collector and are indicated as the A or B contained in a red square. While, using half-shot circuits to control your LEDs adds to the parts count, there are definite benefits in my opinion. If you would rather just use the basic multivibrator in Schematic 3 as your flasher, simply add 1 or more LEDs to each collector of Circuit 1 or 2,  maybe tweak each current limiting resistor and your done.
To the top right of Schematic 4 is the basic half-shot circuit. While it is subjective and dependent on your timing resistor and capacitor values, the flash effect is more pleasant. Instead of one LED being OFF while the other is ON, they have a slight overlap or independence, which to my eyes at least, better simulates an emergency vehicle flash.
To augment this effect further, an additional half-shot was added for a total of 2 LEDs connected to each 1/2 of the basic timer multivibrator. The second half-shot circuit was altered to improve its output waveform by adding a diode and a resistor. In a traditional multivibrator, the leading edge of the output waveform is usually not square. This is because the collector voltage does not immediately jump to its highest potential (at or approaching 12 VDC in our example) when the transistor is switched OFF. The capacitor connected to the collector must charge through the collector resistor and this causes a delay due to the time constant (the product) of the collector R and the C connected to it. By adding a diode to block the normal capacitor charging current path and a separate resistor to charge the capacitor, the collector voltage will instantly rise to its supply level when the transistor is switched OFF. This provides a more rectangular shaped output waveform which may improve switching (especially when cascading more 1/2 shots). On the lower part of Schematic 4, you can see the cascaded half-shot circuit. Connect either the Basic 1/2 shot or the Cascaded 1/2 shot circuit to each collector of your basic timer multivibrator (such as # 1 or # 2 from Schematic 3). Additional stages can be added. Protective diodes are shown in blue.
Above. A breadboard of the the basic timer multivibrator with a cascaded half-shot circuit on each side. Two LED colors were used to better illustrate this. The emitter-base break down protection diodes were connected to the transistor emitters in this bread board. Subsequent circuits used protection diodes on each transistor base per Schematic 3, circled Section 2. This circuit was connected to a well filtered 12.2 VDC power supply.

Above. A schematic topology provided by W7ZOI. I suggest using a 1K to 2.2K ohm resistor for the base resistor of Q3, Q4, Q9 and Q10. In Section A, Q3 and Q4 are timer controlled switches which turn their LEDS ON and OFF. You may have to temporarily ground one of the transistor bases get obtain proper function when you string up several sequential transistor-LED stages.

Sequential Flasher Using ICs

A 20 LED flasher was built using the 4 ICs shown to the right; a 555 timer, a 4017 decade counter and two 4049 hex (NOT) inverters. This circuit has 2 rows of 10 LEDs. The top row walks from right to left with only one LED on at any given time. The bottom row has 9 LEDS on at once and also walks right to left. In the bottom row, the LED that is OFF corresponds to the LED which is ON just above it in the top row. Use static precautions with the 4000 series CMOS integrated circuits. This circuit requires a power supply as the current  draw is substantial.

All my circuits are built using Ugly Construction. Copper boards are buffed lightly with fine steel wool before use. The LEDs were mounted in a small piece of PC board that was drilled to accommodate them. Board sections and separations were cut using a hobbyist motor tool while outdoors. The section with LEDs was isolated from but soldered to the main copper board surface. You can test your LEDs for correct polarity using a 9 volt battery, a 1 K resistor and a coupled of wire clips.

To the right is the schematic for the 555 timer astable multivibrator used to drive the 4017 decade counter input. This is straight-forward. There are countless web sites with information regarding the 555 timer. Use your favorite web search engine to locate some of them. To tweak the flasher speed up, reduce the value of the 10K R. To tweak for slower triggering, increase the 10K R value. You can also choose other R values for the 10K and 2K2 resistors or add in a potentiometer for variable speed. The 100 uF and 0.01 uF capacitors also provide filtering for the other 3 ICs.

The above photo shows all of the ICs grounded and wired up to the B+. The 100 uF and 0.01 caps are also connected to a small "island" for the B+. The 555 timer wiring is completed. A 47K R Speed resistor is shown in this 555 circuit, but was later reduced to 10K as triggering was too slow. Pins 8, 13 and 16 are soldered to the copper to well anchor the 4049 ICs. In a 4049, pin 8 is connected to ground and pins 13 and 16 are not internally connected so it is okay to also solder them to ground. I also connected pin 3 of the 555 to pin 14 of the 4017. At this point, I connected a 9 volt battery clip and battery and voltmeter tested the circuit to ensure the 555 and 4017 were firing properly. Test as you go is my biggest recommendation to budding electronic experimenters. The circuit as photographed above was later modified to have a single current limiting resistor for each of the LEDs on the bottom row as shown in Schematic 7.

Shown above is main schematic; the ten 4017 outputs into the two 4049 inverters. There are 6 inverters per 4049, but only 5 were used on each. For LEDS #1-10, the measured current was 10.42 mA for the ON LED. The current draw of each of the 9 ON LEDs #11-20 was 9.24 mA. The total current draw of this project at 12.22 VDC was 131 mA. The peak 4049 NOT gate output voltage was 10.84 VDC. Consider always measuring the current flowing through your LED flasher projects to prevent exceeding the current rating for your circuit components.

A completed flasher prototype is shown above. It takes concentration to keep your pin numbers straight and correctly wire this circuit. I connected a 9 volt battery clip to the B+ and to ground and after wiring up 2 LEDS, I snapped on the 9 volt battery to ensure I had wired correctly by observing the LEDs function. The battery was then disconnected and 2 more LEDS were wired up. The 9 volt battery was then again connected for a short while. This was repeated every 2 LEDs until it was finished. Thus, I knew the project worked even before I had completed it. If you just wire up all the LEDs and need to find and correct a mistake afterward, it is often more difficult debugging/repairing than testing as you go. If you happen to accidentally bridge 2 adjacent IC pins together with solder, heat up the solder bridge and drive a small screwdriver between the 2 shorted pins. This happened once when soldering the circuit above, but the problem was mitigated as described. Ground the 2 unused 4049 inverter inputs.

Above in Schematic 8 is a depiction of how I wired up my inverters for reference only. There is no one right way.

The above schematic describes how the 4049 inverters switch ON and OFF the LEDs.

A side view of a completed prototype 20 LED flasher is shown above.  Good luck with your own LED flasher experiments!

Contribution by Ned, K7ELP

This next section is a contribution by Ned, K7ELP. This miniature Christmas tree has one steady white LED at the peak and 6 red and 6 green blinking LED’s. It uses a standard 9V (PC1604) battery for power. In addition to supplying power for the miniature Christmas tree, the 9V battery acts as a stand.

Theory of Operation

Referring to the schematic, an astable multivibrator is used to blink the red and green LED’s. The red and green LED’s are grouped in four series strings of three LED’s and a 220W current limiting resistors. The single white LED shows two 470Ω in series. One resistor of 1k would have been adequate, but during the construction of the circuit I needed to jump some traces on the PCB. With a fully charged battery the individual strings of the red and green LED’s draw approximately 8mA from the battery and the white LED draws approximately 6mA. The colored LED’s continue to flash when the battery is discharged below 7V. The oscillation frequency with the base resistors of 560KΩ and 0.47uF capacitors between the base and collectors is approximately 5Hz. The 560KΩ base resistors were calculated to provide 12 uA of base current for the Darlington transistors, and the 0.47uF capacitors were found experimentally.


The miniature Christmas tree was constructed on a 3” X 6” piece of fiberglass single sided copper clad board that was positive resist sensitized. The layout is of the bottom of the board. The PCB image has 2 extra triangles of resist on either side of the sloping sides, and two extra rectangles on either side of the stand . I did this to save a little on etching solution, since these parts of the board would be trimmed off and discarded. The stand portion of the pattern also shows three rectangles. The two smaller ones are used to solder the battery clip to the circuit board. The larger one was used to save a little on the etching solution and is covered by the battery when it is in place. Once the board was exposed, developed, etched, drilled and then tin plated I painted the top of the board with some spray paint.

When I was laying out the PCB, I decided to have the resistors on the top of the board with the LED’s, as the resistors are colorful and could appear as other decorations on the tree.

I installed all the parts on the board except for the 0.47uF capacitors. By not installing the capacitors and applying power, all the LED’s should light up. When I applied power there was one string of 3 LED’s that didn’t light and the white LED didn’t light. The string of LED’s had both ends connected to the collector of its associate transistor and had two current limiting resistors in the string. This problem was solved by removing the extra resistor and installing a jumper on the bottom of the board. The white LED had both ends of its circuit connected to the 9V positive trace. The negative side of the white LED circuit was connected to ground thru a 470Ω resistor to jump some traces. The 1KΩ resistor in the positive lead was changed to 470Ω. The shown PCB layout was corrected for these errors. The picture of the back of the completed PCB shows a jumper and one of the 470Ω resistors.

Shown above is the N7ELP schematic. There are no worries about emitter-base reverse breakdown with these Darlington pair transistors. On this web site click on schematics or parts lists for printable versions.
Shown above is the complete project parts list.
Above is the PCB trace. Click on the image for a higher resolution version. Big Thanks to Ned for this great contribution.