best antenna

[ccz]

Not giving up!
I'm not yet dismissing the previous experiment. One explanation could be that the dongle was overwhelmed by the new signal, hence the poor performance. So I tested with all gain settings and it still performs the same. Putting a 68.8mm wire brings back the good reception.


Will also try next to build the following antennas as I didn't find any mentioning of them here:

Slim Jim - http://jeroen.steeman.org/Antenna/Sl...nna-Calculator
PCB antenna - http://f5ann.pagesperso-orange.fr/An...MHz/index.html
PCB colinear - http://f5ann.pagesperso-orange.fr/An...MHz/index.html

[abcd567]

Not giving up!
I'm not yet dismissing the previous experiment. One explanation could be that the dongle was overwhelmed by the new signal, hence the poor performance. So I tested with all gain settings and it still performs the same. Putting a 68.8mm wire brings back the good reception.


Will also try next to build the following antennas as I didn't find any mentioning of them here:

Slim Jim - http://jeroen.steeman.org/Antenna/Sl...nna-Calculator
PCB antenna - http://f5ann.pagesperso-orange.fr/An...MHz/index.html
PCB colinear - http://f5ann.pagesperso-orange.fr/An...MHz/index.html

Great. Happy experimenting!

It will be nice if you post some photos of your test setup and test antenna.

[abcd567]

Please see this page also with 6 different antenna designs:

http://jeroen.steeman.org/Antenna/An...gn-Calculators


jeroen.steeman.collinear-dipole-antenna-2.jpg .. jeroen.steeman.j-pole-antenna-2.jpg .. jeroen.steeman.slim-jim-antenna-2.jpg .. jeroen.steeman.coax-collinear-antenna-2.jpg .. jeroen.steeman.dual-bi-quad-antenna-2.jpg

[abcd567]

J-Pole & Variants

j-pole and variations-b.jpg

[ccz]

Thanks! Will try them asap.

Yesterday I got some lathe time and finished a collinear project I begun a week ago.
Some brass tubes cut to 0.8 x 1/2 wavelength with 1/4 base and tip, not shorted. I added a very nice wire ground plane between antenna and cable that is not seen in the pictures.

It came out looking nice but performance was extremely poor, under 50nm. Even placing the amplifier right after the antenna did not help.


This tells me that whatever precision I can muster in parts making is not going to help me a lot. I need to tune the antenna and no specialized equipment at hand.

Considering my current results I will also order one antenna from ebay, to add it as a reference point.

Until next time, I have stumbled upon this video, tuning the antenna using an oscilloscope and signal generator:

[SpaxmoidJAm]

i thought from your previous post you werent going to have any dielectric on the antenna?

[ccz]

That's right, but since I didn't have time to print the spacers, I assembled first the one with dielectric.

The version without will follow soon. First I want to try the oscilloscope method for tuning, it seems pretty bad that I can only do 1/4 wave antennas. I have to find out where I am making mistakes.


Are our RTL sticks 75 or 50 ohms? I saw it mentioned somewhere but can't put my finger on it right now.

Ahh, got it:

What is the RTL-SDR input impedance?

Since these dongles are intended for TV, all dongles will have an input impedance of about 75 Ohms. However, the mismatch loss when using 50 Ohm cabling will be very minimal at around 0.177 dB.

The 75 Ohm impedance for the R820T can be checked on the datasheet which can be downloaded here.

[Rooster]

ALL TV sticks are 75 Ohm... but I wouldnt loose much sleep over the impedance, the rest of the homebrew antennae are going to be nowhere near calculated !!!!
As for using a 'scope - you are going to need very deep pockets for a 1 gHz version ! not to mention a sig gen that covers the frequencies.

If you are intent on spending money - you wont go far wrong with an analyser... such as :

http://miniradiosolutions.com/54-2/

[abcd567]

miniVNA €450
and a Signal generator, may be same price!

I can buy 20 very good quality commercial antennas for this money.

Purchase a 100 ft/30 m coil of cheap TV or Satellite coax, and a tape measure. Then cut 6 lengths of 1.50 meters each from this coil, and start cutting pieces.

8 pieces x 110mm (vf=0.8),
8 x 112mm (vf=0.81),
8 x 113 mm (vf=0.82),
8 x 114 (vf=0.83),
8 x 116mm (vf=0.84), and
8 x 117mm (vf=0.85).

Assemble the 6 cocos and try. At least one of the 6 will be good! Use the balance coax in the reel as your feeder coax. Cost < €10

[SpaxmoidJAm]

your plan wont work, because if you cut to those lengths by the time you have trimmed of the outer needed to push the elements together they will non longer be the right length.

see image, length.jpg

that's your corrected wavelength

[abcd567]

your plan wont work, because if you cut to those lengths by the time you have trimmed of the outer needed to push the elements together they will non longer be the right length.

see image, [ATTACH=CONFIG]7875[/ATTACH]

that's your corrected wavelength

Yes, sure the lengths given by me do not include the pins on both ends. This is the way element lengths are specified in all designs on internet. The maker has to add required extra length for pins as it suits him. Pin length is not critical to design, and can be anything from say 10mm to 30mm.

Have you noted I mentioned to cut 6 lengths of 1.5 meters each? The length required without pins is 8 x 117 = 940 mm, say 1 m only. This additional 0.5 meter is for pins (8x2x0.025=0.4m).

[SpaxmoidJAm]

just wanted to make that clear to anyone reading the post.

[abcd567]

just wanted to make that clear to anyone reading the post.

A wise step to cover my mistake of not mentioning "add to these lengths, the length of pins at both the ends" . Thanks.

[ccz]

Time flies... but in the meantime I managed to work a little on the air gap coco, ordered one antenna from ebay, and soldered my best spider until now.
The ebay coco is worthless, will try to sort it out with the seller.

The new spider antenna works really great, over 200nm coverage, bulk of received messages is at around 100-130nm.
It's a ~50nm improvement over my normal spiders/cantennas.

For this spider I used one male F connector and one F female joiner. RTL side is covered by a MCX connector adapted (soldered) to the RG59/U cable I used.
The cable is all copper with 0.6mm dia center wire. Attenuation is 13.02dB/100m at 100MHz, I expect it to be a lot over 30dB at 1000MHz.
Since the cable length is about half meter, it should be fine.
The second spider uses 11m of the same cable with identical reception. Weird, right?

The only different thing that I did comparing to other spiders is that the central coax wire was left longer by an extra F joiner + whip length.
The male F connector was mounted on the cable and the central wire was protruding by 10-11cm.
This was mounted through the F joiner with the central wire coming through and exiting for about 70-80mm. I trimmed it to 68.8mm.
On the F joiner I added some copper wire to allow the fixing nut to sit flush on the open end.
4 copper wires (I used 1.5mm appliance cable here) were soldered on the fixing nut and trimmed to 68.8mm length each.

[abcd567]

@ccz
At last, after long wait...
Nice to read detailed report. At microwave, the hardware often behave unexpectedly. The net result is combination of many factors, and combination of little difference in each can result in substatial difference in overall performance.

Glad to see you broke the tradition of using SO239 connector (good only for VHF) and N connector (for UHF) and used F connector which is designed for microwave frequencies. The idea of soldering the radials to the nut is good. Others who used F connector, have soldered radials on the washer.

Since in the new spider you have placed few turns of wire below the nut, raising radials to the level where the whip starts, this might have improved the performance of your new spider.

I have once tried to avoid SO239 and use an F connector, but also avoided soldering as I have very poor soldering skills. Click here (post #5) to see it.

I also have tried to make a Spider without any connector at all. Click here to see.

For outdoor use put few drops of hot melt glue or a sealent to prevent rain water or moisture entering the coax and corroding its braid and center conductor.