I really don't understand why you need to create electrical contact between the n connector and the can (eg. i've seen some just scrape the paint off and screw on and others solder the n connector to the can)?
The copper that sticks up in the can is hooked up to the center connector of the n connector, then the can has electrical contact on the outside of the n connector(not sure what that part is called). The outer coax shielding connects in my knowledge to that outer part of the n connector and the inner core of the coax connects to the center contact in the n connector.
Does this not mean that ultimately the can is being connected to ground on the wireless card?
What is the point of the can being connected to the ground? Is this to shield the copper wire inside the cantenna from signals other than what is coming in the front?
Sorry if this is not clear. I have only a very little bit of knowledge on the topic and i'm trying to figure out the basic logic of the design. I'm basically looking to find out how the can and copper wire in the can are interacting in basic terms.
The can acts as a ground plane for the 1/4 wave driven element. The ground plane should be at least 1/4 wave radius around the feed. The flange on the coax connector isn't anywhere near big enough.
That's called the driven element or feed antenna. It's 1/4 wavelength long.
Electrical contact at 2.4Ghz means soldered all the way around.
Yep. The coax is the ground return path for the RF signal. The idea is to keep all the RF inside the coax cable and eventually radiated into the can. If there's a break ANYWHERE in the ground system, the RF will literally leak out. That why you want a good soldered ground between the coax connector and the can (ground).
Well, yes. But this isn't DC or low frequency we're playing with. It's 2,400,000,000 Hertz, which acts quite differently than DC or audio. If you have a break or thin section anywhere in the ground system, it will act like an inductor instead of a shield causing signal to be either blocked, reflected, or both. Either way, the desired signal doesn't make it to the antenna.
See first paragraph. It's the ground plane for the driven element.
Well, that's a non-trivial exercise starting from zero. However, I'll give it a try. The purpose of any antenna is to radiate a signal in some desired manner. To do this efficiently, it has to match the impedance of the transmitter (usually 50 ohms) with that of free space (377 ohms).
If we start with a simple 1/4 wave vertical radiator over a very large ground plane, we have the start of an antenna. It's 35 ohms instead of 50 ohms, but that's close enough. The minimum radius of this large ground plane is 1/4 wavelength, but can be larger.
If we place a reflector approximately 1/4 wave behind the 1/4 wave vertical radiator, we now have a directional antenna. The reflector can be almost any size larger than 1/4 wave, but some specific sizes work better than others. It can also be a solid piece of reflecting metal, such as the bottom of the coffee can.
The ground plane also can be bent into a cylinder, as in the coffee can, and still have roughly the same effect. Combine the bottom reflector, with the wrap around ground plane, and we have a coffee can.
There are plenty of complications, interactions, calculations, and effects I've omitted, but that should be sufficient to cover the basics.
's a bit thin. The diameter of the driven element largely determines the useable bandwidth of the antenna. It also has a slight effect on the length of the driven element. If you could find some bare copper (or silver plated copper) wire of larger diameter, it will be an improvement. Basically, get the thickest wire you can fit into the center pin of the Type N socket.
If you want to get fancy, try using a conical driven element. If you look carefully at the picture at:
'll notice that the driven element is cone shaped affair which improves the impedance match and bandwidth. See:
shows a photo and construction details of the N connector with conical driven element (on a different antenna). This might be overkill for a first attempt. If I have time, I'll run side by side simulations of a wire feed, versus the conical feed. I expect the gain to be about the same, but the bandwidth of the cone to be much larger.
No book or web site is going to tech you antenna theory in a relatively short time. The various ham radio publications from the ARRL and other organizations are a good start. Online, I suggest:
you seriously want to dive into antennas, methinks it best that you download and run an NEC2 based antenna modelling program, such as
4NEC2. There are plenty of antenna sample NEC2 "decks" available to download or purchase that will provide various samples of different antennas. That's how I created the screen dumps of the coffee can antenna at:
you can play with the values and dimensions and see what the effects of longer waveguides, weird driven elements, and horns do to the antenna. I've been working on a modified version of a spreadsheet to generate NEC2 "decks" for can antennas based upon the spreadsheet at:
Good question. Different animals. The "gap" between ground planes and reflectors can be quite wide and still be effective. Let me simplify things using the common ground plane antenna. It has a vertical 1/4 wave driven element, and 4ea 1/4 wave radial ground plane elements space 90 degrees around the driven element. The gap between ground plane elements is HUGE yet, it still works.
However, if you take the same ground plane antenna, and somehow install a thin conductor between the coax cable shield ground and the
4 ground radials, nothing will work. That's because the thin conductor acts more like an inductor than a conductor of RF. Think skin effect where all RF conduction occurs on the surface. Small diameter equals small surface equals lousy conductivity at RF.
It doesn't take much of a thin conductor to cause problems. For example, the inductance of straight wire:
is 2mm long and 0.5mm diameter is 0.86nH (nano Henrys). At
2.4Ghz, the reactance of this piece of wire is: 2 * Pi * freq(GHz) * inductance (nH) 2 * 3.14 * 2.4 * 0.86 = 13 ohms That 13 ohms would be in *SERIES* with the 50 ohms drive impedance resulting in a loss of approximately (I don't wanna do the vector math): 10 log ((50-13)/50) = -1.4dB loss That's a fair bit of loss for a lousy 2mm piece of thin wire.
The thin wire represents the inductance that I would guess would be present if the connection between the coax ground (N-connector) and the ground plane (cantenna can) were run through a thin wire. It also doesn't include the mismatch loss (reflection coefficient) which will increase the loss somewhat.
Note: This loss is also why you don't want any exposed (unshielded) coax center conductor when soldering a coax cable to a circuit board. It doesn't take much exposed center conductor to create significant loss.
Incidentally, you can build an can feed driven element that is NOT sensitive to the connection between the coax connector and the coffee can. Instead of coming in from the side of the can, insert the probe through a hole in the bottom of the can. Instead of a single 1/4 wave radiator, use either a dipole, folded dipole, or full wave loop. Each of these form a complete antenna, where the can acts something like a dish reflector and does not require a ground contact to the coax cable shield. There are some complications to this design, but the advantage is that the cable and connector can be insulated from the coffee can.