[SOLVED] 915Mhz Antenna design

Status
Not open for further replies.
A

amirke

Member level 5
Joined
Mar 31, 2012
Messages
92
Helped
0
Reputation
0
Reaction score
0
Trophy points
1,286
Activity points
1,927
Hello all,
I have a communication system based on a 915MHz carrier wave.
The communication is for less than 1 meter.

No need for PA in the system.
What are the options for the antenna in the circuit
10 simple wire soldered to the PCB transceiver (wire - 50 ohm)
2) Chip antenna
3) PCB antenna (wires routed in the PCB)

I want to use the first option - simple wire.
Can you tell me what would be the difference in the quality of the transmitted signal if I use wire instead of a chip antenna or PCB antenna?

Thanks
 

Not sure what you mean with "wire - 50 ohm"? A λ/4 monopole has a length of 82 mm at 915 MHz and near to 50 ohm real input impedance (with sufficient ground plane), matches a 50 ohm transceiver output directly and makes an effective antenna, a shorter wire antenna need a matching network and is generally less effective.

Chips and electrical small (< λ/4) PCB antennas are generally more sensitive to detuning by hand effect and typically a bit less efficient.

1m range will be usually spanned without any antenna, just due to antenna effect of a not perfectly shielded transceiver circuit. Any small wire or "PCB antenna" attached to the RF port will further increase the range.
 

This is real sinful truth. I've seen many situations when a simple wire works better than an antenna developed by an "army" of antenna engineers.
So, as was already stated, a wire antenna it will work fine, but for sure is not a very "professional" option.
In your situation, to get reliable and consistent result, I would go for an Inverted-L or an Inverted-F PCB antenna. Beware that you need a proper ground plane for this type of antennas (actually the same as you need if use a simple wire antenna).
 


You say it is for transmission for 1 metre or less so there is no need for anything complex. A wire or a track on a PCB would be all that is required


Dave
 
Reactions: amirke

    A

    amirke

    Points: 2
    Helpful Answer Positive Rating
Thank you, guys.
This is enlighting.

The reason I am asking of a wire-based antenna is that of placement issues.
I have a minimal PCB design, and thus I want the antenna will be placed out of the PCB.

So, the conclusion - an 82mm length wire Will do the job?
Is a specific wire is recommended for used as an antenna?
What is the effect of the shape of the wire (straight, curled)?
 

82 mm assumes a straight wire. However as stated by Davenn and myself, it's not actually necessary for 1 m range.
 
Reactions: amirke

    A

    amirke

    Points: 2
    Helpful Answer Positive Rating
How do I take into consideration wire length, if it is not straight? (in case of long-distance transmission)

What are the possible damages in case of not using an antenna at all (or not fitted antenna - longer/shorter)
 

FvM said:
82 mm assumes a straight wire.
Click to expand...

... at 90° angle from the ground plane. And the ground plane must be large (much larger than the wire). In many cases, both assumptions are violated and EM simulation is needed for really good antenna efficiency.

However in this case, you seem to have a lot of margin in received power, so you can try the straight wire. At the usual power levels (1mW) there is no damage to devices if the antenna is mismatched.
 
Reactions: amirke

    A

    amirke

    Points: 2
    Helpful Answer Positive Rating
Thanks for the assitance
 


volker@muehlhaus said:
It's not that simple. The trick with these small "antennas" is that they require a very well defined large ground, and depend on that ground. Changes to the ground size will detune the antenna.
Click to expand...

What is the amount of ground needed?
the rectangle around the chip?
On 1 layer? 2 layers? more

Is there any thumb rules for placing the chip antenna?
 

Large means large, e.g. > λ.

That's a principle statement about monopole antenna operation, you can review antenna theory text books to understand how it's substantiated. Most real mobile transmitters don't fulfill the condition, thus the actual antenna impedance is different from the theoretical value and has to be considered in the matching network design.

However, as mentioned several times, the discussion is completely irrelevant for your 1 m link. Anything goes in this case.
 
Reactions: amirke

    A

    amirke

    Points: 2
    Helpful Answer Positive Rating
volker@muehlhaus said:
It's not that simple. The trick with these small "antennas" is that they require a very well defined large ground, and depend on that ground. Changes to the ground size will detune the antenna.
Click to expand...

The occupied area is just roughly 10mmx20mm, so it's not critical thought.
 
Reactions: amirke

    A

    amirke

    Points: 2
    Helpful Answer Positive Rating
BigBoss said:
The occupied area is just roughly 10mmx20mm, so it's not critical thought.
Click to expand...

What do you mean by "occupied area"? I refer to the required ground plane size for such devices (much larger than device size), not to the device size. From my epxerience in designing miniaturized 915MHz antennas, you always co-design these antennas for a given ground plane size. 10mm x 20mm total sounds a bit small to me.
 

chip if you have little room,
printed ONLY if you have plenty of board room.
Due to the close proximity, you could easily detune the printed style antenna with object proximity. The chips, being very high dielectric constant material, are more immune to proximity detuning.
 



the device is small, but its performance depends on that large PCB ground for a total size of 20mm x 49.5mm. You cannot use that small device to build a small PCB.
Click to expand...
All monopole antenna designs involving a small ground plane can be better analysed as asymmetrical dipoles. The actual impedance depends on the geometry of both parts. That's obviously also true for the suggested ceramic antenna, the manufacturer requires a specific ground plane shape to achieve 50 ohm impedance. A different shape gives a different impedance and thus requires a matching network for optimal performance.
 

This is a common discussion and it hinges on the performance of a conventional antenna and one that is electrically small. It is further complicated because your particular separation is relatively short at one meter.

Conventional antennas are not well characterized when they are separated by only a few wavelengths. The radiation pattern of an antenna will not solidify until there are many wavelengths of separation. On the other side the separation is too great for capacitive or inductive coupling to play a strong part. Those mechanisms do not function well until separation is well under a wavelength. I would expect to see significant variation between the devices as they are moved relative to each other and a prediction of the absolute coupling even if they are not moved is difficult to anticipate.

As a WAG but I would anticipate that a chip antenna would bb on the order of -10 dB worse than a dipole or monopole. The ground plane or enclosure will play a big part in this. I would not be surprised to see coupling on the order of -30 dB regardless of the antennas or their orientation.

You do not say how much power you intend to radiate nor what your receiver needs for good connectivity. Those are key numbers to keep in mind. If you have plenty of link margin go ahead with a chip antenna.

I would do some testing with a VNA to see what I was likely face in a real system. These dimensions and the relatively low microwave frequency would not be difficult to mock up in a lab. The result would go a long way to giving you a good characterization of your particular situation. One experiment is worth a thousand expert opinions.

By the way could you just use a good transmission line, like a flexible coax to make your life easier?
 

Status
Not open for further replies.

Similar threads

Menu
Log in
Register
Cookies are required to use this site. You must accept them to continue using the site. Learn more…