Prev: More items that don't deserve patents
Next: Ceramic Capacitor Failure
From: Joerg on 30 Mar 2010 18:04
Nico Coesel wrote:
> Joerg <invalid(a)invalid.invalid> wrote:
>
>> Nico Coesel wrote:
>>> Every now and then you find something new on your path...
>>>
>>> I'm working on a wireless device which is going to use a PCB trace as
>>> an antenna. The bitrate is quite high (approx 250kbit) and we decided
>>> to use the 433MHz band because it has little restrictions. The main
>>> problem is the antenna. I managed to put a 115mm (4530 mil) long
>>> trace (1/4 labda monopole) onto the 50mmx80mm (2" x 3.15") board to
>>> form an L shaped antenna. Its fed from a 50 Ohm transmission line
>>> which runs over a reasonable big ground plane (top and bottom stitched
>>> ). So far I was able to gather some info from applications notes and
>>> so on.
>>>
>>> Ofcourse there is more on the board than just the antenna although I
>>> made sure the antenna runs far away from the dense populated areas.
>>>
>>> Now the real problem is going to get the antenna tuned. As far as I
>>> can see that takes two steps: getting the antenna to resonate at the
>>> desired frequency and matching the impedance. The gear I have
>>> available is a spectrum analyzer, an oscilloscope, a directional
>>> coupler and an HF generator.
>>>
>>> I know a vector network analyzer would be the right tool especially
>>> for determining the mismatch, but I'm wondering if I could do without.
>>> If not, I've found this kit.
>>>
>>> http://www.sdr-kits.net/
>>>
>>> Looks nice and affordable any comments?
>>>
>> Since it was designed by a ham it's probably alright. But you don't
>> necessarily need one for this job. Resonance can be achieved by trial
>> and error, probably you are going to add inductors until the antenna has
>> the correct electrical length.
>
> Yes.
>
>> For matching you might want to borrow a
>> UHF-wattmeter or SWR-bridge from a ham operator. But it must work on the
>
> I was planning to use the directional coupler and the spectrum
> analyser (with tracking generator) to measure the reflected power
> (sort of SWR). From what I understand the SWR doesn't say whether to
> add a capacitor, resistor or inductor. OTOH, if the antenna is in
> resonance it is supposed to be resistive. The transmission line is
> supposed to be resistive as well so the most logical thing to do would
> be to add a resistor.
>

A resistor will burn off power but if you have excess you can afford
that. You can use the directional coupler but since you seem to have no
cable or significant length of trace from xmit chip to antenna why
bother? As long as you don't load the xmit chip too much. Mostly they
tell you, by leaving a blister on your finger ;-)


>> 70cm band which not all of them do. If he has a dipmeter that would be
>> nice, too, makes finding resonance a breeze.
>
> Sounds like a good idea. The problem is the 'ham operator' part :-)
>

Look for roofs with "strange" antennas on there :-)

--
Regards, Joerg

http://www.analogconsultants.com/

"gmail" domain blocked because of excessive spam.
Use another domain or send PM.
From: Dave Platt on 30 Mar 2010 18:31
In article <4bb251f3.96445937(a)news.planet.nl>,
Nico Coesel <nico(a)puntnl.niks> wrote:

>I was planning to use the directional coupler and the spectrum
>analyser (with tracking generator) to measure the reflected power
>(sort of SWR).

Yes, that will work.

> From what I understand the SWR doesn't say whether to
>add a capacitor, resistor or inductor.

Correct... a single SWR measurement cannot tell you this. Multiple
SWR (or reflected-power) measurements at different frequencies *can*
tell you a lot about what you need to do.

You really *never* want to actually add a resistor... they're lossy.
It's possible to change impedances without adding substantial loss...
it's done by creating a reactive impedance transformer (e.g. an
L-match), and that's a much better approach.

> OTOH, if the antenna is in
>resonance it is supposed to be resistive. The transmission line is
>supposed to be resistive as well so the most logical thing to do would
>be to add a resistor.

That's a poor choice... it may be logical, but you won't really like
the results. It may lower the SWR, it may raise the SWR, and
in any case it will dissipate RF power as heat (wasting transmit
power, and reducing receiver sensitivity).

If you can adjust the frequency that you're feeding into the test
setup, and tweak it a bit on either side of your endpoint frequency,
you can use something along the lines of the following approach.
You'll want to trim to resonance, and then match (probably).

- Trim the antenna to something close to the correct resonant
frequency.

- Tweak the frequency back and forth to find the SWR minimum.

- If the SWR minimum occurs below your design frequency, then your
antenna is too long. Trim it, or add a few pF of capacitance in
series with the feedpoint.

- Conversely, if SWR minimum is above your desired frequency, the
antenna is too short. Lengthen it, or add a bit of inductance in
series with the feedpoint.

Lather, rinse, repeat, until you've got the SWR minimum very close to
the design frequency. At this point, the antenna (plus any reactance
you added) will be resonant at the desired frequency - it will present
a resistive impedance.

The next step will be to match it. My guess is that the type of
antenna you've specified (an "L", not far from the ground plane) is
going to have a feedpoint impedance which is significantly below 50
ohms. As a result you'll need to raise its effective impedance
somewhat.

The way to do this will probably be with an L-match circuit, which
will take only two components... one inductor and one capacitor. What
you would do is:

(1) Add some reactance (either capacitive or inductive) in series with
the antenna feedpoint - that is, between the transceiver and the
antenna system.

(2) Add some reactance of the *opposite* type (e.g. an inductor, if
you stuck a cap in series with the antenna) between one side of
the series reactive component, and ground... that is, in shunt.

If you're matching a too-low antenna impedance (as I suspect you
are), the added shunt reactance would go on the transceiver side
of the series reactance. If you're matching a too-high antenna
impedance, stick the shunt between the antenna side of the series
reactance, and ground.

It's often possible to combine the series matching reactance, with the
reactance that you added to tune the antenna to resonance... the
impedances add together.

The exact amounts of reactance (positive and negative) which you would
need to add, could be calculated exactly if you knew the actual
antenna impedance. Since you don't, you'll probably have to
cut-and-try. They'll be somewhere in the range of zero to 50 ohms
(reactive) at your design frequency. SMT components glues to
toothpicks or matchsticks, and then pressed down onto pads on your
prototype PC board, can let you iterate through multiple combinations
of components fairly quickly.

If you do everything right, what you can end up with is a
two-component network (one reactance in series with the antenna, one
in shunt to ground) which matches the antenna to a 50-ohm-resistive
(or whatever you choose) feedline.

If you wish (and if you have PC-board space) you can then replace the
lumped SMT components with PC-board stripline equivalents. so that the
matching network is actually built into the PC board itself.

--
Dave Platt <dplatt(a)radagast.org> AE6EO
Friends of Jade Warrior home page: http://www.radagast.org/jade-warrior
I do _not_ wish to receive unsolicited commercial email, and I will
boycott any company which has the gall to send me such ads!
From: Al Borowski on 30 Mar 2010 22:54
On Mar 31, 4:11 am, n...(a)puntnl.niks (Nico Coesel) wrote:
> Every now and then you find something new on your path...
>
> I'm working on a wireless device which is going to use a PCB trace as
> an antenna. The bitrate is quite high (approx 250kbit) and we decided
> to use the 433MHz band because it has little restrictions. The main
> problem is the antenna.

[...]

I own one of these: http://www.timestechnology.com.hk/ - they're a
neat little tool for this kind of work. Personally I'd just run the
trace and leave room for 3 surface-mount components in a pi-
configuration for tuning - as well as an RF connector for prototyping.
Then you can plug the T100 and come up with a good matching network.

Or, just try one of these: http://www.antennafactor.com/Products/Embeddable/SP-Series-The-Splatch

Cheers,

Al
From: oopere on 31 Mar 2010 05:19


Nico Coesel wrote:
> Every now and then you find something new on your path...
>
> I'm working on a wireless device which is going to use a PCB trace as
> an antenna. The bitrate is quite high (approx 250kbit) and we decided
> to use the 433MHz band because it has little restrictions. The main
> problem is the antenna. I managed to put a 115mm (4530 mil) long
> trace (1/4 labda monopole) onto the 50mmx80mm (2" x 3.15") board to
> form an L shaped antenna. Its fed from a 50 Ohm transmission line
> which runs over a reasonable big ground plane (top and bottom stitched
> ). So far I was able to gather some info from applications notes and
> so on.
>
> Ofcourse there is more on the board than just the antenna although I
> made sure the antenna runs far away from the dense populated areas.
>
> Now the real problem is going to get the antenna tuned. As far as I
> can see that takes two steps: getting the antenna to resonate at the
> desired frequency and matching the impedance. The gear I have
> available is a spectrum analyzer, an oscilloscope, a directional
> coupler and an HF generator.
>
> I know a vector network analyzer would be the right tool especially
> for determining the mismatch, but I'm wondering if I could do without.
> If not, I've found this kit.
>
> http://www.sdr-kits.net/
>
> Looks nice and affordable any comments?
>

Maximizing power to a fixed load is the same as maximizing voltage to
the load. If you have a spectrum analyzer you may easily probe the
signal at the antenna feed with a "high valued" series smd resistor (you
can even calculate the associated loss). Next, you can try different
matching networks and see if you can get near your transmitter
specifications. Of course, this is... time consuming.

Otoh,the poor man's vector network analyzer is the "slotted line". If
you are able to set up a "long" (>lambda/2) transmission line of known
Zo, you may record the level at some (say 10) points and compute the
reflection coefficient ro: abs(ro) from the vswr and angle (ro) from the
positions of the minimum (or minima). A matching network is then
straightforward. Comment: there are (better) alternatives to resonate
(first) and match (later), such as matching the real part of R o G
(first) and the cancelling jX or jB.

An alternative to the kit would be to try to reproduce this
http://saturn.uni-mb.si/~dogsa/ne_old/gradivo/Low%20Frequency%20Circulator-Isolator.pdf
with "better" devices.

Pere
From: John Walliker on 31 Mar 2010 06:05
On 30 Mar, 19:11, n...(a)puntnl.niks (Nico Coesel) wrote:
> I managed to put a  115mm (4530 mil) long
> trace (1/4 labda monopole) onto the 50mmx80mm (2" x 3.15") board to
> form an L shaped antenna. Its fed from a 50 Ohm transmission line
> which runs over a reasonable big ground plane (top and bottom stitched

Try to keep the gap between the trace and the ground plane as wide as
you can, especially at the end furthest away from the feed point, as
this will make matching easier. As already mentioned, first trim the
length and only then use matching components. Again, as already
suggested, put the footprints for a pi network on the pcb so that all
options are available to you.

>The gear I have
> available is a spectrum analyzer, an oscilloscope, a directional
> coupler and an HF generator.

I have used this combination many times (using a tracking generator)
and it will certainly tell you when you have got it right. Knowing
the direction in which to make changes is harder though without a VNA.
(I now have a second hand 8753A and using this does save a lot of
time.) Aim to get a return loss of better than -10dB, but if you can
only achieve around -6dB you will still probably be happy with the
system performance.

Don't forget that the housing, even if plastic, may affect the tuning.

A useful check after you have matched the antenna is to wave your hand
around nearby while watching the reflected signal. You should see
significant variations in return loss as the signal reflected from
your hand is received by the antenna. If this doesn't happen, a lot
of the output power may be being absorbed by matching network losses
rather than being radiated usefully.

John
First  |  Prev  |  Next  |  Last
Pages: 1 2 3 4 5
Prev: More items that don't deserve patents
Next: Ceramic Capacitor Failure