CC3000 Matching Circuit Design?

[evilheart]

i am currently on a design using Texas instruments CC3000 wifi module and i am facing problem with the matching circuit , using antenna.
too bad i am not experienced RF designer so designing my own matching circuit can be a bit risky and time consuming,
there's a matching circuit provided in the datasheet with the ACX AT8010-E2R9HAA antenna , which seems it's not on the market anymore so i used the johnson 2500AT04M0400 which is equivalent (ACX and johnson companies are partners) ,
but i have no idea why those values are used or why we are using matching circuit at the first place while the (antenna - track - CC3000) are already matched.

They say in the data sheet that the values given are for the EM board , but i don't know what that exactly means , does that just mean i need another values if i changed the antenna?
or that every circuit is unique and that these values depends other factors that i don't know yet?

if i need to design my own matching circuit , can any one please recommend a good guide/tutorial/tool for that ??

regards

 

[jossion]

Hi,

All RF devices in the market are manufactured to work for 50 ohms load impedance for some particular reasons. But some semiconductor devices made for multi purposes will not be able to give you 50ohms impedance at the output. So they say it has to be matched. Matching means converting a non 50 ohm impedance to 50 ohms , so that maximum power will be transferred. If they have given the output impedance at a particular frequency; using smith chart you can do the matching circuit. You can use online smith charts for that. Tutorials are plenty in the web for doing the same.

 

[evilheart]

thanks jossion for your reply

well , in the datasheet they specified that the device impedance is 50 ohm , they also recommend that the RF track should be 50 ohm , still in example schematic (actually it's the CC3000 evaluation board schematic).

so according to their recommendations and data there should be no matching circuit , but they uses it

 

[jossion]

Hi,

In that case the antenna may not be 50 ohms. Matching may be for the antenna not for cc3000

 

[E Kafeman]

8010 antenna have a recommended matching network to be tuned to approximately 50 Ohm and that part is integrated in the TI demo board for CC3000, together with a minor impedance adjustment to compensate for mismatch of PCB transmission-lines and chip output impedance.
If a chip is stated to have 50 Ohm impedance, so yes it is possible but it can be wide variations. Add 50 Ohm lines that also are calculated to 50 Ohm, about. Serial connected errors sums up so in the end you have practically any impedance, but not 50 Ohm.

if i need to de+sign my own matching circuit , can any one please recommend a good guide/tutorial/tool for that ??

https://www.antune.net/calibration.html
Also check the impedance matching tool at same site. I used that tool to revert existing implemented chip antenna to a general 50 Ohm interface. It is however implemented by inserting existing component values which are assumed to be ideally correct, which they never is.
If possible feed the impedance matching software with real measurements. Instruction how to do a full impedance matching, check https://www.antune.net/demo/bluetooth for an antenna matching instruction. Then do the same thing for active circuit TX and RX.

Inserted parameters gives curve to the left and by AnTune automatically calculated network resulting in a 50 Ohm interface is showed to the right.

 

[evilheart]

thanks kafeman for the great answer :grin:
, i will check this out

 

[evilheart]

Kafeman

well , it took some time , as i needed to finish other layout issues ,

i downloaded the software , and i am trying to play with it now , but i am still a bit of lost here.

what i should do exactly , what's the parameter i should maximize or minimize ??

as i understand , the yellow curve is the curve of impedances at the frequency sweep range done by the simulator , right ??

so , should i try to make the impedances at my frequency range (2400-2500 GHz) as close as possible to 50 ohm , right??
or i should to minimize the VSWR at the frequency range ??

it's my first RF design so i lack experience , on the same time i should deliver the design with minimum issues as i won't do the hardware prototype , so changes should be minimized.

that's my last layout by the way , is should follow most of the layout recommendations , the RF track length is about 8 mm , i wonder if that will make things much easier and compensate for errors on other parts.
s15.postimg.org/mi4ar3gqj/RF_layout.jpg
s29.postimg.org/5p2bh0b7r/RF_layout2.jpg

 

[E Kafeman]

The software is very flexible and can measure a lot of aspects. As a beginner, most common problem is poor calibration, really bad calibration and no calibration at all.
See instructions at https://www.antune.net/calibration.html

Conjugate matching:
Turn of automatic matching network.
1. Solder a calibrated test-cable and measure impedance in one direction, Typical from the pad for matching->active TX circuit. Save as a mem.trace and set that trace as impedance-ref.
2. Measure in the other direction in PCB. Typical pad for matching->antenna.
3. Set marker for frequency range of interest.
4. Select component size/model.
5. Set you matching circuit topology.
6. Turn on "Constant tune".
A network topology with componet-values will be shown.
Solder these according to instructions in www.antune.net.

A bit simpler matching is to just try to match for 50 Ohm.
1. Set system impedance at 50 Ohm
2. Solder a calibrated test-cable and measure in direction of interest at PCB.
3. Set marker for frequency range.
4. Select component size/model.
5. Set you matching circuit topology.
6. Turn on "Constant tune".
A network topology with componet-values will be shown.
Solder these according to instructions in www.antune.net.

Minimum check that calibration is correct by temporary shortcut at place of impedance matching component with a knife or solder blob.
At Smith-chart should then whole curve just be a single dot at 9 o'clock.