Shifted Resonant Frequency between HFSS simulation and PCB antenna

:grin: Thank you Sir :-o, but tomorrow I have classes for the whole day, so I am going to do that on the next 2 days. Have a nice day Sir.

These are my circuit for the multiplier voltage and the impedance value of it. I cut off the trace to antenna, the capacitor in the end of circuit has a value 0.1uF.





- - - Updated - - -

The impedance of multiplier is 26.9Ohm.

- - - Updated - - -

May I use this for the impedance matching circuit Sir (RL is the impedance of multiplier and Rg is the impedance of antenna).

You are back? Welcome!
I will look at your data in 2 or 3 hrs, need to finish some work here.
Meanwhile have some questions:
Where do you study?
Are you the only guy doing RF projects or more students work with RF?
Would you guys benefit having volunteer instructor there?
I want to take sabbatical and go traveling. In particular was thinking to visit Vietnam. If your school would organize me place to stay and cover food I can stay for a semester and help your school as volunteer instructor. Can you talk with your professors about me?
Since you have good equipment there we can make very good antennas and rf circuits.
: )

I am very glad to here about it.

1. I am Vietnamese but now I am studying in Pukyong National University, Busan, Korea Sir. My professor, he has being taken sabbatical in Texas USA for one year.
2. Just only me doing RF in laboratory Sir. In my laboratory, the fund is limited, one month for one time odering PCB and compnents but in that time we can buy necessary things for the whole month.
3. I can say to my professor about you but I don't know how to say.
Anyway, I'm really thanks for your support.

Sure, ask your Professor. In Busan? I would prefer to visit HoShiMing ... but it can still be good time spent in Busan.

I looked at your plots of the multiplier - very good, it will be easy to match.
Matching network calculation - calculation is correct but Rg and Rl numbers? I looked at Antenna plot form 2 days back it was 56Ohm + 7pF. Unless you retuned it better after that and i do not know. Please confirm. If you still have 7pF in need to be compensated with inductance ~4.7nH to do conjugated match. So the total network will be T: starting from multiplier 4.7nH inductor in series, 3.3pF capacitor shunt to Gnd and another 4.7nH inductor in series. Then Antenna.
Bare in mind component values are not very accurate at these frequencies. I suggest you install the network you calculated, see what NA shows and adjust. You can send me plots again.
Where to install the components. Series components - make cuts in trace and place it there. Do not overheat traces or they will come off. Shunt capacitor - clean green mask from traces and solder it between the traces.

- - - Updated - - -

Sure, ask your Professor. In Busan? I would prefer to visit HoShiMing ... but it can still be good time spent in Busan.

I looked at your plots of the multiplier - very good, it will be easy to match.
Matching network calculation - calculation is correct but Rg and Rl numbers? I looked at Antenna plot form 2 days back it was 56Ohm + 7pF. Unless you retuned it better after that and i do not know. Please confirm. If you still have 7pF in need to be compensated with inductance ~4.7nH to do conjugated match. So the total network will be T: starting from multiplier 4.7nH inductor in series, 3.3pF capacitor shunt to Gnd and another 4.7nH inductor in series. Then Antenna.
Bare in mind component values are not very accurate at these frequencies. I suggest you install the network you calculated, see what NA shows and adjust. You can send me plots again.
Where to install the components. Series components - make cuts in trace and place it there. Do not overheat traces or they will come off. Shunt capacitor - clean green mask from traces and solder it between the traces.

- - - Updated - - -

You may wan to test your multiplier before connecting antenna. If you have signal generator you can apply 915MHz signal to the connector (Antenna disconnected), set amplitude low as you expect to be on antenna terminals, and measure multiplier output. This way you can calculate efficiency of multiplier and see what kind of voltage it outputs. Just do not overload the diodes.

>> I can say to my professor about you but I don't know how to say.
If you want I send you my resume. I am not a professor, instead I have my little company do antennas and rf design. Send me you Email ad I send it to you.

That's right Sir, +7pF. So I will soldering the matching circuit but now I not have the equivalent components (not have 4.7nH inductor). My email is dangtrungnam2809@gmail.com.
I am not have a signal generator Sir. I will find it in another lab.

So you need the components. Bare in mind that you do not know the exact values of components that will work for you. This last step is always done empirically: you solder the component and measure result. Hence you need to acquire ranges of components. It is called matching kits.
For example:
https://www.digikey.ca/product-detail/en/johanson-technology-inc/L603DC/712-1451-ND/1915219
- inductors
https://www.digikey.ca/product-detail/en/johanson-technology-inc/S603DS/712-1159-ND/1786608
- capacitors
This is what I use last 3 months. There are many others from different manufacturers.
When selecting components make sure the components work at your frequency (generic capacitors will not work at 900MHz) I suppose you know that.

I will send you EMail with my resume and answer questions.

Ok, I understand Sir.

I has justed received some components on today. These are my results. I have the impedance of antenna is: 55-j24.85(7pF).





I hope the result will be like that 55+j24.85 but I cannot get it.

The plot looks good. You are not too far off.
I suggest:
Calibrate NA at the end of the cable again:
Remove all matching components.
Re-measure input impedance of multiplier. Confirm it did not change much since last time.
Design matching network with only two components C and L2
Design it so R – real part of the impedance - is 55 – 57 Ohm
Use 3d component – L1 to adjust X - reactance - to be complex conjugate to that of antenna (~+j30)
If the calibration is correct addition of L1 should not result in much change of R part of the Impedance, only X.
Send me the picture how you installed matching components and NA plots of steps of matching network design

I apologized because I forgot calibration before measuring. That is my big mistake.

This is an impedance and a return loss value of antenna:

Impedance is 41.2+j17.787

Impedance of Rectifier circuit (removed all of matching components)

Impedance of rectifier: 99.169 – j30

This is how I calibrate the Port 1 of NA


I use this website to calculate for me the impedance circuit:
**broken link removed**

I tried so many cases and I cannot control the results, it worse and worse.

This is one of my matching circuits (in the new one when Zload > Zsourse)


I just use Csource = 1.8pF and this is the result.


I will keep going and try to matching it.

- - - Updated - - -

My circuit board and the result with C = 1.8pf and Lload = 6.8nH.



The real part is nearly 41Ohm. When I add Lsource the value changed so much and I cannot control it. How can I adjust Sir?

- - - Updated - - -

How about this result Sir, I think I can use it 44.3 - j36. I achieve it when just use only C=2.2pF.

All together it looks good.
The plot 1p8and6n8.PNG is the best,to the point of being acceptable. It is very good match to 50. You wanted 56 Ohm according to your last measurements but what is the most recent measurement of Antenna?
Please re-do this measurement and plot Smith Chart and S11.

- - - Updated - - -

On Picture P_20160411_105319.jpg I see lots's of goop ... resins - wash it all out. Make sure your outside surfaces of your circuit are conductor or PCB coating.

Thank you Sir for the valuable suggestions.
In #51 I measured the impedance of antenna again and the result is 41.2+j17.787. In the previous measurement I forgot to calibrate the NA, this lead to wrong evaluation. The privious results is 55-j29. Now it changed to 41.2+j17.787 after calibration. Sorry Sir about that inconvenience.
In my opinion, I think the plot with C=2.2pF is the best because of it is 44-j36 and nearly 41.2-j17.787 (impedance of antenna).
If I make a new PCB board, I will use the same size of board, the same antenna after tunning and the same size of the traces or I just keep the size of tunning antenna and get the new board shape then after that measure again and choosing the proper matching components?
I learned a lot of things from that experiment. I cannot say thank you enough.

Another detail: what is the geometry of your feed cable these days? Remember two weeks back we wanted to make it strictly perpendicular - If feed cable is not perpendicular anymore you can get any kind of readings without knowing it.

The same measurement Sir but in the last time I forgot calibrate NA. I just press the "PRESET" button and measure. But now before measuring I calibrate first and then measure, the same way Sir instructed me.

To confirm your NA measurements you can add another type of measurement: connect Antenna, Matching Network, Multiplier all together. Bard is not operational and you can measure output power. Use signal generator.

Fix the measurement spatially: make support/holder for your board so it is always mounted the same way. Fix position of radiating antenna.

Vary components of matching network 2 - 3 values away from optimal values you determined. Determine components for which output power peaks. Does it correspond to optimal matching network as seen by Network Analyzer?

What is the intended application of your board in real life - where it should be mounted in real operation? Mount the board as it supposed to and re-measure impedance. You find out that impedance of your antenna will change significantly.

Also I can propose physical changes to your board that will improve it's efficiency from what you got now.

- - - Updated - - -

I predict you will improve performance of your board if you orient circuit as shown on attached picture (also bunch it up so it fits on the board) and further improve performance if you remove solder mask(green paint) from antenna radiators. - on both sides of the board

... can not upload picture. website does not work.

I see Sir. I will re-draw the circuit. I still do not understand "draw antenna electrodes on both sides of the board". In my understanding, that mean will I draw the antenna in both side of the board?

I intend to use this rectifier to supply power for sensors system board (IC tag chip, MCU and acclerometer sensor). Energy Harvesting ---> IC DC-DC step up (IC boost - LTC3105) ---> sensors system board. As you mentioned the impedance will be changed significantly so that I need to re-measure impedance after connect them together and I know how to do it.

- - - Updated - - -

The result is good after I connect antenna, matching circuit and Multiplier together. Before tunning antenna and have a matching circuit, the output voltage is 1.6V and the distance from signal generator (I used RFID reader 1W output) is ~80cm, but now the distance is 3.5m and the output voltage is 1.4V. Around 2m, output voltage of the rectifier is 1.6V.

>> "draw antenna electrodes on both sides of the board"
Right now your antenna is drawn only on top side of the PCB. RF currents run on both surfaces of the foil and this generates E-field. From the underside of copper foil E-field passes through the thickness of PCB, through FR4 material. This results in some (small) loss. If you make exact same antenna pattern on the other side of PCB and connect both sides with vias (use many small vias, every 2 - 4 mm) there will be no field between top and bottom foil, hence you recover small loss you have now when field passes through PCB.


3.5 m - that's a long range.

It looks like your project is done. Or do you need even better performance? What is your goal in this project? (other than getting good marks.)

Yes, the performance is good 3.5~4m. A long range. But I think I am going to try your recommendation in the next version. It is very exciting Sir.

Did you estimate the expectable received RF power with 1 W transmitted at 3.5 m distance using free space path loss? I get about 100 mV into 50 ohms (with dipole gain at both sides). Even a lossless rectifier can't source more than 0.45 V into 1 k load with this input.