Working on rf energy harvesting(915MHz). I had try to match the impedance of antenna to the rectifier using a mini VNA. After measuremet, I got the impedance of the antenna. (42.4+22.5j),rectifier Z1=(17-61j). Before matching, Vout1=320mv(-7dbm input power from signal generator), 600mv(using antenna to receive powe, 0.3meter from the signal source) . I try to match both the real part and the imaginary part. Using a impedance matching calculator , I select a serial C(1.7) and shunt L(8.2) as matching network. However, the impedance of rectifier become (7.9-15.9j), I think this maybe caused by the quality of the L,C. So, I change the value of them to make it as matched as possible according to the smith chart. However, it hard to reach the matching, so I finally match the rectifier to Z2=(43.7-13j) and the Vout2 is 200mv(-7dbm), 300mv(by antenna).
I think the reason is Z2Vout1, though Z2 is more close to 50Ω.
If I want to maximize the output power, Zsource=Zrec, so that Vout=Vsource/2. If I want more voltage, Zrec should larger than Zsource, which increase the loss of power.
Higher output voltage means higher voltage the capacitor can reach, which means the more energy can be stored into the capacitor( as the higher the capacitance, the more energy at the same voltage level)
So, In order to obtain the most voltage at output, what might be the best matching:
1.matching the real part only
2.matching the real part and imaginary part
3.matching the Z, no matter the 2 parts.
Power and voltage, what should be more important?
thanks
Note that matching assumes Zsource=Zrec* , not Zsource=Zrec
https://en.wikipedia.org/wiki/Impedance_matching
I guess you are dealing with rather narrow frequency range, therefore you can try simply 1) create strong resonance 2) minimize capacitance in parallel to diodes - that should maximize voltage at the diodes and efficiency of the rectification.
Nikolay Pavlov Yeah, I am using a 915MHz transmitter as source. I had successfully match the impedance of the rectifier to antenna but it took me lots of time. I tried to use the LC network calculate by the ADS, however, it totally not work, the mismatching even worse. So, I had to modify the serial capacitor and shunt inductor according to the smith chart, 'luckily' I got the desired impedance. Is there precise way to find the right parameter of LC network?
I would use some adjustable inductor (just coils that you can squeeze mechanically) to create the resonance.
First you guess the inductor value from estimate of relevant capacitances.
Then, you can use sweep generator to catch some resonance first e.g. at 700 MHz or whatever you managed, then you can try to bring it to 915MHz.
You may look at, I guess either series or parallel inductance might be considered.
Article Enhanced RF to DC converter with LC resonant circuit
Assuming we have same reactive power circulating in resonant circuitry, the RF voltage across the diodes will be higher if relevant (diodes+extra circuitry) effective capacitance is lower .
Higher RF voltage improves efficiency at low input power.
Nikolay Pavlov I also found that the improvement is not significant after impedance matching. In my current design, there is a strong and stabel signal source(4w), the desired distance is about 1m, at such distance the recieved power >0dbm. Is it still worth to do impedance matcing?
I would rather try to create resonance. In fact, it usually brings you reasonable matching at near the resonance. What are the diodes that you are using?
Nikolay Pavlov sms7630. When I try to matching according to the smith chart, it doesn't follow the regular. For example, if I add a serial capacitor, the point should rotate on the impedance circular anti-clockwise, but it didn't. I spent whole day trying it but still not success. Is that because the non linear behaviour of the rectifier?
Ll Lee I guess you may have very high RF amplitude. Can you try 100mV or so? Also, you don't have any output load and perhaps unnecessary high output capacitance - it may result in long simulation time.
Nikolay Pavlov Thanks, I am not doing simulation. The picture in the question is just for illustration. I am using a real circuit (PCB), same circuit and component to the picture, and doing impedance matching with real component. Not simulation. Do you mean add output load can improve the situation? I had thought about it, but I didn't try to add a load.
Hi, see this paper may be is useful: https://ieeexplore.ieee.org/document/8639080