Continue to Site

Welcome to EDAboard.com

Welcome to our site! EDAboard.com is an international Electronics Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.

Problem: S11 of LCR circuit through directional coupler

Status
Not open for further replies.

Zackus

Newbie level 4
Joined
Feb 19, 2018
Messages
5
Helped
0
Reputation
0
Reaction score
0
Trophy points
1
Activity points
90
I am building a circuit to measure the S11 characteristics of an LCR tank circuit. The LCR tank circuit is connected to the next stage, so in order to probe the LCR circuit, I am using a directional coupler. The circuit is shown in the attached pdf. As shown there, S11 of the tank circuit shows the resonance by a dip. Even after the bias tee, the resonance is clearly there. However, when I measure via the coupler, the resonance is present but there are several attenuation throughout the S11 spectrum. When I measure the S11 of the coupling port of the coupler separately (shown in the right upper corner) with out port floating and in port terminated.

The coupling is close to 12 dB of the coupler, so maybe that explains -22 dB in S11. But I should still obtain the resonance dip superimposed on this S11 characteristics. Can anyone explain why do I not find the resonance of LCR circuit through the coupler?

Note: All components are from Mini-Circuits
 

Attachments

  • Directional coupler problem.pdf
    330.7 KB · Views: 109

-Bias Tee must be connected after the coil with a Low ESR Capacitor, otherwise self resonance frequency of the internal coil will be appeared as seen there.Ground connection must be done as short as possible with a high quality.low ESR decoupling capacitor.
-Terminal-C must also be terminated by 50 Ohm to prevent malfunctioning of the directional coupler.( even-tough there is a some isolation )
 

Thanks for the reply! But please clarify a few things for me because I am new to this. What are you referring to as the coil? Is it the inductor L? Are you saying that I have to put a capacitor between L and the bias tee? Actually I don't see any additional resonance in terminal B when compare to terminal A as in the graphs. What I am interested in is to see the exact same resonance frequency in terminal C. Since my VNA is connected to terminal C for this, I can't ground terminal C. By the way, how do I determine the value of the capacitance you are suggesting? Thanks again..

-Bias Tee must be connected after the coil with a Low ESR Capacitor, otherwise self resonance frequency of the internal coil will be appeared as seen there.Ground connection must be done as short as possible with a high quality.low ESR decoupling capacitor.
-Terminal-C must also be terminated by 50 Ohm to prevent malfunctioning of the directional coupler.( even-tough there is a some isolation )
 

As you see, Terminal-a and Terminal-B measurements are almost same.So, the problem is between them.
The inductor which is in the Bias-Tee must be decoupled to GND by a low ESR capacitor otherwise ( if it's not well decoupled ) it might show a resonance as seen in your figures.
 

Sorry, but I didn't understand your set-up. Bias-tee is used to supply some active part, but I don't see any: I think the only reason is that C is a varactor diode, then I suppose the circuit is a VCO.

In any case I see your graphs are labelled as S21, not S11 so it seem you measured a transmission parameter and not a reflection parameter.
You cannot connect the port 1 of a network to terminal C to measure the S11 of the tank circuit.

Please specify a little bit better how did you do to obtain the three graphs (how did you connect the ports of the network, and the content of the "next stage gray box) and what's the purpose of your measurement
 

i agree. are the graphs mislabeled?

Sorry, but I didn't understand your set-up. Bias-tee is used to supply some active part, but I don't see any: I think the only reason is that C is a varactor diode, then I suppose the circuit is a VCO.

In any case I see your graphs are labelled as S21, not S11 so it seem you measured a transmission parameter and not a reflection parameter.
You cannot connect the port 1 of a network to terminal C to measure the S11 of the tank circuit.

Please specify a little bit better how did you do to obtain the three graphs (how did you connect the ports of the network, and the content of the "next stage gray box) and what's the purpose of your measurement

First I have to clarify that yes, the graphs are mislabeled. They are all S11. I am very sorry for that mistake.

Second let me explain the purpose. I am interested in measuring the noise in the device under test which has a resistance R. In order to extract it out, I have installed R in an LCR circuit which will have a resonance that can be determined by heterodyne method that is hidden in the 'next stage'. The next stage has an amplifier, then band pass filters, frequency mixer, low pass filter and finally a read-out using a diode to achieve this purpose.

The only option for me to locate the resonance peak is to use the S11 mode of the vector network analyzer (VNA). Each of the graphs shown has been obtained by connecting the port 1 of VNA to that terminal as titled and measuring S11. This has been done in all the terminals labeled and titled in the graphs. Sorry again for the S11 and S21 confusion.

But as part of my design, it is very important that I be able to determine the resonance frequency at all times without dismantling the circuit. Therefore I use a directional coupler. I plan to use the coupler terminal to measure the S11 of the LCR circuit by hooking up the port 1 of VNA to terminal C. This is not working for me. The directional coupler itself has a strange S11 as shown in the top right corner of the page, which is already low in value. I am getting my resonance in terminal A and terminal B so I was expecting the resonance spectrum to superimpose on the S11 of the coupler. But no, there is no resonance there at 92 MHz as I was expecting (see the graph titled Terminal C).

Why am I using the bias tee? That is to satisfy another requirement - be able to measure the device resistance R at all times. The DC part (inductor part) of the bias tee goes to a constant DC current source with a voltmeter. By using the bias tee, I make sure that the DC doesn't go to the RF part of my circuit, but only to my device R.

I hope it is clear what I need. Just to reiterate, I need to see the resonance frequency at 92MHz that corresponds to my tank circuit AT TERMINAL C. Without removing the bias tee.

As you see, Terminal-a and Terminal-B measurements are almost same.So, the problem is between them.
The inductor which is in the Bias-Tee must be decoupled to GND by a low ESR capacitor otherwise ( if it's not well decoupled ) it might show a resonance as seen in your figures.

I am embarrassed to confess that I still don't understand the logic of BigBoss. Since terminal A and terminal B are almost same as I was expecting them to be, then I would think that there is NO problem between them. The resonance in the graph titled Terminal C could be coming because the combination of bias tee with the directional coupler could be leading to a new resonance. But it still doesn't explain why the main resonance at 92 MHz has vanished in the terminal C.

Thanks to all of you!
 

OK. As I said you cannot measure the S11 f the load simply connecting the port of the NA to the terminal C.

If you connect the NA to terminal C, part of the power will be reflected by the directional coupler, according to its return loss performances. Part of the signal will reach the load, attenuated by the coupling factor, then the reflected portion of the power will reach the port of the NA attenuated again by the coupling factor,. Then the reflected power by the directional coupler will sum with that reflected by the load taking also into account their relative phases. This means that the method could works using an ideal directional coupler or possibly after the calibration at the input port of the directional coupler.

You could try to calibrate the NA using the open-short-match method connecting the standards to the input port of the directional coupler (removing the bias-tee). You could also try to calibrate after the bias-tee to take into account also its contribute. The network analyzer should be set to generate an "high power" like 0 dBm in order to stay in its dynamic range.

However take into account that some power generated by the NA will reach the next stage of your circuit. I don't know if this could be a problem for you.
 

For non-euclidean complex transform reasons on the smith chart, i do not think you can tolerate the 20 dB loss of a directional coupler and see what you want to see. You will not have the accuracy. Try a 3 dB power spliltter.

ckt.JPG
 

For non-euclidean complex transform reasons on the smith chart, i do not think you can tolerate the 20 dB loss of a directional coupler and see what you want to see. You will not have the accuracy. Try a 3 dB power spliltter.

View attachment 144913

If I understand right, the splitter will send power to the right side as well. The right side is a delicate amplifier which I would prefer to keep isolated during this measurement. I still can use it as long as the power is less than -40 dBm.

OK. As I said you cannot measure the S11 f the load simply connecting the port of the NA to the terminal C.

If you connect the NA to terminal C, part of the power will be reflected by the directional coupler, according to its return loss performances. Part of the signal will reach the load, attenuated by the coupling factor, then the reflected portion of the power will reach the port of the NA attenuated again by the coupling factor,. Then the reflected power by the directional coupler will sum with that reflected by the load taking also into account their relative phases. This means that the method could works using an ideal directional coupler or possibly after the calibration at the input port of the directional coupler.

You could try to calibrate the NA using the open-short-match method connecting the standards to the input port of the directional coupler (removing the bias-tee). You could also try to calibrate after the bias-tee to take into account also its contribute. The network analyzer should be set to generate an "high power" like 0 dBm in order to stay in its dynamic range.

However take into account that some power generated by the NA will reach the next stage of your circuit. I don't know if this could be a problem for you.

I have to study more about the calibration of VNA for this purpose. I tried to make a different tank circuit and reduced the cables and adapters, and hooked everything directly to the VNA. It eventually gave better results by showing a resonance this time. That is attached below. This time I have also connected the output of the coupler to the S11 mode of VNA as well. One problem is that the resonance in Terminal C is shifted by about 15MHz from the original frequency. That is just at the tolerance, but is undesirable. After all the frequency determination is the main purpose of the coupler. So I will try to find a coupler with stronger coupling but still good isolation with the output port.

Also, while I am on it, I am curious about something in VNA since I am new to using the VNA. After connecting my device, sometimes, the correct spectrum emerges after a certain 'relaxation' time. Attached is the video. Could you tell me why that happens? Here's the video:


Thanks!
 

Attachments

  • Directional coupler problem update2.pdf
    271.4 KB · Views: 72

well, do not use a splitter at all if you want to isolate the DUT. Use a SP2T RF switch instead.
 

Status
Not open for further replies.

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Back
Top