Continue to Site

# Designing a broadband matching network

Status
Not open for further replies.

#### bingjiang99

Dear all,

I have a circuit block ---R---C---L---, where R (5 ohm) is small, L is large, and C is small. The total impedance Z = R at 915 MHz (L and C resonate). Since the imaginary part can be very large, how can I design a broadband matching network? I have to point out that we can't do anything to the block itself.

Best regards,

How wide bandwidth do u want to achieve for the matching? and the S11? For achieving above 10% band width bandpass filter maybe need for the matching..
Since R is so small, it's hard to do matching..

### bingjiang99

Points: 2

Around 10% is good enough for me. S11 should be less than -10 dB. The big problem is that how to design this bandpass filter with such a load? The imaginary part can be significantly bigger than the real part.

cooleyn said:
How wide bandwidth do u want to achieve for the matching? and the S11? For achieving above 10% band width bandpass filter maybe need for the matching..
Since R is so small, it's hard to do matching..

smith.exe smith chart

You could refer to some filter design/synthesis book. You have to put L and C as part of bandpass filter.Since R is so small(Q valueis high), I think it's difficult.If you accept loss matching, it becomes more easy. What're the specific values for L and C?

### bingjiang99

Points: 2

Yes, I know that it can be vey difficult. L is 35 nH, and C is 0.1 pF. The problem is that we can't seperate L and C from the block.

cooleyn said:
You could refer to some filter design/synthesis book. You have to put L and C as part of bandpass filter.Since R is so small(Q valueis high), I think it's difficult.If you accept loss matching, it becomes more easy. What're the specific values for L and C?

Can you put elements in parallel with the filter?
Or add a resistor in series? More BW and more
passband attenuation.

First plot you Filter out on a Smith chart then put
Q lines on the Smith Chart and add components that
1) keeps the impedance in the Low Q region and that
will bring your filter to a Zo match.

### bingjiang99

Points: 2
dsl hybrid matching network

yes, I can put elements in parallel with the filter, but I can't add a resistor.
I can't afford to lose power.

I have no problem with these principles. I need a method to design this filter. Thanks.

Element_115 said:
Can you put elements in parallel with the filter?
Or add a resistor in series? More BW and more
passband attenuation.

First plot you Filter out on a Smith chart then put
Q lines on the Smith Chart and add components that
1) keeps the impedance in the Low Q region and that
will bring your filter to a Zo match.

Yes, I know that it can be vey difficult. L is 35 nH, and C is 0.1 pF. The problem is that we can't seperate L and C from the block.

L = 35 nH, C = 0.1 pF, R = 5 ohms does not resonate at 915 MHz. Resonance is at 2690 MHz.

smith.exe matching

Sorry, I made a wrong calculation for C, it should be around 0.9 pF.

VSWR said:
Yes, I know that it can be vey difficult. L is 35 nH, and C is 0.1 pF. The problem is that we can't seperate L and C from the block.

L = 35 nH, C = 0.1 pF, R = 5 ohms does not resonate at 915 MHz. Resonance is at 2690 MHz.

low q wide bandwidth matching

Hi all,

Find this topic interesting, however i not really sure why the matching is diffcult. From my basic understanding, it would just be made with the movement on the smith chart and its corresponding passive values. I understand that for maximum voltage transfer, we would like a large load, but since we are mentioning about Sparameters, we are interested in power transfer. As long as the matching is done up we should get the maximum power transfer. Maybe i'm wrong or i have taken enough factors into considerations? Would like to hear the views of the gurus out there.

Thanks and wish everyone have a nice day.

Regards
Derek

Because the reactive part is much bigger than the real part (i.e., Q is big), the circuit itself offers very narrow bandwidth. Theoritically, we need to cancel the reactive part all the way along the required bandwidth during the matching. It is not the problem of power matching. It is the problem that how we can remain the maximal (or reasonable) power matching within the required bandwidth.

derek_lkm said:
Hi all,

Find this topic interesting, however i not really sure why the matching is diffcult. From my basic understanding, it would just be made with the movement on the smith chart and its corresponding passive values. I understand that for maximum voltage transfer, we would like a large load, but since we are mentioning about Sparameters, we are interested in power transfer. As long as the matching is done up we should get the maximum power transfer. Maybe i'm wrong or i have taken enough factors into considerations? Would like to hear the views of the gurus out there.

Thanks and wish everyone have a nice day.

Regards
Derek

What type of circuit you are using?
If it is for example microstrip you my try design some matching network using transformer lines, rather then discreet elements.

### bingjiang99

Points: 2
matching pi network at first tune the

It is a similar antenna. However, I have no space for transformer lines. By the way, what is the advantage of transformer lines?

ND said:
What type of circuit you are using?
If it is for example microstrip you my try design some matching network using transformer lines, rather then discreet elements.

genesys automatic matching

By my opinion, it is easier and cheaper (especially on higher freq.).
In your case is even more distinctive, because you will need to use elements with very high Q, although your are design circuit on very high freq.

double matching network mhz

I have no problem for narrow band matching. How to get the broadband is my headache.

ND said:
By my opinion, it is easier and cheaper (especially on higher freq.).
In your case is even more distinctive, because you will need to use elements with very high Q, although your are design circuit on very high freq.

bode-fano

you can try and use double L-matching network or even triple L-matching network. this will result in lower Q and thus bigger bandwidth

I already tried. Because the reactive part is so large, I can't get bigger bandwidth, even with 3 L-matching network.

Jim cage said:
you can try and use double L-matching network or even triple L-matching network. this will result in lower Q and thus bigger bandwidth

broadband matching of electrically small circuits

you can simulate it with agilent ADS which include the tuning function you can get the values you need ,it's important to pay attention to the RF PCB layout. for broadband matching ,i think the PI network is the best ,you can use 2rd PI network

### bingjiang99

Points: 2
bode fano criterion example

Thanks. However, I have no ADS available,

MW+RF=Netpig said:

you can simulate it with agilent ADS which include the tuning function you can get the values you need ,it's important to pay attention to the RF PCB layout. for broadband matching ,i think the PI network is the best ,you can use 2rd PI network

I you can use an additional matching network out of the block without doing any changes to the original matching circuit, and the matching circuit may be anylzed by a smith chart tool(smith.exe).

Status
Not open for further replies.