Bandstop Filter

[rf997]

Hello, I am trying to make a bandstop filter that suppresses between 1710 MHz and 1880 MHz.
I tried many different topologies but did not get any results.

What kind of topology do you think I can use to provide these values?

My S11 features:

• 1690MHz: -15dB
• 1710MHz : -3dB
• 1880MHz: -4dB
• 1900MHz: -16dB

My S21 features:

• 1690MHz: -4dB
• 1710MHz : -23dB
• 1880MHz: -20dB
• 1900MHz: -4dB

 

[BigBoss]

Specifications are too tight. It might be realized by microstrip structures but a Cavity Filter is better option.
It will be so sensitive to manufacturing tolerances.

 

[D.A.(Tony)Stewart]

 

[rf997]

Specifications are too tight. It might be realized by microstrip structures but a Cavity Filter is better option.
It will be so sensitive to manufacturing tolerances.

Unfortunately, I have to do it with a microstrip.

 

[volker@muehlhaus]

Unfortunately, I have to do it with a microstrip.

Unfortunately it will not work. It's just impossible using real world PCB materials.

(It might be possible using superconductors.)

 

[rf997]

Unfortunately it will not work. It's just impossible using real world PCB materials.

(It might be possible using superconductors.)

I understand. So, is it possible to make a bandstop filter with a bandwidth of 170 MHz at the center frequency of 1795 MHz without taking into account the parameters given above?

 

[BigBoss]

I guess what you're looking for.

Transistor Arrays

Renesas' portfolio of transistor arrays include NPN, PNP and NPN-PNP combination transistor arrays based on our complementary UHF-1 SOI process.

www.renesas.com

If you don't like, use an "Integrated RF Amplifier" from Minicircuits, Texas Insturments etc.

RF gain block amplifiers product selection | TI.com

Select from TI's RF gain block amplifiers family of devices. RF gain block amplifiers parameters, data sheets, and design resources.

www.ti.com

I understand. So, is it possible to make a bandstop filter with a bandwidth of 170 MHz at the center frequency of 1795 MHz without taking into account the parameters given above?

If you have 0.4 meter !! space, yes it's possible. But it will be very sensitive to manufacturing tolerances, believe me.
All you have to do is to design/find a filter with Adjustment Feature. Without adjustment, it will be very heavy headache.

 

[volker@muehlhaus]

But it will be very sensitive to manufacturing tolerances, believe me.

And the other question is: does this give acceptable insertion loss in the passband?
Maybe you can zoom in S21 to the range 0 to -10dB, to show what rf997 can expect.

 

[BigBoss]

There is 3-4dB difference between the target and simulated. It might be acceptable. I do not know.\

 

[rf997]

If you have 0.4 meter !! space, yes it's possible. But it will be very sensitive to manufacturing tolerances, believe me.
All you have to do is to design/find a filter with Adjustment Feature. Without adjustment, it will be very heavy headache.
View attachment 186604
View attachment 186605

Thank you, you are right. When I do tolerance analysis, there are huge shifts even for 1% tolerance. So, is it possible to realize the following specs in practice when I loose the s21 values?
new s21 specs:
1690MHz: -4dB
1710MHz: -10dB
1880MHz: -10dB
1900MHz: -4dB

 

[BigBoss]

Thank you, you are right. When I do tolerance analysis, there are huge shifts even for 1% tolerance. So, is it possible to realize the following specs in practice when I loose the s21 values?
new s21 specs:
1690MHz: -4dB
1710MHz: -10dB
1880MHz: -10dB
1900MHz: -4dB

Your filter's first response was too steep so you should loosen these specifications. For instance

S₂₁
1690 MHz=-6..-8dB
1710 MHz=-15..-18dB
1880 MHz= -15..18dB
1910 MHz=-6..-8dB

Perhaps more relaxed. Tolerance for a Microstrip Line Filter is really tight if you work with a professional PCB manufacturers. Otherwise, they will be very troublesome.
Do not trust to ordinary Chinese PCB manufacturer because they use silk screen method for prototyping and obviously this brings very large mfg. tolerance errors.
Also they use "fake substrates".