Narrow Band Filter design @ 40MHz

[adnan012]

hi,

I need a sharp band (few 10Khz) 40 MHz filter with simple design (3rd order). I always find it difficult to tune the filter properly.

Regards

 

[pragash]

are you looking for chip filter solution or looking to design the filter yourself? if you looking to design it yourself, what help do you need?

 

[adnan012]

I want to make filter using discrete components. I want help in layout, shielding etc. Can i tune the filter using oscilloscope and a DDS function generator ?

 

[BigBoss]

It will maybe an old technique but I was designing such filters using with EM coupled high Q values ( silver plated ) mechanically tunable inductors or fixed inductor and a piston trimmer variable capacitors
Because here the coupling is very important.It won't be neither high nor low.I obtained few MHz Bandwidth but few kHz is roughly possible..
In fact, 40MHz is a similar to Amateur Radio Communication Band so if you search Amateur Radio Enthusiasts' websites, you will find lot of docs to be inspired.

 

[adnan012]

Thanks for reply

Do simulation works (ADS or any other tool)? I was able to make filter using smd inductor and capacitors (0805). Q of inductor was around 60 to 80.
Another issue is hand effect while tuning ver-caps.

 

[BigBoss]

Thanks for reply

Do simulation works (ADS or any other tool)? I was able to make filter using smd inductor and capacitors (0805). Q of inductor was around 60 to 80.
Another issue is hand effect while tuning ver-caps.

It's possible but there are some constraints.

-SMD Inductors are not Adjustable
-Coupling between -especially inductors- will create trouble ( layout problem )

There are other special techniques such as Printed Coupled Inductors ( like an transformer ) resonated at desired frequency by adding extra capacitors.
But the Bandwidth is really narrow, I'm afraid the BW will be larger than expected.

 

[volker@muehlhaus]

I need a sharp band (few 10Khz) 40 MHz filter with simple design (3rd order).

Have you checked if this bandwidth is possible in theory with 3rd order, and what component Q factors you need?

 

[adnan012]

Thanks for reply.

There is no requirement for bandwidth but i need maximum attenuation out side the required band (40MHz).

How can i shield inductors?

 

[pragash]

hi adnan,
you can get initial guess component values from ADS. you can download s-parameter data of lumped element (L and C) from supplier website and simulate/tune in ADS. it will give you accurate results.

 

[adnan012]

I have already followed the same procedure as you mention and achieved some results. Do i need to shield the inductors ?

 

[pragash]

there is a way of laying out RF components to minimize the coupling between them. i have nvr heard of shielding for filters or any inductors. you can use RF layout technique or Printed Coupled Inductors suggested by bigboss.

 

[volker@muehlhaus]

To see design/topology options, you can use Nuhertz Filter Free, that's state-of-the art filter synthesis software. The free version can design up to 3rd order.
**broken link removed**

 

[adnan012]

Thanks for reply

Is there any example for "Printed Coupled Inductors"?

 

[volker@muehlhaus]

I worked a lot in printed inductors: this is NOT a solution to your needs, to bad Q and not tunable.

Instead, look at the tunable shielded inductors that BigBoss mentioned. This was a typical radio component for such applications when filters were designed from discretes.

 

[adnan012]

What are the typical Q values for standard through hole ceramic capacitors?

 

[FvM]

I need a sharp band (few 10Khz) 40 MHz filter with simple design (3rd order).

In other words Q ≈ 1000. That's not feasible with intended design techniques. L quality factor is the critical point, C with Q>1000 is at least feasible (NP0 can do, mica is optimal).