[Lowfer] That filter can't be right!!!

[Jay Rusgrove]

Bill

With the same inductor values on both the input and output arm of the 3 pole
filter,  it is designed for the same impedance in and out. If one is trying to
effect an impedance transformation from say 1 ohm up to 50 ohms you would want
to use a "T" matching network instead.

A "T" network for the above transformation would have the following values:


Q          XL1          XL2          XC1

8            8             27.39          7.6

10         10            50.5            9.17

values courtesy Motorola RF Data Manual.

In order to get the transformation to 1 ohm a minimum Q of 8 is required. The
"T" match design has the added benefit of pretty good harmonic suppression. If
further suppression is needed an additional 3 pole 50 ohm in/50 ohm filter out
could be tied onto the end of the "T" network.

Jay

Bill Ashlock wrote:

> John A,
>
> >Not sure how bleary you were when typing that message, but the diagram >I
> >have of Lyle's transmitter shows .018 uF, which is 18 nF. The >diagram
> >posted shows 18 nF. Where did you get the 1.8 nF from?
>
> Wow. Am I glad to find out 1.8nF is the wrong value!!  The schematic that
> I've been using for about four years originally came from Lyle's Web site.
> This copy has been in my original, well worn, Lowfer notebook all this time
> and clearly shows 1.8nF (and believe me, I looked at this about 10 times
> last night, just to be sure). Well, at least I'm a bit more versed in the
> filter area for my efforts.
>
> BTW: This N=3 filter is not a real good match for a 0 ohm output PA like we
> are using. All odd-numbered filters including this N=3 are supposed to have
> the same input and output resistance, IE: 50 ohms, in and out, for the
> Lyle's values. In my Pspice simulation last night I replaced the 50 ohm
> generator resistor with a 1 ohm  resistor and this produced a 20% ripple in
> the pass band. This ripple is not a serious problem unless you are running
> experiments at different Lowfer frequencies and need a constant output. An
> N=4 or other even-numbered filter with a final capacitor to gnd is a better
> choice for this application. The basic three component values will need to
> be adjusted slightly as well. Even an N=2 (just a 63uh inductor and a 8.7nf
> capacitor) filter can be used if your stop band attenuation needs are not
> too great.
>
> Bill
>
> _________________________________________________________________
> Join the world�s largest e-mail service with MSN Hotmail.
> http://www.hotmail.com
>
> _______________________________________________
> >From the Lowfer mailing list
> Send messages to: Lowfer@mailman.qth.net
> To sub/unsub visit: http://mailman.qth.net/mailman/listinfo/lowfer