Scott wrote:
I want to build a low pass filter for 1296 and have scaled a design out
of the handbook. I would like to use chip inductors and chip capacitors
for the filter. I was looking at thin film chip inductors (Digi-Key)
and they list the self resonant frequency. For an inductor passing
1296, what should the minimum self resonant frequency be? Is 10 GHz
enough? The filter is a 5 element Chebyshev Low Pass Filter, inductive
input (3 coils in series with a cap at each inner junction to ground),
Fco=1290 MHz, F-20dB=2600 MHz. Any other component specs I need to be
aware of?

Scott
N0EDV

Adding to what others have already posted, you can simulate the filter,
with parasitic resistances and reactances and the inevitable
transmission line effects, in the freeware RFSim99. You have to be
careful to think about the physical layout, and see everything longer
than a couple millimeters as a transmission line. The data sheets on
the discrete parts should give you a decent idea about the parasitic
R/L/C of each package. RFSim99 includes the ability to model
self-resonance and Q in inductors and capacitors without adding more
parts to the schematic; just check the "use physical model" box for
those parts you want to specify for Q and SRF.

And I'd echo what Tim wrote, too: you might want to consider a
distributed filter, especially if you can make a PC board with
reasonable accuracy. Also, if the goal is to kill harmonics of a
single narrow band, you may do better with notches at the harmonic
frequencies.

Cheers,
Tom

That's exactly what I'm trying on my present project at 902. I have a
50 microstrip line on the circuit board where the RF connector hooks up.
I will etch it to be a 1/4 wavelength at 1802 MHz (2nd harmonic) and
see how that works. If it works OK, I may try the same at 1296...I
might also go with a helical filter...they work nice but can be a bit
pricey!

Scott
N0EDV

K7ITM wrote:


Also, if the goal is to kill harmonics of a
single narrow band, you may do better with notches at the harmonic
frequencies.

Cheers,
Tom

Scott wrote:
That's exactly what I'm trying on my present project at 902. I have a
50 microstrip line on the circuit board where the RF connector hooks up.
I will etch it to be a 1/4 wavelength at 1802 MHz (2nd harmonic) and
see how that works. If it works OK, I may try the same at 1296...I
might also go with a helical filter...they work nice but can be a bit
pricey!

Isn't it possible to make a helical resonator with a helix of wire
inside a box made from soldered-together pieces of PC board?(*) A filter
could be made from a series of these. The Q might not be optimum, but
should be good enough for a lot of purposes.

(*) I actually made a two-resonator filter for 450 MHz using this method
long ago, but it didn't work well. I came across it recently and checked
it with a network analyzer -- I had no decent test equipment back when I
made it -- and found that the sections were severely undercoupled. But I
haven't tried modifying it to see if it could have been made to work as
intended.

Roy Lewallen, W7EL

Yes, that's all a helical filter appears to be...a coil inside a metal
box. I have no idea how many turns, turn spacing and anything else that
might be important to the design of one. Fortunately, Downeast
Microwave sells a bunch of different helicals for a LOT less than
Digikey, but I need the 1305 MHz set and I don't see them at DEMI, so
I'll have to fork out $26 each from Digikey (and probably will end up
using two) ;(

Scott
N0EDV




Roy Lewallen wrote:
Scott wrote:

That's exactly what I'm trying on my present project at 902. I have a
50 microstrip line on the circuit board where the RF connector hooks
up. I will etch it to be a 1/4 wavelength at 1802 MHz (2nd harmonic)
and see how that works. If it works OK, I may try the same at
1296...I might also go with a helical filter...they work nice but can
be a bit pricey!


Isn't it possible to make a helical resonator with a helix of wire
inside a box made from soldered-together pieces of PC board?(*) A filter
could be made from a series of these. The Q might not be optimum, but
should be good enough for a lot of purposes.

(*) I actually made a two-resonator filter for 450 MHz using this method
long ago, but it didn't work well. I came across it recently and checked
it with a network analyzer -- I had no decent test equipment back when I
made it -- and found that the sections were severely undercoupled. But I
haven't tried modifying it to see if it could have been made to work as
intended.

Roy Lewallen, W7EL

Offhand, I'd be tempted to use an interdigital filter. There was a nice
article about this, including a computer program in basic for the
caculations in Ham Radio magizine, Ummm... November 1985 (?).

I used this in a 2400 MHz filter I needed one time and I found the skirt
selectivity to be pretty good. Of course, like a lot of wavelength
related filters, it had harmonic responses, but a low pass filter fixed that
quite nicely.



Jim Pennell
N6BIU

--

21:05 Pacific Time Zone
Oct 20 2006

International Time
04:05 UTC
21.10.2006