| |
G8MNY > TECHNI 05.09.19 17:32l 111 Lines 4264 Bytes #999 (0) @ WW
BID : 17903_GB7CIP
Read: GUEST
Subj: Lowpass filter for 5MHz
Path: IZ3LSV<IV3SCP<IW0QNL<OK0NBR<OK2PEN<GB7CIP
Sent: 190817/0858Z @:GB7CIP.#32.GBR.EURO #:17903 [Caterham Surrey GBR]
From: G8MNY@GB7CIP.#32.GBR.EURO
To : TECH@WW
From Brian g3yrh @ GB7BMX Updated Aug 04
The following lowpass filter was built using the tables in the ARRL manual, it
is a 5 section one & seems to work ok, no definitive measurements were taken
other than measuring power at 5MHz & 7MHz, seems to cut off at 6MHz.
l1 l2
-------/\/\/\/\/\------/\/\/\/\/\-------
| | |
| | |
--- c1 --- c2 --- c3
--- --- ---
| | |
---------------------------------------
c1 = 330pF 1% silver mica
c2 = 500pF + 250pF silver mica
c3 = 330pF 1% silver mica
l1 = 17t 22swg on t50-2 toroid
l2 = 17t 22swg on t50-2 toroid
rgds
------------------------------------------------------------------------------
G8MNY comment...
Doing some analysis on a 50ê source & load, the Ls needs to be about 2uH.
Excluding losses the responce is...
MHz Loss dB
5 0.1
7.4 3
10 16
15 34
If there is a lot of 2nd harmonic to remove from the Tx (e.g. not a push pull
PA) a "M" derrived filter should be considered. This is where a capacitor
across one of the coils is used to resonate it near 10MHz to improve the
rejection.
l1 l2
-------/\/\/\/\/\------/\/\/\/\/\-------
| | |
| |_______||______|
| | || |
--- c1 --- c4 --- c3
--- --- c2 ---
| | |
---------------------------------------
c1 = 330pF 1% silver mica
c2 = 500pF + 250pF silver mica
c3 = 330pF 1% silver mica
c4 = 170pF
l1 = 17t 22swg on t50-2 toroid
l2 = 15t 22swg on t50-2 toroid
With L2 = 1.5uH & tuned with a 170pF to 10MHz then responce...
MHz Loss dB
5 0.1
6.6 3
10 >50
15 35
--------------------------------------------------------------------
Comments from Tony G3NXC@GB7SOL.#29.GBR.EU
The frequency response figures that you quote based on a 2uH inductor value are
in agreement with my own calculations. Unfortunately, though, 17t on a T50-2
toroid does not yield 2uH! The AL figure for this core, taken from the
Micrometals web site & other places, is 4.9nH/N**2 & this results in an
inductance value of 1.42uH for 17t.
Using 1.42uH as the inductor value actually results in a lower input VSWR at
5MHz than 2uH - 1.05:1 compared with 1.1565:1. The cut off frequency, as you
would expect, is higher with the -3dB point being at 8.13MHz; the attenuation
being 12.1dB at 10Mhz & 29.5dB at 15MHz.
If it is felt necessary for the inductance to be increased to 2uH then the
number of turns should be increased to 20. Another solution would be to leave
the inductors as they are & increase the capacitor values with C1 & C3 becoming
360p & C2 becoming 820p. This would result in the input VSWR being 1.12, the
3dB point being at 7.85 MHz and the 10MHz and 15MHz attenuation figure
becoming 13.8dB & 31.1dB.
I'm not convinced about the need for a notch in the response characteristic at
10MHz. The attenuation characteristics calculated for the basic filter rely on
the load remaining at 50 Ohms independent of frequency. This is not, of course,
the usual situation in practice. Assuming that the load is a dipole, the feed
impedance will be around 50 Ohms at resonance: at the second harmonic, though,
the feed impedance will rise to around 5000 Ohms. This load will transformed
by the feeder but whatever value appears as the filter load, the result will be
that would be difficult to transfer any power to the aerial through the filter.
Apart from anything else, the 100:1 VSWR would cause the feeder losses to rise
significantly. So even if some second harmonic were to 'escape' the filter,
very little would actually reach the aerial - assuming 100ft of UR67, the
attenuation of the second harmonic resulting from feeder loss alone would be
around 9dB.
All of the above apart, the physical size of the inductors means that the
filter is only really suitable for moderate power applications!
73 de Tony, G3NXC@GB7SOL
Why Don't U sent an interesting bul?
73 de John G8MNY @ GB7CIP
Read previous mail | Read next mail
| |
