PI Filter

Ivan thanks for the reply. And thanks for ponting the very good software for LC filters.

As regards to the actual output filter I found a circuit that used a parallel LC tuned to the crystal freq instead of a pi low pass filter.
The collector of the final transistor goes to the top of the LC through dcblocking cap and the antenna goes to a capacitive tap.

I tried it and worked ok for me, at least better, now the output is in the range quoted for the pixie, about 150mW 12V and 100mW at 9V.

Col---C--+--+--+
             !    !   !
            L    C   C
             !    !   !----ANT
             !    !   C
             !    !   !
G--------+--+--+

Hi Bernie,
I used http://www-users.cs.york.ac.uk/~fisher/cgi-bin/lcfilter to calculate the filter. For cutoff frequency 31 MHz I obtained C1 = C2 = 102 pF and L1 = 0,51 uH. Your coil seems to have one half of the right value. Go there and try it yourself.

BR from Ivan OK1SIP

1+


I found internet pages where it is said the 2 caps and the coil have to be 50 ohm at a “cutoff” frequency (around 20% higher than operating frequency).


But others said that the caps shall be 50 ohms and the coil 100 ohm (as it is two low pass L type filters in series)


 


Due to the low power output, I prefer less attenuation of the 28Mhz signal.


Any suggestion of which one is  correct, or best?


 


 


2+


I decided to go for the “all 50 ohm” and connected one 100 pf cap at either side of an airwound coil (which per diameter, turns, length, etc calculte to about 0.26 uH).

Then if I “dip” the coil/caps "ensemble" with my GDO I see a resonance point at around 44 Mhz.


 


I can figure out that, as the two caps are in series through the groundplane, making a single 50pf cap (resonance -> 160/sqrt(50*.26)= aprox 44 Mhz).


Is that an indication of the filter having cutoff at 44 * 0.707 = 31 Mhz?


 


 


3+


When the filter is not connected, the transmitter develops 5.6V rf on a 56 ohm resistor as a load (through a 2n2 cap as dc block). Measured with homebrew diode probe and dvm.


When filter is inserted between the dc block cap and the 56 ohm resistor load, the rf volts developed at the resistor is 1.6 V.


I think this translates to aprox 23mW/280mW, that is -11 dB. Whereas I was waiting 0 dB attenuation with this filter.


 


I thought that maybe the high power included power of harmonics, but not that much.


Any suggestion of what is wrong?



 

Thans


Bernie