QRP Labs 10W PA (3) -- Designing 40m LC Low Pass Filter (Updated)

 To design the 40 m LC filter for the 10W amp output,  I decided to use the 7 pole Chebyshev table found on Page 6-43 of the 1981 ARRL Handbook.  Click the diagram to enlarge.  

The information can also be found in the 1979 QST article "Low-Pass Filters for Amateur Radio Transmitters", by E. E. Wetherhold, W3NQN.  The design for 40 meters supposedly uses standard capacitor values for C1, C3, C5 and C7, but I don't have any 300 pF caps in my parts collection, so I'll use 82 pF plus 220 pF to get close. I have T50-6 cores in my parts store, so I'll them.  The table of parts is: 

• C1 & C7 -- 300 pF NPO, approximate with 82 pF in parallel with 220 pF
• C3 & C5 -- 680 pF NPO
• L2 & L6 -- 1.37 uH, 18T on T50-6 core, using 26 AWG enamel wire
• L4  -- 1.62 uH,  20T on T50-6 core, using 26 AWG enamel wire

The figure below shows the LTSPICE circuit for simulation (click to enlarge). The AC voltage source is set up with a source impedance of 50 ohm. 

The response of the filter is shown in the plot below (click to enlarge).  The attenuation at the 14 MHz, the lowest possible 2nd harmonic is -40.5 dB.

At the fundamental frequency of 7 MHz, with 10W power at the load, the current through the inductors will be: 

$I_L=\sqrt{10W/50\Omega}= 447 mA$

Using the micrometals.com design page the losses in the core and wiring are predicted to be 100 mW for each core.  This will generate a minor and tolerable 300 mW of loss in the filter. 

 

Some useful LC filter links: 

kitsandparts.com toroid turns calculator

Online filter design tool

Elsi filter design download page

GQRP Club app note on LP filters

Update: Construction and Testing

The photo below shows the 7 pole Chebyshev LC filter as constructed.  I soldered the ground end of the shunt capacitors to a one-sided PC board.   I don't have any way of cutting the board right now.  You can see that the end capacitors are two capacitors in parallel (82 and 220 pF). I wound the inductors and then tested them on my inductance meter at 100kHz. I fiddled with the spread of the wires on the inductor to get near the required inductance, then fixed the windings down with Q-dope.

The full power test of the filter is shown in the photo below.  Click to enlarge. The signal generator is set at 4.2 volts @ 7.0307 kHz.  Peak to peak voltage at the input of the amp is measured by the scope as 3.8V but you can see in the scope photo, the yellow trace, that there's a fair bit of noise. The output after the LC filter is the blue trace at 63.2Vpp, which works out to 10.2W.  I can see some distortion on the output, a little bit of symmetric flattening at the peaks.  This is probably 3rd and 5th harmonics.  I don't have a way to measure harmonics  right now.  Gain is:

$=20\cdot \log(63.2Vpp/3.8Vpp) = 24 dB$

 

Now onto designing and building the 20m filter.