Updating the hanbook BC reject filter to handle 100 watts

Transcription

1 Updating the hanbook BC reject

2 I did a crude analysis of the currents in the filter. The coupling capacitors handle the RF current and then some for 100 watts at 50 ohms. 1.4 amps. The coil current is greatest at 1.8 MHz, L2 sees 1.96 amps, L4 sees 2.83 amps, L6 sees 2.14 amps. 24 gauge wire on a T50-2 core isn't enough copper to not overheat. At 60 Hz neglecting skin effect, 24 gauge copper is rated at 0.8 amp in a least conservative design. Polystyrene capacitors aren't rated for RF current.

3 Neglectingskin effect 18 gauge seems big enough. While searching through the moving boxes for my magnet wire I didn't find any. And with skin effect that's not enough to keep the temperature rise down, going bigger I found a roll of #14 building wire. So I used that. I had already chosen to use T130-2 toroids. The computed turn counts were 18 and 21 turns. 21 turns filled the cores to the point that only a 1/4 support would go through.

4 When winding toroids there is no fractional turn, only whole turns. The design values for L2, L4, and L6 are 3.68, 4.91 and 4.82 uh. In whole turns 18, 21 and 21, giving computed values of 3.564, and uh on the T130-2 toroids. I machined a fiberglass rod (yard marker raw material) to 1/4 for the toroids with 21 turns and left it near 5/16 for the 18 turn toroid. I threaded the ends and punched holes in the box with double nuts on the large end.

5 Capacitor selection is another problem. Most aren't rated for RF current handling capacity. I chose capacitors from the Murata DEBB33F family for the larger values and the DEA1X3F family for the small values. I had to use capacitors in parallel for nearly all the circuits. When I received the capacitors I wished they had fatter conductors. The temperature coefficients for these families are small. The DEA1X3F being the better of the two. General purpose ceramics can change 85% down at their highest rated operating temperature. Intolerable in a filter.

6 Capacitor shopping list, I bought from 3x 1500 pf DEBB33F152KA3B 2x 330 pf DEA1X3F331JA3B 5x 3300 pf DEBB33f332KA3B 2x 2200 pf DEBB33F222KA3B 3x 1000 pf DEBB33F102KA3B 1x 220 pf DEA1X3F221JA3B These run less than a dollar each. And have decent temperature stability.

7 These capacitors I chose for very high voltage rating 3.15 Kv to have larger bodies to handle RF current better. Some 1KV I used in FD filters have died and will be replaced with capacitors in the DEA1X family.

8 For C1 I used 1x 1500 and 1x 330 in parallel to get 1830 pf. For C2 I used 4x 3300 and 1x 2200 in parallel to get 15,400 pf. For C3 I used 1x 1000 and 1 x 220 in parallel to get 1220 pf. For C4 I used 2x 1500 in parallel to get 3000 pf. For C5 I used 1x 1000 and 1x 330 to get 1330 pf. For C6 I used 1x 1000 and 1x 3300 to get 4300 pf. For C7 I used a single 2200 pf.

9 These variations in the component values from the design values mostly move the notch frequencies. Test results. The reject band is at least as good as the handbook chart. I've not measured it closely, I ran out of dynamic range with my scope and its 10:1 probe and my signal generator.

10 RF power tests. I hooked the filter between an FT857 and a dummy load. I ran key down on 1.8 MHz for ten minutes, showing a bit over 100 watts on the MFJ tuner wattmeter. The dummy load globars reached degrees C. The capacitors in the filter warmed up to the 40 to 65 degree C range. They are rated to 85 degrees C. On CW or SSB the heating would be less. The temperature rise for the shunt capacitors in series with the coils was less at 1999 khz.

11 On 3.5 Mhz the capacitors warmed up only 10 to 15 degree C above ambient. Less on 10m. On the higher bands the filter does raise the SWR as much as 2:1 on 28 MHz but ten minutes key down on 10m didn't warm the shunt capacitors much or the series capacitors as much as on 160m. Primarily because the shunt inductors limit their currents. I say its a success and should survive quite a while at 100 watts, not at 200 though.

12 If I was I was starting over, I'd consider T157-2 or T200-2 toroids for easier winding with the 14 gauge house wire and I still will look for sturdier capacitors. The core of L2 warmed up slightly on 160m. A T157-2 would take 16 turns to make uh and 19 turns to make uh. A T200-2 would take 18 turns to make uh and 20 turns to make 4.80 uh.

13