Caution: This project is in the design and simulation stage. Changes will be made as prototyping, building, testing, evaluation and commissioning proceeds.
I started with the W6PQL filter design as discussed and modified by Jeff, K6JCA. That caused inefficiencies in certain parts of the band as I discussed in part 2. I redesigned them by simulating the W6PQL filers on LT Spice to find the shape (they are 7th order, shunt first Chebychev filters) and the cut off frequency. Then I used rf-tools L-C filter design program to recreate the same filters. I set the component tolerance to 5%. It yielded somewhat different component values and better reflected its terminating impedance at its input.
Below is a side by side comparison of the W6PQL and RF-Tools LC filters. The table following the side by side schematic is the -3 dB frequency and the input impedance of each filter. In some cases, the RF-Tools filter is only slightly better, but better enough that in all cases, I have switched to these designs.
For capacitor power consumption, I will take a slightly different approach from Jeff's and see if it yields the same results.
The maximum voltage swing available at the primary of the output transformer is two times the supply voltage or 96 volts. This translates to 288 volts at the secondary. If 100% of it is reflected back, the biggest voltage available at any capacitor is 576 volts peak or 410 volts RMS (407.3 actually). There are two capacitor values for each band filter. Maximum current through the capacitor occurs at the top of each band. The dissipation factor for the Kemet 1,000 volt capacitors is 0.1% (using 1,000 volt capacitors per Jeff's example). A table below summarizes this.
I will assume maximum capacitor temperature of 125 degrees C (solder melts at 190 degrees C) and enclosure ambient temperature of 50 degrees C (it will be hot in there). This implies a temperature rise of 75 degrees. The Kemet typical multilayer chip capacitor thermal resistance when surface mounted with small traces leading from it is 77 degrees C/watt. So, I will limit the power dissipation in each capacitor to 1 watt. These capacitors need to be close to each other in value to equally share the power more or less evenly.
I essentially got the same number of parallel capacitors as K6JCA. So, I will use the new design values discussed above, but the capacitor count from Jeff's design.
The other change that I have made to the design is to move the relays relay drivers to the low pass filter board.
W6PQL and RF-Tools Filters Side by Side
No comments:
Post a Comment