Purpose and Function The band pass filter reduces the potential for noise and unwanted signals from out of band frequencies. Because the filter is for 3 ham bands it is of necessity somewhat wider than what might be used for a single band. The incoming RF is band-pass filtered in T5/L4/C39, with the RF output at T5's secondaries in antiphase. 1 William R. Robinson Jr. AJ4MC p1of 12
Theory and Design Although the assembly manual states that the filter is made up T5, L4 and C39 I claim these components make up a poor non-symmetrical band-pass filter and a design I was not able to analyze. However, if we add C27, which is also band dependent in the assembly manual, a symmetric designable filter results. I claim this is an oversight in the assembly manual. One method to design the band pass filter design is the method summarized in 2(p60) 1. Transform the band-pass requirements into equivalent low-pass requirement using equation 3-14 BW f a. 2(Eq 3-14) (normally used to select number of elements required, BWc fc but not used as we know the design uses two elements) 2(Eq 3-15) b. Fo Fa * fb 2. Refer to the low-pass attenuation curves. (Table 3-1 Butterworth Equal Termination Low Pass Element Values) 3. Transform the low pass network into a band-pass configuration a. Resonate each element with an element of the opposite type but same value i. All shunt elements become parallel resonant and all series elements become series resonant 4. Scale the band-pass configuration In both impedance and frequency using equations 3-16 3-19 William R. Robinson Jr. AJ4MC p2of 12
Calculated I used the band pass filter design method summarized in 2(p60) Transform the band-pass requirements into equivalent low-pass requirement using equation 3-14 o The lowest frequency for the 40M band is 7.0 MHz 3 o The highest frequency for the 20M band is 14.35 MHz 3 o If we add about 2 Mhz to each side to keep the frequencies of interest in the un-attenuated part of the band-pass filter o BW f 2(Eq 3-14) BWc fc and used in table 3-1 2(p40) Simulation shows that at 3db bandwidth to be 78.2 1.05 MHz f 16.3 5 fc 70.7 1.05 f 0. 188 (normally used to select number of elements required, fc but not used as we know the design uses two elements) o The geometric center frequency, Fo 2(Eq 3-15) Fa * fb Fo 16.3* 5. MHz Fo 9.04 Mhz Refer to the low-pass attenuation curves. (Table 3-1 Butterworth Equal Termination Low Pass Element Values o The design is a two element design with the following coefficients Device L1 C2 Coefficient 1.414 1.414 Transform the low pass network into a band-pass configuration o Resonate each element with an element of the opposite type but same value All shunt elements become parallel resonant and all series elements become series resonant o Now the coefficients are Device L1 C1 L2 C2 Coefficient 1.414 1.414 1.414 1.414 Scale the band-pass configuration In both impedance and frequency using equations 3-16 3-19 Device C27 L4 C39 T5 Coefficient 1.414 1.414 1.414 1.414 Value pf/uh 313 0.993 397 0.78 Actual values 330 0.90 470 0.69 The actual values are listed above also and have reasonable correlation William R. Robinson Jr. AJ4MC p3of 12
Cutoff Frequencies Fc_lower = 5.0 MHz (design input) Fc_upper = 16.3 MHz (design input) Ripple Ideal Butterworth has no ripple Ripple = 0 dbv Attenuation Slope Using Fig3-9 in reference 2(p40) for 2 elements o Slope = db @8- db@4/octave o Slope = 36-24/octave o Slope = -12 db/octave Insertion Loss An Ideal filter has no insertion loss as ideal indictors and capacitors only store energy. However there is of course the loss caused by the source and load resistances. This is simply that of a voltage divider, 20 log(rl/rs+rl) o Rsource = Rload = 50 ohms o Insertion loss Rx = 20 log(50/50+50) Insertion Loss = -6 dbv INDUCTANCE of T5 (for 40, 30, 20 M bands) 16T/2x8T bifilar #30 on T25-6 L = (AL*Turns 2 )/1000 uh 4 AL=2.7 +/- 5 % Lprimary =(2.7*16 2 )/1000 uh 4 T5 Lprimary = 0.69 uh L1/2secondary =(2.7*8 2 )/1000 uh 4 T5 1/2secondary = 0.17 uh T5 Frequency response T5 Fc_lower = 3 Mhz 4 T5 Fc_Higher = 40 Mhz 4 INDUCTANCE of L4 (for 40, 30, 20 M bands) 18T #30 on T25-6 L = (AL*Turns 2 )/1000 uh 4 AL=2.7 +/- 5 % L = (2.7*18 2 )/1000 uh 4 L4 = 0.87 uh William R. Robinson Jr. AJ4MC p4of 12
Simulation Rload is assumed to be 50 ohms see reference 5 for calculation.. The design is somewhat sensitive to Rload especially the really low values o The plot below shows the frequency response with Rload form 10 to 100 ohms in 10 ohm steeps The bottom trace is 10 ohms. VDB(out)[0] VDB(out)[1] VDB(out)[2] VDB(out)[3] VDB(out)[4] VDB(out)[5] VDB(out)[6] VDB(out)[7] VDB(out)[8] VDB(out)[9] rload.resistance rx band pass-small Signal AC-22-Sweep-Graph 0.0-5.000-10.000-15.000-20.000-25.000-30.000-35.000-40.000-45.000 1.000M 10.000M 100.00 Frequency Cutoff Frequencies Fc_lower = 5.10 MHz Fc_upper = 15.94 MHz Ripple Ripple = 0 dbv Attenuation Slope Using points at 1Mhz and 2 Mhz for the lower slope o Slope = db @1Mhz-db@2Mhz/octave o Slope = 40.5-28.1/octave o Lower Slope = 12.4 dbv/octave Using points at 200Mhz and 400 Mhz for the upper slope o Slope = db @400Mhz-db@200Mhz/octave o Slope = 68.4-56.3/octave o Upper Slope = 12.1 dbv/octave Insertion Loss An Ideal filter has no insertion loss as ideal indictors and capacitors only store energy. However there is of course the loss caused by the source and load resistances. This is simply that of a voltage divider, 20 log(rl/rs+rl) o Rsource = Rload = 50 ohms o Insertion loss Rx = 20 log(50/50+50) Insertion Loss = -6.03 dbv INDUCTANCE of T5 (for 40, 30, 20 M bands) William R. Robinson Jr. AJ4MC p5of 12
T5 Lprimary = N/A T5 1/2secondary = N/A T5 Frequency response T5 Fc lower = N/A 5 T5 Fc Higher = N/A 5 INDUCTANCE of L4 (for 40, 30, 20 M bands) L4 = N/A William R. Robinson Jr. AJ4MC p6of 12
Real Circuit Results below are with the entire receiver section was completed. Below is the input and output of the when all of the receiver section was completed. o Local oscillator is at 7.000 MHz o Antenna is at 7.010 MHz Antenna Input amplitude was adjusted for near clipping at audio output Time Domain o Channel 1 is the input at C24/L4 o Channel 2 is the output at the input to T5 Note the decrease in the harmonic content from input to output Note the attenuation through the filter Frequency Domain o FFT of the output of the o Similar to the output from the Antenna Low pass filter but less margin between desired signal and unwanted signals There is the expected peak at 7.0 Mhz There is a peak at 12 MHz which is -20 dbv from the expected signal Perhaps this is from the USB section There is a peak at 16 MHz which is -12 dbv from the expected signal I have no explanation for this signal There is a peak at 21 MHz which is -12 dbv from the expected signal This is probably the third harmonic of the square wave LO leaking back into the antenna as it disappears with removal of the power William R. Robinson Jr. AJ4MC p7of 12
Cutoff Frequencies Fc_lower = 4.1 MHz Fc_upper = 18.9 MHz Results below are before the entire receiver section was completed unless otherwise noted. Ripple Ripple = 8.7 6.9 dbv Ripple = 1.8 dbv Attenuation Slope Using points at 1Mhz and 2 Mhz for the lower slope o Slope = db @1Mhz-db@2Mhz/octave o Slope = 35.4 25.7/octave o Lower Slope = -9.7 dbv/octave My signal generator did not allow me to measure the upper slope Insertion Loss Insertion Loss = -6.9 dbv INDUCTANCE of T5 (for 40, 30, 20 M bands) T5 Lprimary = 1.1 uh T5 L1/2secondary = 0.48 uh T5 Frequency response T5 Fc lower = 2.5 Mhz T5 Fc Higher = 15.8 Mhz William R. Robinson Jr. AJ4MC p8of 12
RX TX Ensemble T5 Using A-B on secondary 0 0 5 10 15 20 25-5 -10 dbv -15 dbv -20-25 Frequency MHz INDUCTANCE of L4 (for 40, 30, 20 M bands) L4 = 1.2 uh William R. Robinson Jr. AJ4MC p9of 12
Comparison Response curve with real load is significantly different than calculated and simulation results o I was unable to mimic this with patristic components in simulation o I note that at least the lower hump at about 6 Mhz for the real load is also noticeable in T5 s response o The real curve is probably sufficient for its application SoftRock RX TX Ensemble 0 0 5 10 15 20 25-5 -10 Gain dbv -15-20 -25-30 Measured 50 Ohm Load Measured Real Load Simulated 50 ohm load -35-40 -45 Frequency MHz The table below compares the results o Both inductors measured significantly ( greater than the 20% AL tolerance) so less turns were used to compensate o T5 Fc_upper is also disappointing Real-Measured Simulation Calculated Fc_lower MHz 4.1 4.37 5 Fc_upper MHz 18.9 18.45 16.3 Ripple dbv 1.8 0 0 Attenuation slope lower dbv/oct -9.7-12.1-12 Attenuation slope upper dbv/oct N/A -12.4-12 Insertion Loss dbv -6.9-6.03-6.0 Inductance T5 Primary uh 1.1 N/A 0.69 Inductance T5 ½ Secondary uh 0.48 N/A 0.17 T5 Fc_lower MHz 2.5 N/A 3 William R. Robinson Jr. AJ4MC p10of 12
T5 Fc_upper MHz 15.8 N/A 40 Inductance L4 uh 1.2 N/A.87 William R. Robinson Jr. AJ4MC p11of 12
References 1. Robson, Richard R. Sr., WB5RVZ, Ensemble RXTX Project, http://www.wb5rvz.com/sdr/ensemble/index.htm, online, accessed 2011. 2. Bowick, Chris, RF Circuit Design 2 nd Edition (Elsevier Inc 2008), p60 3. ARRL, US Amateur Radio Bands, http://www.arrl.org/files/file/hambands_color.pdf, online, accessed 2011. 4. kitsandparts.com, Specs for T25-6 RF toriods, http://toroids.info/t25-6.php online, accessed 2011. 5. Robinson, William AJ4MC, USDR RxTx Ensemble Antenna Low Pass Filter (another of these short papers). http://bellsouthpwp2.net/w/r/wrobinson/, online, accessed 2011. William R. Robinson Jr. AJ4MC p12of 12