***************************************************************** * ARRL Radio Designer 1.0 File EXRF9501.TXT V 1.1 (23 Dec 1994) * * Exploring RF -- January 1995 *QST*, pp 80-82 * * compiled by David Newkirk, WJ1Z (dnewkirk@arrl.org) * * * * Parts One and Two of this file will run as analyzable ARD * * netlists--SEPARATELY. ARD won't like it if you just copy this * * entire file, containing Parts One AND Two, into * * Circuit Editor, and will remind you of this with the error * * message "Duplicate label name: ONEPORT."* * * * * Part One: * * * * Using ARRL Radio Designer 1.0's Optimizer * * to adjust a C-L-C T-Network * * for a 10+j0-Ohm to 50+j0-Ohm transformation * * at 1.9 MHz * ***************************************************************** * * In this example, optimization adjusts a C-L-C T network * to transform a 10-ohm resistive load to 50 ohms, resistive * (50+j0 ohms), at 1.9 MHz. To do this, we describe the network as * a 1-port (the ONEPORT block below), in which the netlist * line "RES 3 0 R=10" specifies the load we want to transform. Question * marks bracket the optimizable values--its 20- to 1000-pF input * capacitor and its 0.1- to 35-uH inductor. (Note that we've constrained * the optimizable values to these limits. If the optimizer can't reach * its goal within these constraints, we learn something quite useful: * that the load we want to transform is outside the transformation limit * of the network's components.) Since our goal is to simulate * manual adjustment of a T-net serving as an antenna tuner, we won't let * optimization adjust the network's output capacitor (also a 20- to * 1000-pF unit). Instead, we begin by setting the output capacitor * to its maximum value (as an experimenter would likely do in adjusting * the network), since keeping the network's output capacitance as high * as possible minimizes the network's loss for a given transformation. * * Note: Figure 1 of January 1995 QST's Exploring RF shows the network * capacitors as 20- to 240-pF parts. Upping their maximum capacitance to * 1000 pF is one of the "few additions and changes" implied by the text's * Figure 1 callout. With its capacitors (especially its output capacitor) * constrained to 240 pF, the Figure 1 network can do no better in * in transforming 10+j0 ohms to 50+j0 ohms at 1.9 MHz than the lower * three curves in the column's Figure 3. You can experiment with such * limitations at will just by changing the constrained values, or * by hard-coding one or more values, as I've done with C/OUT\in the block * that follows. Fixing C/OUT\'s value in this way lets you answer this * question: "Can the C-L-C network in Figure transform 10+j0 to 50+j0 ohms * at 1.9 MHz with a C/IN\ between 20 and 1000 pF, an L between 0.1 and * 35 microhenries, and C/OUT\ set to 240 pF?" * BLK CAP 1 2 C=?20PF 82.1733PF 1000PF? Q=1000 ; this is C/IN\ IND 2 0 L=?0.1UH 30.2528UH 35UH? Q1=200 F=2.52MHZ ; this is L CAP 2 3 C=240PF Q=1000 ; this is C/OUT\ RES 3 0 R=10 ONEPORT:1POR 1 END * Once the optimizer has found a set of values that work, we can determine * the network's loss and frequency response by reconfiguring the T-net * as a 2-port. This lets ARRL Radio Designer determine the network's forward * transmission gain (MS21, the magnitude of the S parameter S21). To do this, * we remove the * network's 10-ohm load resistor (we'll specify a * 10+j0-ohm load in ARRL Radio Designer's Report Editor) and replace it with * an ideal transformer (TRF) configured to look like 10 ohms with 50 ohms a * its output. (Because the TRF is ideal, it is lossless and does not skew the * network's frequency response.) * * Removing the contrained optimizable values and their bracketing question * marks is a good thing to do in this new block because we'll want to use the * ONEPORT block over and over to see if the T-net can successfully transform * various loads, and because ARRL Radio Designer's optimizer attempts * to adjust ALL the question-mark-delimited values in a.CKT file--even * those outside the circuit named as optimizable in the OPT block--during * optimization. * BLK CAP 1 2 C=477.616PF Q=1000 IND 2 0 L=4.90764UH Q=200 F=2.52MHZ CAP 2 3 C=1000PF Q=1000 TRF 3 4 0 0 R1=10 R2=50 OUT1000P:2POR 1 4 END * The remaining three 2-port blocks show the results of optimization done * at other output capacitance values. Each succeeds in transforming * 10+j0 ohms to 50+j0 ohms, but plotting their response indicates * that the network's minimum loss varies inversely with output capacitance * --at least for this 10+j0- to 50+j0-ohm transformation. In using this * T network for feeding transmitter power to an antenna at 1.9 MHz, we'd * want to minimize the network's loss to avoid component overheating. BLK CAP 1 2 C=122.811PF Q=1000 IND 2 0 L=19.3847UH Q=200 F=2.52MHZ CAP 2 3 C=240PF Q=1000 TRF 3 4 0 0 R1=10 R2=50 OUT240PF:2POR 1 4 END BLK CAP 1 2 C=104.752PF Q=1000 IND 2 0 L=23.063UH Q=200 F=2.52MHZ CAP 2 3 C=200PF Q=1000 TRF 3 4 0 0 R1=10 R2=50 OUT200PF:2POR 1 4 END BLK CAP 1 2 C=82.1733PF Q=1000 IND 2 0 L=30.2528UH Q=200 F=2.52MHZ CAP 2 3 C=150PF Q=1000 TRF 3 4 0 0 R1=10 R2=50 OUT150PF:2POR 1 4 END * The FREQ block sets relative high simulation resolution (1-kHz STEPs) * from 1.8 to 2.0 MHz for maximum report (graph and table) resolution in * Amateur Radio's 1.8- to 2.0-MHz (160-meter) band. Specifying 300 * exponentially stepped (ESTP) frequencies from 0.5 to 30 MHz is * appropriate for logarithmically graphing the network's frequency * response across this wider range. FREQ STEP 1.8MHZ 2.0MHZ 1KHZ ESTP 0.5MHZ 30MHZ 300 END * The OPT block specifies our target 50+j0-ohm load in terms of the * 1-port impedance parameters RZ11 and IZ11. The TERM=1E-5 line tells * the optimizer to consider the job done--to TERMinate optimization-- * when it has adjusted the circuit's optimizable values for an error * function of 1E-5 rather than chasing the error function all the way * down to 0 (the optimizer's default goal). * * Because the SWR-driven power-foldback circuits in modern solid- * transmitters usually don't kick in until SWR hits 1.5 to 2.0, 1E-5 is * much farther than we need to go in achieving a match if all we want the * network to do is make its load look enough like 50+0j0 to make a * transmitter happy. With this in mind, we could modify * the optimization goal along the lines of * * RZ11=33 75 ; see "Two Value Goals" on page 19-17 of the ARD Manual * IZ11=0 * * This spec lets the optimizer stop when it causes the resistive * portion of the net's input impedance to fall within the RZ11 range * that corresponds to an SWR of 1 to 1.5, while keeping the reactive * portion of the net's input Z at 0. (Power in a reactance doesn't do * useful work, so we want a transmitter to see a purely resistive * load even if that means settling on a resistive load that * diverges a bit from 50 ohms.) OPT ONEPORT F=1.9MHZ RZ11=50 IZ11=0 TERM=1E-5 END * ************************************************ * ARRL Radio Designer 1.0 File EXRF9501.TXT * * Exploring RF -- January 1995 *QST* * * * * Part Two: * * * * Using ARRL Radio Designer 1.0's Optimizer * * to adjust a C-L-C T-Network * * for a 60+j60-Ohm to 50+j0-Ohm transformation * * at 7.2 MHz * ************************************************ * * Instead of terminating the network with a resistor or an LC, RL or RC * network, we use ARD's ONE black-box element to specify a complex * load directly, as shown in the * * ONE 3 0 ANTENNA * * netlist line below. Rather than specify the load on this line, we * point to the ANTENNA entry in netlist's DATA block. * BLK CAP 1 2 C=?20PF 198.234PF 240PF? Q=1000 IND 2 0 L=?0.1UH 1.56629UH 35UH? Q1=200 F=2.52MHZ CAP 2 3 C=?20PF 122.386PF 240PF? Q=1000 ONE 3 0 ANTENNA ONEPORT:1POR 1 END FREQ STEP 7.1MHZ 7.3MHZ 1KHZ ESTP 1MHZ 30MHZ 300 END OPT ONEPORT F=7.2MHZ RZ11=50 IZ11=0 TERM=1E-5 END DATA ANTENNA: Z RI ; see "Data for One" on page 17-6 of the ARD Manual 60 +60 END