MMIC MADE AT HOME PY2WM Joćo Kolar De Marco, PY2WM - Brazil (SK) original design Translated from Portuguese at http://www.lelac.com.br/py2wm/MMIC/mmic.html (now flags error 404!) MMIC MADE AT HOME Feb 5, 2006 In some magazine and book circuits we find an amplifier device called MMIC (Monolithic Microwave Integrated Circuit). It is usually small in size, has definite gain and impedance, does not require external components in addition to input / output coupling capacitors and a supply resistor. It is therefore very easy to use. Because it has a definite impedance (usually 50 ohms) and is stable, it can serve as a load for devices that require adequate termination such as diode mixers, LC and crystal filters, etc ... One series that has become common is MAR-xx from Minicircuits. To understand this table (simplified from the original of the manufacturer), we take the MAR-1 and following the table we see that its usage range is 1000 MHz DC, we have next the gain in 100 and 1000 MHz (in some also the 2000 MHz). Maximum power means that at 1.5 dBm output (about 1.4 mW) the signal already has a compression of 1 dB. The dynamic range gives the value of the noise factor and the output power for IP3 (which is when the 3rd order intermodulation products are equal to the output itself), we also have the ROE at the input and output at 50 ohms, dissipation limits and the supply current and voltage used to calculate the supply resistor. The typical internal scheme of a conventional MMIC is thus: Fig 1 - Notice that it has only input, output and ground. Fig 2 - Connections of an MMIC to the circuit. You can connect another in cascade for greater gain. The resistor R1 is calculated by the value of current and voltage of the table, considering the voltage Vcc used. For example, for MAR-1, with 17 mA and 5 V, if we use 12-volt Vcc, R1 must have 7 volts between terminals, then R1 will be: R1 = 7 / 0.017 = 411 ohms. Because R1 is effectively in parallel with the output, it is common to employ a series RF choke in order to avoid the loss that would occur. MAKING A MMIC AT HOME We can make an MMIC with discrete parts. We take the BC548 transistor and simulate it in software. The task is arduous because it is necessary to combine several aspects at the same time: currents of polarization, gain, noise and impedances of input and output, mainly .. Since the typical Ft of a transistor like this is around 300 MHz, then assuming that it projects to a gain of 10 (20 dB), the maximum frequency will be at 30 MHz, above that the gain will decrease. Using higher Ft transistors can extend the frequency range of use. To make the "homemade MMIC" I initially simulated it with the Multisim PSPICE then passing to the Ansoft Serenade Harmonica . Both have student versions, are functional and free. Multisim has an excellent tutorial that takes the beginner by the hand, step-by-step (in English). The new version of Ansoft also has an excellent tutorial, unlike the Harmonica I use, which I had to learn hard! Fig 3 - home MMIC scheme. ATTENTION, the resistors have been long searched ! The polarization, gain and impedances of input and output are quite interdependent and from the indicated values, it is not recommended to change without knowing it . The circuit needs 3.4V at 40mA, with these values ? the load resistor Rc is calculated. To increase the gain a little, you can put a RF choke in series with this resistor. In the model I used Vcc = 12V and 220 ohms resistor. RESULTS With the transistor BC548 the results were as follows: Fig 4 - Gain and noise in dB Fig 5 - Input impedance (Zin) and output (Zout). Fig 6 - Output versus dBm input for both BC548 and BFR96. The home MMIC goes well up to about 10 dBm at the output (10 mW). This homemade MMIC can be used in receivers and transmitters, such as RF amplifier for general use up to 30 MHz. For use at higher frequencies it is necessary to use a more suitable transistor. For example the BFR96: Fig 7 - Gain and noise for the BFR96. Note that the horizontal axis now goes up to 400 MHz! Fig 8 - input and output impedance. The circuit was NOT optimized for this transistor however, it would certainly be possible to improve it. MMIC amplifiers are like the "lego" of RF electronics, allowing you to assemble circuits in blocks, with inputs and outputs in 50 ohms defined. The proposed models can be used in input RF amplifiers in receivers, in the IF channel, as buffers in oscillators and input stages of frequency meters and other instruments, and driving stages in transmitters. MMIC MADE AT HOME 2 - ADDITIONAL EXPERIENCES 4 / Feb / 2006 Fig 9 - Using 2N2222. It is useful up to 10 MHz only. But it is not interesting as receiver RF amplifier (first stage receiver). Fig. 10-2N2222. Above 10 MHz the impedances are reactive. MMIC MADE AT HOME MORE ADDITIONAL EXPERIENCES 5 / feb / 2006 In this home MMIC I have tried to redefine the circuit optimizing for transistors BFR90, BFR91 or BFR92, all are similar enough. The BFR91 has a small margin for best in Noise and maximum power for IP3. The circuit requires 4.3V and 35 mA. The value of the 220 resistor at the output can be recalculated to other supply voltages. To increase the gain slightly a RF choke can be placed in series with the resistor. It is not recommended to feed with less than 8V, which would make the resistor of very low value, making a choke compulsory, but mainly for destabilizing the DC bias point. Fig 16 - 4dB noise, 17 dB gain falling to 13 dB at 1 GHz, S11 and S22 (reflection at the input and output) excellent up to 300 MHz (taking -15 dB as reference), and reverse transmission of -22 dB (S12, read: insulation against passage of signal from the output to the input). Fig 17 - Another way of expressing input and output impedances, which are absolutely flat to 100 MHz. Fig 18 - Another way, the SWR (ROE) reaches 2: 1 in 500 MHz. Fig 19 - Response to input power sweep. The circuit was not optimized for this aspect, however the two transistors were polarized to divide similar currents, 17 mA in the first and 13 mA in the second. Fig 20 - Stability index K. The circuit is conditionally stable, depending on what is connected at the input and output. Examining up to 5 GHz, if K = 0.6 @ 1.6 GHz, and S11 <0 of DC at 5 GHz. NORTON RF AMPLIFIER 30 / jan / 2010 Here is a typical Norton Noise-free RF amplifier (also called "without loss"). The "noiseless" or "lossless" designation comes from the fact that RF feedback is taken from the transformer, made for strong coupling, rather than the use of resistors, which introduce noise and loss . To mount this amplifier, a BFR91 was used. The RF transformer was calculated for a gain of 12 dB and 50 ohm input / output impedances. The simulation was done with an FT50-43 toroid, the top winding (in the scheme) with 11 turns (46uH), central winding 4 turns (6uH) and lower winding with 1 turn (0.38uH). It was not taken into account the losses in the transformer, the results can be a bit optimistic. But considering the efficiency of the mixers of the Minicircuits where two transformers are used, it is very probable that the real income is very close. Repair the correct connection! In this chart are the main aspects of interest. SWR (ROE) below 2: 1 from 2 to 400 MHz. The noise factor here can best be viewed, the only change was on the Y axis, in dB. Stability factor K. The circuit is unconditionally stable. The 1 dB compression point is approximately at Pin = 7 dBm with Ic = 10 mA IP3 at approximately 25 dBm per input tone. References: Wes Hayward, Introduction to Radio Frequency Design, ARRL, 1995.