Circuit Modelling Design Notes for use with Modelling Software

( Microwave Office and other Modelling Packages )

When designing with a modelling package it is advisable that when using " Discrete " components where ever possible to use the so called " Prefered Values " , if the design value of a capacitor is " non standard " then the use of multiple values to arrive at the design value is a reasonable route to take, however try to avoid using parts with too dissimilar a value .

The choice of package size also affects the performance of your design so reference to full manufacturer's data sheet / S parameters is advisable , try to choose the size of component to minimise any discontinuities in circuit paths , eg if your transmission line is 2mm wide DO NOT use a " 1206 " package device in it , its "3 .175 " mm wide ! use an " 0805" . When you use a shunt component from a transmission line to earth it is advisable to model the component coming from a junction to allow you to get the correct length of transmission line the junction size being equal to the package width of the shunt element , this also allows for the discrete width of the components end cap to appear to be distributed along the transmission line not just at a point as is the case in the real circuit .This means that when you are optimising the track lengths using the inbuilt tools , the capacitor / inductor is simply slid up / down the transmission line as in real life .When using a capacitor in the transmission line as a DC block try using a value that is series resonant you will be suprised at the improvement in power transfer , this also applies when you are decoupling a supply rail as it is more effective than just increasing the value as you would think from 1st principles , with decoupling you may have to use several values about a decade apart to ensure the supply is decoupled at ALL frequencies where the active device has gain .

The following part will detail the design & construction of a " Low Noise Amplifier for 1.8 - 2 Ghz " based on the application note for the Avago ( Agilent / HP ) for the ATF54143 .

From the various ways in which the circuit is drawn you can see the Significant change in results , the discrete components used DO NOT CHANGE only the interconnections / physical spacing .

The initial circuit as copied from the application note & drawn as a conventional circuit ( Schematic ) the interconnections in Red have NO physical length nor do they have any losses THEY ARE PERFECT !. The second circuit as redrawn to include pads & transmission lines that are neccesary to make the real amplifier including the relevant subcircuits from the package library or the manufacturers website . This shows the benefit of using the relevant data as the actual design was used on the " ACE Active Integrated Antennas " Course held at the University of Birmingham in September 2008 .The calculated gain was 16dB , the measured gain of the prototype being 15.7 dB , the course members who were all new to designing & building SMD RF amplifiers managed to get gains of between 14.5 & 15.6 dB gain ( A Sample of 10 items ) the variation being due to the placement of the degenerative feedback links in the source path . The individual components were selected from the online libraries ( Having checked the supplier's websites for full component part No's ) the relevant layout was taken from the manufacturer's application note .The necked in transmission line was created to allow for the use of additional components on the original application note layout with the narrow track being removed to allow for the addition of an inductor or resistor ( or both ) should the preamplifier be designed to work at the lower end of the devices range ( It covers 100MHz - 6 GHz ) the gain being higher at lower frequencies could lead to stability problems hence the provision in the board layout to add these additional components.

The drawing above shows the way in which many people would draw an amplifier from first principles for low frequencies , however with a little thought you will realise that this is not what you can build in real life the tracks between components DO have a finite length & width !

The input match of the idealised amplifier is not matched to the system impedance as can be seen from the polar plot above .

The same input match as shown on a Return Loss plot , note the match is < 10dB flat above 2 GHz

The Gain plot of the Idealised amplifier using the component values from the Manufacturer's data sheet with the maximum gain shown as 14.27 dB at 1725MHz.

Note the same component values are used in this " Real " amplifier just the realisitic application of Microstripline tracks & junctions in accordance with the layout copied from the manufacturer's data sheet , see how the matching , gain & frequency response have ALL been altered , just by the ADDITION of real interconnections

The input match is certainly not perfect at the desired frequency

The input Return Loss does show an improvement in as much as there is a dip showing a limited amount of selectivity at the desired frequency

Note the Significant change in the maximum gain value & the frequency at which the maximum gain occurs !

The Values of gain & return loss will be altered by the amount of degenerative feedback present in the source of the active device ,these in turn will also adjust the frequency slightly at which maximum stable gain and best input & output match occur this can be seen when using the " Tune " facility in Microwave office viewing the interaction between all these on their relevant graphs .This can be seen more readily by putting all four " S2P " graphs on a single plot and by putting both input & output polar responses on a single plot .

Graph of S11 , S12 , S21 & S22

Picture of the " Real " Amplifier

In addition to the design for the Gasfet amplifier shown above there is a more detailed design for a two stage wideband MMIC amplifier along with some other basic projects on this Link

Back To RF Design Notes

This page last updated 2nd Sept 2010