Summer 2008 Issue

Low-Noise Pre-amplifiers for the
1.3, 2.3, and 3.4 GHz Amateur Bands

A goal for the microwave operator is to reduce the
noise introduced by pre-amps. In this article G4DDK
describes low-noise pre-amps for three of the
popular microwave amateur bands.

By Sam Jewell,* G4DDK

Photo A. The completed 2.3-GHz version of the LNA.

Invariably, 1296-MHz moon-bounce (EME) requires the use of a very-low-noise pre-amplifier (LNA) to receive the weak signals that are often encountered. This is especially true when only a small TVRO dish can be used as the antenna. I have successfully used a 7.5-foot KTI TVRO dish for both 1.3- and 2.3-GHz EME using the pre-amplifier designs in this article.
My initial EME activity was with the well-known 1.3-GHz LNA design published by Tommy Henderson, WD5AGO1, whilst the 2.3-GHz LNA was a modification of a design by Al Ward, W5LUA2.

Following requests for help and information. I made a number of PC boards for radio amateurs in Europe. In general these worked well, which is a testament to the solid designs from Tommy and Al.

Because of ongoing demand and difficulties obtaining new ATF10135 MESFETs (metal epitaxial semiconductor field effect transistors) for use in the second stage of the 1.3-GHz pre-amplifier, I decided to investigate an alternative second-stage device. I also decided to house the pre-amplifier in a readily available tin-plate box. The new design achieves a lower 1.3-GHz noise figure, and higher gain than the original pre-amplifier design.

Noise-figure and gain measurements at various microwave events in the UK, The Netherlands, Germany, as well as at Central States VHF Conference 2007 have shown that a stable, repeatable, noise figure of around 0.25–0.27 dB, with an insertion gain of 36 dB, is achievable with the 1296-MHz version.

It was apparent that the same pre-amplifier board also had the potential to work at 2.3 GHz, especially with the air-supported input components as used in Al’s design. This necessitated some component-value changes to optimize performance at the higher frequency. After installing components with the calculated values (and some inspired empirical substitution!), the result was a noise figure of around 0.35 dB and an insertion gain of about 26 dB. The reason for lower gain at 2.3 GHz is partly due to the second-stage device and the use of a non-optimum microstrip line, which is part of the 1.3-GHz design. However, for EME work, even an insertion gain of 26 dB may be enough to eliminate the usual second pre-amplifier unit.

Further work showed that the pre-amplifier would also produce an acceptable noise figure and gain in the 3.4-GHz amateur band. A 3.4-GHz noise figure of between 0.5 and 0.55 dB with an insertion gain of around 28 dB is easily achieved.

Versions of the pre-amplifier have been successfully tuned for use at 1090 MHz, 1240–1296 MHz, 1420 MHz, 2200–2290 MHz, and 2302–2320 MHz, all with excellent results. Work has commenced on a 432-MHz version of the pre-amplifier.

Circuit Description

The circuit schematic is shown in figure 1.This is the same for all three versions of the pre-amplifier. Component values are shown in Table 1. Where component values are different for each of the various bands, these are shown in Table 2.

Two different low-noise GaAs FETs have been specified for use in the 1.3-GHz pre-amplifier. The NE32584C gives the lowest noise figure, but these are no longer available from NEC. However, there are still large stocks of the NE32584C and other package variants available as surplus stock in the U.S. and Europe.

The Avago ATF36077 has been shown to work extremely well in the TR1 position, but it has a marginally higher noise figure at 1.3 GHz compared to the NE32584. The NE32584 is therefore the preferred device for 1296-MHz EME.

The second-stage device is an Avago ATF54143 in all cases. I have been unable to get the newer NE3210 HEMT to work well in the first stage at these frequencies.
The input circuit consists of a “T’ match with suitable low-loss capacitors and inductors. These components are air supported, rather than soldered to PC board pads, in order to keep losses due to parasitic strays to a minimum.

Low-noise matching is achieved by careful adjustment of the spacing of the turns of L1. Adjustment is critical in order to achieve the very lowest noise figure. This will not coincide with maximum gain. In these designs lowest noise always occurs on the high-frequency side of the maximum-gain frequency.

Click here to return to Summer 2008 highlights

Click here to subscribe to VHF


© Copyright 2008, CQ Communications, Inc. All rights reserved. This material may not be reproduced or republished, including posting to a website, in part or in whole, by any means, without the express written permission of the publisher, CQ Communications, Inc. Hyperlinks to this page are permitted.