Comparing several L LNA.
(May 22 2013)

Method for NF measurements.

The NF measurement system is described here: Using the HP8970A with circulators under computer control.

Five amplifiers of type L LNA made by AD6IW were measured at an ambient temperature of 26.7 degrees. No cooling was applied other than conduction through the cables. For each amplifier +7 V was first applied. The NF measurement was made after thermal equilibrium was reached. Then +12 V was applied and the NF again measured after the new thermal equilibrium was reached. A typical measurement is shown in figure 1.

Table 1 gives the results for five different L LNA units as well as for a G4DDK amplifier.

Figure 1. The amplifier was powered up with 7 V at sample 600. Thermal equilibrium is reached after about 100 samples (about 10 minutes.) 12 V was applied at sample 1300 and the new thermal equilibrium is reaced after about 10 minutes.

Unit    DC     Te      NF
        (V)    (K)    (dB)
G4DDK   12    30.99  0.441
L LNA1   7    36.86  0.520
L LNA2   7    36.87  0.520
L LNA3   7    36.82  0.519
L LNA4   7    37.06  0.522
L LNA5   7    36.58  0.516
L LNA1  12    38.19  0.537
L LNA2  12    38.04  0.535
L LNA3  12    37.84  0.533
L LNA4  12    38.15  0.537
L LNA5  12    37.69  0.531

Table 1. Noise temperatures and NF for several L LNA and one G4DDK amplifier. Note this is a relative scale. The true noise figures may be 0.25 dB lower. The influence of the second stage is not taken into account.

The system noise temperature at the output of the amplifier under test is not accounted for. It is 41.51 K (0.581 dB) on the same relative NF scale as in table 1, but the zero point is unknown when this is written.

The G4DDK amplifier has been measured in Orebro 2012 with the result NF = 0.187 dB (12.8 K) while the specification on the L LNA claims Te = 21 K. If one assumes an error of 18.2 K in table 1 the system Te at the output of the DUT would be 23.3 K which after division by the gain which is about 2000 would add 0.012 K for the L LNA and much less for the G4DDK amplifier. With an assumed error of 18.2 K in table 1 the typical noise temperature for L LNA would be 36.8-18.2 = 18.6 K which is 2.4 K below the value given in the product specification.

The biggest uncertainty seems to be the temperatures of the amplifiers. Figure 2 shows a measurement of L LNA 5 at an ambient temperature of 25.4 degrees where Te is 36.45 K.

Figure 2. A table fan is started at sample 6070.

A table fan directed towards the LNA causes the temperature to drop by about 5 degrees with a reduction of Te by 1 K as a consequence.