OwenDuffy.net |
|
An exploration of the performance improvement of a real 144MHz station using a mid priced kit form low noise amplifier. |
LNAs are widely used on 144MHz SSB stations to improve station receive performance. Indeed it is popularly held that a masthead LNA is mandatory for any "serious" station.
This article is a follow-on to the article Effective use of a Low Noise Amplifier on VHF/UHF which introduced a G/T based model for analysis of performance improvement and a theoretical example of a 432MHz station. This article explores the performance improvement (or otherwise) of a real 144MHz station using a mid priced kit form low noise amplifier. The analysis is specific to the configuration and noise environment, but the factors and the methodology are applicable more generally.
At the manufacturer's express request, this article does not identify the manufacturer or the LNA product.
Item | Value | Comments |
Active device | BF988A | Dual gate MOSFET |
Gain (dB) | 22 | Measured gain: 28dB at 0dBm out |
Noise Figure (dB) | <1.5 |
Table 1 sets out the key specifications for the LNA. The LNA has a tuned input filter and broadband transformer coupled output. The input filter consists of a parallel tuned circuit using an air cored inductor connected to g1 of the MOSFET, and a series tuned circuit (toroidal ferrite cored coil) from the input to a tap on the coil mentioned earlier. The LNA which included T/R relay switching to 100W was assembled in a sealed die cast box with N connectors.
The LNA was constructed as per the kit instructions and circuit diagram and assembled in a sealed die cast aluminium box with N connectors for the input and output.
Alignment of the input filter is critical to achieving gain, best Noise Figure, and more importantly, best front end selectivity for minimisation of IMD in the LNA. The LNA was set up on a sweep generator with a crystal detector on the output and the input filter aligned. Adjustment of the filter for narrowest response using the sweep generator was quite critical. Amplifier gain at 0dBm out is 28dB and bandwidth at -3dB is 4MHz. Fig 2 shows the response of the amplifier (remembering that the HP423A crystal detector produces -ve DC output), the middle graticule line is 3dB below the peak response, and sweep is 1MHz per division (graticule labels are not relevant to this display).
The input filter was initially aligned using an SSG and SINAD meter on the receiver which gave a fairly broad response. The use of the sweep generator enabled a narrower frequency response which should reduce LNA susceptibility to IMD noise.
The LNA gain, centre frequency and bandwidth are temperature sensitive, mostly attributable to variations in the ferrite cored inductor in the input circuit with temperature.
Noise figure was measured using the technique described at Noise Figure Y factor method calculator .
Noise source | |||||
ENRS | 42.7 dB | Th | 5.43e+6 K | Tc | 290 K |
Instrument (CAL) | |||||
ATT12in | 0.00 dB | ATT22in | 20.0 dB | ||
Th2in | 5.45e+4 K | Tc2in | 290 K | ENR2in | 22.7 dB |
Y2 | 13.2 dB | T2 | 2.42e+3 K | NF2 | 9.70 dB |
Measurement (DUT) | |||||
ATT12in | 30.0 dB | ATT212in | 20.0 dB | ||
Th12in | 5.71e+3 K | Tc12in | 290 K | ENR12in | 12.7 dB |
Y12 | 8.48 dB | T12 | 607 K | NF12 | 4.90 dB |
Results | |||||
G1 | 27.2 dB | T1 | 90.3 K | NF1 | 1.18 dB |
Notes:
DUT is a low noise 144MHz LNA (Gain~=28dB, NF~=1dB). |
Table 2 sets out the results from the Y Factor noise figure measurement.
Fig 1a shows the test using NFM.
Amateur equipment is often designed with receivers that exhibit exceptional sensitivity in a test environment, but that sensitivity is not realised when connected to an antenna. There are two factors that commonly limit the sensitivity that can be realised when connected to an antenna:
The following explores the realisable sensitivity of the LNA in an on-air scenario.
Fig 2 shows the configuration used for determination of realisable sensitivity. The LNA forms part of the receiver in the block diagram. The test measures sensitivity in the presence of the real world noise and signal environment at a place and time.
Item | Value | Comments |
Identification | TS2000 | |
Bandwidth setting (Hz) | 300-2400 | |
Effective Noise Bandwidth (Hz) | 1450 | |
Noise Figure (dB) | 7.4 |
Table 3 sets out the key performance characteristics of the test receiver.
Row | Antenna | Filter | Attenuator (dB) | Receiver | SSG (dBm) | Realisable sensitivity at attenuator input (dBm for 10dB SINAD) | Receiver Noise Figure (dB) | Receiver sensitivity improvement with LNA (dB) | Total equivalent noise at rx input (dBm) | ||
LNA | TS2000 | ||||||||||
Attenuator | LNA | ||||||||||
1 | Dummy Load | NO | 0 | NO | NO | YES | -74.4 | -125.4 | 7.4 | -134.9 | |
2 | Dummy Load | NO | 0 | YES | NO | YES | -81.0 | -132.0 | 0.8 | 6.6 | -141.5 |
Table 4 sets out the measurements and calculated results using the test configuration in Fig 2 with a dummy load in place of an antenna.
The noise observed in Row 2 is due mainly to the receiver internal noise, and the thermal noise generated in the 50Ω dummy load. IMD noise would be so low as to be insignificant. Note that total equivalent noise is 6.6dB lower than the TS2000 alone (Row 1), so the combination TS2000 with LNA has a much lower internal noise.
Note that this method of determining Noise Figure becomes unreliable at very low Noise Figures where Noise Figure is very sensitive to error in SSG output level. The remainder of this article uses a more realistic LNA + receiver NF of 1.2dB (as reported by the Y-factor noise test above).
The Diamond X-50 is a dual band omnidirectional vertical antenna with a little more gain than a half wave dipole.
Row | Antenna | Filter | Attenuator (dB) | Receiver | SSG (dBm) | Realisable sensitivity at attenuator input (dBm for 10dB SINAD) | Receiver Noise Figure (dB) | Receiver sensitivity improvement with LNA (dB) | Total equivalent noise at rx input (dBm) | ||
LNA | TS2000 | ||||||||||
Attenuator | LNA | ||||||||||
3 | X-50A | NO | 0 | NO | NO | YES | -69.2 | -120.2 | -129.7 | ||
4 | X-50A | NO | 0 | YES | YES | YES | -50.1 | -101.1 | -19.1 | -110.6 | |
5 | X-50A | NO | 10 | YES | YES | YES | -67.8 | -118.8 | -1.4 | -128.3 |
Table 5 sets out the measurements and calculated results using the test configuration in Fig 2 using the Diamond X-50A antenna.
The noise observed in Row 3 is due to the receiver's internal noise, the external in band noise, and IMD noise generated in the receiver from all sources arriving at the antenna terminals.
Despite the fact that the receiver internal noise is much lower with the LNA, in Row 4 the total noise has increased by almost 20dB over that in Row 3, indicating the IMD noise generated in the LNA is much higher
Note the difference between Rows 4 and 5, introduction of a 10dB attenuator has reduced receiver input by 10dB, yet the sensitivity has improved by 17.7dB, because the total equivalent noise at the receiver input has dropped by 17.7dB. This result indicates the noise is principally due to third order IMD in the LNA.
Note also the increase in "Total equivalent noise at rx input" with the LNA, this increased apparent noise is due principally to IMD noise generated in the LNA.
The LNA severely degrades system performance when used with a wider band antenna (eg amateur FM with omni directional and multi-band antennas , weather satellite).
The 4 element Yagi is a simple home constructed 4 element antenna with gamma match. The combination of multiple tuned elements and the tuned gamma match result in narrower bandwidth than a simple half wave dipole antenna.
Row | Antenna | Filter | Attenuator (dB) | Receiver | SSG (dBm) | Realisable sensitivity at attenuator input (dBm for 10dB SINAD) | Receiver Noise Figure (dB) | Receiver sensitivity improvement with LNA (dB) | Total equivalent noise at rx input (dBm) | ||
LNA | TS2000 | ||||||||||
Attenuator | LNA | ||||||||||
6 | 4el Yagi | NO | 0 | NO | NO | YES | -67.4 | -118.4 | -127.9 | ||
7 | 4el Yagi | NO | 0 | YES | NO | YES | -69.2 | -120.2 | 1.8 | -129.7 |
Table 6 sets out the measurements and calculated results using the test configuration in Fig 2 using the 4 element Yagi antenna.
Measurements reported in Rows 6 and 7 show the improvement with the LNA using a 4 element Yagi antenna which is horizontally polarised and has more selectivity than the X-50A. The increase in apparent Total equivalent noise at rx input with the LNA suggesting some IMD noise in the LNA.
The slight increase in "Total equivalent noise at rx input" with the LNA is also evidence of IMD noise generated in the LNA.
At the time of measurement of Rows 6 and 7, the ambient noise was assessed using the method described in Ambient noise calculator with the TS2000 alone. The measurements indicated Fa of 13.5dB, which considering the the ENB of the receiver and line loss would cause an ambient noise power at the receiver terminals of -129.9dBm.
Fa on 144MHz at this location varies from about 6dB to about 18dB, with around 14dB being the common level at the time of the morning aircraft enhancement activity.
G/T is the ratio of antenna gain to equivalent system noise temperature (including external noise), usually expressed in dB/K.
The G/T ratio is a meaningful indicator of system performance, unlike quoting receiver noise temperature or noise figure in isolation of sky temperature and antenna gain. The ability to receive weak signals is directly related to G/T, the higher G/T, the better. Signal/Noise ratio is proportional to G/T (Signal/Noise=S * λ2/(4*π) * G/T / (K*B) where S is power flux density, K is Boltzman's constant, and B is receiver effective noise bandwidth).
The cost basis used for the charts below is the cost of components and materials for construction, and excluded labour costs.
The median ambient noise figure (Fa) observed during the AE net activity when most stations are heard is 13.5dB. This is a realistic design criteria for AE contacts.
Fig 3 shows the G/T ratio for the base configuration and several options for improvement. The base configuration is a TS2000 receiver with 10m of RG213 to a 4 element Yagi antenna, with an Fa of 13.5dB. The LNA option is the LNA described in this article, and the higher gain antenna option is for a 4.5m boom Yagi with a gain of about 14dBi.
Fig 4 shows the estimated cost of improvement of G/T (or S/N) vs cost for the improvement options. The graph tells the story!
Whilst the LNA has a gain of 28dB and a noise figure of around 1.2dB, it improves S/N by just 1dB in the station scenario described when used on a sufficiently selective antenna.
In this scenario, there is less than 0.1dB of improvement in locating the LNA near the antenna rather than near the receiver.
The lowest ambient noise observed for an hour or two on some days in the afternoon is 6.0dB. This is an extremely optimistic design target, it considers the very lowest noise environment that occurs at times when other stations almost never active.
Fig 5 shows the G/T ratio for the base configuration and several options for improvement. The base configuration is a TS2000 receiver with 10m of RG213 to a 4 element Yagi antenna, with an Fa of 6dB. The LNA option is the LNA described in this article, and the higher gain antenna option is for a 4.5m boom Yagi with a gain of about 14dBi.
Fig 6 shows the estimated cost of improvement of G/T (or S/N) vs cost for the improvement options. Again, the graph tells the story!
Whilst the LNA has a gain of 28dB and a noise figure of around 1.2dB, it improves S/N by just 3.5dB in the optimistic scenario described when used on a sufficiently selective antenna.
In this scenario, there is 0.2dB of improvement in locating the LNA near the antenna rather than near the receiver.
To summarise the analysis of G/T impact in the realistic and unrealistically optimistic scenarios:
The following conclusions are made:
Term | Meaning |
dB | decibel - power ratio |
dBm | decibels - power wrt 1 mW |
ENB | Effective Noise Bandwidth |
Fa | Ambient noise figure |
FET | Field Effect Transistor |
G/T | Gain to Noise-temperature ratio |
IMD | Inter Modulation Distortion |
LNA | Low Noise Amplifier |
MOSFET | Metal Oxide Semiconductor Field Effect Transistor |
RFI | Radio Frequency Interference |
SINAD | signal-plus-noise-plus-distortion to noise-plus-distortion ratio |
SSG | Standard Signal Generator |
S/N | Signal to noise ratio |
VOX | Voice operated switch |
Version | Date | Description |
1.01 | 28/05/2007 | Initial. |
1.02 | ||
1.03 |
Use at your own risk, not warranted for any purpose. Do not depend on any results without independent verification.
© Copyright: Owen Duffy 1995, 2021. All rights reserved. Disclaimer.