NF MEASUREMENTS AT THE OREBRO EME MEETING 2013.
(June 27 2013)

The different kinds of measurements performed at the meeting.

Standard measurements were made on all microwave bands using an Agilent 8973A noise figure meter with a N4000A noise head. Those measurements were done by Mart, SM0ERR and they are (will be) reported here http://www.moonbouncers.org

The data from the standard measurements for 1296 MHz is listed in table 1.


   Unit             NFmeas  Adapter   NF
(No) (name)          (dB)                (dB) 
 4  RW3BP-13        0.125         Nf-SMAm       0.116
 5  SM5QA           0.292          none         0.292
 8  RW3BP-6         0.117         Nf-SMAm       0.108
 9  RW3BP-12        0.144         Nf-SMAm       0.135
10  SM7FWZ          0.139         Nf-SMAm       0.130
11  SM0ERR          0.189         Nf-SMAm       0.180
14  PA3DZL-13okt    0.499         Nf-SMAm       0.490
15  PA3DZL          0.200         Nf-SMAm       0.191
16  G3LTF           0.196         Nf-SMAf       0.187
17  SM0DFP          0.261         Nf-SMAm       0.252
18  SM0DFP_3        0.846         Nf-SMAf       0.837 
19  SM0DFP_2        0.871         Nf-SMAf       0.862
20  SM7GEP-0.217    0.217         Nf-SMAf       0.208
21  SM7GEP-0.562    0.562         Nf-SMAm       0.553
22  SM7GEP-0.484    0.484         Nf-SMAm       0.475
23  SM5BSZ          0.219         Nf-SMAm       0.210
Table 1. Measurements on 1296 MHz by SM0ERR using a 8973A with a N4000A noise head. The N to SMA adapters are assumed to degrade the NF by 0.009 dB according to [3]. For details about the amplifiers see the details section.


Precision relative measurements on 1296 MHz were done by me, SM5BSZ, Leif using the setup described here: Using the HP8970A with circulators under computer control. The raw data from those measurements is listed in table 2.


   Unit             Teraw   Troom   Teraw26
(No) (name)          (K)    (C)     (K)
 4  RW3BP-13         8.77   26.2   8.74
 5  SM5QA           33.31   26.2  33.28 
 8  RW3BP-6          8.96   27.2   8.83
 9  RW3BP-12         8.65   27.3   8.51
10  SM7FWZ          13.60   27.6  13.39 
11  SM0ERR          13.60   27.8  13.37
12  SM4IVE          11.66   27.8  13.43
14  PA3DZL-13okt    33.62   27.7  33.40
15  PA3DZL          14.15   27.6  17.82 
16  G3LTF           11.34   27.5  11.14
17  SM0DFP          16.94   26.5  16.65
18  SM0DFP_3        61.60   26    61.60
19  SM0DFP_2        62.12   26    62.12
20  SM7GEP-0.217    25.59   26    25.59
21  SM7GEP-0.562    32.94   26    32.94
22  SM7GEP-0.484    33.22   26    33.22
23  SM5BSZ          11.80   26    11.80
24  PA2DW           14.17   26    14.17
25  PA2DWrelay      22.60   26    22.60
Table 2. Raw data from precision measurements. Teraw26 is corrected using the temperature dependence of a typical LNA 0.13 K/degree C to compensate for the differences in the ambient temperature to give the data that would have been obtained at a room temperature of 26 degrees with the LNA in the air flow of a table fan.


We also made measurements to compare the zero points of the 8973A/N400A setup and the 8970A setup with circulators with the zero point established by Serge, RW3BP by use of radio astronomy. The conclusion is that the measured Te in table 2 is too low by 0.5 K.


   Unit              Te      NF(US)   NF(Japan)    SM0ERR     Diff
(No) (name)          (K)     (dB)       (dB)        (dB)      (dB)  
 4  RW3BP-13         9.24   0.1362     0.1348      0.116     0.0202
 5  SM5QA           33.78   0.4785     0.4739      0.292     0.1865
 8  RW3BP-6          9.33   0.1375     0.1361      0.108     0.0295
 9  RW3BP-12         9.01   0.1329     0.1315      0.135     0.0021
10  SM7FWZ          13.89   0.2032     0.2012      0.130     0.0732
11  SM0ERR          13.87   0.2029     0.2009      0.180     0.0229
12  SM4IVE          13.93   0.2038     0.2017        -          -
14  PA3DZL-13okt    33.90   0.4801     0.4755      0.490    -0.0099
15  PA3DZL          18.32   0.2660     0.2634      0.191     0.0750
16  G3LTF           11.64   0.1709     0.1692      0.187    -0.0161
17  SM0DFP          17.15   0.2495     0.2470      0.252    -0.0025 
18  SM0DFP_3        62.10   0.8427     0.8348      0.837     0.0057
19  SM0DFP_2        62.62   0.8490     0.8411      0.862     0.0130 
20  SM7GEP-0.217    26.09   0.3741     0.3705      0.208     0.1661
21  SM7GEP-0.562    33.44   0.4740     0.4694      0.553     0.0790
22  SM7GEP-0.484    33.72   0.4777     0.4731      0.475     0.0027
23  SM5BSZ          12.30   0.1804     0.1786      0.210    -0.0296
24  PA2DW           14.17   0.2072     0.2051
25  PA2DWrelay      23.10   0.3329     0.3298
Table 3. Final results from the 1296 MHz LNA measurements at the Orebro EME meeting 1013. Te and the noise figures are corrected to an ambient temperature of 26 C. Each amplifier is run long enough to reach thermal equilibrium in the air flow of a table fan. The reference temperature is 290 K in the US and 293 K in Japan. A comparison with the data from table 1 is included.


It is interesting to note that the L LNA with number 17 in table is measured correctly with the N4000A noise head on the 8973A. These amplifiers are well matched to 50 ohms so impedance variations on the noise head should be relatively harmless. The amplifiers 18 and 19 are also well matched to 50 ohms and they also show small errors due to the noise head impedance variations.

Some of the non-matched amplifiers show large discrepancies. Maybe a careful study of amplifier 5 would reveal something interesting. Maybe it oscillates...

Details of precision measurements in chronological order.

A table fan was used to direct an air flow on the tested amplifiers and thermal equilibrium was awaited before measurement started. This took typically 10 minutes for each amplifier.


[1] The first amplifier tested was by RW3BP. A modified G4DDK design. It carried a label NF=0.145, RL=13 dB. The NF after 50 averages of the 64 averages from the 8970A was 0.1337 dB. A second measurement yielded exactly the same result. The corresponding temperature is 9.07 K. The standard deviation with this amount of averaging is 0.026 K corresponding to 0.0004 dB. The ambient temperature was 26.2 degrees. A selected N(female)-SMA(male) adapter which carries a colour marking was used between the N(male) of the circulator system and the DUT which is RW3BP-13 in the tables above.


[2] The second test was with a Spinner N(female)-APC3.5(male) to replace the adapter for the RW3BP-13. Quite unexpectedly the NF was significantly degraded. NF was 0.1415 dB (Te=9.60 K) The Spinner adapter was cleaned by use of isopropylic alcohol and cotton swabs (Johnson & Johnson) and dried with compressed gas just before it was inserted.

The selected adapter was then cleaned the same way. After cleaning the NF had increased from 0.1337 dB to 0.1382 dB. Unmounted and again mounted it showed 0.1374 dB. The cleaning was repeated with professional swabs intended for optics on the suspicion that the swabs intended for human usage contain something that gets deposited in the connector. That made no difference. NF=0.1375 dB. Then the reference connector was dried by holding it quite some time above a hot soldering iron. That brought the noise down to 0.1351 dB (9.16 K) After heating the Spinner adapter it showed NF=0.1346. (Te=9.13 K) We decided to avoid isopropylic alcohol and to make a study on connector cleaning when back home. The original 0.1337 differs from 0.1351 by 0.0014 dB. That can not be explained by statistical variations and is probably an effect of heat transfer from the adapter that was put into the circuit while it was still significantly above room temperature.


[3] Then we inserted two adapters, SMA(female)-N(male) connected to N(female)-SMA(male) between the Spinner adapter and the RW3BP-13 LNA. That yielded Te=10.39 K, a change by 1.26 K from the reading just before this adapter pair was inserted. From this we concluded that the dissipative losses of the selected and colour marked N(female)-SMA(male) adapter rises the system temperature by 0.63 K.


The next day the regular amplifier measurements started after the 1296 MHz filter was re-tuned for a perfectly flat passband.


[4] RW3BP-13 Te=8.77K at an ambient temperature of 26.2 degrees.


[5] SM5QA This amplifier is made with a MMIC SKY67151-396LF. The IIP3 is specified as +16 dBm with 5V DC supply and a current of 55 mA. The result: Te=33.31 K.


[6] PA2DW a G4DDK amplifier. Te=16.23 K. This amplifier was re-tuned later to a better NF and tolerance to high WSVR without oscillations. See below.


[7] PA2DW with relay It was impossible to connect thefemale SMA connector of the relay to the selected colour marked N(female)-SMA(male) adapter for mechanical reasons. It was necessary to use a N(female)-SMA(female) and a SMA(male)-SMA(male) adapter. The observed Te was 28.49 K.

With the same N(female)-SMA(female) and SMA(male)-SMA(male) adapter the RW3BGP-13 showed Te=13.02 K which is 4.25 K worse than the temperature observed with the colour marked adapter. Assuming that the N to SMA adapters have equal losses one can conclude that the SMA(male)-SMA(male) adapter degrades Te by 4.25 K. (=0.063 dB on the NF)

If the PA2DW amplifier could have been connected to the colour marked adapter, the observed Te would have been 28.49-4.25=24.24 K. This means that the relay and the short cable degrade system performance from 16.23 K to 24.24 K. A loss of 8.01 K (0.118 dB on ther NF.) On an antenna with T=30K S/N would change by a factor of (30+24.24)/(30+16.23) = 1.17 (=0.68 dB) due to the relay losses. On an ideal antenna with T=3.5K the S/N loss would be 1.4 times (=1.46 dB)

This amplifier was re-tuned later to a better NF without the relay. See below.


[8] RW3BP-6 This amplifier was marked NF=0.15 dB, RL= 6dB. The ambient temperature was 27.2 degrees and the Te value 8.96 K.


[9] RW3BP-12 This amplifier was marked NF=0.15 dB, RL= 12dB. The ambient temperature was 27.3 degrees and the Te value 8.65 K.


[10] SM7FWZ A G4DDK amplifier with serial number 1705. The result was Te=13.60 at an ambient temperature of 27.6 K.


[11] SM0ERR A G4DDK amplifier with MGF4919G. Te=13.60 at an ambient temperature of 27.8 degrees.


[12] SM4IVE A G4DDK amplifier. Te=11.66 K at an ambient temperature of 27.8 degrees.


[13] SM4IVE with relay and adapter Te=21.80 K at an ambient temperature of 27.8 degrees. The relay/adapter combo results in a N(male) connector. Measurement is made with a N(female)-N(female) colour marked adapter replacing the N to SMA adapter.

The N(male)-SMA(male) and the N(female)-N8female) was replaced by the N(female)-SMA(female) and SMA(male)-SMA(male) adapter that was used in [7]. That gave Te=22.16K for SM4IVE with relay but without adapter. Correcting for the loss of the SMA(male)-SMA(male) Te with relay would be 22.16-4.25=17.91 dB.

The problem with the SM4IVE with relay and adapter is that the SMA connector of the N(male)-SMA(male) does not fit mechanically. The SMA connector is well above 1 mm from tighted for mechanical reasons. Most probably there is some loss because of this, but since the loss of the N(female)-N(female) colour marked adapter is currently unknown it is impossible to say how much loss the improper SMA connection causes.


[14] PA3DZL-13okt This amplifier is for 2320 Mhz and was measured by mistake. Te=33.62 at an ambient temperature of 27.7.


[15] PA3DZL A G4DDK amplifier with serial number 905. Te=14.15 at an ambient temperature of 27.6 degrees.


[16] G3LTF This amplifier was measured with a N(female)-SMA(female) adapter, the colour marked reference. It is reasonable to assume the loss is the same as for the N(female)-SMA(male) but whether this is true is not known. The ambient temperature was 27.5 degrees and Te=11.34 K.


[17] SM0DFP An L LNA amplifier by AD6IW. Te=16.94 at an ambient temperature of 26.5 degrees.


[18] SM0DFP_3 A dual amplifier with a hybrid input. Te=61.60 at an ambient temperature of 26 degrees. The result from NF measurements 2012 were NF=0.84 dB, Gain=27.5 dB.


[19] SM0DFP_2 A dual amplifier with a hybrid input. Identical to SM0DFP_3. Te=62.12K at an ambient temperature of 26 degrees. The result from NF measurements 2012 were NF=0.87 dB, Gain=29.0 dB.


[20] SM7GEP-0.217 This is a G4DDK design with a label "0.217/35.5" The device is a MGF4919G. The input is a SMA male so the measured Te = 25.59 K is directly comparable to [15]. The adapters do probably not differ significantly however so the noise temperature can reasonably well be compared with the noise temperature of other amplifiers having a female SMA as input.


[21] SM7GEP-0.562 This amplifier carries a label "0.562/14.5" The device is a ATF10135 and the measured Te is 32.94. Since the gain is only 14.5 dB the effect of the system NF of the test setup can not be neglected. The system has Te=20.5 K with the attenuator setting needed for this particular amplifier so the contribution is 20.5/28.18 = 0.73 K. The corrected Te is thus 32.21 K. The input is SMA female.


[22] SM7GEP-0.484 This amplifier carries a label "0.484/17.6" The measured Te is 33.22 which with the correction for the second stage is 32.67 K. The input is SMA female.


[23] SM5BSZ An experimental design with the MGF4919g. Te=11.80K.


[24] PA2DW retuned The amplifier had oscillation problems when high SWR was presented to the input. For details, look here. The final Te was 14.17K, an improvement of 2.06 K.


[25] PA2DW retuned with relay It was again necessary to use a N(female)-SMA(female) and a SMA(male)-SMA(male) adapter. Tuning was without the relay, [24]. The observed Te was 26.86 K, 1.6 K better than before modification and re-tuning. Corrected for the SMA(male)-SMA(male) adapter the Te for this amplifier with the relay is 26.86 - 4.25 = 22.60 K.