linrad support: High performance hardware for linrad: The 70 to 10.7 MHz converter.
(Sept 07 2003)

Hardware to convert from 70 MHz to 10.7 MHz

The local oscillator

The local oscillator uses quartz crystals to provide very low noise sidebands. Click here for details of the local oscillator

The mixer

The A balanced mixer with J310 junction-FET transistors running as switches is used. To get the dynamic range required to fit the 10.7 MHz input of the 10.7 to 2.5MHz converter it is necessary to use several transistors in parallel.

The RF amplifier and filter

The RF amplifier and filter has 2dB gain alltogether. The MOS-FET serves as a buffer that isolates the filter from the input and provides an input impedance of 50 ohms. The filter has a flat bandwidth of about 1MHz and provides 75dB attenuation at the mirror image frequency. The filter should give about 120dB attenuation at the mirror frequency but there is a substantial leakage around it. This leakage could be eliminated by screening walls but I have not introduced any such screens since it is easier to provide some selectivity in the converters that preceed the RX70 unit which, after all, is just an IF frequency.

Click here for details of the 70 MHz amplifier and filter

The complete 70 to 10.7 MHz converter

The schematic diagrams contained in the links above as well as the DC supply schematics are laid on a double sided PCB. The unit has two channels so there are two RF amplifier/filter units and two mixer units.

All the layout files for the EDWIN CAD system as well as all PCB production files can be down loaded here CAD files for 70 MHz unit There is also a photo of the assembled unit.

The 70 MHz to 10.7 MHz converter described here is a free design. All the information on this page is free and may be used by anyone for any purpose

Testing and tuning

The RX70, the 70 to 10.7 MHz converter can be tested and tuned by means of simple standard instruments if the unit is connected to a RX10700 unit, in turn connected to a RX2500 unit and a computer with a Delta44 board installed running Linrad. Look here for details: Testing and tuning the RX70

Sensitivity and noise figure

The NF (noise figure) of the RX70 itself is 8 dB, but when the unit is used together with a RX10700 a RX2500 and a modified Delta44, the system noise figure is 15 dB at the 70 MHz inputs when the Delta44 is run at minimum gain. Table 1 illustrates how the system noise figure (sensitivity) and inband dynamic range is affected by the A/D converter performance.


A/D converter System NF Noise floor in Saturation Dynamic range (dB) 500Hz(dBm) (dBm) noise floor(dBc/Hz) Ideal 11.5 -134.5 -12 -149.5 Delta44(mod) 12.0 -135.0 -25 -137.0 at max gain Delta44(mod) 14.8 -132.2 -13 -146.2 at min gain
Table 1. The system performance depends on the A/D converter used. This table shows typical data for the dynamic range within the visible passband.

For the RX70 unit, the 1dB compression point is at +17dBm. The unit is limited by the J310 mixer. The noise figure of the 70MHz input amplifier itself is 5dB but after loosing about 5dB in the filter the net gain from the input to the mixer is 2dB.

All frequency mixers produce spurious responses at very high signal levels. The RX70 + RX10700 combination has two frequency mixers in series and therefore the spurious responses of both of them are added. Besides the spurious responses that are produced by mechanisms involving overtones of the desired signal, there are mirror image frequencies and responses produced by overtones of the local oscillators. Table 2 shows the first order responses from 35 MHz to 310MHz. These are the responses that grow linearly with the input signal.


Frequency Level Mechanism (MHz) (dB) 70.04 0 RF-LO (Desired response) 70.06 ~70 RF-LO (Baseband mirror image) 48.56 -75 LO-RF (Mirror image) 75.06 -120 RF-LO (RX10700 mirror image) 107.86 -107 2xLO-RF 129.34 -99 RF-2xLO 167.16 -117 3xLO-RF 188.64 -114 RF-3xLO 226.46 -104 4xLO-RF 247.94 -104 RF-4xLO 285.76 -106 5xLO-RF 307.24 -99 RF-5xLO . . . . . . 900.24 -88 RF-15xLO
Table 2.First order responses of the RX70 + RX10700 + Delta44 combination. Level is dB below desired response at 70.04MHz.

The data of table 2 shows channel 1, the worst channel. The stop band attenuation of the 70MHz filter could be greatly improved by the addition of more screen walls that would have to be screwed to the lid of the box. Just removing the lid will improve the mirror image by 10dB and taking the board out of the box improves by anothewr 10 dB. Bearing in mind that the RX70 unit is an IF input, the first order false responses should not affect system performance. The converters preceeding the RX70 unit should have an IF amplifier to make the mixer see a low noise figure. This IF amplifier will reduce the first order spurs by at least 30dB and with some selectivity at the antenna input all first order spurs belonging to the 70MHz IF will be insignificant.

The second order spurious response, the overtone of 35.020 MHz generated by the RF amplifier has a level of -33dB at an input level of 0dBm. This signal drops twice as fast as the input signal (in the dB scale) With a system noise figure of 15dB the noise floor is at -132dBm in 500 Hz bandwidth so the second order spur is at the noise floor at an input level of -49.5 dBm at 35MHz. The corresponding second order intercept point is at +33dBm. It is important that the converters used in front of the RX70 unit do not deliver more than -50dBm below 40MHz or so. Second order intermodulation between signals, the sum frequency of which the second harmonic is a special case will then be well below the system noise floor. Remember that the first amplifier at the antenna input will raise the noise floor by at least 10 dB.

There are several other second order spurious responses between 35 and 100 MHz. They are listed in table 3.


Frequency Level at 0dBm Mechanism (MHz) (dB) 35.020 -33 LO-2xRF 53.93 -125 2xLO-2xRF 64.67 -132 2xRF-2xLO 83.58 -150 3xLO-2xRF 94.32 -146 2xRF-3xLO 71.295 -128 RF-LO (RX10700 response)
Table 3.Second order responses of the RX70 + RX10700 + Delta44 combination at 0dBm input power. Level is dB below desired response at 70.04MHz.

At 0dBm input power the second order responses are close to the noise floor in 500Hz bandwidth except for the very strong second harmonic response at 35 MHz discussed above. Only the RX10700 response, 2 x LO(13.275) - 2 x RF(11.995) will survive the filters in the preceeding converter that will produce the 70MHz signal but the level is low.

Table 4 shows the strongest higher order spurs in the 35 to 100 MHz range.


Frequency Level at 0dBm Mechanism (MHz) (dB) 35.953 -156 2xLO-3xRF 70.973 -132 6xRF-7xLO 71.440 -102 5xLO-4xRF 71.058 -150 11xRF-13xLO
Table 4.The strongest higher order responses of the RX70 + RX10700 + Delta44 combination at 0dBm input power. Level is dB below desired response at 70.04MHz.

Spurious response data is not often published in receiver tests. As a comparison, the IC706MKIIG was tested on 50 MHz. With preamp off the noise figure is 9.5dB giving it a 5 dB lower noise floor than the RX70 system. The IC706 has 1dB compression at about -16 dBm. Close to the 1 dB compression point the IC706 has a large number of spurious responses but already at -30dBm it has about half a dozen well audible responses within +/- 1 MHz. The strongest of them reaches the noise floor in 500Hz bandwidth at an input level of -52dBm. I do not know how better receivers perform but I do think spurious responses is an important aspect of receiver performance.

Two signal dynamic range

A weak and a strong signal are combined in a 3dB hybrid and the result is sent to channel 1 of the RX70 + RX10700 + RX2500 + Delta44 (mod) signal processing chain. The weak signal is placed at 70.040 MHz and a notch filter at the same frequency is inserted between the strong signal and the hybrid.

The level of the strong signal is adjusted until the S/N of the weak signal is degraded by 3dB. Table 5 shows the result of this test with the Delta44 board in maximum and in minimum gain mode.


Frequency        Level             Level   
separation       [+4dB]           [-10dB]
  (kHz)      (dBm) (dBc/Hz)   (dBm) (dBc/Hz)
-2000          27     186        27    189
-1000          21     180        20    182
 -500          18     177        15    177          
 -300          15     174        13    175
 -200          13     172        11    173
 -150          13     172        10    172
 -100          12     171         9    171
  -80           7     166         7    169
  -70           6     165         5    167
  -60           8     167         8    170        
  -50         -13     146       -24    138
    .           .      .          .     .
 visible        .      .          .     .
 passband     -13     146       -24    138  
 -50 to +20     .      .          .     .  
    .           .      .          .     .
   20         -13     146       -24    138
   30           2     161       -19    143        
   40           4     163         2    164            
   50           6     165         5    167             
   60           9     168         6    168            
   70          10     169         7    169    
   80          11     170         8    170    
  100          12     171         9    171    
  150          13     172        10    172    
  200          13     172        11    173    
  300          13     172        12    174
  500          15     174        13    175
 1000          17     176        15    177
 2000          27     186        27    189
Table 5. Two signal dynamic range of a RX70 + RX10700 + RX2500 + Delta44
combination.
The Delta 44 is run at low gain [+4dB] and at high gain [-10dB]

An inspection of table 1 shows that the dynamic range is better by 8dB for offending signals within the visible passband if the Delta44 is operated in minimum gain mode. One can also see that the dynamic range is not much better for signals outside the visible passband if the Delta44 is operated in minimum gain mode. A better soundcard than the Delta44 will not improve performance much unless the gain is reduced in the converters. Reducing the gain while using the Delta44 is a bad idea because lower gain will not reduce the noise floor much, the Delta 44 already contributes most of the noise floor. Less gain in any of the converters will reduce the noise figure and bring the noise floor closer to the point of saturation.

The parts of the RX70 + RX10700 + RX2500 + Delta44 system are matched to each other to optimise the performance for a 500 kHz wide IF strip at 70 MHz. At frequency separations above 100 kHz performance is limited by the sideband noise generated by the RX70 unit. At 2 MHz frequency offset, the system is limited by saturation. As can be seen from the table 0.5W is required at the input for a 3dB loss of S/N for a very weak signal. The noise floor then drops by about 0.5dB while the signal drops by 3.5dB.

Three signal dynamic range

For two signals within the visible passband, the third order intercept point, IP3, is +17 to +23 dBm depending on the frequency separation. For two signals outside the visible spectrum IP3 is higher. Table 6 shows the level of the third order intermodulation signal at different signal levels and frequency separations. The received signal or intermodulation product is at 70.040 MHz
 Input       Signals       IM3(30kHz)   IM3(300kHz)
 level       S-meter        S-meter      S-meter
 (dBm)        (dB)            (dB)         (dB)
   0          140             ***          82.2
  -5          135             71.3         66.8
 -10          130             59.0         51.8
 -15          124.9           39.0         36.9
 -20          120.0           25.6         22.0
 -25          114.9           14.9          7.2
 -30          110.0            1.5         -7.0
Table 4. Signals and third order intermodulation products. The levels are from the Linrad S-meter. Levels above 124.9 are extrapolated.

The third order intercept point can be obtained from the data in table 4:

IP3(30kHz) = +26dBm
IP3(300kHz) = +29dBm

Compared to the RX10700 + RX2500 + Delta44 combination, adding the RX70 converter degrades the third order intercept point by 7dB while the noise figure is degraded by 1 dB. The total loss of intermodulation free dynamic range is thus 8 dB. By replacing the J310 FETs with a GaAs FETs it would be possible to improve the mixer IP3 from about 31dBm to about 40dBm according to i.e. U.Rohde QEX Jan/Feb 2003 p 27. I have made a few experiments with MGF1302 GaAs FETs but they were not encouraging. With IP3 = 29dBm at a noise figure of 15dB the IM3 free dynamic range in 500 Hz bandwidth is 107 dB. This is good enough for my needs, the J310 mixer is just about good enough to match the sideband noise of the crystal oscillators. Improving the RX70 mixer will not make any significant improvement to the two signal dynamic range, table 5 above, so I see no reason to replace the J310 transistors.