H-Mode mixer - Transformers - TT4-1A

Fairchild FSA3157 + 74AC04

Because the FSA3157 results are so promising this switch is also put to the test with a fundamental mode 74AC04 squarer which allows for tweaking the mark/space ratio.

The following table lists the results with the FSA3157 and the 74AC04 squarer adjusted for best RF-IF isolation at 9MHz:

FSA3157 + TT4-1A + 74AC04 Vdd = 7.0V, Vbias = 2.5V Best bias
Band BPF mixer MDS -dBm IF rej spur mixer IIP3 (dBm) @ 20KHz spacing Bias IIP3
IL CL - + avg -6 -3 0 +3 +5 0
-dB -dB BPF BPF -dB dB dBm dBm dBm dBm dBm V dBm
160 3,3 5,00 133 129 49,2 42,4 3,0 42,4
80 1,4 4,92 133 131 49,2 45,2 46,7 48,0 51,2 49,5 2,0 50,5
40 2,3 4,92 133 130 49,6 44,7 2,3 44,7
30 3,4 4,88 133 129 49,6 43,0 2,5 43,0
20 3,4 4,88 133 129 49,2 41,4 2,7 41,4
17 3,0 4,84 133 129 48,0 40,3 2,7 40,3
15 3,9 4,88 133 128 48,0 4,5 39,6 2,8 39,6
12 3,7 4,88 133 129 44,8 37,6 2,7 37,6
10 2,3 4,92 133 130 42,8 36,8 2,8 36,8
6 5,40 131 42,4 35,5 3,0 35,5
4 5,60 131 28,4 33,9 2,5 33,9

The RF-IF isolation is very good up to 12M from where it starts to drop off. When adjusted separately on each band all bands made around 49dB RF-IF isolation at 9MHz.

Conversion loss is slightly lower with this adjustment and 15M spur level is rather bad at 4,5dB above MDS average!

The following table lists the results with the FSA3157 and the 74AC04 squarer adjusted for best spur rejection on 15M. This adjustment is about 3 quarters of a turn away from the best RF-IF isolation point with the 20 turn mark/space ratio trimpot of the squarer:

FSA3157 + TT4-1A + 74AC04 Vdd = 7.0V, Vbias = 2.5V Best bias
Band BPF mixer MDS -dBm IF rej spur mixer IIP3 (dBm) @ 20KHz spacing Bias IIP3
IL CL - + avg -6 -3 0 +3 +5 0
-dB -dB BPF BPF -dB dB dBm dBm dBm dBm dBm V dBm
160 3,3 5,00 133 129 31,2 42,4 3,0 42,4
80 1,4 5,00 133 131 30,9 45,2 46,2 47,7 50,5 50,2 2,0 49,2
40 2,3 5,00 133 130 31,2 45,4 2,3 45,4
30 3,4 4,96 133 129 30,8 43,8 2,5 43,8
20 3,4 4,96 133 129 32,8 41,9 2,5 41,9
17 3,0 4,96 133 129 30,8 40,3 2,5 40,3
15 3,9 4,96 133 129 31,6 -5,6 39,6 2,7 39,6
12 3,7 4,96 133 129 32,0 38,1 2,8 38,1
10 2,3 5,00 133 130 32,0 36,8 3,0 36,8
6 5,48 132 33,2 35,2 3,0 35,2
4 5,64 131 22,8 33,6 2,0 33,6

Conversion loss is back at the usual level however the average spur level is now reduced by 10dB to only 5,6dB below MDS! This means that most spurs are gone. Most of the loud ones are less loud too. The same adjustment seemed to be successful at all other bands too, although without a detailed spur map it is difficult to quantify exactly. This point requires a bit more investigation. It could be that a PIC controlled per band adjustment is the best solution.

IMD did not change much with the adjustment for best spur reduction.

The price to pay for the spur null-out adjustment is a noticeable decrease in RF-IF isolation at 9MHz being only slightly better than 30dB on all bands except 4M.

The very good IP3 performance of the FSA3157 combined with the spur-nullout with the 74AC04 squarer makes this configuration a very interesting choice. The fundamental squarer also makes 6M and even 4M a possibility with subtractive mixing without a great increase in conversion loss. The LO will still remain below 80MHz which is an advantage as the wideband SFDR is better far below the DDS nyquist frequency.


Back to Transformers

Back to the TOC