2001-09-24
Yet Another Comb Generator
DSE had a special on a range of clock oscillator modules, so I purchased 20 of each available frequency on the off chance that one day they'd be useful. I didn't have to wait too long before I needed one. While putting together a S-Band dish antenna for P3D I found the need for a signal source around the 2.4 GHz region. I could have purchased one of those nifty kits from SSB Electronic or Down East Microwave, but instead I decided to make something much cheaper.
The 14.91 MHz clock modules from DSE were just about perfect, the 161st harmonic of 14.91 MHz is 2400.51 MHz. Murphy was nice enough to give me a break here, all the other clocks had harmonics at least 15 MHz away. This device isn't nice and clean spectrally like the 'weak signal source' kits are, instead it produces a nice comb of harmonics starting from the 3rd (44.73 MHz) up to beyond what I can measure.
Here are the first 3-200:
44.73 59.64 74.55 89.46 104.37 119.28 134.19 149.10
164.01 178.92 193.83 208.74 223.65 238.56 253.47 268.38
283.29 298.20 313.11 328.02 342.93 357.84 372.75 387.66
402.57 417.48 432.39 447.30 462.21 477.12 492.03 506.94
521.85 536.76 551.67 566.58 581.49 596.40 611.31 626.22
641.13 656.04 670.95 685.86 700.77 715.68 730.59 745.50
760.41 775.32 790.23 805.14 820.05 834.96 849.87 864.78
879.69 894.60 909.51 924.42 939.33 954.24 969.15 984.06
998.97 1013.88 1028.79 1043.70 1058.61 1073.52 1088.43 1103.34
1118.25 1133.16 1148.07 1162.98 1177.89 1192.80 1207.71 1222.62
1237.53 1252.44 1267.35 1282.26 1297.17 1312.08 1326.99 1341.90
1356.81 1371.72 1386.63 1401.54 1416.45 1431.36 1446.27 1461.18
1476.09 1491.00 1505.91 1520.82 1535.73 1550.64 1565.55 1580.46
1595.37 1610.28 1625.19 1640.10 1655.01 1669.92 1684.83 1699.74
1714.65 1729.56 1744.47 1759.38 1774.29 1789.20 1804.11 1819.02
1833.93 1848.84 1863.75 1878.66 1893.57 1908.48 1923.39 1938.30
1953.21 1968.12 1983.03 1997.94 2012.85 2027.76 2042.67 2057.58
2072.49 2087.40 2102.31 2117.22 2132.13 2147.04 2161.95 2176.86
2191.77 2206.68 2221.59 2236.50 2251.41 2266.32 2281.23 2296.14
2311.05 2325.96 2340.87 2355.78 2370.69 2385.60 2400.51 2415.42
2430.33 2445.24 2460.15 2475.06 2489.97 2504.88 2519.79 2534.70
2549.61 2564.52 2579.43 2594.34 2609.25 2624.16 2639.07 2653.98
2668.89 2683.80 2698.71 2713.62 2728.53 2743.44 2758.35 2773.26
2788.17 2803.08 2817.99 2832.90 2847.81 2862.72 2877.63 2892.54
2907.45 2922.36 2937.27 2952.18 2967.09
And an image of a section of the comb in the UHF region around 500 Mhz (yes my sweep isn't linear, I know):
Because it is used for antenna testing, the power connection (the RCA) is well decoupled with caps and ferrite beads. There is very minimal leakage via the power lead, I checked this carefully. Another thing that I did't require, but is a nice feature, is the LO rejection. Check out this image of my VR-500 right alongside the unit, tuned to 14.91 MHz with antennas parallel I expected a max scale reading, instead I got only a few bars. The LO is undetectable 1 metre away, even if you coil the power lead around the antenna!
How does it work? The few dBm signal out of the clock module is amplified strongly by an over-driven BFR91, we don't care about linearity, in fact we want harmonics, that is the point. Next this RF mess is blasted into another BFR91, this time biased into class-C operation. The collector circuit is a shorted 2.4 GHz 1/4 wave stub in the hope of favouring S-Band harmonics. Finally there is a high pass filter with a break point of about 30 MHz and a low value output coupling cap.
I didn't 'design' this device, I just built it as I went on gut instinct. So please don't email me saying my value choices are insane (they probably are). NB: I believe I raised the 100 Ohm resistor in the final collector circuit to 560 Ohms to reduce the power consumption and output a bit. Next time I have the unit apart I'll check.
3 comments.
Attachments
| title | type | size |
|---|---|---|
| circuit postscript source | application/postscript | 13.571 kbytes |