WEBVTT NOTE Keying Defects, IMD, and Spread Spectrum 1 00:00:00.000 --> 00:00:01.579 Key clicks. 2 00:00:01.579 --> 00:00:10.972 Question. What is the primary effect of extremely short rise or fall time on a CW signal? 3 00:00:11.872 --> 00:00:15.845 Answer. The generation of key clicks. 4 00:00:16.445 --> 00:00:22.125 Question. What is the most common method of reducing key clicks? 5 00:00:23.025 --> 00:00:28.761 Answer. Increase keying waveform rise and fall times. 6 00:00:29.361 --> 00:00:43.126 A sharp edge is broadband, Fourier again. Clicks are the harmonics of that edge, audible as sharp thumps for kilohertz either side of a signal that is otherwise 50 Hz wide. 7 00:00:43.126 --> 00:00:48.488 The fix is deliberate shaping: a few milliseconds of rise and fall. 8 00:00:48.488 --> 00:01:04.014 Too slow and the keying sounds mushy and limits speed; a few milliseconds is the compromise every modern radio implements, and the reason a badly-adjusted one is immediately obvious on the band. 9 00:01:04.014 --> 00:01:18.979 Note this is the same phenomenon as the PSK zero-crossing rule and the sinusoidal pulse shaping from the previous lesson. Sharp transitions cost bandwidth, in every mode. AFSK distortion. 10 00:01:18.979 --> 00:01:26.264 Question. What is a common cause of overmodulation of AFSK signals? 11 00:01:27.164 --> 00:01:31.564 Answer. Excessive transmit audio levels. 12 00:01:32.164 --> 00:01:41.123 The most common digital-mode fault there is, and the vertical lines on a waterfall from the General track are its signature. 13 00:01:41.123 --> 00:01:50.811 Question. What parameter evaluates distortion of an AFSK signal caused by excessive input audio levels? 14 00:01:51.711 --> 00:01:57.017 Answer. Intermodulation Distortion (IMD). 15 00:01:57.617 --> 00:02:05.617 Question. What is considered an acceptable maximum IMD level for an idling PSK signal? 16 00:02:06.517 --> 00:02:10.477 Answer. -30 dB. 17 00:02:11.077 --> 00:02:23.276 A concrete target, which is what makes this actionable. Software such as fldigi measures IMD on a received idling PSK signal and reports it. 18 00:02:23.276 --> 00:02:37.653 Worse than −30 dB means reduce your drive, and it is a measurement rather than an opinion. If you can hear your own signal on a WebSDR, you can measure your own IMD. 19 00:02:37.653 --> 00:02:45.408 Few operators do, and the bands would be noticeably cleaner if they did. Codes, compared. 20 00:02:45.408 --> 00:02:52.048 Question. What are some of the differences between the Baudot digital code and ASCII? 21 00:02:52.948 --> 00:03:07.890 Answer. Baudot uses 5 data bits per character, ASCII uses 7 or 8; Baudot uses 2 characters as letters/figures shift codes, ASCII has no letters/figures shift code. 22 00:03:08.490 --> 00:03:15.279 Question. What is one advantage of using ASCII code for data communications? 23 00:03:16.179 --> 00:03:22.979 Answer. It is possible to transmit both uppercase and lowercase text. 24 00:03:23.579 --> 00:03:43.514 Five bits gives 32 symbols, not enough for letters and digits, hence Baudot's shift codes and its inability to send lowercase. ASCII's 7 or 8 bits gives 128 or 256, which covers everything. 25 00:03:43.514 --> 00:03:49.474 Question. What is the advantage of including parity bits in ASCII characters? 26 00:03:50.374 --> 00:03:54.734 Answer. Some types of errors can be detected. 27 00:03:55.334 --> 00:04:03.051 Some. A single parity bit catches any odd number of bit errors and misses every even number. 28 00:04:03.051 --> 00:04:17.814 It is the weakest useful error detection there is, which is why serious protocols use a proper checksum or CRC, and why the pool's wording is carefully hedged. Spread spectrum. 29 00:04:17.814 --> 00:04:28.025 Question. What spread spectrum communications technique uses a high-speed binary bit stream to shift the phase of an RF carrier? 30 00:04:28.925 --> 00:04:32.205 Answer. Direct sequence. 31 00:04:32.805 --> 00:04:38.205 Question. Which describes spread spectrum frequency hopping? 32 00:04:39.105 --> 00:04:47.733 Answer. Rapidly varying the frequency of a transmitted signal according to a pseudorandom sequence. 33 00:04:48.333 --> 00:05:01.608 Two ways to spread a signal over far more bandwidth than its information needs: Direct sequence multiplies the data by a fast pseudorandom code, spreading the spectrum. 34 00:05:01.608 --> 00:05:14.013 The receiver multiplies by the same code to recover it. Frequency hopping moves the carrier around a pseudorandom pattern, with the receiver hopping in step. 35 00:05:14.013 --> 00:05:20.413 Question. Why are received spread spectrum signals resistant to interference? 36 00:05:21.313 --> 00:05:28.253 Answer. Signals not using the spread spectrum algorithm are suppressed in the receiver. 37 00:05:28.853 --> 00:05:44.463 The de-spreading operation concentrates the wanted signal and spreads everything else. An interfering carrier, however strong, is smeared across the whole bandwidth and drops below the recovered signal. 38 00:05:44.463 --> 00:05:53.196 That is why GPS works at signal levels below the noise floor, and why spread spectrum shares spectrum gracefully. 39 00:05:53.196 --> 00:06:10.120 The amateur limit is 10 watts PEP, from the General rules lesson, deliberately low, because a spread-spectrum signal occupies a great deal of spectrum by design. Check yourself. Your CW sounds clicky to other operators. 40 00:06:10.120 --> 00:06:22.253 What do you adjust, and in which direction? Your PSK IMD measures −22 dB. Acceptable? Why can a spread-spectrum receiver ignore a strong interfering carrier? 41 00:06:28.253 --> 00:06:46.744 Increase the keying waveform rise and fall times. Sharp edges produce the harmonics heard as clicks. No. −30 dB is the acceptable maximum for an idling PSK signal, and −22 dB is worse than that. Reduce your drive. 42 00:06:46.744 --> 00:06:59.071 De-spreading concentrates the wanted signal while spreading the interferer across the whole bandwidth, dropping it below the recovered signal.