تجربة عملية: تصوّر الموجات الكهرومغناطيسية
علاقةٌ واحدة، والطيف كلّه
ما ستفعله
- طول الموجة عند 100 MHz
- ارفع f 24×، فينكمش λ 24×
- طاقة فوتونٍ أخضر
- التأخير عبر كيلومترٍ واحد
صفحات تفاعلية لهذا المقرر. كل رقم فيها يُحسَب في متصفّحك — فليست رسوماً مسجّلة، ولا تتّصل بأي خادم.
صفحات العروض بالإنجليزية في الوقت الحالي؛ أما الدرس المرتبط بكل عرض فهو بالعربية بالكامل.
علاقةٌ واحدة، والطيف كلّه
نغمتان، ونبضةٌ واحدة بطيئة
مؤشّرٌ واحد يضبط كامل إشارة AM
يواجه AM وFM الضجيجَ نفسه
بتات أكثر لكل رمز، وأخطاء أكثر
السقف على كل وصلة: سعة شانون
وجِّه حزمة بالـطور، لا بالحركة
راقب الإشارة تضعف مع المسافة
اجمعها: هل تُغلَق الوصلة؟
قناة واحدة، مقسّمة إلى حوامل فرعية متعامدة
فكّك إطار 802.11 بايتًا بايتًا
صمّم وصلة واحدة، من طرفٍ إلى طرف
Move one knob — the frequency — and watch the wavelength λ = c/f, the period T = 1/f, the photon energy E = h f and the band all recompute, with a transverse E-and-B wave that packs in more cycles as the frequency climbs and a marker that walks the spectrum.
Add two or three sinusoids and watch their sum, added point by point. Slide two frequencies close and a slow beat appears at |f₁−f₂|; set two equal tones in phase and they add to A₁+A₂, anti-phase and they cancel. The beat frequency, carrier, envelope and combined peak amplitude are all measured off the samples.
Build a standard AM signal s(t) = A_c[1 + m·cos(2πf_m t)]cos(2πf_c t) and watch its envelope carry the message, an envelope detector recover it, and the spectrum show a carrier with a sideband at f_c ± f_m. Raise the modulation index m past 1 and the envelope goes negative — the detector clips and distorts. The sideband frequencies, envelope minimum A_c(1−m) and power efficiency m²/(2+m²) are all computed.
Build an FM signal s(t) = A_c cos(2πf_c t + β sin(2πf_m t)) and read its modulation index β = Δf/f_m and Carson bandwidth 2(Δf + f_m). Add the same seeded noise to an AM and an FM signal and watch a frequency discriminator recover a clean message where an envelope detector gets a noisy one — the 3β²(β+1) SNR advantage above the ~10 dB FM threshold. Widen Δf for more noise immunity but more bandwidth.
Drag walls and the transmitter across a floor plan; every wall crossing, mirror reflection and knife-edge diffraction is computed, and the coverage map is redrawn from the geometry.
Build a multipath channel tap by tap and watch five linked views: impulse response, phasor sum, envelope over time, frequency response and the distribution the envelope obeys.
Pick a modulation, set the noise, and watch simulated bit errors land on the theoretical curve. Gain error, phase error and Gray coding are separable.
Steer an N-element array by phase alone. Beamwidth, side lobes, grating lobes and directivity are all measured off the computed pattern, never looked up.
Where the power went, how far the link reaches, and how much the choice of propagation model changes the answer. Effective aperture is shown beside the gain knobs.
One time-frequency rectangle shared four ways — FDMA, TDMA, CDMA, OFDMA — with real Walsh codes and a scheduler fairness comparison.
Build an OFDM symbol from subcarriers with a real IFFT, add a cyclic prefix, push it through a channel and read the EVM and ISI back off the receiver.
Take an 802.11 MAC frame apart field by field — Frame Control, Duration, Address 1/2/3, Sequence Control, optional Address 4, body and FCS — and watch the contiguous offsets sum to the total length. Decode the Frame Control word bit by bit into type, subtype and flags; compute the FCS as a real CRC-32 (0xEDB88320) over the bytes and flip one to see it read corrupt; and read an MCS index as a computed 802.11n PHY rate.