14. Things to Try

A few exercises and explorations to lock the concepts in.

1. Compute NA and acceptance angle. A standard SMF has $n_1 = 1.4675$ and $n_2 = 1.4625$. Find NA. Find the acceptance angle. (Answer: NA = 0.121, acceptance angle = $7.0°$.)
2. Compute V-number. For the same fiber with core radius $a = 4.5$ µm at $\lambda = 1310$ nm, compute V. Is the fiber single-mode at this wavelength? At 850 nm? (V = 2.61 at 1310 nm, multimode but barely — this is why "single-mode at 1310 nm" fibers are often called G.652 with a cutoff around 1260 nm.)
3. Plan a 200 km link. Launch power 0 dBm, fiber loss 0.20 dB/km, two splices per km, two connectors total, receiver sensitivity -25 dBm coherent. Will it work? Where do you put EDFAs?
4. Plot Rayleigh and IR absorption. Use the Python snippet in Section 12 and find the loss minimum numerically. It should land in the 1550-1580 nm range.
5. Calculate dispersion limit. For 10 Gbps NRZ over standard SMF at 1550 nm with a DFB laser of 0.05 nm linewidth, find the maximum reach before dispersion penalty exceeds 1 dB (rule of thumb: when pulse spreading equals about 0.3 of bit slot).
6. Visit a fusion splicer. Most networking trade shows have demo benches. Watching two glass fibers melt into one in 8 seconds is genuinely awe-inspiring.
7. Open up an SFP. A retired 10 GbE transceiver costs 5 dollars on eBay. Inside you will find a tiny laser TO-can, a photodiode, a TIA, and a small microcontroller speaking the I$^2$C management protocol. Reading the EEPROM contents teaches you everything about optical transceiver inventory protocols.
8. Trace the path of an email. Send yourself a message between two cloud regions. Look up which submarine cable carries traffic between those regions. The latency you measure (ping in ms) divided by 5 µs/km gives you a rough sanity check on the route length.