Continues from Larmor precession. M already walks around B0 at ω₀.
A coil along z hears nothing from M precessing around z. You have to tip M off z so a piece of it sits in xy and keeps rotating. The tip is an RF pulse: a small field B1 in the xy plane, oscillating at ω₀. Same frequency as the walk. That matching is resonance. Push a swing at its natural period. Off-frequency B1 mostly misses.
In the frame that rotates at ω₀, B0 drops out and B1 looks still. M then precesses around B1 instead of around B0. How far it tips is the flip angle. For a constant-amplitude pulse of duration t_p:
α = γ B1 t_p
A 90° pulse parks M in xy. A 180° pulse sends it to −z. The scanner talks in α, not in B1.
If B1 is off-resonance, M sees an effective field that is not B1, so the tip is smaller and the axis is wrong. Slice selection later uses a gradient so only one slab is on-resonance. That is a later note. What the coil actually records is the rotating xy component, not Mz.
Next note: T1 and T2 relaxation.
中文
接拉莫爾進動。M 已經在 ω₀ 繞 B0 走。
線圈沿 z 放，聽不到 M 繞 z 進動；得把 M 從 z 推開，讓一塊待在 xy 裡繼續轉。這一推是 RF pulse：xy 平面上一個小磁場 B1，以 ω₀ 振盪，跟走路的頻率一樣，對上就是共振；盪鞦韆要照它自己的週期推，頻率不對的 B1 多半推空。
在以 ω₀ 轉的座標系裡，B0 消失，B1 看起來是靜的；M 改繞 B1 進動，不再繞 B0。推多遠是 flip angle；振幅固定、長度 t_p 的 pulse：
α = γ B1 t_p
90° pulse 把 M 停在 xy，180° pulse 把它送到 −z；掃描器講的是 α，不是 B1。
B1 偏離共振時，M 看到的有效場不是 B1，所以推得比較小，軸也不對；後面選層會用梯度，只讓一層在共振上，那是後面的筆記。線圈真正記到的是轉著的 xy 分量，不是 Mz。
下一則：T1 和 T2 弛緩。