Visual walkthrough — Control moment gyroscopes (CMG) — high torque, singularity

3.5.49 · D2 · Physics › Guidance, Navigation & Control (GNC) › Control moment gyroscopes (CMG) — high torque, singularity

---

Step 1 — Angular momentum ek picture ke roop mein kya hota hai?

KYA HAI. Ek bhaari disk (flywheel) ko fast spin karte hue imagine karo. Physics use ek single arrow deta hai jo capture karta hai "kitni spin hai, kaunsi line ke baare mein, aur kaunsi direction mein." Is arrow ko hum angular momentum kehte hain aur likhte hain $\vec{h}$. $h$ ke upar chhota arrow-hat sirf matlab hai "yeh ek vector hai — iska ek direction bhi hai, sirf size nahi."

Arrow kyun, sirf ek number kyun nahi? Kyunki spin ka ek direction hota hai: disk kisi axis ke baare mein spin karti hai. Hum arrow ko us axis ke saath point karte hain (right-hand rule: apni right fingers ko spin ke saath curl karo, thumb $\vec{h}$ ke along point karta hai). Arrow ki length bataati hai kitni spin hai.

PICTURE. Disk flat padi hai; burnt-orange arrow $\vec{h}$ seedha uske centre se upar nikalta hai.

---

Step 2 — Length freeze karo, sirf direction ghuma o

KYA HAI. CMG kabhi wheel ko speed up ya slow down nahi karta. Yeh poore spinning wheel ko pakadta hai aur uski axis ko tilt karta hai. Toh arrow $\vec{h}$ apni length $h$ maintain karta hai lekin uska head ek naye direction mein swing karta hai.

Length fixed rakhna kyun important hai? Kyunki CMG ka poora magic isi par tika hai. Hum arrow ko grow nahi karne wale (woh ek reaction wheel hoga, jo ${\omega }_w$ badlta hai). Hum sirf ek already-lamba arrow rotate karte hain — yeh karna sasta hai aur, jaise hum dekhenge, ek bada effect deliver karta hai.

PICTURE. $\vec{h}$ ki tip radius $h$ ka ek circle trace karti hai. Arrow ek ghadi ka fixed-length haath hai; sirf uska angle badlta hai.

---

Step 3 — Rotation ke liye Newton ka law: torque = $\vec{h}$ ka change

KYA HAI. "Force momentum ko change karta hai" ka rotational version hai: torque angular momentum ko change karta hai. Symbols mein: $\vec{\tau }=\frac{d\vec{h}}{dt}.$

Yeh tool kyun, derivative $\frac{d}{dt}$ kyun? Hum jaanna chahte hain ki CMG kaunsa torque deliver karta hai. Torque literally woh rate hai jis par spin arrow change hota hai. Symbol $\frac{d\vec{h}}{dt}$ exactly ek sawaal ka jawab deta hai: "abhi $\vec{h}$ ki tip kitni fast, aur kis direction mein, move kar rahi hai?" Yahi precise sawaal hai jo hume chahiye — koi aur tool "instantaneous change of a vector" ka jawab nahi deta.

PICTURE. Ek heartbeat ke fark par clock hand ke do snapshots. Old tip se new tip tak chhota green arrow $d\vec{h}$ hai; chhote time $dt$ se divide karo aur torque arrow milta hai.

---

Step 4 — Rotating arrow ki tip kaise move karti hai? $\vec{\Omega }\times \hat{s}$ aata hai

KYA HAI. Gimbal, axis $\hat{s}$ ko ek fixed line ke baare mein ghumata hai jise gimbal axis $\hat{g}$ kehte hain, turning rate $\dot{\delta }$ ("delta-dot") par. Yahan $\delta$ gimbal angle hai aur $\dot{\delta }$ hai kitni fast woh angle badhta hai. Turning ko ek arrow $\vec{\Omega }=\dot{\delta }\hat{g}$ mein package karo: yeh us axis ke along point karta hai jiske baare mein turn ho raha hai, aur uski length turning rate hai.

Kisi bhi unit arrow ke baare mein jo rotate ho raha ho, ek universal fact hai: $\frac{d\hat{s}}{dt}=\vec{\Omega }\times \hat{s}.$

Cross product $\times$ kyun, multiply ya add kyun nahi? Hume ek spinning unit arrow ki tip velocity chahiye. Tip ek circle par move karti hai: uski speed (turning rate)$\times$(us circle ki radius) hai, aur uski direction turning axis $\hat{g}$ aur arrow $\hat{s}$ dono ke perpendicular hai. "Size = dono ka product, direction = dono ke perpendicular" — yahi exactly cross product $\vec{\Omega }\times \hat{s}$ ka job description hai. Koi plain multiplication perpendicular direction nahi deta; sirf cross product deta hai. Isliye yahi tool aur koi nahi.

PICTURE. Gimbal axis $\hat{g}$ (teal, vertical). Spin arrow $\hat{s}$ usse hata hua. Uski tip ek plum-coloured circle par chalt hai; velocity arrow $d\hat{s}/dt$ us circle ke tangent hai — dono $\hat{g}$ aur $\hat{s}$ ke right angles par.

---

Step 5 — Assemble karo: CMG torque law

KYA HAI. Step 3 aur Step 4 ko saath mein rakho. Kyunki $\vec{h}=h\hat{s}$ hai aur $h$ frozen hai, differentiate karo: $\vec{\tau }=\frac{d\vec{h}}{dt}=h\frac{d\hat{s}}{dt}=h(\vec{\Omega }\times \hat{s})=\vec{\Omega }\times (h\hat{s})=\vec{\Omega }\times \vec{h}.$

$h$ ko andar/bahar kyun le ja sakte hain? Kyunki $h$ ek plain constant number hai (Step 2 ne use freeze kiya). Ek constant derivative aur cross product se freely slide karta hai. Upar ki har equality sirf "ek constant ko idhar udhar move karo" hai.

PICTURE. Arrows ka final triangle: $\vec{\Omega }$ (teal) upar, $\vec{h}$ (orange) lean karta hua, aur unka cross product $\vec{\tau }$ (plum) unke common plane ke perpendicular shoot karta hua.

---

Step 6 — Degenerate case: jab tilt kuch nahi karta ($\theta =0$)

KYA HAI. $|\vec{\tau }|=h\dot{\delta }\sin \theta$ dekho aur $\theta \to 0$ karo: agar tum $\vec{h}$ ko apni khud ki axis ke baare mein turn karne ki koshish karo ($\hat{g}$ parallel to $\vec{h}$), toh $\sin 0=0$ aur torque vanish ho jaata hai.

KYUN? Ek arrow ko us line ke baare mein spin karna jiske along woh pehle se point kar raha hai, uski tip ko kahin nahi move karta. Koi tip motion nahi (Step 3) matlab koi torque nahi. Yahi trouble ka seed hai: ek direction jise CMG push nahi kar sakta.

PICTURE. $\hat{g}$ flat $\vec{h}$ ke along rakha hua; tip circle ek point mein collapse ho jaata hai. Torque arrow gone.

---

Step 7 — Kai CMGs: same picture, Jacobian mein stack kiya hua

KYA HAI. Har CMG $i$ ek torque column $\partial {\vec{h}}_i/\partial {\delta }_i={\hat{g}}_i\times {\vec{h}}_i$ contribute karta hai — uski apni "abhi main kaunsi direction mein push kar sakta hoon" arrow. $n$ inhe side by side ek matrix $A\in {\mathbb{R}}^{3\times n}$ mein stack karo, Jacobian. Total body torque hai $\vec{\tau }=-A\dot{\vec{\delta }}$.

Matrix kyun? Humare paas $n$ knobs hain (gimbal rates ${\dot{\delta }}_i$) aur hum ek $3$-D torque chahte hain. Ek $3\times n$ matrix exactly hai "$n$ inputs ko ek $3$-D output mein combine karo." Ek commanded torque hit karne ke liye use pseudo-inverse $A^{+}$ se invert karo.

PICTURE. Chaar chhote plane-arrows (ek har CMG ke liye) fan out kiye hue; jab sab ek common line $\hat{u}$ ke against flatten ho jaate hain, reachable set flat ho jaata hai — Step 6 ki singular wall, ab ek cluster mein.

---

Ek-picture summary

Sab kuch ek chain mein collapse ho jaata hai: frozen-length arrow → use turn karo → tip perpendicular move karti hai → woh motion hi torque hai → jab tip kisi line ke along nahi move kar sakti, woh line singularity hai.

Recall Feynman retelling — plain words mein zor se bolo

CMG ek fast-spinning wheel hai jiske "spin arrow" $\vec{h}$ ki fixed length hoti hai — hum wheel ko kabhi speed up ya down nahi karte. Jab ek chhota motor us arrow ko gimbal line $\hat{g}$ ke baare mein tilt karta hai, arrow ki tip ek chhote circle ke along sweep karti hai. Newton ka rotation law kehta hai torque bas kitni fast tip move karti hai hai, toh torque wahan point karta hai jahan tip ja rahi hai — jo sideways hai, tilt axis aur arrow dono ke right angle par. Kyunki arrow pehle se lamba hai, slow tilt bhi uski tip ko fast whip karta hai: bada torque, thoda effort — yahi amplification hai. Lekin agar tum kabhi arrow ko apni khud ki direction ke baare mein tilt karo, tip kahin nahi jaati aur tumhe zero torque milta hai. Kai aisi wheels line up karo aur woh sab ek hi direction ek saath kho sakti hain: ek wall jise singularity kehte hain, jiske around hum damped (robust) inverse aur torque-free null motion se tiptoe karte hain.

Recall Predict-then-verify checkpoints

Kaunsa tool ek rotating unit vector ki tip velocity deta hai, aur kyun wahi tool? ::: Cross product $\vec{\Omega }\times \hat{s}$ — yahi akela ek aisa result return karta hai jo dono turn axis aur arrow ke perpendicular ho, jis ka size = rate × radius. Step 5 mein $h$ ko derivative se bahar kyun nikal sakte hain? ::: Kyunki $h$ ek frozen constant hai, aur constants derivatives aur cross products se freely slide karte hain. Ek single CMG kab zero torque produce karta hai? ::: Jab gimbal axis $\vec{h}$ ke parallel ho ($\theta =0$), toh $\sin \theta =0$ aur tip move nahi karti. Kaunsa geometric event ek cluster ko singular banata hai? ::: Saare columns ${\hat{g}}_i\times {\vec{h}}_i$ coplanar ho jaate hain, ek common perpendicular $\hat{u}$ share karte hain jiske along cluster torque nahi kar sakta.