WHY does this rule exist? Because addition only makes sense between identical kinds. "5 metres + 3 seconds" has no meaning — there is no number that is both. So if a formula adds two terms, nature forces them to be the same dimension.
Corollaries (memorise the consequences, not the lines):
Arguments of sin,cos,ln,ex must be dimensionless (you can't take the sine of "2 metres").
How it is built from base quantities M,L,T — its kind, independent of the chosen unit.
State the principle of homogeneity.
Every term added or equated in a valid equation must have identical dimensions.
Dimension of force?
[MLT−2] (from F=ma).
Dimension of energy/work?
[ML2T−2] (force × distance).
Dimension of pressure?
[ML−1T−2] (force / area).
Dimension of power?
[ML2T−3] (energy / time).
Why must the argument of sin or ex be dimensionless?
These functions equal sums of powers of their argument; adding x to x2 requires x to be a pure number.
Can dimensional analysis prove an equation correct?
No — it's blind to pure numbers (like 21), signs in sums, and trig/exp forms; it can only disprove or fail to disprove.
A failed dimensional check means?
The equation is definitely wrong.
Derive how pendulum period scales with ℓ and g.
T=kℓagc; matching L,T gives a=21,c=−21, so T∝ℓ/g.
Why is the pendulum period independent of mass?
Matching the power of M gives b=0 since g and ℓ carry no mass.
Three failures of the derivation method?
Can't find numeric constants; fails when unknowns > base dimensions; can't build multi-term (sum) formulas.
Convert 1 N to dyne using dimensions.
[F]=MLT−2⇒1N=103g⋅102cm⋅s−2=105 dyne.
Recall Feynman: explain to a 12-year-old
Imagine every measurement wears a colored shirt: red for length, blue for time, green for weight. You're only allowed to add things wearing the same shirt — two red lengths give a red length, fine; but a red length plus a blue time is nonsense. When grown-ups write a physics equation, you can check it by looking at the shirts on both sides: if they match, the equation isn't obviously broken; if they clash, it's definitely wrong. The cool bonus: if you know what shirts the ingredients wear, you can often guess the recipe — like guessing a swing's wobble time grows with its length and shrinks with stronger gravity — just by making the shirts match up.
Dekho, har physical quantity ka ek "type" hota hai — koi length hai, koi time, koi mass. Inko hum dimensions kehte hain: [L], [T], [M]. Asli rule bahut simple hai — sirf same type ki cheezein add ho sakti hain. 5 metre aur 3 second ko jod nahi sakte, kyunki dono alag-alag kind ke hain. Isi rule ko principle of homogeneity bolte hain: equation ke dono side, aur jo bhi terms plus/minus se jude hain, sabki dimension same honi chahiye.
Pehla use: equation check karna. Jaise v2=u2+2as — teeno terms ka dimension L2T−2 nikalta hai, toh equation theek hai. Lekin agar koi term mismatch kare (jaise s=ut+21at mein at ka dimension velocity ban jata hai), toh equation pakka galat hai. Yaad rakho: fail hona matlab definitely wrong, par pass hona matlab "shayad sahi" — kyunki dimensional analysis ko pure numbers (jaise 21, 2π) dikhte hi nahi.
Doosra use, jo exam mein points dilata hai: relation derive karna. Maan lo pendulum ka period T depend karta hai ℓ, m, g par. Likho T=kℓambgc, dono side ki M,L,T ki powers match karo, aur tumhe a=21, b=0, c=−21 mil jaata hai — yaani T∝ℓ/g, aur mass ka koi role nahi! Constant 2π yeh method nahi dega, woh experiment se aata hai.
Bas limitation samajh lo: numeric constants, plus/minus signs, aur multi-term formulas yeh method nahi de sakta. Toh ise ek fast "sanity check" aur "first guess" tool ki tarah use karo — galti pakadne mein aur quick derivation mein yeh superpower hai.