Thermal design power (TDP): Question bank — Hinglish deep dive

6.4.3 · D5 · Hardware › Power, Thermal & Reliability › Thermal design power (TDP)

Yeh ek rapid-fire conceptual workout hai Thermal design power (TDP) ke liye. Yahan koi heavy arithmetic nahi hai — har item ek misconception ya ek boundary case ko target karta hai jo TDP ki definition quietly invite karti hai. Prompt padho, ek saans mein zor se jawab do, phir reveal karo.

Definition Is page pe use kiye gaye symbols

$R_{θ}$ = thermal resistance: ek material kitni strongly heat flow ko oppose karta hai, °C/W mein (temperature rise in degrees per watt of heat pushed through). Bada $R_{θ}$ = heat pile up hoti hai aur cheezein hot ho jaati hain; chota $R_{θ}$ = heat aasani se escape kar jaati hai. Subscript $θ$ (Greek "theta") sirf yeh flag karta hai ki yeh ek thermal resistance hai, electrical nahi.
$R_{θ, JC}$ = Junction-to-Case resistance: silicon die se chip ke metal lid ke bahar tak.
$R_{θ, CS}$ = Case-to-Sink resistance: lid aur heatsink ke beech thermal paste layer ke across.
$R_{θ, SA}$ = Sink-to-Ambient resistance: heatsink se aas-paas ki hawa mein (fins aur airflow pe depend karta hai).
$R_{θ, total}$ = poore heat path ki total thermal resistance add karke: $R_{θ, total} = R_{θ, JC} + R_{θ, CS} + R_{θ, SA}$ (neeche series-path figure dekho kyun yeh add hoti hain).
$T_{j}$ = junction temperature: silicon die (the "junction") ka actual temperature abhi, °C mein. Yahi woh cheez hai jo chip kaam karne par aur heat pile up hone par badhti hai.
$T_{j,max}$ = maximum junction temperature: sabse zyada hot $T_{j}$ ko damage ya throttling se pehle pahunchne diya jaata hai (typically 90–105 °C). Toh $T_{j}$ live reading hai, $T_{j,max}$ woh red line hai jise cross nahi karna chahiye.
$T_{ambient}$ = ambient temperature: aas-paas ki hawa ka temperature (ek room usually ~25 °C hota hai).

Upar ka figure dikhata hai kyun teen resistances add hoti hain: heat die se nikalti hai aur har layer se sequence mein travel karti hai — die → case → paste → heatsink → air — jaise paani ek pipe mein teen narrow sections ke saath force hota hai. Ek path, order mein liya gaya, isliye resistances series mein add hoti hain, $R_{θ, total}$ deti hain.

Formula Kyun cooling inequality hold karti hai —

$Δ T = P \cdot R_{θ}$ chain Pehle, Ohm's-law mapping, term by term unpack ki gayi. Electricity tum already jaante ho: ek voltage difference ek current ko ek resistance ke through push karta hai, aur Ohm's law kehta hai $Δ V = I \times R$ . Heat bilkul usi shape ka law follow karti hai, characters ki ek-to-one swap ke saath:

ek temperature difference $Δ T$ voltage ka role play karta hai — yeh woh "pressure" hai jo flow drive karta hai (heat sirf tab move karti hai jab ek side hotter ho, jaise current sirf tab flow karti hai jab ek side higher voltage pe ho);
ek heat flow $P$ (watts mein) current ka role play karta hai — yeh kitna actually per second move ho raha hai;
ek thermal resistance $R_{θ}$ electrical resistance ka role play karta hai — path us flow ko kitna strongly oppose karta hai.

Un swaps ke saath, $Δ V = I R$ ban jaata hai $Δ T = P \times R_{θ} .$

Upar ka figure isse ek actual circuit ki tarah draw karta hai: ek heat "source" (die) power

P

ko teen series thermal resistors ke through ambient "ground" tak push karta hua.

Ek zaroori caveat — yeh ek steady-state law hai. $Δ T = P \cdot R_{θ}$ sirf tab hold karta hai jab temperatures settle ho chuke hon aur change hona band ho gaya ho (thermal steady state). Yeh thermal capacitance ko ignore karta hai — yeh fact ki die aur heatsink heat store karte hain aur warm up hone mein time lagte hain. Ek brief spike ke dauran, woh stored-heat buffer $T_{j}$ ko lag karne deta hai, aur yahi wajah hai ki turbo boost momentarily TDP exceed kar sakta hai bina instantly overheat kiye. Simple formula long-run balance point hai, moment-to-moment truth nahi.

Ab substitution. Woh temperature jise heat cross karni hai woh die se air tak hai: $Δ T = T_{j} - T_{ambient}$ . Heat flow $P$ ke liye hum TDP plug in karte hain — lekin yeh allowed kyun hai? Kyunki TDP ko sustained worst-case heat ke roop mein define kiya gaya hai, aur "sustained" precisely woh steady-state condition hai jo formula require karta hai: long run mein, bahar flow hone wale watts generate ho rahe watts ke equal hone chahiye, aur woh generation TDP hai. Dono substitute karne par: $T_{j} - T_{ambient} = TDP \times R_{θ, total} .$ Die ko safe rakhne ke liye hum $T_{j} \leq T_{j,max}$ demand karte hain. Cooler ke liye resistance allowed hai use rearrange karne par: $R_{θ, total} \leq \frac{T _{j,max} - T _{ambient}}{TDP} .$ Yeh ek upper bound hai: koi bhi cooler jis ki lower resistance hai woh die ko limit se cooler rakhta hai, jo theek hai.

Kuch prerequisite ideas jo fresh rakhni chahiye: Power consumption in CMOS circuits, Heatsink design and thermal resistance, Thermal throttling, aur Turbo Boost and power states.

True or false — justify karo

TDP ek din mein CPU ke average power draw ke barabar hota hai.

False. TDP ek sustained worst-case thermal target hai cooler ke liye; real usage ~10 W idle se short turbo spikes TDP se upar tak swing karta hai, toh din ka average usually TDP se kaafi neeche hota hai.

Ek chip kabhi bhi, briefly bhi, apni TDP se zyada power draw nahi kar sakti.

False. Modern CPUs deliberately TDP exceed karte hain seconds ke liye Turbo Boost ke through; die ki thermal mass extra heat absorb kar leti hai pehle ki sustained average TDP ke paas settle ho.

Identical TDP wale do CPUs har instant identical heat dissipate karte hain.

False. Identical TDP ka matlab sirf yeh hai ki unke coolers usi maximum sustained load face karte hain; instant-to-instant heat workload, clocks, aur boost behaviour par depend karta hai, jo different hote hain.

CPU consume karne wali saari electrical power eventually heat ban jaati hai.

True. CPU koi mechanical work nahi karta, isliye energy conservation se essentially 100% power jo yeh draw karta hai heat mein convert ho jaati hai jise remove karna padta hai.

Path ke saath thermal resistances (junction→case→paste→sink→air) series mein resistors ki tarah add hoti hain.

True. Heat ek path se sequence mein flow karti hai, isliye

R_{θ, total} = R_{θ, JC} + R_{θ, CS} + R_{θ, SA}

, bilkul series electrical resistance ki tarah.

65 W CPU hamesha 125 W CPU se slower hota hai.

False. Speed IPC × frequency × cores hai; ek newer, efficient 65 W part ek older 125 W wale ko beat kar sakta hai. TDP heat measure karta hai, performance nahi.

Supply voltage kam karne se power par barely effect padta hai kyunki power mostly leakage hoti hai.

False. Dynamic power

V^{2} f

ke roop mein scale karti hai (jahan

V

supply voltage hai aur

f

clock frequency hai), isliye voltage ka outsized (squared) effect hota hai — undervolting heat par sabse strong levers mein se ek hai.

TDP tumhe laptop ka total wall-plug power batata hai.

False. TDP sirf CPU ka thermal figure hai; display, GPU, storage, aur PSU inefficiency sab wall power add karte hain CPU ke TDP se aage.

Agar tumhara heatsink exactly TDP requirement pe rated hai, toh thermal paste aur package resistance matter nahi karte.

False.

R_{θ, JC}

(die-to-case) aur

R_{θ, CS}

(paste layer) usi total budget mein se eat karte hain; unhe ignore karna

R_{θ, total}

ko limit se upar push kar sakta hai aur throttling cause kar sakta hai.

Ambient temperature badhane se ek fixed heatsink kam TDP handle kar paata hai.

True. Usable budget

T_{j,max} - T_{ambient}

hai; zyada hot room numerator shrink karta hai, isliye wahi

R_{θ}

ek smaller safe TDP support karta hai.

Error dhundho

"Mere chip ka 95 W TDP hai, isliye mera power bill 95 W continuous CPU use reflect karta hai."

Error yeh hai ki ek cooling design target ko ek continuous consumption figure treat kiya ja raha hai; actual draw usually idle/light load pe kaafi kam hota hai, isliye bill kaafi kam hota hai.

"Mujhe exactly

R_{θ} = 0.6

°C/W wala heatsink chahiye kyunki 75 °C budget divided by 125 W gives 0.6 °C/W."

'75' allowed temperature rise

T_{j,max} - T_{ambient} = 100 - 25 = 75

°C hai, aur 0.6 °C/W total path budget hai; heatsink alone lower honi chahiye kyunki

R_{θ, JC}

aur

R_{θ, CS}

already iska part consume kar lete hain.

"Maine frequency 20% se overclock ki toh power 20% badhti hai."

Overclocking ko usually zyada voltage bhi chahiye, aur dynamic power

V^{2} f

ke roop mein jaati hai; squared voltage term real rise ko 20% se kaafi zyada bana deta hai.

"Turbo 1.5× TDP push karna matlab mera 125 W cooler instantly overwhelm ho jaata hai."

Turbo transient hai; die ki heat capacity short spikes buffer karti hai, aur cooler ko sirf sustained TDP handle karna hai, har millisecond peak nahi.

"TDP woh peak instantaneous power hai jo chip kabhi bhi draw kar sakti hai."

TDP ek sustained thermal steady-state figure hai, na ki ek instantaneous ceiling — peaks regularly briefly isse exceed karti hain.

"Cheap thermal paste theek hai; heatsink toh same hai."

Poor paste

R_{θ, CS}

(case-to-sink term) ko 0.2–0.5 °C/W inflate kar deta hai, jo TDP se multiply hone par

T_{j}

mein dozens of degrees add kar deta hai — Heatsink design and thermal resistance dekho.

"Higher TDP directly matlab higher clock speed."

TDP architecture, process node, aur voltage se heat reflect karta hai; clock sirf ek factor hai, aur efficient designs zyada speed per watt produce karte hain.

Why questions

TDP ek realistic workload se kyun define hota hai na ki ek synthetic stress test jaise Prime95 se?

Kyunki cooling un heat ke liye size ki jaati hai jo customers actually generate karte hain; ek synthetic power-virus load spec inflate karega aur unnecessarily expensive coolers force karega.

Cooling constraint

R_{θ} \leq (T_{j,max} - T_{ambient}) / TDP

use kyun karti hai equals sign ki jagah?

Humein

T_{j}

limit ke at or below rehna chahiye; koi bhi lower resistance chip ko cooler rakhti hai, isliye requirement

R_{θ}

par ek upper bound hai.

Thermal equation

Δ T = P \times R_{θ}

sirf steady state mein kyun hold karta hai?

Kyunki yeh thermal capacitance ignore karta hai — woh heat jo die aur heatsink mein store hoti hai. Sirf jab temperatures change hona band kar dete hain tab heat generated aur heat removed balance karte hain, aur tabhi formula exact hota hai.

P = TDP

ko steady-state formula mein rakhna valid kyun hai?

Kyunki TDP by definition sustained heat load hai, aur "sustained" exactly woh settled, steady-state condition hai jo formula require karta hai.

Undervolting same percentage se clock speed reduce karne se zyada heat kyun reduce karta hai?

Dynamic power voltage par squared depend karta hai lekin frequency par sirf linearly, isliye voltage cuts dominant

V^{2}

term ko attack karte hain — Overclocking and voltage scaling dekho.

Newer process nodes often ek CPU ko lower TDP par zyada kaam kyun karne dete hain?

Chote transistors lower voltage aur capacitance par switch karte hain,

α C V^{2} f

cut karte hain — jahan

α

activity factor hai (fraction of transistors switching each cycle) aur

C

total switched capacitance hai — isliye zyada computation same heat budget mein fit hoti hai.

Ek chip cooling inadequate hone par simply fail karne ki jagah throttle kyun karta hai?

Throttling frequency (aur often voltage) lower karta hai, jo dynamic power aur isliye heat drop karta hai, die ko performance aur safety ke beech trade karke protect karta hai.

Data-center designers ko ambient temperature ki desktop users se zyada chinta kyun karni padti hai?

Dense server rooms zyada warm chalte hain aur bahut saare chips pack karte hain,

T_{j,max} - T_{ambient}

budget shrink karte hain aur us heat ko multiply karte hain jise extract karna padta hai.

Edge cases

Jab

T_{ambient} \to T_{j,max}

hota hai toh safe-TDP budget ka kya hota hai?

Numerator

T_{j,max} - T_{ambient} \to 0

ho jaata hai, isliye koi bhi finite heat junction limit exceed kiye bina remove nahi ho sakti — cooling impossible ho jaati hai chahe heatsink kitna bhi accha ho.

Agar ek workload essentially zero power draw kare (deep idle), toh TDP kya batata hai?

Us moment ke baare mein kuch nahi — TDP ek maximum sustained design figure hai, isliye near-zero load par actual heat TDP se kaafi neeche hoti hai aur instantaneous state se irrelevant hoti hai.

Kya ek chip ki instantaneous power kabhi negative ho sakti hai (heat waapas andar flow ho)?

Nahi — ek CPU hamesha dissipate karta hai, kabhi cold generate nahi karta; heat sirf bahar ambient ki taraf flow karti hai kyunki operation ke dauran

T_{j} > T_{ambient}

hota hai.

Ek fanless passively-cooled laptop ke liye, usable TDP ko sabse zyada kya limit karta hai?

Sink-to-ambient resistance

R_{θ, SA}

bina forced airflow ke bahut high hoti hai, ek low sustained TDP force karti hai aur brief turbo plus throttling par heavy reliance hoti hai.

Agar

R_{θ, JC}

akela already total budget

(T_{j,max} - T_{ambient}) / TDP

exceed kar le, toh koi bhi external cooler kya kar sakta hai?

Kuch nahi — heat package ke andar trapped ho jaati hai sink tak pahunchne se pehle, isliye koi bhi external heatsink, chahe kitna bhi perfect ho,

T_{j}

ko spec mein nahi rakh sakta.

Agar measured sustained power full load par rated TDP se below baithe, toh iska kya matlab hai?

Chip ya toh ek lower configured limit se power-limited hai, thermally throttled hai, ya simply us workload par efficient hai — TDP cooling ke liye ek ceiling hai, guarantee nahi ki chip ise hit karegi.

Exactly

R_{θ, total} = (T_{j,max} - T_{ambient}) / TDP

par, operating margin kya hai?

Zero —

T_{j}

precisely

T_{j,max}

par land karta hai, isliye koi bhi zyada hot room, dust, ya degraded paste immediately throttling trigger kar deta hai.

Recall Quick self-test

Woh ek-sentence distinction do jo yeh poora page train karta hai. Answer ::: TDP woh sustained heat hai jo ek cooler remove kare (ek thermal design target), na ki woh energy jo chip wall se draw kare (ek fluctuating power measurement).