RPM to RPS for Drive Platters and Case Fan Speeds

Platter and Fan Speeds Restated as Turns per Second

Everything that spins inside a computer case is labelled per minute and behaves per second. A drive advertises 7 200 RPM, but what determines whether a request waits is how long one revolution takes and how much of one you sit through before the wanted sector reaches the head. A fan is sold at 1 200 RPM, but the note it produces depends on how many times a blade passes a fixed point each second.

Conversion factor: divide by 60, one minute holding sixty seconds. A 7 200 RPM platter therefore turns 120 times per second, so one revolution takes 8.333 ms and the average rotational latency — half a turn — is 4.167 ms. A 1 200 RPM fan manages exactly 20 turns per second.

What One Revolution Buys You

Half a Turn Is the Wait a Benchmark Reports

Once the head is on the right track, the wanted sector is somewhere ahead of it. Averaged over many random requests that is half a revolution, which is why rotational latency is quoted as half the period rather than the whole one.

Blade Count Times Turns per Second Sets the Pitch

A nine-blade fan at 800 RPM sends 120 blades a second past any fixed point, heard as a low 120 Hz hum. Push it to 2 000 RPM and the same fan sings at 300 Hz, a far more noticeable pitch.

Sequential Speed Is Track Capacity Times Turns per Second

A head parked over one track reads whatever that track holds, once per revolution. Two megabytes on an outer track passing 120 times a second is roughly 240 MB/s before overhead is subtracted.

Spin-Up Is the Heaviest Moment on the 12-Volt Rail

Getting a stack of platters from rest to full speed takes several seconds and draws well over double the running current, which is why a chassis full of drives staggers their starts.

From a Label Speed to a Waiting Time

A short sequence that takes a nameplate figure through to something you can hear or measure.

Enter the speed printed on the drive label or fan spec

Use the rated figure for a drive, since platter speed is fixed, but use the speed a fan is actually running at rather than its maximum, because a curve-controlled fan spends most of its life well below the number on the box.

Invert it for the period, then halve it for latency

One divided by the turns per second gives the time for a full revolution; half of that is the average rotational wait. At 120 turns a second those are 8.333 ms and 4.167 ms, a gap a 5 400 RPM drive can never close.

Multiply by the blade count to predict the tone

Count the blades and multiply. The result is the fundamental frequency of the hum that fan makes at that speed, which tells you whether it will sit under the ambient noise of the room or stand out above it.

Reverse it when a controller reports per second

Tachometer and monitoring outputs sometimes arrive as turns per second while every fan curve is drawn in RPM. The swap arrows flip the pair, typing in the second box works identically since both fields accept input, and copying gives the bare number.

Revolution Times and Latency Across Drives and Case Fans

The last two columns are the ones that matter. For a drive, half a turn is the average rotational latency added to every seek; for a fan, the same figure is simply how long the blade pattern takes to repeat.

Rotating part	RPM	Turns per second	One revolution	Half a turn
Quiet 140 mm case fan	800	13.33	75.00 ms	37.50 ms
Case fan under moderate load	1 200	20	50.00 ms	25.00 ms
Static-pressure fan at full speed	2 000	33.33	30.00 ms	15.00 ms
2.5-inch laptop and archive drive	5 400	90	11.11 ms	5.56 ms
3.5-inch desktop drive	7 200	120	8.33 ms	4.17 ms
Enterprise SAS drive	10 000	166.67	6.00 ms	3.00 ms
Fastest platters ever shipped	15 000	250	4.00 ms	2.00 ms

The diminishing returns are stark. Moving from 5 400 to 7 200 RPM saves 1.39 ms on every random access, but the far bigger jump to 15 000 RPM saves only another 2.17 ms while roughly doubling power draw and noise. That ceiling is much of why platter speeds stopped climbing: half a revolution puts a floor under mechanical latency that extra spin cannot break through.

Two Drive Speeds Compared by the Wait They Impose

Turning both candidate speeds into milliseconds per half turn shows exactly how much latency a faster spindle really buys before the purchase is made.

Fan Pitch Estimated Before the Panel Goes On

Running the intended curve speeds through and multiplying by blade count predicts the tone a build will hum at, which beats swapping fans afterwards.

Figures Copied Into a Build Sheet or Benchmark Log

Copying gives digits with no unit or spacing attached, which is what a spreadsheet column of measured spindle and fan speeds actually needs.

Questions About the Spinning Parts in a Case

Why is average rotational latency exactly half a revolution?

Because once the head has settled on the correct track, the sector you want could be anywhere around the circumference. Over many unrelated requests those positions are spread evenly, so sometimes the sector arrives at once, sometimes a whole revolution passes, and the mean lands in the middle. That gives 4.17 ms at 7 200 RPM and 5.56 ms at 5 400 RPM. Seek time is added on top, but latency is the part that follows from spindle speed alone.

Where did 5 400 and 7 200 RPM come from as standard speeds?

Early drives ran at 3 600 RPM, the synchronous speed of a two-pole motor on a 60 Hz supply, and once the industry had tooled up around it the increments that followed were simple multiples: 5 400 is one and a half times that, 7 200 is double. Faster drives were built for servers, but they need stiffer bearings, smaller platters and much more power, repaying only a couple of milliseconds. Once solid-state storage took the latency-sensitive work, capacity drives settled back where noise and power are comfortable.

Why does a fan hum at one particular pitch rather than just sounding like air?

Every blade sweeping past a fixed obstruction — a strut, the frame, a grille, a radiator fin — produces a small pressure pulse, and those pulses repeat at the turns per second multiplied by the blade count. The ear hears that regularity as a tone sitting on the broadband rush of moving air. Seven blades at 1 200 RPM give 20 × 7 = 140 Hz; nine blades at the same speed give 180 Hz. Tones annoy far more than equally loud hiss, which is why a fan bolted against mesh sounds worse.

Why does a drive read more slowly as it fills up?

Sequential throughput is the data on one track multiplied by how often that track passes the head. Turns per second never change, but track capacity does: outer tracks are physically longer and hold considerably more sectors, so a drive reading 240 MB/s near the rim might manage half that near the hub. Drives fill from the outside in, so early files land on fast tracks and later ones on slower ground — which is why a full drive drags despite spinning at the same rate.

How much extra power does a drive draw while it is spinning up?

Enough to matter as soon as there is more than one. Accelerating a platter stack from rest to 7 200 RPM takes several seconds, and a 3.5-inch drive commonly pulls around two amps from the 12-volt rail during it, against a few tenths when idling — a peak near 24 watts each. Across an eight-bay enclosure powering up, that surge easily exceeds what the supply will deliver, hence staggered spin-up bringing drives online a second or two apart.

RPM to RPS