Feeds and Speeds Explained: A Practical Machinist’s Guide

Feeds and speeds are the two cutting parameters behind every machining cut: the speed is the cutting speed (how fast the tool edge moves across the material, in SFM) and the feed is the feed rate (how fast the tool advances, in IPM or IPR). To get a perfectly clean part, versus a broken bit, it helps to know about SFM and IPM. In feed, we’ve two parameters: cutting speed (FPM / SFPM) and feed rate. If we hit the two of them at the right rates, we can obtain good results, such as a perfect finishing, minimal waste of the bits used in the task or quick operation times, but if we fail, the result will be a noisy finish, damaged surfaces and broken tools. Here you’ll find out, the formulas to calculate: Speeds and Feeds FPM and feed rates (in IPM or revolutions per minute)

What Are Feeds and Speeds?

Feeds and speeds are the two motion settings behind every cut: the speed (cutting speed, in surface feet per minute) is how fast the tool edge moves across the material, and the feed (feed rate, in inches per minute or per revolution) is how fast the tool advances. You pick the cutting speed from the material and tooling, convert it to spindle RPM, then set a feed rate that gives each tooth the right bite.

In our experience building CNC lathes and mills, the operators who struggle most treat these as one setting: they pick a feed rate that looks safe and let spindle speed fall wherever it lands. That goes wrong because the cutting edge then runs at the wrong surface speed for the material, and a machine engineered to hold 50–4,000 RPM cannot rescue a number chosen for the wrong reason. University teaching labs such as Boston University’s Engineering Product Innovation Center publish deliberately conservative speeds and feeds for the same reason.

For job shop machining, there are only two motion parameters on which a cutting task is based: Feed rate and cutting speed. Cutting speed (the “speed”) dictates how quickly the tool’s cutting edge moves across the workpiece, usually listed in inches of cutting action per minute. On a lathe and generally with rotating tooling on the mill, it’s expressed in IPR (inches per revolution), or with linear motion on a mill, as SFPM (surface feet per minute). The feed rate (the “feed”) determines how rapidly the cutting tool move through the work, typically specified in IPM (inches per minute) on the mill, or IPR on a lathe, as Wikipedia’s Speeds and feeds entry puts it, “as they occur simultaneously…feeds and speeds are often discussed as a related, and frequently specified together, item.”

Why this distinction is important: The cutting speed will be responsible primarily for generating most of the heat and your tooling wear. The feed rate has responsibility for the chip thickness, cutting forces and surface finish. You’ll pick a cutting speed based on your material and tooling, then convert that to the spindle RPM for the specific diameter of your cutting tools and select a feed rate to provide a suitable bite per tooth. If this machine aspect of it’s entirely new to you, our primer on the basics of lathe and mill operation is an excellent starting point covering the hardware that these numbers interact with.

Cutting Speed and SFM: Surface Speed by Material

Machinists often misread a chart speed as an RPM. A 1/8″ tool and a 2″ face mill at the same 600 SFM turn at very different RPM, so a shop that skips the conversion will burn the small tool and risk a snapped cutter, because the wrong surface speed is the most expensive mistake in the cut. We size every spindle, built around that conversion, so that a 30% usable-RPM gap matters as much as the top speed.

The surface speed is a feature of the machine setup and the materials – not a property of the tool. For instance you can get better surface speed from aluminum that stainless and titanium. As you see, you figure the SFM your looking to achieve, you translate that to spindle RMP – and the machine has control over RPM – not surface speed.

How do you convert SFM to RPM?

Divide the surface speed by the tool diameter, then scale to a minute. The clean shop relation is RPM = (3.82 × SFM) ÷ D, where D is either the cutter diameter (milling) or the workpiece diameter (turning) in inches. For a 1/2″ tool at 600 SFM that is (3.82 × 600) ÷ 0.5 = 4,584 RPM, and most tooling handbooks state the same relation.

MIT’s Fab Lab adds an important real-world limitation to their speeds-and-feeds chart: a spindle might have to go above 2000 RPM just to get a decent surface speed out of small cutters, which explains why hobby routers can barely cut a chip of steel. Work the calculation backward (SFM = (RPM × D) ÷ 3.82) to determine the surface speed your desired RPM produces.

Honesty with the numbers here. The U of Florida machine-shop instructions clearly tell you, “These Recommended Surface Speeds are on the Conservative Side…” Similarly, most technical and handbook tables lean that way. Real shops that use coated carbide tools often push these rates far beyond what any of these tables suggest.

“The table below contains recommended surface speeds for common materials… These values are conservative.”

University of Florida Department of Mechanical & Aerospace Engineering, Design & Manufacturing Lab

Feed Rate, Chip Load, and Feed Per Tooth

By far the most common feed mistake we see on machines sent back for “spindle problems” is a feed rate set too low. The tool rubs instead of cutting because each tooth takes almost nothing, and the heat that follows wears the edge and work-hardens stainless. A machine engineered for a 0.004″ chip load still fails at 0.0005″ because the physics, not the machine, is the limit.

Once you set RPM, feed rate determines how much material each cutter tip is taking off in a bite. The milling feed formula, stated identically by MIT’s Fab Lab and a lot of others:

Feed rate (IPM) = RPM × chip load per tooth × number of teeth

This feed rate formula pairs with the RPM formula from the last section: the RPM formula sets spindle speed from surface speed, then the feed rate formula sets advance from chip load and the number of teeth (the flutes) on the cutter. Whether you run high-speed steel or carbide, the math is identical; only the surface speed you plug in changes, since carbide tolerates far higher cutting speeds than high-speed steel.

Chip load (feed per tooth) is the thickness of material each cutting edge removes per revolution. It is the single most misunderstood number in feeds and speeds. Too high and you overload the tool, causing chatter and broken edges. Too low and the edge rubs instead of cutting, which builds heat, forms a built-up edge on the tool, dulls it fast, and causes work hardening in stainless steel. A widely used shop rule of thumb is roughly 0.001″ of chip load per tooth for every 0.25″ of cutter diameter, which lines up with the per-diameter chart later in this guide. Online tools such as the Machining Doctor chip load calculator use the same feed-per-tooth basis.

The Feeds and Speeds Formulas: A Worked Example

This is where a shop gets a nasty surprise: the calculator returns 6,112 RPM and the machinist sends it to a mill engineered to spin 4,000, so the tool never reaches the planned surface speed. We flag this risk during machine selection because a process that needs high RPM on small tools belongs on a high-speed spindle, not a standard one.

Charts and calculators are fine, but you should be able to do the math by hand to sanity-check any number a machine or app hands you. Here is the full chain, which we call The 4-Input Feeds and Speeds Method, because four inputs (surface speed, tool diameter, chip load, and flute count) give you both your spindle speed and your feed rate, the same calculation university machine shops like Olin College’s shop teach.

The last point is where most shop calculators lied to students. The variable-speed machines that we build have spindle speed ranges of some where between 50-4000 rpm so the small cutter asking for high surface speed may have required a spin speed higher than most spindles of that type are capable of. An 1/8th diameter cutter ask for at 1000sfm would want about 30560 rpm which will take a high-speed router spindle to make it happen. On the standard mill/lathe, you bite back on the surface speed and the numbers don’t need adjusted the same. Want to see the code that produce these numbers, on our site you’ll find a reference to our g/m-codes list where S and F codes are direct indicators.

Feeds and Speeds Charts by Material

Charts tempt you to treat one value as the single right answer, and that is a costly mistake. A machinist cutting 6061 on a rigid VMC and another on a worn bench mill must not run the same number, because pushing the wrong value risks chatter and broken tools; rigidity, not the table, sets the ceiling, and a 30% cut in feed is sometimes the only fix. We publish ranges, not single values, for exactly that reason, the same way the University of Florida Design & Manufacturing Lab labels its surface-speed table conservative.

Use these as approximate, initial guide lines. The surface speed ranges below approximate conservative academic values and the more optimal values for modern coated carbide will attain, pick the lower value for solid setups in which you’ve some question and build on from here. These are cross referenced from University shops data, tool manufacturer speed and feed data, and engineering data bases.

The range values shown were taken from the University of Florida Design and Manufacturing Lab, tool manufacturers’ suggestions and engineering reference books, and these are to be considered minimum appropriate ranges.

Feeds and Speeds for Milling and End Mills

Classic milling error: one feed for everything. A profiling pass and a full-width slot are not the same cut, so a machinist who slots at profiling feed will snap the tool because the cutter jumps from 30% engaged to 100%; using one feed everywhere is the mistake that breaks more end mills than any wrong speed. On the rigid mills we build you can push the slot, but only after you derate the number the chart gave you; MIT’s Fab Lab feeds-and-speeds reference makes the same point about matching feed to engagement.

Milling is where the formulas above get used most. Pick the surface speed from the chart, convert to RPM for your end-mill diameter, then set feed from chip load and flute count. Two extra variables matter on a mill: radial depth of cut (how much of the cutter width is engaged) and axial depth of cut (how deep). A full-width slot engages the whole cutter and generates the most heat and force, so derate surface speed and feed compared with a light profiling pass. Aluminum usually runs 2–3 flutes for chip clearance; steel and stainless run 4 or more flutes for edge strength and finish, and a higher helix angle shears more smoothly for a better finish.

Direction matters too. Climb milling (the cutter rotation matches the feed direction) gives a thick-to-thin chip and a better finish on a rigid machine, and it is the default for most modern carbide work. Conventional milling (rotation opposes feed) starts each chip thin, which rubs and work-hardens some materials but tolerates backlash on older manual machines. Keep separate numbers for a rough cut, where you push depth and feed for material removal rate, and a finish cut, where you lighten the chip for surface quality.

What feeds and speeds should I use for a 1/4″ end mill?

Let’s take a 1/4″ (0.25″) 3-flute carbide end mill in 6061 aluminum. Start with 700 SFM: RPM = (3.82 × 700) ÷ 0.25 = 10,696 RPM. At a 0.0025″ chip load, feed = 10,696 × 0.0025 × 3 = 80 IPM. If your spindle caps at 4,000 RPM you are RPM-limited again, so cap there and the feed becomes 4,000 × 0.0025 × 3 = 30 IPM.

For the same tool in mild steel, drop to roughly 400 SFM (carbide) and a 0.0015″ chip load. Production milling like this is the daily work of our CNC milling machines, and a rigid vertical machining center is what lets you actually run the higher end of these ranges without chatter.

Drilling, Turning, Reaming, and Tapping

Each operation punishes the wrong feed differently. Push a reamer at drilling speed and it rubs and loses size by 0.001 in; feed a tap off its pitch and it strips the hole, because a tap has no choice but to follow the thread. We see both failures on machines that were fast but not rigid, which is why workholding matters as much as horsepower on a turning center.

That same surface-speed math applies across operations, but the feed units and adjustments change. Here’s how each differs from milling.

Turning is the operation where the controller does the most work for you. A lathe running G96 holds a constant surface speed by raising RPM as the cutting diameter shrinks, so a facing cut keeps the same SFM from the rim to the center. Switch to G97 (fixed RPM) for threading and for drilling on the lathe. A tap is the strictest case of all: its feed is locked to the thread pitch, so a tap fed even slightly off ratio will strip the thread, which is why synchronized tapping exists. This logic live on every machine in our CNC lathe line, and the same surface-speed rules apply to manual work on a universal lathe. Operation-specific speed factors, like running a reamer or counterbore slower than a drill, are documented in university shop charts such as Olin College’s milling speed reference. Choosing the right cutter for each job is its own topic, covered in our guide to essential lathe cutting tools.

Adjusting Chart Values for the Real World

Charts assume an ideal setup. Your machine, fixture, and tool length rarely match the lab. Use this decision sequence to move off the starting numbers in the right direction. Peer-reviewed work on cutting-parameter selection shows the same trade-off between feed, speed, and tool life. We tune published numbers down on lighter machines because rigidity, not the chart, sets the limit, and a process built around a 0.004″ chip load still needs a 20% haircut when the tool is long.

Most people forget about chip thinning: as a resource notes, chip load equals feed/tooth only where depth of cut is equal to or exceeds the cutter diameter. Once we turn the dial down – use less than half the cutter diameter engaged (radial depth of cut) – we “thin the chip” and the true chip load decreases. As our education section says, this “chip thinning phenomenon is the primary reason why high-speed machining produces such high feed rates when used at light engagement.”

6 Feeds-and-Speeds Mistakes That Break Tools

Each mistake below ties cutting parameters to tool life, a relationship documented in peer-reviewed machining research.

There aren’t many different reasons why we’re snapping tools, but most of them go to a familiar half dozen – drawn from years in shop rooms and toolholder catalogs:

Industry Outlook: Adaptive and AI-Driven Feeds and Speeds

For machine builders this shift sharpens the same skill rather than replacing it. An operator who does not understand SFM will trust an adaptive system blindly and miss the moment it is masking a setup problem, because the software holds force, not judgment. We build the spindle sensing that feeds these systems, and a machine still depends on a machinist who can sanity-check 4,000 RPM against the cut; on the shop floor that judgment is what separates a 6,112 RPM target from a 4,000 RPM reality.

The biggest change to feeds and speeds isn’t a new formula; it’s the person setting the number. Adaptive control systems now read real-time spindle load, vibration and temperature and automatically march the feed override up or down in real time to maintain a constant cutting force, rather than a constant programmed feed.This is at the heart of high-speed machining (HSM), taking very light radial cuts at very high feeds.The patent literature already recognized this. US20020091460A1,for example, describes an adaptive control that adjusts the effective feedrate based on cutting forces detected by sensors on the spindle. US8880212B2 (High performance milling) likewise disclosed controlling feed rates to keep machining load within a predetermined range. CAM-software companies,too, now provide adaptive and trochoidal toolpaths that keep the tool’s engagement with the material fairly constant, allowing the tool to be fed much faster.

The main judgment,for a shop owner looking to machine with any proficiency,is that adaptive control turn tables of starting feeds and speeds into jumping-off points rather than landing strips. This has the effect of, rather than decreasing, increasing the importance of understanding the base physics behind cutting tool performance. Anyone familiar with SFM, chip load and chip thinning, after all, will quickly learn when a machine’s self-adjustment is making sensible, economic decisions and when it’s masking a fundamental problem. The research into machining economics -such as this peer-reviewed paper examining cutting-parameter choice for a piece of a composite aluminum-has continually kept reaching the same conclusion: to get the highest rate of material removal that you still have room for within the framework of tool life and cost-you’re going to make that trade-off according to the very same principles of speed and feed that have guided us for a century.For further insight from machine to CAM, see our ANTISHICNC machining blog.

Frequently Asked Questions