Chip Chad
Most feeds and speeds tools give you a number. Type in your diameter, pick a material, get an SFM. Done. The problem is that number assumes a machine that probably isn't yours — a rigid knee mill with a quality spindle and zero backlash. Run it on a desktop router and you'll either break an end mill or spend the next hour wondering why everything chatters.
Version 0.3.0 of Chip Chad is about closing that gap. Here's what changed.
Your machine, your profile
The old preset machine list is gone. Users kept telling us the presets never matched their actual machine closely enough to be useful — a "hobby router" can mean a $400 gantry kit or a $3,000 industrial-grade desktop. Presets can't capture that difference.
Custom machine profiles replace them. You fill in the numbers that describe your machine — power, RPM range, rigidity — and Chip Chad saves the profile locally. Next time you open it, your machine is waiting.
Along with this, we added a new Ultra Light rigidity class specifically for high-RPM desktop routers and gantry mills running trim routers or spindles under 1.5 kW. These machines have a fundamentally different stiffness profile than even a small benchtop mill, and the recommendations now reflect that.
Empirical engagement caps
This one matters most for side milling on lighter machines.
The chip load formulas are mathematically correct. But on a flexible machine, the ratio of width-of-cut to depth-of-cut the math allows can be mechanically problematic — the cutting forces couple in ways that cause deflection, chatter, or both. The formula doesn't know your machine flexes; it just knows chip area.
Empirical caps are an editorial layer on top of the hard math. For Ultra Light, Very Light, and Light rigidity profiles, Chip Chad now enforces maximum WOC and DOC limits derived from observed behavior on real machines of those classes. When a cap is active, the app tells you which parameter is being limited and why — it doesn't silently adjust numbers and leave you guessing.
You can turn the caps off if you know what you're doing. They're on by default because most of the time the formula is the one that's wrong.
30+ new materials
The material library expanded significantly. You can now get recommendations for:
Wood and engineered wood — softwood, hardwood, dense hardwood, MDF, plywood, particleboard. These have different chip load requirements and surface finish behavior than metals, and the recommendations now account for that.
Plastics — HDPE, UHMW, acetal, nylon, acrylic, polycarbonate, ABS, PVC, and PEEK. Each has different heat tolerance and chip-clearing characteristics. PEEK in particular runs more like a tough metal than a typical plastic.
Composites — PEEK-CF (carbon-filled PEEK) and G10/FR4 fiberglass. Abrasive, tool-wearing, and unforgiving of wrong chip loads.
More metals — 17-4PH and 15-5PH precipitation-hardening stainless, H13 hot-work tool steel, Class 40 gray iron, Ti-64 titanium, Monel, and expanded coverage of tool steels and stainless grades by subtype.
We also added Kienzle chip-thickness exponents to every material in the library. The short version: cutting force doesn't scale linearly with chip thickness, and the scaling is different for every material family. Aluminum and 316 stainless don't follow the same curve. Now they don't pretend to.
The results page grew up
Three new estimates appear on the results page alongside the feeds and speeds recommendation:
Tool life — a minutes-per-edge range based on your material, cutter type, surface speed, and chip load. This isn't a guarantee, but it gives you a reference point. Running aluminum at the right carbide SFM should get you 120–240 minutes per edge. Running steel with HSS at the right conditions gets you 15–40. Pushing beyond the recommended range tells you how much life you're trading away.
Deflection — a stickout-aware estimate of how much your tool tip is moving under load. If the number is red, the problem probably isn't your feeds; it's your setup.
Surface quality — for ball-nose and corner-radius end mills, Chip Chad now estimates the theoretical scallop height from your current feed and stepover. It tells you whether feed or stepover is the constraint on finish quality, which tells you which one to change first.
Smarter math under the hood
Corner-radius and bull-nose end mills now have proper geometry handling. The corner radius affects how much axial depth the edge can handle before it overloads, and the recommendations now account for that directly.
Coupled engagement tuning — the WOC and DOC manual sliders now enforce mutual limits. You can't drag both into a ratio that breaks the physics. If one is constraining the other, the UI says so.
Four-zone risk bands — the recommendation status moved from a binary ok/warn to a four-zone system: amber caution → green working range → amber caution → red. The difference between "a little conservative" and "in the good zone" is real information.
Six articles in the shop notes
The articles section launched alongside this update with six pieces covering the things that come up most in hobbyist machining:
- Understanding Chip Load — why it's the number that matters most and what it actually controls
- How to Tune a Hobby CNC for Better Cuts — tighten the machine before you touch the numbers
- Getting Better Surface Finishes on a CNC Mill — stepover, climb direction, and the four fixes to try in order
- Tool Life vs. Cutting Speed — the steep curve most people don't know about
- Choosing the Right End Mill — flute count, coating, and geometry for the job
- Setting Up a 3D Touch Probe with GRBL or FluidNC — workpiece zeroing that actually repeats
More coming. The goal is one article per question that keeps showing up in forums.
Try it
Chip Chad is free, runs in the browser, and doesn't require an account. If you've been using a lookup table or a tool that doesn't know what machine you're on, it's worth fifteen minutes to see what a machine-aware recommendation looks like.