CNC machinist technical exam – how deep do the G-code questions actually go on the written test?
I've been a CNC machinist for six years running mills and lathes, mostly Haas and Mazak machines, and I'm preparing for the CNC Computer Numerical Control Machinist Technical certification exam. My practical skills are solid but I'm honestly not sure how much the written exam focuses on G-code and M-code knowledge versus blueprint reading, metrology, and tooling selection. I want to make sure I'm not spending most of my study time on programming when the exam might care more about something else entirely.
I've been studying for about five weeks, maybe an hour to an hour and a half on most weeknights. My current practice scores are around 69%, which doesn't feel like enough. The blueprint reading and GD&T sections are fine — that's second nature from daily work. Where I'm struggling is with questions that feel more academic than practical, like theoretical cutting speed and feed calculations, tool geometry angles, and some safety standard questions that don't map cleanly to how things actually work on the shop floor.
I've talked to two guys at my shop who took the exam a couple years back. One said the G-code section was deeper than he expected and he wishes he'd reviewed canned cycles more thoroughly. The other said the programming questions were pretty basic and the metrology section was the real challenge. I don't know who to believe. If anyone's taken this recently I'd really appreciate a current read on what the exam actually emphasizes.
Your 69% with six years of shop experience is close to passing territory already — the gap is almost certainly the academic-style questions on theory and safety standards, not your practical knowledge. Find a question bank that emphasizes those areas and do two or three focused sessions on them. That's probably worth 8–10 percentage points by itself before your test date.
I took a similar CNC machinist certification about eighteen months ago and the G-code questions were moderate — not deep programming knowledge but you need to recognize what common codes do and be able to read a simple program and identify errors. Canned cycles like G81, G83, and G84 are worth reviewing. You don't need to write programs from scratch.
Metrology and inspection content was heavier than I expected on my exam. Understand how to read a micrometer and vernier caliper to tenths, know your gage R&R concepts at a basic level, and review common CMM terminology. That section rewards study time more than the G-code portion does for experienced machinists who already live in the code.
The speeds and feeds calculation questions are annoying if you're used to dialing things in by feel or trusting the software. Memorize the basic formulas for surface feet per minute and chip load — there were maybe four or five calculation questions and they're straightforward if you know the formula going in. Don't let them catch you unprepared.
So I was in a similar spot about eight months ago, working full-time on a Haas VF-2 and trying to squeeze in study time whenever I could. Honestly the G-code questions aren't as deep as I expected. They're not asking you to write a full program from scratch or anything like that. It's more like "what does G41 do" or "which code initiates a canned drilling cycle" -- recognizing codes rather than programming them. What tripped me up more was the coordinate system stuff, so I'd spend some time on cnc/questions/coordinate systems datum setup before your test date because that showed up more than I anticipated.
For fitting it in around a busy schedule, I didn't do long study sessions. I'd do 20-30 minutes on my lunch break or right after the kids went to bed. It's not glamorous but it adds up fast. The practical experience you already have is a huge advantage because a lot of the questions just click when you've actually run the machines.
Failed mine the first time because I spent way too much time drilling G-code syntax and barely touched the theory side. The written test isn't really asking you to write programs from scratch -- it's more like "what does G41 do" or "which M-code starts the spindle" and then a bunch of questions around speeds and feeds, tool offsets, and why you'd use certain canned cycles. I didn't take that stuff seriously enough and it showed.
Second attempt I shifted my focus toward the conceptual stuff -- coordinate systems, the difference between absolute and incremental positioning, how tool length compensation actually works. You still need to know your common G and M codes cold, but don't let that eat all your study time. Once I balanced it out I passed pretty comfortably. It's less about whether you can write a program and more about whether you understand what the machine is doing and why.
Just passed mine last month so I can actually answer this. The G-code questions aren't nearly as deep as I expected — you're not writing programs from scratch or anything like that. It's more like "what does G41 do" or "which code cancels cutter comp," conceptual stuff. I've been running Haas for years and honestly some of the questions felt almost too basic if you've got real floor time.
The thing that made the biggest difference for me was drilling the M-codes and modal groups, not just G-codes. I kept mixing up which codes were modal and which were non-modal and that tripped me up on practice questions way more than the actual G-code definitions did. Spend a focused hour on that distinction and you'll be in good shape.
Failed it the first time, honestly. I went in thinking my shop floor experience would carry me and completely underestimated how specific the written questions get. G-code wasn't as deep as I feared in terms of memorizing every code, but the coordinate system stuff got me bad — I wasn't solid on how they set up datum references and expected you to work through offset problems on paper. Did a bunch of practice on cnc/questions/coordinate systems datum setup 3 before my second attempt and it made a real difference.
Second time around I passed. What changed was I stopped assuming "I know how to do this on the machine" meant I could explain it on paper. They'll give you a scenario and you have to walk through the logic, not just punch in numbers. If you've got six years of Haas and Mazak time you're already ahead, just make sure you can articulate why you're choosing G54 over G55 or how a tool length offset actually works conceptually, not just mechanically.
Quick update since I posted last week -- I finally sat down and did a full practice run last night and pulled a 74, which honestly surprised me (in a good way). The G-code questions weren't as brutal as I expected, mostly parametric basics and canned cycles, but coordinate systems tripped me up more than I thought they would. I actually found this practice set really useful: cnc/questions/coordinate systems datum setup 3 -- it's got the kind of datum offset questions that keep showing up.
I'm planning to sit the real exam in about three weeks. Feels tight but I think it's doable if I keep the momentum going. Good luck to everyone else prepping -- you've got this.