CNC Repair Tools: What Actually Earns Its Place in a Technician’s Bag
A servo alarm can come from a loose 24 V terminal or from a dead amplifier, and the two look identical on the screen. What separates a ten-minute diagnosis from a two-day parts swap is almost never experience alone — it is which instrument is in your hand and where you put the probe. This page covers the CNC repair tools that actually earn their place in a maintenance technician’s bag for FANUC and SINUMERIK electrical work: what to buy, what specification matters and why, and what to leave on the shelf.
It is also the tooling hub of this site. Every instrument below is tied back to the troubleshooting guides where it gets used, so you can go from “I need to measure this” to “here is the procedure” in one click.
1. The four instruments that cover most CNC electrical faults
Field experience across hundreds of alarm calls keeps landing on the same four tools. Everything else is a specialist addition.
1.1 Digital multimeter — True RMS, CAT III, 0.1 mV resolution
What matters: True RMS (not averaging), a DC millivolt range that resolves 0.1 mV, continuity with a fast beeper, diode test, and a safety rating of CAT III 600 V or better for work inside a cabinet. A magnetic hanger and a backlight are not luxuries when you are holding a probe in one hand and a ladder printout in the other.
Why True RMS is not optional: averaging meters are calibrated for sine waves. The output of a spindle drive, the PWM chopping on a servo amplifier and the ripple on a switching supply are all non-sinusoidal, so an averaging meter reads a number that is simply wrong — and wrong in a direction that makes a healthy supply look marginal. When you are deciding whether a 24 V rail is sagging under load, that difference decides the call.
What it is good at: 24 V control circuits, ±10 V analogue command levels, contactor coil continuity, fuse and rectifier checks, battery voltage on the memory backup, and confirming that a signal reaches a terminal before you blame the ladder.
What it cannot do: read an encoder’s differential A/B pair, or show you whether a pulse train is clean. A meter gives you an average of a signal that may be missing every second edge. That is what the scope is for.
1.2 Clamp meter — DC capable (Hall effect), not AC-only
What matters: a DC clamp function. An AC-only clamp is the single most common wrong purchase: it reads zero on a DC servo bus and tells you the drive is dead when it is not. Look for 10 mA-class resolution on the low range and a jaw that fits your motor supply cable.
Why it earns its place: three-phase current imbalance on a spindle motor, the difference between a mechanically overloaded axis and an electrically failing one, and verifying that a servo draws roughly what the load predicts before you start changing parameters.
1.3 Insulation resistance tester — 500 V / 1000 V megohmmeter
What matters: 500 V and 1000 V test voltages, a range that reaches the GΩ decade, and a test button you can hold while watching the reading settle. Polarisation index readings are a bonus for motors that have been sitting in a humid workshop.
Why it is the tool that finds the fault nobody else finds: motor windings and cables that fail to earth only when warm or wet. A multimeter will read a perfectly sane resistance to earth; a megohmmeter will break down the insulation and show it.
Before you press the test button: disconnect the amplifier and the encoder from the circuit under test. A megohmmeter applies several hundred volts — exactly the kind of overvoltage that destroys a drive input stage or an encoder. Isolate first, test second, reconnect third. This is the most expensive mistake in this page.
1.4 Oscilloscope — two channels and a differential probe
What matters: at least two channels, 100 MHz class bandwidth (encoder and PWM signals are well inside this), and — far more important than bandwidth — a differential probe. Clipping a scope ground lead onto a live machine reference is how people create a second fault while diagnosing the first.
Why it earns its place: encoder A/B/Z integrity, a missing or inverted edge, PWM output symmetry, supply ripple, and the noise that shows up when machine grounding is poor. If your machine throws a family of detection-related servo alarms that will not clear, the scope is the instrument that proves whether the problem is the scale, the cable or the earth.
Table 1 — Which instrument answers which question
| Question in front of you | Instrument | What a wrong reading looks like |
|---|---|---|
| Is the 24 V rail holding under load? | True RMS multimeter | Averaging meter reads healthy while the rail sags on PWM load |
| Is the axis mechanically loaded or electrically weak? | DC clamp meter | AC-only clamp reads zero and looks like a dead drive |
| Is the motor winding breaking down to earth? | Megohmmeter | Multimeter reads fine; fault only appears warm or wet |
| Is the encoder pulse train clean and complete? | Oscilloscope + differential probe | Meter averages missing edges into a normal-looking voltage |
| Is a terminal or contactor running hot? | Thermal camera | Nothing — a meter cannot see heat before failure |
2. Specialist items worth owning
These are the tools you buy once and reach for ten times a year. They are what separates a technician who can only diagnose from one who can also confirm and hand over.
Table 2 — Second-tier kit and the job it does
| Item | What it is for on a CNC |
|---|---|
| Thermal camera | Finding a heating terminal, a failing contactor pole or an amplifier with a blocked heatsink before it alarms out |
| Encoder signal test box / breakout | Injecting or tapping A/B/Z without disturbing the machine harness |
| Anti-static wrist strap | Mandatory whenever a board, drive or memory module leaves the cabinet |
| Insulated hand tools and terminal crimpers | Repairing 24 V circuits so the fix survives vibration, not just the shift |
| Borescope or inspection camera | Looking inside a drag chain, a motor terminal box or behind a gantry without a two-hour strip-down |
| USB / Ethernet adapters for the control | Pulling a full backup before you change anything — the cheapest insurance in this trade |
Two of these have a software half. If you are editing or reverting ladder logic on a FANUC control, you need the programming package as well as the cable — see the FANUC LADDER III guide for version selection and installation. If you are chasing vibration or tuning a loop, FANUC SERVO GUIDE turns the scope into a measurement instrument for the servo itself.
3. What to leave on the shelf
- An averaging (non-True-RMS) multimeter. It will give you a confident wrong answer on every drive output on the machine.
- An AC-only clamp meter. Useless on a DC bus, which is most of what a CNC has.
- A “complete” network and cable tester kit. You will use two of the twelve adapters, once.
- A bench LCR meter. Genuinely useful in a repair shop with a soldering station; dead weight in a field bag.
- Cheap probes. The probe is half the measurement. A cracked lead adds resistance that shows up as a fault you then spend an hour chasing.
4. Selection notes that save money
Match the CAT rating to the cabinet, not to the catalogue. A CAT III 600 V meter is the floor for work on machine power circuits. The rating is about transient survival, and it is the one specification where going cheap is a safety decision rather than a budget one.
Calibration is a consumable. Once a year for a meter that lives in a van, and immediately after any drop onto concrete. A meter that reads 0.3 V low will send you looking for a fault that does not exist.
Buy the probe and the fuse with the instrument. Both are consumables and both are the usual reason a “dead” meter comes back to life.
Second-hand is fine for the scope, risky for the meter. A used scope either shows a trace or it does not. A used meter can be out of calibration in a way you cannot see until it has already misled you.
Where to buy: for reference, the categories above are easy to search on the usual marketplaces —
True RMS multimeters,
DC clamp meters,
insulation testers,
oscilloscopes with differential probes,
thermal cameras.
Buy from a seller with a returns path and check the safety rating on the actual listing — marketplace listings often show a photo of the professional model and ship a different one.
5. Where each instrument shows up in the troubleshooting guides
This is the part most tool pages skip: the instrument is only useful once you know which point to put it on. Each of the following guides names the measurement and the expected reading.
- FANUC Diagnosis Numbers (DGN) Quick Reference — the diagnosis screen is your first instrument. Read it with the meter still in the bag.
- FANUC No-Alarm Fault Diagnosis — DGN000 / DGN003 / DGN005: the machine is waiting for a signal, and the multimeter confirms whether that signal is present at the terminal.
- CNC Servo Vibration, Overshoot and Noise — classify the phenomenon before touching a parameter; the scope and the tuning software tell you which class you are in.
- FANUC Servo Alarms SV0381/0384/0385/0386/0387 — a grounding and noise family. This is where a scope and an earth resistance check beat any amount of parameter editing.
- FANUC Servo Amplifier Replacement — the discharge and isolation procedure before you put a meter on a DC bus.
- FANUC Battery Alarm 306/307 — a meter reading on the backup battery, taken while the power is on, is what tells you whether you have minutes or hours.
6. If you would rather not buy the instrument for a one-off job
Not every fault justifies a purchase. If you are a dealer inspecting a machine before delivery, an integrator with a single stubborn axis, or a shop that has the meter but not the scope, that is exactly what remote technical support is for: you measure, I read the numbers with you and tell you where to put the probe next.
And if you are starting from the alarm itself rather than from the tool, go to the FANUC troubleshooting library — 48 field guides organised by symptom, all cross-linked to the instrument you need.
7. Tooling questions that come up most
Do I really need a True RMS multimeter for CNC work?
For anything that touches a drive output, yes. Spindle and servo outputs are non-sinusoidal, and an averaging meter is calibrated for sine waves, so it gives you a confident wrong number exactly when you are deciding whether a supply is marginal. If you only ever check 24 V control circuits and continuity, an averaging meter mostly works — but the first time you put it on a drive output it will mislead you.
Can a normal AC clamp meter measure servo current?
No. The servo and spindle buses are DC, and an AC-only clamp reads close to zero on them, which looks exactly like a dead drive. You need a Hall-effect clamp with a DC range. This is the most common wrong purchase in this trade.
What is the minimum oscilloscope for encoder and PWM diagnosis?
Two channels and a 100 MHz class bandwidth covers encoder A/B/Z and PWM signals comfortably. The probe matters more than the bandwidth: use a differential probe rather than clipping a ground lead onto a live machine reference. A portable USB scope with a proper differential probe is more useful in a cabinet than a large bench scope with the wrong front end.
Disclosure: some links in section 4 are affiliate links. If you buy through them this site may earn a commission at no extra cost to you. Instruments are listed by category and specification, not by sponsorship — nothing here is placed because a brand paid for it. Always confirm the safety rating and current specification on the seller’s page.
