MTTR (mean time to repair) is the average time it takes to get a CNC machine running again after a breakdown. If your machines are down for 40 minutes on average each time they fail, your MTTR is 40 minutes.
This guide shows how to calculate MTTR from the stop data your shop already collects, what counts as repair time, and how to use the number to cut downtime. It is the companion to our guide on MTBF for CNC machines. MTBF tells you how often a machine breaks down. MTTR tells you how long each breakdown lasts.
MTTR measures the time between two moments: the machine stops because of a failure, and the machine is back in production. Everything in between counts. That includes waiting for a technician, waiting for a part, finding the cause, doing the repair, and restarting.
The “R” can stand for repair, recover, restore or resolve. People use all four. What matters is that you pick one rule and apply it to every machine, so the numbers can be compared.
The formula is simple: MTTR = total repair time ÷ number of repairs. Here, “repair time” means the whole stop, from the failure to the return to production.
Use only unplanned stops (breakdowns). Setups, breaks, changeovers and planned maintenance are not failures, so they stay out of the calculation.
One machine had five breakdowns in a month. Here is how long each one lasted, from the moment the machine stopped to the moment it was producing again.
| Breakdown | Cause | Time stopped |
|---|---|---|
| 1 | Spindle alarm | 30 min |
| 2 | Coolant pump | 45 min |
| 3 | Chuck sensor | 20 min |
| 4 | Hydraulic leak | 90 min |
| 5 | Door switch | 15 min |
| Total | 200 min |
MTTR = 200 minutes ÷ 5 repairs = 40 minutes.
Look at the middle value too. Sorted from shortest to longest, the five stops are 15, 20, 30, 45 and 90 minutes, so the middle one is 30. The single 90-minute stop pulls the average up. When one long breakdown drives your MTTR, look at that breakdown first.
You do not need a big project to start. You need the same five steps every month.
If stop times are still written on paper, they will be late and incomplete. Our guide to automated downtime detection explains how to capture them without extra paperwork.
Capture the stop data MTTR depends on
JITbase records stop reasons at the machine, so breakdowns are separated from planned stops from the start.
Most disagreements about MTTR come from three questions. Settle them once and write the answers down.
Very rare events, such as a crash that takes a machine out for days, are better reported on their own. Mixed into an average, they hide how the machine behaves on a normal month.
A 90-minute breakdown is rarely 90 minutes of repair. Here is one stop split into its four parts.
| Part of the stop | What happens | Time |
|---|---|---|
| Waiting | The operator calls for help, and the technician or part arrives | 35 min |
| Diagnosis | Someone finds the cause | 15 min |
| Repair | The fault is fixed | 30 min |
| Restart | The machine is restarted and the first part is checked | 10 min |
| Total | 90 min |
Only 30 of the 90 minutes were hands-on repair. In many shops, waiting and diagnosis take a large share of the stop. That is why asking technicians to work faster does not always lower MTTR. The gains usually come from shortening the wait and the search for the cause.
You will not see this split in a single average. Record at least the waiting time separately on a few breakdowns, and the pattern shows up quickly.
The two numbers answer different questions. MTBF is the average running time between breakdowns. MTTR is the average time to get the machine running again. A machine can break down rarely but take a long time to fix, or break down often but restart fast. Each case needs a different action.
Together they give availability, the share of time a machine can run: availability = MTBF ÷ (MTBF + MTTR). With an MTBF of 100 hours and an MTTR of 4 hours, availability is 100 ÷ 104 = 96.2%. Availability is also the first factor of OEE, although OEE counts all stops, not only breakdowns. Our OEE guide for CNC shops explains the full calculation.
The formula is easy. The data is the hard part: stop times written down after the fact, planned stops mixed in with real failures, and breakdowns nobody recorded. JITbase is built to remove that manual step.
On the screen at the machine, operators record the reason for stops. That separates a breakdown from a planned stop such as a setup, a break or a changeover. Each stop has a start and an end, so the duration of every breakdown is already there for the MTTR calculation.
JITbase supplies the data for MTTR: the breakdown stops, how long each lasted, and which machine they belong to. You can then do the calculation in a spreadsheet or in a report built on this data. Where the stop reasons allow it, agree on a separate reason for waiting on maintenance. The split between waiting and repair then shows up in your data without extra paperwork.
Check that your machines can send this data
See which CNC machines JITbase connects to, so stop times are captured automatically.
Start from the split in the previous table and attack the biggest part first.
To choose which fault to work on first, rank breakdowns by the total time they cost. Our guide to downtime root cause analysis with Pareto charts shows how to turn that ranking into a fix.
Change one thing on one machine, then compare its MTTR over the next month with the month before. That tells you whether the change worked.
Estimate what slow repairs cost your shop
Put a number on the hours lost to breakdowns and see what a shorter repair time is worth.
MTTR is easy to calculate and easy to get wrong. Set one start and one end rule, count only breakdowns, and measure from the stop to the return to production. Then look at where the time goes. In many shops, the fastest gains come from the waiting and the diagnosis, not from the repair itself.
Reliable stop data makes all of this possible. Once breakdowns are recorded at the machine, MTTR and MTBF stop being a monthly spreadsheet exercise and become numbers you can act on.
MTTR most often stands for mean time to repair. Some shops read it as mean time to recover, restore or resolve. The idea is the same: the average time from the moment a machine stops because of a failure to the moment it is back in production.
Because the letters can mean different things, write your definition down. Decide when the clock starts, when it stops, and whether waiting time counts. A clear definition matters more than the exact wording of the acronym.
There is no single good number. It depends on the machine, the kind of faults it has and how fast help can reach it. A simple sensor fault and a spindle problem take very different times to fix.
The useful comparison is with your own history. Calculate MTTR for each machine over several months, and watch whether it goes up or down. Then compare by type of fault, so you can see which breakdowns take the longest and are worth fixing first.
Yes, in most shops it does. The machine is not producing during that time, and the cost to production is the same as during the repair itself. Leaving the wait out makes MTTR look better than the reality on the shop floor.
This is also the reason to record the waiting separately when you can. If a third of each stop is spent waiting, you know where to act: who gets called, how fast they come, and which spare parts are on the shelf.
Start by finding out where the time goes. If a large share of each stop is spent waiting, more technicians will help only a little. Clear instructions on who to call, spare parts kept close to the machines and maintenance coverage on every shift often save more time than extra staff.
Then work on the diagnosis and the restart. A short list of recurring faults and what fixed them, and a clear check before production resumes, shorten two parts of the stop that no one counts as repair. Change one thing on one machine, and compare its MTTR over the following month.
No. MTTR describes unplanned failures. A scheduled oil change, a planned inspection or a setup is work you chose to do at a time you chose, so it is not a breakdown.
Counting planned stops would blur the picture. A shop that does more preventive maintenance would see its MTTR move for the wrong reason. Recording planned and unplanned stops under different reasons at the machine keeps the two apart.
There is no fixed minimum, but a handful is not much. With two or three breakdowns in a period, one unusual stop can change the average a lot. Treat the number as a first indication, not as a result.
Show the number of repairs next to the MTTR, and look at several months together. If a machine rarely breaks down, group similar machines to get a larger sample, and say clearly that the figure is for the group.