PLCs.ai
The PlatformAutomation ContextPricingYour DataBlogDeveloperLog in
Start free trial
Downtime math

What does it cost to find the fault?

Not the cost of downtime. The cost of the part before the repair even starts — the hours your people spend working out why the line stopped. Put your own numbers in.

Every default comes from published industry data. Sources are listed at the bottom.

What gets priced here

IEC 60050-192 splits fault repair into three elements. This calculator prices the first one: the time spent finding the cause.

Fault localizationWorking out what failed, and why.PRICED HERE
Fault correctionPhysically putting it right.not counted
Function check-outProving the line runs again.not counted
01

Your line

Presets are illustrative line-level starting points, not benchmarks. Replace every one of them with your own figures.

Currency
Converted at ECB rates, 2026-08-07.
02

The stops

Split the stop in two. Everything before you know the cause, and everything after. Only the first half is what this calculator prices.

03

What a stopped hour costs

Two ways in. Build it up from your line rate, or enter a figure your finance team has already signed off.

Is this line capacity-constrained?

Sold out means the units are gone for good, so the full contribution margin is charged against the stop.

04

Your assumption

How much of the time-to-cause do you think is removable? This one is yours, not ours.

Reduction in time-to-cause30%
We have no audited efficacy figure to offer you, so we are not going to invent one. This slider is an assumption you set and own. Everything above it is sourced; this is not.
Cost of diagnosis time
$386,551

per year, spent working out why the line stopped — before anyone picks up a tool.

126 h
Hours a year finding causes
$1,333
Per stop
40%
Of repair time is finding it
$966k
All unplanned downtime
Lost output$2,700
Idle line crew$232
Responders$144
One stopped hour$3,076
At your 30% assumption, 38 h a year comes back:
$115,965

Against the published averages

Your own numbers, checked against what has actually been published. We would rather flag a figure than quietly print a big one.

Unplanned stops
Reference: 25 per month
25 per month
Mean time to repair
Reference: 65 min
65 min
Finding the cause, as a share of repair time
Reference: 30–40% of repair time is hands-on work
40%
How this is calculated

No black box.

The whole model is four lines of arithmetic. If you disagree with a number, change it — and if you disagree with the method, here it is in full.

cost of one stopped hour= lost output + idle line crew + responders
lost output= units/hr × contribution marginif capacity-constrained= crew × loaded rate × 0.5if you make it up later
fault-finding hours/yr= stops/week × weeks/year × (minutes-to-cause ÷ 60)
cost of diagnosis time= fault-finding hours/yr × cost of one stopped hour

Lost contribution margin only counts when the units are genuinely gone. If the plant has slack and makes them up on a later shift, the real cost is the incremental premium on that make-up hour — the 0.5× of time-and-a-half — which is usually an order of magnitude smaller. Charging full margin to a line with spare capacity is the most common error in downtime maths, and it is why most vendor calculators can only produce large numbers.

A controls engineer can tell you how long it took before anyone knew why the line stopped. Nobody can tell you what share of MTTR that was. So the calculator takes the two durations and derives the share itself, then shows it back against the published composition, where hands-on repair accounts for only 30–40% of repair time and the rest is detection, diagnosis and verification.

Siemens reports up to $2.3M per hour for automotive. That is a whole-plant figure: at roughly 60 vehicles an hour it implies $38,300 of loss per vehicle, several times the contribution margin on a car, because it bundles supply-chain penalties and downstream plants. It cannot describe one line. We use the report for what it does support — how often plants stop, and for how long.

Detection delay, waiting on parts, and the repair itself all sit outside this number, even though they are real downtime. There is also no audited figure for how much time-to-cause PLCs.ai removes, so the reduction slider is set by you rather than by us.

Siemens publishes 25 stops and 27 hours per month across all sectors, not per sector. So every preset starts from the same stop frequency, even though an FMCG line realistically sees more and shorter stops than a mill. Change it to match your own stop log.

Loaded rate means the total cost to employ someone, including benefits and payroll taxes. The US Bureau of Labor Statistics put that at $46.60 per hour worked across private industry in March 2026. Skilled trades and controls engineers sit above it, which is why the responder rate defaults higher than the line crew rate.

Rates come from the European Central Bank’s daily reference rates and are refreshed once a day. Switching currency converts every money field on the page, including our own list pricing in the payback panel. If the rate lookup is unavailable the page falls back to indicative rates and says so.

Sources
IEC 60050-192:2015Active repair time = fault localization + correction + check-outMIL-HDBK-472Fault location as a separately estimated repair elementEN 13306:2017European maintenance terminologyISO 22400-2:2014Standard availability and OEE definitionsSiemens, 202425 stops and 27 hours per plant per month, a 65-minute MTTRARC Advisory Group~5% of production capacity lost to unplanned downtimeUS BLS, ECEC, March 2026Loaded labour cost of $46.60 per hour worked

Cut the time it takes to find it.

Those hours go into one question: why did the line stop? That is a question about the logic — which routine holds this conveyor stopped, and what has to be true for each one. Upload a project and ask it directly.

Start free trial →