Fault diagnosis on the car – from the fault code to the cause
Between an entry that has been read out and a cause that has been proven lie several test steps. They are usually skipped when the vehicle shows a symptom, a diagnostic device names a fault code – and a component is already on the table to be replaced. This is exactly where expensive mistakes arise: the code records which monitor responded; the trigger behind it stays open. Fault diagnosis on a car is therefore not a look-up but a sequence of test steps.
This page explains what that sequence looks like, what each level of the data says, and how you can tell that a cause is merely assumed rather than proven. If you want your own data placed in context, MotorScope assesses it independently of repair work and parts sales – usually within one hour of complete data arriving, around the clock, at night and at weekends too.
Have your diagnostic data checked
When systematic fault diagnosis becomes necessary
Not every entry in the fault memory needs methodical work. But as soon as the symptom and the code do not fit together, the order of the test steps decides the cost:
-
The engine warning light is on, and the code read out does not explain how the car drives.
-
The vehicle goes into limp-home mode although no component has obviously failed.
-
Several control units report faults at the same time and nobody relates them to one another.
-
The entry is stored as sporadic and cannot be reproduced at the test appointment.
-
The DPF regeneration fails to run or repeats unusually often.
-
The same fault returns after the repair – so the replaced part was not the cause.
-
The garage cannot find the fault, or two businesses name two different causes.
-
You are asked to approve the replacement of an expensive component and want to understand the data first.
Safety first. If the engine warning light is flashing, if there is a severe loss of power, smoke, a smell of fuel, unusual noises, or anything unusual about the brakes or steering, do not keep driving the vehicle. Have it checked on site first – a data analysis is no substitute for an inspection of the vehicle.
The six levels of fault diagnosis
Fault diagnosis on a car is a chain. Each level answers a different question, and each level presupposes the one before it. If a stage is skipped, what emerges is not a diagnosis but an assumption – and assumptions are paid for on the vehicle in replaced components.
1. Symptom – what the vehicle actually does
At the start stands not the code but the behaviour: when does it occur – on a cold start, under load, on the overrun, only after a longer drive? Since when, how often, under which conditions? The symptom determines which entries are relevant at all.
If this level is skipped: the fault memory is read without any reference. An old, long-inactive entry moves to the centre while the connection to the current behaviour is never made.
2. Fault memory – which system reported a deviation
The entry names the control unit, the fault code and the fault status. It says where a limit was exceeded – not by what. It also matters which entries occur together and in what order they arose: follow-on entries look like faults in their own right.
If this level is skipped: a single code, typed out by hand, is declared to be the diagnosis. The accompanying entries that would narrow the cause down never appear in the assessment at all.
3. Freeze frame data – the conditions under which it happened
Freeze frame data is the snapshot from the moment of storage: engine speed, coolant temperature, engine load, vehicle speed. It shows the operating situation. A fault at cold idle is a different case from the same code at full load and operating temperature.
If this level is skipped: the check takes place in an operating situation in which the fault never occurs. Every value is unremarkable – and the result is wrongly recorded as “no fault found”.
4. Live data – how the system behaves in operation
Live data shows the values continuously, not as a single frame: how does the reading respond to load changes, how quickly does it follow, does it jump, does it collapse? Only here does it become visible whether a sensor is delivering its signal or whether the control behind it is already working at its limit.
If this level is skipped: a component is judged on a single static value. Drop-outs that occur only under load or when warm go undetected.
5. Target/actual comparison – whether the value fits the operating situation
A value within tolerance is not yet a plausible value. What decides it is the comparison with the value that would be expected at this engine speed, this load and this temperature, and with publicly available manufacturer specifications and reference values. Only this comparison separates “plausible” from “unremarkable”.
If this level is skipped: everything looks normal because nothing lies outside the limits. But the deviation lies in how the values relate to one another – and that is precisely what goes unnoticed.
6. Test step – the measurement that confirms or rules out a cause
At the end stands not an order but a check: a measurement, a comparison, a targeted test on the vehicle whose result proves or discards a cause. A solid fault diagnosis ends with a verifiable question – not with a part number.
If this level is skipped: parts are replaced instead of checked. If the symptom remains, the next component follows – the case known as swapping parts on suspicion.
If you get no further at this point, you can have your diagnostic data assessed independently.
The levels build on one another: without the symptom the fault memory is arbitrary, without freeze frame data live data is hard to place, and without the target/actual comparison every irregularity remains a claim. That is exactly why MotorScope works through the levels in this order – and states explicitly which level is missing from your data.
Systematic diagnosis or swapping parts on suspicion
Systematic diagnosisSwapping parts on suspicion
Starting pointSymptom and dataFault code and experience
Next stepA measurement that proves or rules out a causeA component that gets replaced
Result if it worksCause proven, repair justifiedSymptom gone, cause unknown
Result if it does notOne hypothesis is ruled out, the circle narrowsCosts incurred, situation unchanged
TraceabilityEvery step can be checkedOnly the result is visible
The difference lies not in the effort but in what you are left with: a properly executed test step yields knowledge even when the assumption is not confirmed. A component replaced without a prior check yields none in that case.
What a fault code says – and what it does not
A standardised fault code has five characters. The first is a letter and says in which layer of the vehicle the entry arose: P for the powertrain, B for body and interior systems, C for chassis and brakes, U for communication between control units. The letter therefore determines where the search begins – not which component is affected. How the remaining characters are structured, and what distinguishes standardised from manufacturer-specific codes, is set out under the structure of a fault code.
Fault status: static or sporadic
Alongside the code, the control unit stores the status. Static means the fault is present at the moment of the read-out. Sporadic means it was present but is not active right now – the entry remains as history.
A sporadic entry is therefore not incidental: it belongs to a behaviour tied to a condition, and together with the matching freeze frame data it indicates the circumstances under which the fault can be reproduced.
The counter also matters, where the control unit keeps one: how often has the fault occurred, how many driving cycles have passed since it last did? A one-off entry from twenty driving cycles ago has to be judged differently from the same code being set on every journey.
Clearing is not a test step: secure the data before clearing, otherwise status, frequency, freeze frame data and the order of the entries are lost.
What data makes a fault diagnosis solid
Where to get this data and what a read-out costs is described under online vehicle diagnostics. For the assessment itself, what counts above all is that every level of the diagnosis is covered:
Required
-
Vehicle data: VIN (17 characters; the letters I, O and Q do not appear in it), model, engine, year of manufacture and mileage.
-
Description of the symptom: when the behaviour occurs – right after starting, under load, on the motorway, with a cold engine or only after a longer drive. Since when, how often, under which conditions.
-
Diagnostic data: diagnostic tester export, fault memory print-out or app export with control unit, fault code and fault status.
Helpful, if available
-
Freeze frame data with engine speed, coolant temperature, engine load or vehicle speed.
-
Live data and target/actual comparisons.
-
Screenshots from an OBD app and photos of the warning lights.
-
Invoices for work already carried out and a list of replaced parts.
-
Any indication that the fault memory was cleared, that the battery was weak or that a connector was recently disconnected.
How to make your data usable: send the complete log wherever possible rather than a code typed out by hand. Screenshots only help if every value is legible and not just the top line of the screen is shown. Photos should be sharp and show the entry in full. A special diagnostic tester is not a requirement – existing print-outs, photos or app exports are often enough, provided the fault description is meaningful.
How the analysis works
-
Describe the problem – Give the vehicle model, engine, year of manufacture and mileage – and describe when the symptom occurs.
-
Send the data – Upload fault codes, diagnostic logs, freeze frame data or screenshots through the form.
-
Assess the connections – MotorScope compares the data with the symptoms you described and the work carried out so far.
-
Receive the analysis report – Usually within one hour of complete data arriving, you receive your analysis report with a traceable assessment and a prioritised list of possible checks.
What the analysis report contains
-
Systematic assessment of the fault memory and – where available – the live data and sensor values
-
Plausibility check of the values and assessment of the technical connections
-
Comparison with publicly available manufacturer specifications and reference values
-
VIN-based check for publicly available recalls and service campaigns
-
Assessment of the irregularities and recommendations on how to proceed
What the online assessment does – and what it does not
What the online assessment does
-
Puts fault memory, live data and symptoms into context
-
Shows which measurement is most likely to bring clarity next
-
Makes a repair recommendation verifiable before you approve it
-
Prepares the garage visit with a clear question
-
Is independent of repair orders and parts sales
What it does not do
-
Repairs and parts sales
-
Approving repairs or certifying roadworthiness
-
Replacing a personal inspection of the vehicle
-
Swapping parts on suspicion, or a guarantee of success
-
Replacing the German MOT (HU), the emissions test (AU) or any other legally required inspection
If the data does not show a clear picture, the report says so explicitly. Rather than settling on a cause prematurely, it shows which check will bring the most clarity next.
Frequently asked questions about fault diagnosis on the car
Is the fault code alone enough for a diagnosis?
No. The code names which control unit registered which deviation. Whether the cause lies in the reported component, in its supply, in the wiring or in a mechanical condition upstream is decided only through freeze frame data, live data and the target/actual comparison.
What do the letters P, B, C and U in the fault code mean?
They name the area in which the message arose: P for the powertrain, B for body and interior systems, C for chassis and brakes, U for communication between control units. The letter shows where the search begins – not which part is affected.
Why does the same fault come back after a part is replaced?
Because the replaced component reported the deviation but did not cause it. As long as the actual trigger – low voltage, unmetered air or a signal path, for instance – remains unchanged, the entry is set again. More on this under the fault returns after the repair.
What do I do if two garages name different causes?
Test both statements against the same data. Usually it is not the readings that differ but their interpretation. An independent assessment shows which of the two hypotheses can be confirmed or ruled out by which test step – see the garage cannot find the fault.
Should I have the fault memory cleared before the appointment?
Better not, without securing the data first. Clearing loses the fault status, the frequency, the freeze frame data and the order of the entries – exactly what makes an assessment possible.
How much data do I need for a solid assessment?
Required are the vehicle data, a concrete description of the symptom and a read-out with control unit, fault code and fault status. Everything beyond that – freeze frame data, live data, earlier invoices – adds to what can be said. Where to get this data is set out under online vehicle diagnostics.
Does the assessment name the component to be replaced?
No. Approving a parts replacement is not part of the assessment. The report places the available indications in context and names the checks that can prove or rule out a component as the cause.
You will find further answers under All questions and answers.
Understand the data first. Then act with purpose.
Upload your diagnostic data and receive an independent assessment usually within one hour – before you order the next part.