Redundant brake system: categories, PLr, MTTFd, DC and CCF according to ISO 13849-1 – how to size brakes that still hold safely when one path fails.
AI This article was created with the help of AI.

Table of contents
- What makes a brake redundant: categories according to ISO 13849-1
- Determining the performance level: risk graph and PLr
- MTTFd, diagnostic coverage and common cause failure
- Homogeneous vs. diverse redundancy in practice
- SIL according to IEC 61508 compared with the performance level
- Typical mistakes when implementing redundant brake systems
- Frequently asked questions
- Sources
What makes a brake redundant: categories according to ISO 13849-1
With brakes, redundancy is not a state but a structural property that can be demonstrated. A brake is redundant if the safety function, for example reliably holding a vertical load or stopping an axis, is maintained even when one braking path fails. DIN EN ISO 13849-1 describes this capability through categories B, 1, 2, 3 and 4, which define the architecture of the safety-related parts of a control system. Only from category 3 onwards are two functional channels with mutual monitoring required; only there does the standard actually speak of redundancy[1].
| Category | Structure | Basic characteristic |
|---|---|---|
| B | Single-channel | Basic principles applied, a fault can lead to loss of the safety function |
| 1 | Single-channel | Well-tried components and well-tried safety principles |
| 2 | Single-channel with test channel | Fault detection in the functional channel by a test device, test rate at least 100 times higher than the demand rate |
| 3 | Two channels, redundant | A single fault must not lead to loss of the safety function, mutual channel monitoring |
| 4 | Two channels, redundant | Additionally robust against the accumulation of up to two undetected faults, DC of 99 % by calculation |
It is important to distinguish this from the passive brake: a spring-applied brake that engages automatically on loss of energy works on the fail-safe principle but, strictly speaking, is not a redundant system but a single fail-safe function. We use passive brakes where loads act on the drivetrain even at standstill or where emergency stop functions have to protect people and machine.
The categories are meant to be technology-independent: they apply equally to the electrical, pneumatic, hydraulic and mechanical parts of a safety function. For the design engineer this means that the category describes only the structure. Whether a redundant brake system actually achieves its performance level is decided by the quantitative variables in the next section.
Determining the performance level: risk graph and PLr
The assessment does not start at the brake but at the risk assessment. From it you derive the required performance level PLr via the standard’s risk graph, based on three parameters: the severity of injury (S1 slight, reversible injuries, S2 serious, irreversible injuries or death), the frequency and duration of exposure (F1 if the cumulative exposure time is at most one twentieth of the operating time and not more often than once per 15 minutes, otherwise F2) and the possibility of avoiding the harm (P1 possible under certain conditions, P2 not possible)[3].
MTTFd, diagnostic coverage and common cause failure
The achieved performance level PL then results from four variables that you demonstrate for every safety function:
- Category as a structural requirement, from category 3 onwards with two-channel redundancy and mutual monitoring.
- MTTFd per channel, classified as low (3 to less than 10 years), medium (10 to less than 30 years) and high (30 to 100 years); in category 4 the value can be raised to up to 2,500 years[3].
- Diagnostic coverage DC as the percentage of dangerous failures detected; category 4 requires 99 percent by calculation[2]. How this proof can be provided in practice is shown in our article on sensor-monitored brakes.
- Common cause failure: measures against failures with a common cause, demonstrated with at least 65 out of 100 points according to Annex F of the standard[4].
From these variables you calculate the probability of a dangerous failure per hour and thus the PL, which must at least correspond to the PLr. Tools such as the free SISTEMA from the IFA lead from the safety-related block diagram to the achieved PL and document the proof in a traceable way[4].
Homogeneous vs. diverse redundancy in practice
Two identical brakes on the same shaft look like redundancy but often share the same causes of failure: a common power supply, identical friction lining from the same manufacturer, the same mounting environment, the same thermal and contamination load. This homogeneous redundancy is susceptible to common cause failures, and that is exactly why Annex F of the standard demands explicit measures such as separation of energy paths, different lining qualification and independent control. The ZVEI also points out that subsystems of the same, homogeneous design are generally to be excluded, because questions of systematic suitability and of the required fault detection often cannot be answered sufficiently[3].
Diverse redundancy goes one step further and combines different operating principles, for example a spring-applied safety brake as the electromechanical channel with a hydraulic clamping unit as the second, independent channel. Different physics, different actuators and separate media paths measurably reduce the risk of common causes of failure and are the way to robust proof in category 3 or 4. For the sizing of the spring-applied brake as a fail-safe channel, we refer to our technical article on the electric motor with brake.
SIL according to IEC 61508 compared with the performance level
Alongside PL there is the SIL system according to IEC 61508, which is applied in the process industry and for safety functions with a low demand rate, less often than once per year; the 2023 edition of ISO 13849-1 explicitly refers to the DIN EN 61508 series of standards for this operating mode[2]. Both systems can be compared approximately via the failure probability:
| Performance level | PFH (dangerous failures per hour) | corresponds approximately to |
|---|---|---|
| a | 10^-5 to less than 10^-4 | no SIL |
| b | 3 x 10^-6 to less than 10^-5 | no SIL |
| c | 10^-6 to less than 3 x 10^-6 | SIL 1 |
| d | 10^-7 to less than 10^-6 | SIL 2 |
| e | less than 10^-7 | SIL 3 |
This assignment is an approximation based on the PFH limits, not a normative equivalence[5].
What does this mean for a brake with SIL 1? A SIL refers to a safety function: the operator determines it in the hazard assessment for each safety function, after which suitable devices are selected and combined into a system[7]. A redundant design is not always necessary for this: under test principle GS-MF-28, the proof for a proven-in-use brake (category 1, PL c) can be certified[8], i.e. at a level that according to the table above corresponds approximately to SIL 1. If staying under a vertical axis cannot be prevented, however, the DGUV provides for a redundant fall-protection device according to category 3, PL c[8].
Typical mistakes when implementing redundant brake systems
In practice, redundant brake systems rarely fail because of the technology but because of the proof. These are the mistakes we encounter most often:
- Installing two brakes and claiming redundancy: without separate control paths, independent energy supply and demonstrated CCF measures according to Annex F, the structure is not category 3, even if two brake calipers are mounted.
- Safety function defined vaguely: if it remains unclear which components lie in the functional channel between the triggering event and the reaction, no robust calculation is possible. The standard requires an exact definition for every safety function, with a short description, triggering event, necessary reaction and permitted response time[1].
- Proof not provided: the achieved PL must at least correspond to the PLr, and this has to be substantiated by calculation, not merely claimed. SISTEMA documents this step from the block diagram to the achieved PL[4].
- Ignoring an insufficient result: if the achieved PL is not sufficient, there are exactly three levers: choose components with a higher MTTFd, improve fault detection or raise the category, for example from category 3 to category 4 with a DC of 99 percent.
A fourth point, often overlooked: a PL e must not rest on fault exclusions alone, and undetected failures with a very low probability may be disregarded in category 4 only if this fault consideration is documented and verified[6].
At ATEK Drive Solutions, we develop redundant brake systems as part of the solution for the complete drivetrain: precisely matched to your safety function, ready for your project including the sizing proof, and integrated from a single source with gearboxes, motors and controllers. For the safety-related operation of servo axes, we combine our drives with spring-applied safety brakes and short release times, as our article on servo gear motors with safety brake shows. Talk to us before the first brake is ordered: the proof decides safety, not the number of brakes.
Frequently asked questions
What is a redundant brake system?
A redundant brake system maintains its safety function even when one braking path fails (fail-operational principle). This is achieved through two functional channels with mutual monitoring, as required from category 3 onwards according to ISO 13849-1. A prerequisite is proof against common cause failure with at least 65 out of 100 points according to Annex F.
From which category according to ISO 13849-1 is a brake control considered redundant?
From category 3 onwards. There, two functional channels are provided, and a single fault in a component of the first or second channel must not lead to loss of the safety function (single-fault tolerance). Category 4 reinforces this further: even the accumulation of two undetected faults must not defeat the safety function.
How is the required performance level PLr determined?
Via the risk graph of ISO 13849-1 with its three parameters: severity of injury (S1 slight reversible injuries to S2 serious irreversible injuries or death), frequency and duration of exposure (F1/F2) and possibility of avoiding the hazard (P1/P2). Alternatively, the PLr can be taken directly from a machine-specific type C standard.
What do MTTFd and diagnostic coverage DC mean?
MTTFd is the mean time to dangerous failure per channel and is divided into three classes: low 3 to less than 10 years, medium 10 to less than 30 years, high 30 to 100 years. Diagnostic coverage DC describes, as a percentage, what proportion of dangerous component faults is detected; category 4 requires 99 percent by calculation.
When is SIL according to IEC 61508 used instead of the performance level?
In the process industry and for safety-related controls with a low demand rate, i.e. less often than once per year, reference is made to the DIN EN 61508 series of standards. Both systems are approximately comparable: PL c corresponds to SIL 1, PL d corresponds to SIL 2, PL e corresponds to SIL 3.
Why is homogeneous redundancy with two identical brakes problematic?
Two identical brakes share identical causes of failure: common power supply, the same friction lining, the same mounting environment. Such a common cause failure can disable both channels at the same time and completely negate the redundancy. Diverse redundancy with different operating principles, for example a spring-applied brake plus a hydraulic clamping unit, reduces this risk.
Which mistakes are common when implementing redundant brake systems?
The most common mistake: installing two brakes and automatically claiming redundancy without demonstrating separate control paths and CCF measures. In addition there are imprecisely defined safety functions and missing proof with a tool such as SISTEMA. The achieved performance level must at least correspond to the required PLr; otherwise higher MTTFd values, better DC or a higher category are necessary.