Smart fuse reduces cost, weight of automotive wiring harness

Smart fuse reduces cost, weight of automotive wiring harness

Technology News |
By Christoph Hammerschmidt

From a system point of view, however, the implementation of a more intelligent, electronic fuse offers the potential to cut total cost, as well as to reduce the weight of the vehicle substantially. Now ams has introduced a reference design board which shows the industry a design concept for an accurate yet simple intelligent fuse, and which gives OEMs the opportunity to evaluate the concept, understand its advantages and cost, and simulate its operation in complex power systems.

The traditional fuse: slow, inconvenient, cumbersome

Thermal fuses are slow and inconvenient. The typical fuse in a car is blown after around 20-50ms when subject to 10x its nominal current. And because a fuse cuts out when triggered, it has to be replaced after every over-current event. This is why cars include a special housing at a central point, to provide the easiest possible access when a fuse needs to be replaced.

In fact, when today’s cars have so much modern technology inside them, such as touchscreens, voice recognition and sophisticated driver-assistance systems, the antiquated tool and obscure process for replacing a traditional fuse seem out of place.

A modern electronic fuse does away with this ancient technology. Instead, the dashboard can display diagnostic information on the potential location of the fault that caused the fuse to be triggered, and guidance on the way to repair it.

In addition, an electronic fuse is as much as four orders of magnitude faster, dramatically shortening the exposure of the circuit it protects to damaging current surges. It can also provide a much more precise maximum-current trigger point. This gives the potential to reduce cable diameter when compared to thermal fuse protection, which requires the designer to allow for a broadly defined maximum current range, rather than a specific – and lower – current value.

Another great advantage of the electronic fuse is apparent in the configuration of the power network. Electronic fuses can be placed virtually anywhere in the car; unlike thermal fuses, they do not need to be gathered in one fuse box. One benefit of this is shorter cable runs, and a consequent reduction in cost and weight.

It also enables the power system designer to implement for the first time tree topologies, which are easy to manage, and even ring topologies for advanced safety features. Tree topologies support the use of very much thinner cables, because they allow the power controller to switch off certain portions of a circuit for a certain period to keep the load on the entire circuit below a pre-determined maximum value.

Quantifying the potential for weight reduction

It is clear, then, that replacing thermal fuses with electronic fuses offers the potential to reduce the size and weight of the cable harness. But by how much?

A couple of decades ago, cables in cars mainly carried loads rather than signals. In the case of an indicator, for instance, the stalk at the steering wheel closed a contact which directly connected the indicator’s relays to the battery.

In the 1990s, car manufacturers began to introduce power networks. The goal was to keep load-carrying cables as short as possible, and to use thinner signal cables whenever possible. So today’s indicator stalk does not close a load contact: instead, a sensor detects that the indicator has been switched on, triggering a pair of electronic control units (ECUs) to send this information to the main front and rear body units. It is these body units which switch power on and off to the indicator lights.

Had no other developments occurred in the car, this would have reduced the total weight of the cable harness. In practice, however, cars today include so many more electrical and electronic functions than cars did in the 1990s that the weight saving has been more than eaten up by the requirement for additional network connections. In fact, a modern mid-range car carries more than 1.5km of cable weighing more than 40kg.

And of course, these cables require protection. This often leads the system designer to struggle with the best way to trade off improvements in safety, comfort and functionality against the desire to reduce weight and cost.

Take the example of an electric sun-roof. The design specification requires that the roof will open even when its seal is frozen. Breaking the grip of ice calls for a high current through the sun-roof’s electric motor, as much as 30A.

But the specification for the circuit’s thermal fuse cannot be for 30A, since ageing reduces the fuse’s current rating*: a margin of 20% has to be added. But 36A is not a standard rating for off-the-shelf fuses, so the sun-roof designer is forced to specify a 40A fuse.

This in turn affects the cable specification. For the motor’s 30A maximum current, a 2.5mm² cable would be sufficient. The power network design must also be specified to withstand a maximum 70°C operating temperature. Assuming it takes 50ms to melt the fuse at 400A – ten times the nominal 40A rating – the temperature of the cable might rise far above the maximum temperature rating of 105°C for a standard cable in a car in the event of a current surge.

As a result, the sun-roof designer is forced to specify the next thicker grade of cable: in this case, 4mm², which is 40% heavier, and 40% more expensive.

By contrast, an intelligent, application-oriented electronic fuse, which does not age, could be designed to be triggered at an exact current value of 30A (or higher), enabling the use of 2.5mm2 cable.

The authors’ estimate is that some 5-8kg of copper can be removed from the car by replacing thermal fuses with electronic fuses. At the time of writing, the cost of copper was around €6.50/kg, providing an estimated cost saving in copper alone of up to €52. In addition, the weight saving improves the vehicle’s fuel efficiency, helping the car manufacturer to avoid the proposed €95/g of CO2 levy to be imposed by the European Union when fleet-wide fuel consumption rises above a certain threshold.

But an intelligent fuse can provide even more benefits:

  • it can measure the temperature locally
  • it can enable intelligent power switching

This latter feature lets the system designer reduce even further the weight of copper in multi-function supply cables. Take the example of a front door. This has multiple electrical functions, including:

  • mirror positioning
  • mirror opening and closing
  • mirror heating
  • indicator lighting
  • electric window motor

All of these functions could operate simultaneously; in this case, the door’s power circuit would need to support a high peak load current, and this would call for a cable with a large diameter. The alternative is – at the cost of slightly compromising the user’s convenience – to disable certain functions by default. For example, the mirror heater may be disabled while the window motor is in operation. It might in fact only be necessary to stop the current to the heater for a few milliseconds, when the motor current spikes as it starts up. So the blocking of certain functions might either be minimal or, if prolonged, the user may be informed in the dashboard display.

In either case, precise and continuous current measurement, which is performed by default by an intelligent fuse, ensures that the current flowing through every single electrical sub-system is known, and so the power consumption and peak current requirement may both be managed in an intelligent and granular way. This will, of course, require the development of new and complex software.

Intelligent automotive fuses: a demonstration circuit

Now ams has developed a demonstration board which contains all the functional blocks required to realise an intelligent fuse (see Figure 1). In evaluating the circuit, automotive manufacturers are expected to assess:

  • whether the accuracy and precision of the current and temperature measurements are at the right level
  • whether the circuit offers the right number of channels

The feedback from manufacturers will influence the final specification of a chip-scale version of this demonstration circuit, now in development at ams.

The new fuse consists of a switching element and a current measurement element.

In the IC version of the circuit, the switch will be realised as an internal gate driver and an external MOSFET. The drivers have to operate at high speed, because they need to frequently switch heavy loads, and, in order to avoid excessive switching losses, the MOSFET should spend as little time as possible in the linear region.

Like a conventional thermal fuse, the switch must be on the high side. This means that the driver needs an additional charge pump in order to raise the voltage far enough above the battery voltage to drive the MOSFET.

Accurate current measurement on the high side

A difficult part of the circuit to realise – and the reason why no semiconductor manufacturer before now has implemented this simple but high-performance design – is the high-side current measurement.

Despite being on the high side, this current measurement circuit must offer high accuracy if the circuit is to provide the benefits – such as weight reduction and intelligent power management – described above. Fortunately, ams has developed a technology which can achieve accurate high-side current measurement. What is more, it measures current directly on the board’s copper traces, avoiding the need for an expensive alloy precision shunt resistor.

On the demonstration board, the current measurement components are placed directly on top of the copper traces so that they can measure their temperature as well as the current flow. This eliminates the need for external sensors, while allowing for compensation for the temperature coefficient of the copper. (The dimensions of the copper trace must be specified with reasonable accuracy.)

With this technique for direct measurement on a trace, currents up to around 50A may be measured, using two layers of a four-layer board. Since the circuit is making high-side measurements, there is also a need for level shifters to provide a voltage that the ADC can handle. In the demonstration board, the circuit achieves current-measurement accuracy of ±2% over the operating temperature range. This can be improved if necessary.

The circuit also provides for direct cut-off, bypassing the digital functional block. This ensures that the fuse can switch within a maximum of 20µs when subject to a large current surge. Implementing the comparator and level shifter for this function is another demanding circuit-design challenge.  

In terms of board layout, several current paths are routed together, and covered on the backside with an additional thermal mass (made of copper). This ensures that they are thermally coupled, and so a single temperature measurement may be applied to the entire circuit, with compensation for thermal resistance implemented in the microcontroller (see Figure 2).

The ADC itself has a chopper architecture which produces a zero offset: this enables the circuit to measure current accurately even at low currents. Level shifters drop the analogue voltage over the shunts to the ADC’s voltage domain. Dechopping is implemented in the digital filter of the Sigma-Delta ADC.

A common objection to chopper architectures is that they generate noise. But in this ams design, chopping with ringing cancellation is performed in the analogue domain, and dechopping in the digital (see Figure 3). As a result, noise is negligible, while the offset in the signal path of the current channel is entirely eliminated.


Temperature measurement

In the ams circuit, measurement of the MOSFET’s temperature is performed by external temperature sensors, since most MOSFETs’ internal temperature measurement is not accurate enough. With an additional comparator, this allows for direct over-temperature shut-down, bypassing the digital circuit in the same way as for the over-current shut-down function.

Digital functional block

Figure 1 shows a microcontroller which performs a number of digital functions:

  • Digital de-chopper (demodulator) for current measurement
  • Resistor temperature compensation for the copper shunts
  • DAC shut-down signal for over-current and over-temperature
  • End-of-line storage of calibration values for current paths
  • Software to calculate exact current values for transfer to an ECU
  • Communication interface
  • Management of safety features
  • Programming of over-temperature and over-current trigger characteristics for up to four channels

When implemented as an IC, it is intended that, wherever possible, digital functions should be implemented as state machines; the device will not include a microcontroller. This approach is simpler and cheaper, it consumes less power, and it better supports the requirement for functional safety.

It should be said that the analysis of the functional safety of this circuit is today in its early stages. The plan for implementing the IC product provides for an ASIL A rating, with provision for a higher safety rating if required by the industry. Since the use of intelligent fuses calls for some external software if the full benefits are to be enjoyed, a full functional safety analysis can only be made in the context of the final application.

One fact, however, should not be forgotten: a melted fuse has no ASIL level at all. In terms of functional safety, then, as well as system cost, fuel efficiency, performance and functionality, the intelligent fuse demonstrated by ams is far superior to the traditional thermal fuse.


The steady growth in the number of electronic, electrical and electro-mechanical functions in cars has given rise to many innovations in the design and operation of automotive power systems. In one domain, however, the car remains stuck in a technological Stone Age: the device of choice for circuit protection is still the fusible cut-out (fuse).

Its use continues in spite of its numerous and serious drawbacks. This is because the traditional fuse has one, very powerful attribute in its favour: its unit cost is very low.

Looked at from a system point of view, however, the implementation of a more intelligent, electronic fuse offers the potential to cut total cost, as well as to reduce the weight of the vehicle substantially. Now ams has introduced a reference design board which shows the industry a design concept for an accurate yet simple intelligent fuse. It gives OEMs the opportunity to evaluate the concept, understand its advantages and cost, and simulate its operation in complex power systems.

This article details the drawbacks of the traditional thermal fuse. It then describes the operation of the new demonstration circuit. It shows how innovative high-side current measurement directly on the board’s copper traces, together with fast analogue over-temperature and over-current shut-down circuits, provide a simple, low-power, low-cost and high-performance alternative to the thermal fuse.  

* The ageing of fuses is rarely discussed in the literature, although it is an inherent characteristic of the device. See J. Shi et. Al, Ageing Assessment Condition, Inspection and Lifetime. Journal of Energy and Power Engineering 5 (2011) 892-898

About the authors

Martin Jaiser gained his diploma in electronic engineering from the University of Karlsruhe. He then held several positions in sales and field applications engineering at Analog Devices, Rambus and Elmos. Since 2012, he has been working as a field applications engineer specialising in automotive applications at ams AG, with a particular focus on position sensing, inductive and capacitive sensing, and NFC.

Manfred Brandl is a senior product manager in the mobility sensors division of ams. He began his semiconductor career in 1984 when he joined the then Austria Microsystems as a product engineer. He was then promoted to a role in foundry engineering before joining the automotive business unit in 2000. Since 2003 he has been a product manager for battery management ASSPs. Brandl holds a bachelor’s degree in electrical engineering and a masters degree in mechanical engineering from the Technical University of Graz (Austria). He has been awarded ten patents for inventions in sensors, sensor interfaces, MEMS technology and packaging. He can be reached under the mail address

Further information

All illustrations (C) ams.

If you enjoyed this article, you will like the following ones: don't miss them by subscribing to :    eeNews on Google News


Linked Articles