High-Bandwidth Sensing for Wide-bandgap Power Conversion
Lead Research Organisation:
University of Bristol
Department Name: Electrical and Electronic Engineering
Abstract
This project develops new sensing technology for use in power electronic systems, helping the UK to better compete with global leaders in power electronics. Power electronics is a key electrification technology: it is needed in electric vehicles, renewable energy generation, our electricity grid, and anywhere where the flow of power needs to be accurately dosed. This dosing is carried out by rapidly switching currents on and off to create the desired average. This technology reduces our carbon footprint and contributes nearly £50bn per year to the UK economy and supports 82,000 skilled jobs in over 400 UK-based companies (2016 data).
The power electronics industry is undergoing significant change, as ultra-fast transistors made from silicon carbide (SiC) or gallium nitride (GaN) have recently emerged, to replace silicon transistors.
These new transistors switch 10x faster, which results in 75% less energy being lost in power converters, and enables converters to be shrunk to less than half their previous size. This makes it much easier to build them into other systems, e.g. electric vehicles, resulting in lighter cars with more space for batteries.
This project is about helping to maximise the potential of the new transistors. Many companies are struggling to adopt them, because whilst the very fast switching provides the benefits of improved efficiency and radically smaller system size, it also creates problems with electromagnetic interference, and device and system reliability. The transistors switch current on or off so fast (in less than ten nanoseconds, the time it takes light to travel 3 meters), that engineers cannot accurately measure how the voltages and currents change during this time, even with their best equipment, which means it is difficult to fix problems such as interference. Because of this, even the leading companies are slowing down these new transistors, and losing some of their efficiency potential.
Our project develops small, low-cost sensors, that make these nanosecond-scale changes visible. They will allow engineers to see exactly how the transistors are switching, helping them develop better, smaller, lighter, and more reliable power electronics. They will allow computer-controlled SiC and GaN power converters to sense when they are creating too much electromagnetic noise, and reduce this by switching more intelligently. It will allow power circuits to detect external short circuits and isolate these before they damage the power converter. We are also developing sensors that provide engineers, or control systems, directly with information that they need (e.g. device temperature), rather than having to infer this indirectly from volts and amps, making the measurements faster and more efficient.
The sensors work by detecting electric or magnetic fields via coils, conductive plates, or antennas. The received signal is fed into a chip inside the sensor that computes the required parameter. These new SiC and GaN transistors have made small field sensors on circuit boards viable for the first time, because as signal speeds increase, the wavelengths of these signals become shorter (cm-scale), meaning that their fields can be picked up with millimetre-size coils or antennas.
In order to ensure that we develop what industry needs, we are working with 12 partners across automotive, renewable energy, semiconductors, commercial R&D organisations with deep sector experience, and we are accepting new collaborators on request. Our project provides partners and other UK companies and universities with sample sensors. Their feedback, and discussions with partners helps us prioritise our research, and ensures that we are using our research funds to solve the most important problems. We are providing workshops to help keep engineers up-to-date with advanced measurement techniques, and keeping our results online (publications and a dedicated website) for companies to use as desired.
The power electronics industry is undergoing significant change, as ultra-fast transistors made from silicon carbide (SiC) or gallium nitride (GaN) have recently emerged, to replace silicon transistors.
These new transistors switch 10x faster, which results in 75% less energy being lost in power converters, and enables converters to be shrunk to less than half their previous size. This makes it much easier to build them into other systems, e.g. electric vehicles, resulting in lighter cars with more space for batteries.
This project is about helping to maximise the potential of the new transistors. Many companies are struggling to adopt them, because whilst the very fast switching provides the benefits of improved efficiency and radically smaller system size, it also creates problems with electromagnetic interference, and device and system reliability. The transistors switch current on or off so fast (in less than ten nanoseconds, the time it takes light to travel 3 meters), that engineers cannot accurately measure how the voltages and currents change during this time, even with their best equipment, which means it is difficult to fix problems such as interference. Because of this, even the leading companies are slowing down these new transistors, and losing some of their efficiency potential.
Our project develops small, low-cost sensors, that make these nanosecond-scale changes visible. They will allow engineers to see exactly how the transistors are switching, helping them develop better, smaller, lighter, and more reliable power electronics. They will allow computer-controlled SiC and GaN power converters to sense when they are creating too much electromagnetic noise, and reduce this by switching more intelligently. It will allow power circuits to detect external short circuits and isolate these before they damage the power converter. We are also developing sensors that provide engineers, or control systems, directly with information that they need (e.g. device temperature), rather than having to infer this indirectly from volts and amps, making the measurements faster and more efficient.
The sensors work by detecting electric or magnetic fields via coils, conductive plates, or antennas. The received signal is fed into a chip inside the sensor that computes the required parameter. These new SiC and GaN transistors have made small field sensors on circuit boards viable for the first time, because as signal speeds increase, the wavelengths of these signals become shorter (cm-scale), meaning that their fields can be picked up with millimetre-size coils or antennas.
In order to ensure that we develop what industry needs, we are working with 12 partners across automotive, renewable energy, semiconductors, commercial R&D organisations with deep sector experience, and we are accepting new collaborators on request. Our project provides partners and other UK companies and universities with sample sensors. Their feedback, and discussions with partners helps us prioritise our research, and ensures that we are using our research funds to solve the most important problems. We are providing workshops to help keep engineers up-to-date with advanced measurement techniques, and keeping our results online (publications and a dedicated website) for companies to use as desired.
Organisations
- University of Bristol (Lead Research Organisation)
- Curtis Instruments UK (Project Partner)
- ZF Friedrichshafen AG (Project Partner)
- Dynex Semiconductor (CRRC Times UK) (Project Partner)
- Spark Product Innovation Ltd (Project Partner)
- Toshiba Europe Limited (Project Partner)
- Siemens Gamesa Renewable Energy (Project Partner)
- Austrian Institute of Technology (Project Partner)
- Power Electronic Measurements Ltd (Project Partner)
- GaN Systems Inc (Global) (Project Partner)
- Alter Technology TUV Nord (Project Partner)
- Nexperia UK Ltd (Project Partner)
Publications
| Description | Key findings - updated 12/3/25 See the impact section for more detail. Main findings: 1. First ever sensor that measures power device gate current non-invasively with a bandwidth of over 1 GHz. (Published) 2. First ever gate driver chip for SiC power devices that allows an arbitrary gate current to be programmed, and the means of finding the correct gate current profile. (Published) 3. First ever reported evidence of GaN devices showing signs of aging, weeks or months before they fail. (Published) 4. The first demonstration that we are aware of, of 1.7kV SiC power devices being switched faster than 100 V/ns. In this field, switching speed equals power efficiency. We also show how to measure voltage and current signals that change at these high rates, using custom current sensors. (Published) 5. First ever current sensor embedded into a kA SiC power module terminal with a bandwidth of over 200 MHz (competition is 50 MHz), and a demonstration of it working in a MW system (paper accepted with minor corrections). 6. First ever method for designing non-invasive (magnetic) current sensing in an arbitrary magnetic environment, demonstrated on replica power modules (final edits). 7. First ever demonstration of a gate current profile search method (paper in draft, 1 result outstanding). 8. First ever measurement of switching performance parameters of an electric vehicle drive over the whole automotive operating range: -40C to +150C, 600 to 900V, -450 to +450 A (final edits). 9. World's fastest (2-3x faster than best competition) accelerated artificial aging of power modules with more representative failure modes, enabled by magnetic current sensing (faster and more accurate) (final edits). |
| Exploitation Route | Please see the impact section for a list of examples. |
| Sectors | Aerospace Defence and Marine Electronics Energy Manufacturing including Industrial Biotechology Transport |
| Description | Impact - updated 12/3/25 CURRENT SENSORS FOR RESTRICTED SPACES We have shown how to measure current with a very high fidelity in tight spaces where no commercial sensors fit. We have developed a sensor that is only a few mm wide and 10 mm long, which fits into a notch in a 500A terminal on a SiC power module. These modules are generally bolted strait onto copper electricity conductors, since any unnecessary distance between the module and the conductor increases power loss in the system due to the increased inductance and resulting slower switching. Our sensor is so small that it fits inside the terminal. The competition requires a cable to be looped around the terminal, for which the distance between the terminal and copper conductor would need to be increased. The University of Edinburgh and a UK-based company required a sensor like this, so we designed it for them, and they helped us test the sensor on their 900 A system. The main scientific insight gained here is a process that allows contactless magnetic sensors to be developed for circuits with complex magnetic fields. The downside of this method is that for each type of circuit, a different design may be required. This work has been submitted for publication to IEEE Transactions on Power Electronics, and it is currently pending a 2nd review of our minor edits, so it is likely that this information will be available to the public soon. EXPLORING WAYS OF HELPING COMPANIES DESIGN MORE EFFICIENT POWER MODULES Full electrical characterisation of a prototype module containing latest generation SiC chips. We have developed a way of putting a SiC power electronic semiconductor module into the all different electrical and thermal environments that it would experience in the life of an electric vehicle. This includes running the circuit from -40C to 150C, with the measurement probes thermally insulated to avoid them failing. We have taken 1000s of measurements (millions of datapoints) and written software to help electric vehicle designers understand the data, and enter performance requirements and see visually, when the module is being overloaded. We helped a company check their newly designed prototype power electronic module for electromagnetic interference problems from other power circuitry in the same drive, and we are further developing these methods for wider use. Electromagnetic field study helps to see if a prototype module is susceptive to interference from currents in the system. Interference could potentially turn a device on when it should be off, which would lead to failure of the whole automotive drive and possibly other sub-systems. GATE CURRENT SENSORS FOR ACTIVE GATE DRIVING RESEARCH We have developed the first ever contactless magnetic sensor that measures power device gate current non-invasively with a bandwidth of over 1 GHz. Gate current is difficult to measure via its magnetic fields, since gate current is the small control current that switches on a much larger current in a power transistor. This larger current (100s of amps) creates much larger magnetic fields that would drown out the fields from the gate signal if the sensor is not carefully designed. Over the past 2-3 projects, we have developed the first ever gate driver chip for SiC power devices that allows an arbitrary gate current to be programmed. This work is continuing in this project as it is a very important application for high bandwidth field-based sensing. The sensors in this system provide the means of finding and setting the correct gate current profile, to improve the efficiency of a SiC power converter. CURRENT SENSORS FOR OTHER APPLICATIONS We have also been using prototype current sensors to find the potentially first ever reported evidence of GaN devices showing signs of aging weeks or months before they fail. The first demonstration that we are aware of, of 1.7kV SiC power devices being switched faster than 100 V/ns. In this field, switching speed equals power efficiency. We also show how to measure voltage and current signals that change at these high rates, using custom high-bandwidth gate and source current sensors. CURRENT SENSORS FOR THE DESIGN OF POWER MODULES This is the next aim of the project, to design sensors that help companies design better power modules that are more efficient and rugged. CURRENT SENSORS FOR INSIDE MODULES After this, we will work on sensors that are small enough to embed into modules to help find the limits of a module in an application such as electric vehicles. |
| First Year Of Impact | 2024 |
| Sector | Aerospace, Defence and Marine,Electronics,Energy,Manufacturing, including Industrial Biotechology,Transport |
| Description | Infinity Sensor samples to international research organsations |
| Form Of Engagement Activity | Engagement focused website, blog or social media channel |
| Part Of Official Scheme? | No |
| Geographic Reach | International |
| Primary Audience | Professional Practitioners |
| Results and Impact | The Bristol Energy Management Research Group runs a website called InfinitySensor.com where research organisations apply for samples of our prototype current sensors. We support clients as the integrate these sensors into their experiments, and get feedback in order to be able to improve the performance and quality of the sensors we create. This also promotes interesting discussions with other institutions and industry about instrumentation for power electronics, and helps steer our work towards what industry needs. This route also brings in requests for custom sensors, which we evaluate. If there are more than, say, 4 or 5 companies requesting a particular type of sensor, then we build this into our research roadmap. This ensures that we are putting public money to use in areas where we have evidence that it will help companies develop better products and organisations carry out better research and development. |
| Year(s) Of Engagement Activity | 2019,2020,2021,2022,2023,2024 |
| Description | Summer Internships for female students |
| Form Of Engagement Activity | Participation in an activity, workshop or similar |
| Part Of Official Scheme? | No |
| Geographic Reach | Local |
| Primary Audience | Undergraduate students |
| Results and Impact | Electrical and electronic engineering is male-dominated in the UK. Also, non-UK students can sometimes struggle to find summer placements in companies. Therefore, in the summer of 2023, we provided 3 summer placements for undergraduate students, 2 funded through this project, both female. They worked as a team to develop a new sensor test platform, and then canvassed our industry and academic users of our sensor samples, to assess the research and commercial opportunities in this area. They carried out experiments and wrote technical reports and a business plan. All 3 students felt they had learned a lot about working in research, communicating with clients, and presenting progress in difficult technical subjects. They were supported by this project's Principal Investigator, and by PhD students who appreciated the opportunity to help students climb the steep learning curve that comes with carrying out research and development. |
| Year(s) Of Engagement Activity | 2023 |
