Performance monitoring solution helps provide intelligent control of high power systems

We are also providing a forensic analysis capability so if something critical really happens then there is data available about what happened to the voltage and what happened to the temperature which helps to find a reason or cause for the fault. This enables you in the long run to avoid the problem happening again.

Everything is field upgradable so that helps the customer improve their system availability and reduce energy costs through high efficiency. It also helps reduce maintenance costs.

Richard Ord: We see the pressure on efficiency as being vital for this industry and one of the key elements of efficiency is not necessarily the fine tuning of the turbine blade on the wind turbine it is whether the turbine blade is turning or not.

Often you see a wind farm with a dozen turbines and at least three or four are not turning. Usually about a 33 to 50 percent of the time that failure is due the power electronics. They might not have failed but they have tripped. Somebody needs to go and examine why and reset it. That means going to the wind farm or if it is offshore getting in a boat to find out what happened.

In traditional IGBT gate drivers there are typically only three states – on, off or fault. What we are doing is building that dashboard of high power electronics. We can pose all manner of questions such as:

How many times has this IGBT been used?

How many switching cycles has it made?

Is it starting to suffer?

Is the temperature starting to rise?

What are the conditions that are changing around it that may mean this specific IGBT needs some planned maintenance?

Can you switch the load to some of the other IGBTs?

We are getting that ‘under the bonnet’ visibility right in the core of the power electronics that has been the challenge. Historically if you wanted to measure the temperature of a module you would need a thermistor and a wire which attracts all the interference you can imagine for the high voltage side.

In effect from this side you really don’t know whether your patient is alive or dead. What we have done is put sensors right at the core of the system, digitized it and then we control what signals are passed out using an ARM microcontroller over the fiber optic link. By doing that we are getting more reliable signals with less noise or interference and getting some real information back about how the switch is actually going. In this way people can learn how their systems are operating, modify the mission profile to make it more available and more efficient. Those are the areas Amantys is working on and why we have come into the market.

We are handling the collision of the analog domain and digital knowledge. The techniques we are using have been used in the digital world for 20 years. It is second nature to us.

eeNews Europe: What application areas are you targeting?

Erwin Wolf: The new Power Insight Adapter is a fully engineered solution to add Insight data monitoring and observability to an IGBT module-based power assembly.  We have developed the Adapter to offer the capacity to support existing power systems in locomotive traction, wind turbine, High Voltage DC and industrial drive applications with a performance monitoring solution.

eeNews Europe:  How does the new Power Insight Adapter solution fit into your plans?

Richard Ord: With the Adapter we have created a product that allows people to start straight away using the Amantys Power Insight software platform. The Adapter basically sits across an existing fiber optic link from the gate drive to the central control host system. The Adapter sits on the fiber optic and signals pass through the Adapter so there is no change to the gate drive operation that is still controlled by the host system but it sits there on that fibre optic link to observe the different Power Insight parameters such as the number of switching cycles, different voltage measurements, temperature changes and reports that out independently to a laptop or industrial PC that can compute all the information. We already store some data on the gate drive which can be a rolling snapshot in time of events and signals or it could be a longer sampled set of indicators.

There is some storage in the control card and there is further storage in the Adaptor with a SD card. So when you get a failure event you can look back at some recent history and examine what happened in the run up to the fault. In this way you can start to learn how the system is operating and build a profile of recent events before a failure.

eeNews Europe:  Is there a development kit for the Adaptor?

Richard Ord: We are offering a development kit which uses a GUI interface that allows engineers  to talk to four different gate drives through the fiber optic links. With the GUI you can click on an individual gate drive. You can set your PWM frequencies, you can set the duty cycle and the offset for individual gate drives and you can then watch what is happening. So you will be able to, say, look at the gate drive temperature on an individual gate drive. You will be able to see when the temperature begins to spike.

You can also observe an individual gate drive when you want to look at it in more detail and you will be able to see if there has been a short circuit event or under voltage. The unit will tell you the time stamp of the last event. You will be able to see the temperature start to rise if there is a problem. You can also use the GUI to set thresholds. So you can instruct it to tell you when it goes below -30°C or above +55°C.

You can actually configure the IGBT module and the mission profile you drive it with from outside the cold plate assembly.

The design in cycle in this industry is relatively long. We started developing the Adaptor in November 2012 and created prototypes by the beginning of May 2013 and are launching the product in November 2013. We have a number of customers who have already bought the Adaptor to start developing and commissioning their systems. We are now talking to those customers to discover what the eventual format and footprint they need to integrate the Adaptor in their host systems.

Customers are already coming back with questions such as ‘Can you add humidity?’ to the parameters that can be measured.

The answer is yes if you have a sensor we can measure any parameter and export the data over the fibre optic link to give a resilient communication of a measurement that will help you understand what is happening in the module.

It might be that a particular industrial motor is used on a ship in the Tropics for a winch where it has a relatively constant load when it is switched on and off but that same industrial drive design might then be used in a steel mill where the load has a totally different load profile. Using the Adaptor you can actually ‘tune’ the switching of the individual motor according to its application or according to its environment and that ‘tuning’ can be done ‘in situ’ at the installation.

We are talking about condition monitoring in an area that has not traditionally had it. It is the dashboard of heavy power electronics.

Erwin Wolf: I think the Adaptor is really going to give a system level advantage. It is not so much of an advantage for the module makers but it offers a major advantage for the system makers or somebody who wants to know how their offshore turbines are doing. Or how their trains are doing or a HVDC is doing. You can get the feedback and be able to act before it is too late. The Adaptor enables you to do preventative maintenance or switch an IGBT off before it blows up and the whole system trips. This is also only the first version of the Power Insight which was developed. We have a lot of ideas as to how to refine it.

Richard Ord: Data is easy but information is difficult and knowledge is really hard. Data in power electronics in this sector has not been simple but I think we have made it simple with this technology. As people get more data they will start to receive more information about what is happening they will start to improve their understanding. And then eventually there will be knowledge and that’s prognosis – ‘this’ has happened so ‘that’ may be about to happen. If a wind turbine is about to fail then it will put some sort of signal into the electronics. By monitoring the electronics signals we can find the status of the wind turbine.

I think there are three phases of education in this market. We can give you data to help you start to understand what is happening. From that understanding how can you take that forward? Well eventually that understanding enables you can put that knowledge back into the design of the next generation product. In the meantime, perhaps you can better anticipate what is about to happen in the field.

eeNews Europe: Where are you planning to manufacture your products?

Richard Ord: All of our products are made in the UK. The Power Insight Adapter is made locally in the Cambridge area but the gate drives are sub-contracted out to Jabil in Scotland who have a lot of experience in hi-rel and automotive types of products.

We are on the cusp of series production with a number of significant OEMs. We have of had quite a lot of development programs both with OEMs and IGBT module manufacturers where they have a bespoke architecture and they need a gate drive.

eeNews Europe: Who do you see as your main competitors?

Erwin Wolf: Our own customers’ internal groups are our chief competition because the likes of ABB, Alstom, Infineon, Fuji all have internal activities that drive their modules and systems. On the other hand these are also the people who derive the most benefit out of the new features of the Amantys Power Insight solution because with their own drives they cannot do real-time condition monitoring.

Richard Ord: Each drive that our customers develop tends to be bespoke, fairly expensive and has really limited design re-use or economies of scale. For Amantys these companies offer an opportunity because when they see the capabilities of our condition monitoring solution they start to say things like: "Can’t you monitor this or that parameter?" Our Power Insight is like a communications platform so if you have sensing at one end you can interpret at the other end.  It is simply then a question of configuring the software to add new parameters and new instrumentation.

On the one hand we might appear to be a threat to these customers but we are also helping them liberate their scarce power engineering resources to look more at the system level issues because we give them the platform to look at what is happening at the core of the system.

Essentially we need to be collaborating with all of the IGBT module manufacturers.

Erwin Wolf: We have to differentiate two areas that we work in. We have some standard IGBT drives which do not yet incorporate the Power Insight capabilities and these are pretty straightforward and we do not need to collaborate too much with these but then we have the real partnership situations where Power Insight comes into its own.

The rule is that for each application we will only work with one partner. We cannot cross-contaminate working on one application with two or three partners because it is not possible for us to avoid any contamination. Our partners would not accept that.

eeNews Europe: Is it hard to attract the right calibre of staff at the moment to handle so many projects with a small engineering team?

Richard Ord: Yes it is hard to find them.

Three years ago our youngest engineer was fresh graduate out of Cambridge University and his tutor at the university was one of our founders. In the last few weeks we have recruited a number of other graduates or recently graduated engineers from a range of nationalities. We find that it is hard to locate the right caliber of people but once we find them our conversion rate is very good. What they seem to find is that in a team of 30 they have more scope with Amantys to be involved in different aspects of engineering.

Our new recruits are finding that rather than going into a bigger company where they would have quite a narrow experience they are exposed straight away to the analog engineers, the software engineers and the mechanical engineering team. And they also hear the big bang when the high energy test facility ‘blows’ up a module. So they are getting a very broad and intense exposure. Cambridge is known to be very strong for digital electronics and wireless electronics but high power is fairly unusual. Having said that if you are in electroncis and you are in the UK then Cambridge is a very good place to build your long term career.

Erwin Wolf: I have been involved in the power business since 1979 and power devices were always the unsung heroes. They always made fantastic gross margins and fantastic profits but all the money was given to the digital guys or the analog guys or the wireless guys who didn’t make money but had a large market. Slowly the power semiconductor and the power device sector are coming out of that shadow and are going from being an unsung hero to becoming a real hero because the market is increasing dramatically with energy saving requirements growing and the drive towards an improved efficiency philosophy is expanding.

The power semiconductor and power devices in general are playing a vital role in this trend. If I look at a company such as Infineon they have a lot of experienced people but they also have a bunch of very, very young engineers attracted to that business.

Power is losing its reputation as a slow moving sector which no-one wants to enter. It is a much more attractive area now.

Smart Grid is waking up people. Everyone is talking about Smart Grid and this is attracting new people. In our team I think at Amantys we have a lot of very experienced people with up to 80 percent being really experienced but that is why we could afford to take in some graduates to train up by the more experienced staff members. Now we have a good mixture.

eeNews Europe: Is it a difficult climate to raise finance at the moment?

Erwin Wolf:  "I think general investment climate is tougher now because of the successes of the software Internet groups such as Google, Microsoft and more recently Twitter. This trend has really shifted a lot of the money often given by the venture capitalists towards our kind of business and there are not so many investors left who invest in companies that are doing ‘real stuff’.

That’s the reason why for start-up electronics companies it is relatively difficult to raise money. For us we are free of needing investment for capital equipment purposes. Although we need some capital for our Test Lab that is not a big investment. We have a sub-contractor that is manufacturing for us so we are basically fabless. I don’t think it will be that different between now and three years ago. However, speaking generally for companies that do things like us and who have a concept of manufacturing everything for themselves, it would be very difficult to raise the necessary sums of investment.

If you don’t have the ‘green’ and ‘cleantech’ angle and you are too deep into the technology and need to invest in your own capital equipment it seems very difficult to get money from investors. If you are fabless and you invest in people and software it is easier to raise money. It is not easy but it is easier.

I think we are in a stage where we are very well funded but I think we will need to go for another investment round. We don’t need it now but maybe in a year we will need to go on the road again to raise more funding. 

Richard Ord: "If we weren’t impacting on the efficiency of systems in all sorts of different application areas it would have been extremely difficult to raise the funding. The broad base of the application areas for the Amantys solution is key and it is ‘green’. Our solution has a ‘Cleantech’ badge over the top of it and if we didn’t have that angle and the fact that it can be applied across multiple different industries which each are going through a change dyanmic at the moment then it would be difficult.

Our early funding was critical because it was what gave us escape velocity to go from the shed to an office. We did the classic Cambridge journey of going from a garden shed and pub to a unit for six people in a small Cambridge innovation center. In about a year with the additional financing from Fidelity and ARM we took on another unit at the innovation center. Then nine to ten months on from there we took on our first major office.

If as the primary founding investor you have a good business proposition and a good defensible technology proposition then I think you can raise finance the same now as you could three or five years ago. Ten years ago was a different investment scenario. We have had the Dotcom bubble since then. I think the Silicon industry is a very tough one in Europe to get new investment. It is tougher now in California than it was ten years ago.

eeNews Europe: Has the company’s Cambridge University pedigree been key to raising finance?

Richard Ord: There is an eco-system, an infrastructure and a mentality in Cambridge that ultimately does come from the university and what it has done with the tech business sector over the last 30 to 40 years. To be in Cambridge and also to have links with the university are two big ticks for potential investors. If you are just in Cambridge without links with the university raising finance is not such a problem but if you are based outside Cambridge it can prove very hard to attract funds.

Raising investment is a bit like education. With education you never stop learning. Similarly a start-up never stops needing to raise finance.

Cambridge is the No. 1 hotbed for high technology and there probably only another two or three hotbeds in the UK.

eeNews Europe: What strategy are you planning to use to make sales worldwide?

Erwin Wolf: Distribution is a very important channel especially when you are talking about standard devices like gate drives but it gets trickier when you are talking about product design-in technologies. I think you have to watch it very carefully as a small company you do not have any other real option than to go the distribution route. I think that is not a question of choice but where necessary we will add resources to help to make it easier for the distributors to sell our solutions.

If the business opportunities really support it I think we will aim to have our own sales people in Asia.

Richard Ord: What we found with some of the chip manufacturers and the IGBT module companies that gaining their explicit or implicit approval was important for our credibility so we went and talked to them to a great degree.

We were surprised to find that they would say they had a new architecture coming and they needed a gate drive for it so could we do a dedicated gate drive for them. We quickly found that rather than just being potential partners there were also potential engineering projects to be had.

These kind of companies see that the more scalable the products are the bigger the market can grow. They already have silicon in their genes they are used to getting volume and then you scale up.  That process may perhaps not happen as quickly as some other markets but that is also one of the advantages. For our kind of products if you get a design win you have it for five, ten or fifteen years. There is none of this 18 months design cycle associated with many semiconductor products.

eeNews Europe: Will you need a test lab in Asia?
 

Erwin Wolf: I think with the test lab you have to prove a certain kind of reliability so once you have that like we do now then it is not so important that you have a test lab in China or somehwere else but you do have to provide the data and eventually when it gets serious show the customer that you really do have a test lab where the data has been created and show that it is satisfying the requirement.

Richard Ord: We may continue to scale up the test lab or use external resources if necessary. For the moment our test lab in Cambridge is really valuable to us.

Erwin Wolf: The Cambridge test lab is very valuable because we receive development feedback directly. We also can use our Power Insight solution to monitor during the reliability test so we have feedback right on our own doorstep. This kind of forensic data is so valuable because it enables us to identify any problems that can blow up an IGBT module. We can find out want went wrong and improve our systems. If you have the experience and you have the data once you talk to your customer you can bring your experience to bear.

eeNews Europe:  What does the immediate future hold for Amantys?

Erwin Wolf: We started in Europe so the majority of our design-in projects in Europe are most advanced in the UK, France, Germany and Portugal.

We have development activities with about ten to 15 OEMs – some of those are standard products others are featuring Power Insight.

With some of them we will be very shortly going into serial production and a continuous revenue generating relationship which for a start-up is the first validation of the technology. We see a wide range of potential applications for our technology but we need to maintain our focus on immediate objectives and not go off-the road and become distracted.

The major aims in the short term are to prepare ourselves for serial production to be ready from the quality management point of view and the supply chain management point of view. We need to get ISO 9001 certified – essentially all the unspectacular groundwork needs to be done to be ready for serial manufacturing and generating a continuous revenue stream.

Then we can go on to enlarge the application areas in the medium term.

In the short term the first stage applications are wind power and renewables, locomotive and transportation, industrial motors, power grid and HVDC.

The second stage will look at lower voltages for electric vehicles and hybrid vehicles (HEV) and the UPS.

Richard Ord: Our relationship with Avago is definitely going to help us open doors in the automotive sector and help us to accelerate with some engagements in that area.

eeNews Europe: How do you think different material base technologies such as silicon carbide will have an impact on the market?

Richard Ord: From our perspective the volume markets today are IGBT semiconductors and that is where we are focusing but the techniques are applicable to any base technology such as Silicon Carbide (SiC) or Gallium Nitride (GaN). For example one of our earliest demonstrators that we used to attract our initial financing could use a SiC transistor in there as well as an IGBT. As and when SiC and GaN become viable product technologies we will get involved with them.

Erwin Wolf: The key challenge for silicon carbide (SiC) is to make a FET or IGBT from that material. For diodes I think it is already established for the higher voltages but I personally believe that GaN will take over the market in the long term for high switching transistors up to 1200 V.  GaN promises high frequency switching with such a low RS(on) that the conversion losses are reduced dramatically which is why I think GaN is going to be a very superior technology compared with silicon.

GaN will take some time to enter the market because it is not an easy material to make especially if you want to be cost competitive with silicon but many companies are working on that. Once GaN is available it will make a very fast switching transistor on a very low cost material. SiC is a very expensive material. It has all the great performance characteristics but its performance/cost relationship is not as good as GaN. For us whether the fast switching IGBT is made from SiC or GaN we believe we can apply our technology to all the material options.

Visit Amantys at www.amantys.com

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

---

Share:

---

Linked Articles

From PACE to PQC: The future of secure electronic passports – a conversation with Nouri Alnahawi

November 5, 2025

From System Design to Source Code: Tobias Kästner on Zephyr

October 28, 2025

The HaLoW WiFI booster: Michael De Nil, Morse Micro

October 20, 2025

Hidden Gems and Grand Visions for Zephyr: In Conversation with Benjamin Cabé

October 15, 2025

Security, Scalability, and Future Trends: Pierre Lecomte on Zephyr RTOS

October 6, 2025