Insight & Blogging Archives - Page 3 of 3

1. Who is this “Rohm” again…?

Kyoto based Rohm Semiconductor is the technology leader of SiC power devices. They have been the first to release SiC Schottky diodes, SiC MOSFET in 2010 and Full-SiC power modules in 2012.

They also master the full supply chain since the acquisition of the German SiC wafer manufacturer SiCrystal in 2009. They have access to German made 6-inch wafers since 2013.

According to their roadmap, showed and diffused during conferences in 2014, they are currently working on their 3rd Generation of SiC MOSFET to be released soon.

rohm roadmap on SiC diodes and mosfet releasing generation in 2015 and going up to 1700V

They will use a Trench Gate, together with a Trench Source. This Trench Source will allow reduction of the electric field and thus highly improve performances of SiC MOSFETS.

Roadmaps planned the release of these devices for 2015, and as APEC conference in Charlotte is approaching, the question has been raised.

Rohm kyoto based second generation and third generation of SiC mosfet. planar gate versus trench gate mosfet

Rohm 2nd and 3rd generation of SiC Mosfet

Will APEC be the day when Rohm will make a new step in the Wide band gap power electronics world by releasing its own SiC 1700V switch?

2. A bit of technology:

The 1700V breakdown voltage is very important in Japan. This is the voltage limit where devices can easily be used in Rail traction. You know how rail traction is important in Japan, in terms of sales but also in terms of image. If Rohm is able to gain market shares in auxiliary rail traction converters and light rail in Japan, they will have another advantage over the biggest European semiconductor companies.

They are working on 17A and 50A bare dies.

The 650V and 1200V devices will also adopt the new 3rd Generation U-MOS design, based on trenches.

3. And a tiny bit about competition:

As a reminder, Rohm is already supplying many system makers in Europe, with SiC devices for R&D and advanced design.

Cree is doing a good job too. But we believe that power is not their main target. Cash is coming from LED for them and power electronics is just extra-money.

And you can also note that their communication is getting better as well. They were already doing good products. Now, if they manage to sell them well, they can quickly become a very big competitor for the IGBT dinosaurs!

1. The bounty…

That’s done. Google is officially in power electronics. Well they were already. They did invest into Transphorm (Gallium Nitride based power semiconductors) but that was through the Google Ventures subsidiary. Now that their Little box Challenge is claiming a prize for reduced size inverter, We can’t say anymore it’s not in their targets.

$1M is the reward for that ‘challenge’: Making a small size inverter of 2kW.

Get there for more info: https://www.littleboxchallenge.com

2. The playground…

If you think about it, it’s all in their interest and made sense from the beginning. I mean:

Google is running huge servers, which use huge amounts of energy. Translation: More efficient power electronics means smaller electricity bills…
They are working on laptop designs (and not only the OS) through Chrome books and the $100 laptop project for third world countries.
They also work on self-driving cars. And I’m quite sure that having an electric car that drives itself to the next charging station is a plus for you.
They invested in a top GaN devices manufacturers Transphorm. The company say they target PV inverters, Motor drives and so on… I suspect them to look at laptop adapters too.

All of these are using power electronics at a time.

For those of you who are not familiar (yet…) with this barbarian word: Power electronics in the electronic discipline where you focus on taking electricity in one form, and delivering it in another form (DC to AC, as an example). It’s opposed to analog or digital electronics, where you take one information to deliver it in another (like MP3 to audio analog).

It’s also known now that Google is more and more implied in Grid technologies (transport and distribution of electricity) and working on that fancy trend: SmartGrid. They have very good network knowledge ( errr… Yeah it’s obvious!) which is key for communication between power electronics converters. It avoids islanding (when a PV inverter is disconnected, but still producing, and has no metronome to rely on) or other kind of bad issues, but also great improvement like storage when needed, and use when needed too. Communications is a basic of Smartgrid.

3. The target…

So for those of you who are already in power electronics, here are the specs:

Yes, this is a datasheet or more a specifications sheet maybe, about an inverter, and issued by Google… If you are not either shocked or excited, I am for you!

So, let’s focus. What do they expect: 2kVA with a 240V@60Hz output and 450V DC input. That’s quite common and they are probably thinking about Photovoltaic decentralized inverters. >95% efficiency: The experienced one know it’s standard too in the PV inverter world. So what it so special about that target that it’s worth $1M: The size of course.

Smaller than 40 in.³ and max 20 in. size rectangle. Back in the metric system it’s 655cm³, or about 0.65 liters which is a bit more than a pint in the drinking system… That is very small for such power.

More seriously, it’s a 3.08kW/L.

4. The bounty hunters…

APEI (Arkansas Power Electronics Institute) did very small sized inverter (based on Silicon Carbide of course) to charge an electric car ( a Prius). It’s here and it’s actually matching the requirements, on paper (and on paper only). You will tell me that 6kW in such size is 5kW/L, which is better than the 3.08kW/L asked. The specifications were not the same either.

Though the question is to be able to reduce it to the 2kW and 0.65L demanded. There are uneasily shrinkable parts in there: The passive components. And in addition to that, the “thing” must be air-cooled inside but not higher than 60°C outside.

Another point. It is reserved to Academic groups, University labs.

I’m going back to my database to list and interview the bounty hunters.

Rohm could shake SiC world again with 1700V Mosfet