Subscribe to EDN

GaN community grows: ex-IR CEO launches GaN-based power transistors

March 5, 2010

Hot on the heels of International Rectifier’s announcement last month of its first Gallium Nitride (GaN) -on-Silicon product comes an announcement by start-up company Efficient Power Conversion Corp. (EPC) of its new family of power transistors based on its proprietary GaN-on silicon technology. The transistors’ Vds ranges from 40V to 200V and Rds(on) from 4 to 100m§Ù. The company claims that GaN devices can reduce total server power requirements – including ac-dc and dc-dc losses and fans – by about 18%.

The enhancement mode –normally OFF– GaN technology was explicitly developed to replace power MOSFETs. Says Alex Lidow, EPC’s co-founder and CEO, enhancement mode — rather than depletion mode – is essential for GaN to become a broad-scale silicon power MOSFET replacement. “It’s meant to imitate the functionality of a power MOSFET on supersteroids.” Because the GaN structures can be laid down on a standard 6 in.wafer, the price is comparable to high-end MOSFETS: $0.80 and $5.00 in 1k quantities. (You can buy them today at Digikey.)

Rds(on) for a given device area is a key determinant of cost, and the equivalent Gan transistor is markedly smaller than the equivalent MOSFET as you can see below.


Comparison of MOSFET and EPC

Note that the MOSFET is shown packaged, while the GaN device is a bare die. This is a valid size comparison because since the GaN structure is laid down on a silicon layer that acts as an insulator and the device is hermetically sealed so it requires no package. The GaN transistor can be mounted directly on a heat sink with no surrounding packaging and no thermal resistance between the transistor and its package.

Here are more specs:


EPC GaN specs

How does this new family compare with IR’s iP2010 and iP2011? First off, IR’s devices are native depletion mode – although according to Dr. Mike Briere, responding to a question posed during an IR seminar at APEC, they can be made in enhancement mode. In addition, the EPC transistors are just that – discrete transistors, while IR’s product is a integrated power stage devices that integrates a driver IC matched to a multi-switch GaN transistor, all mounted in a flip chip package.

The Rest of the Story: What about the fact that Alex Lidow of EPC was, until October 2007, the CEO of International Rectifier? Lidow has stated that all of EPD’s intellectual property was developed and owned entirely by them, However, IR filed suit late last year against Lidow, claiming theft of trade secrets. From a PowerPulse article in January:

“IR claims that Lidow stalled IR from making its GaN technology public in 2007, and instead began secretly recruiting the above referenced IR staff (from the R&D team plus senior sales representatives) for his plan to establish a new firm – EPCC, supposedly to provide rival GaN-based products.

Legal representation for Lidow and EPCC claim that the defendants are not using IR’s GaN technology, and that EPCC is developing a different semiconductor product.”

Posted by Margery Conner on March 5, 2010 | Comments (6)

March 9, 2010
In response to: GaN community grows: ex-IR CEO launches GaN-based power transistors
Rob O commented:

Andy T, If you click on the link to EPC early in this story and then click on the PRODUCTS tab on the EPC site you can get to a page with the datasheets for these parts. You will find that: 1) They put the max operating Tj at 125C (same as most commercial Si.) So I wouldn't expect to be able to operate these at a higher Tj than Si. 2) These parts are bumped (flipchip packaging.) So the connections to a board would be through solder bumps. And most of the heat transfer would be through the bumps. I don't believe that you can expect to be able to mount one of these parts directly to a heatsink. If EPC did package the parts in a SOT-89, the die size would (as DaveR suggests) impact the thermal performance. A larger die will have better thermal performance than a smaller die because there is more cross-sectional area for the heat to flow from the back of the die, through the die-attach and out to the case.


March 9, 2010
In response to: GaN community grows: ex-IR CEO launches GaN-based power transistors
John L. commented:

It appears the package is intended to be mounted like any other SMT component (the pic has the contacts exposed).. smallest line/space ~10mils (0.010") shouldn't be a problem for most SMT pick and place machines. Will likely require a bit of thought on the mounting of the heat sink.. (if one is required) asking a lot... 100v voltage across 0.010" termination spacing...(EC1007) not quite sure how that is going to fly with UL. for the lower voltages.. no problem


March 9, 2010
In response to: GaN community grows: ex-IR CEO launches GaN-based power transistors
Rick commented:

There is something I don't understand. Looking at a 60V version (EPC1005), it has a Rdson of 7 mOhms. For a D2Pak like the picture, you can get an IRLS3036 with less than 2 mOhm on resistance. Not really a good comparison for the article. Now based on the size of the EPs and the packaging (or lack thereof), they have potential.


March 9, 2010
In response to: GaN community grows: ex-IR CEO launches GaN-based power transistors
Andy T commented:

@DaveR Larger than what? If you use the same die attach and can put the GaN die on the heat sink directly, you eliminate the theta package-heatsink, so it's actually LOWER thermal resistance and MORE power capability than having an intermediary package. You should be able to buy Si die, so you could take it off the table entirely as an argument. However, if I remember it right, GaN can run at higher Tj than Si as well. You won't argue against this tech thermally, it's worst case par, but there's still the mystery of how to make an electrical connection to die from a board....it's a PITA for most board assembly shops. These guys would be wise to offer a SOT-89 option for us low tech people...


March 9, 2010
In response to: GaN community grows: ex-IR CEO launches GaN-based power transistors
DaveR commented:

Thermal resistance of smaller die will be larger - Less power capacbility! Pkg vs die comparison is not a good comparison.


March 8, 2010
In response to: GaN community grows: ex-IR CEO launches GaN-based power transistors
kartik shukla commented:

What about reliability of this devices. Kartik shukla quality services engineering services div 28 welcome shopping center Baroda-390015 India

POST A COMMENT
Display Name
captcha

Before submitting this form, please type the characters displayed above. Note the letters are case sensitive:

Advertisement
Advertisement
Advertisement
About EDN   |   Site Map   |   Contact Us   |   Subscription   |   RSS
© 2012 UBM Electronics. All rights reserved.
Use of this Web site is subject to its Terms of Use | Privacy Policy

Please visit these other UBM Canon sites

UBM Canon | Design News | Test & Measurement World | Packaging Digest | EDN | Qmed | Pharmalive | Appliance Magazine | Plastics Today | Powder Bulk Solids | Canon Trade Shows