Startup Blurbs

Zilker Labs, a fabless semiconductor startup out of Austin, Texas specializing in digital power integrated circuits announced a couple days ago that a definitive agreement has been signed for the company to be acquired by Intersil. This acquisition comes only five years after the company was founded in 2003. It would be an interesting to examine at what average age semiconductor startups get acquired – maybe when I get some spare time.

Zilker Labs introduced their first product, the ZL2005 Step-Down DC-DC Controller in October 2005. From 2006 through 2008 the company introduced several more integrated DC-DC converters and controllers. The company’s main advantage lay in patents which allowed for designs with efficiency levels in excess of ninety percent. Additionally, Zilker Labs integrated power management functions into their power conversion products, thus reducing the required component count by about sixty percent. While the acquisition amount was not disclosed, the firms that funded the company, namely Sevin Rose Funds, North Bridge Venture Partners, and H.I.G.  Ventures probably made out pretty well on the deal given the relatively quick acquisition. On the acquiring company’s side, Intersil has been on an acquisition spree. In the past five month the company has also acquired D2Audio and Kenet Inc., specializing in intelligent digital amplifiers and FemtoCharge technology, respectively.

Back in August of last year we profiled Tilera when the company unveiled their TILE64 multi-core processor at the Hot Chips conference at Stanford University. And while some of the largest semiconductor players have struggled mightily over the last year, simply take a look at the dismal performance of the top 10 firms over the last year, Tilera has been busy executing their vision of what multi-core processors ought to be. Earlier this year, the company introduced the TILEPro family of processors which are targeted at the advanced networking and digital video market. With the introduction the company claimed a 35x performance per watt improvement over an Intel Quad-Core Xeon processor. It would have been nice for Tilera to have shared some workload and benchmarking data at least for those of us that like to dwell on details.

Anyhow, when you have a complete development environment with chips that deliver on promised performance, in other words the complete package, good things tend to happen. About one year after initial product availability, Tilera now claims to have over 15 design wins and is engaged with over 45 customers. Not a bad year at all, and to top it off the Global Semiconductor Alliance (GSA) just bestowed on Tilera the Start-Up to watch award. Good job guys, keep up the great work! Engineers and entrepreneurs alike need a beacon on the hill to look up to in these less than sunny days.

There is no doubt that obtaining funding for a semiconductor startup in this economy is anything but easy - simply take a look at our post regarding Ambric's fiath. But not all is lost, as Quantance demonstrates; obtaining venture capital is still possible. Just a few days ago, the company managed to obtain $12 million in Series B funding from TD Fund, DOCOMO Capital, Granite Ventures, and InterWest Partners. Quantance intends to use this money to commercialize their patented radio frequency (RF) technology for 3G wireless devices.

We’ve covered quite a few startups targeting wireless devices this year: Black Sand Technologies working on silicon based power amplifiers, WiSpry developing programmable RF products, and finally BitWave Semiconductor and their Softranciever technology - obviously the wireless sector is hot. Quantance’s goal is to improve the efficiency of RF transmit circuits in wireless products, and is shown in the figure below:

As can be seen, the company intends to improve the efficiency in three ways. First, the company proposes a supply voltage loop which modifies the supply voltage of the power amplifier (PA) to track the RF signal modulation. Second, for this to happen, Quantance had to develop an ultra-fast, low noise, power converter. Finally, they also integrated a closed-loop correction system to correct the incoming signal before it reaches the PA. These three things should yield several benefits: Since the voltage to the PA no longer needs to constantly stay at an elevated level, the power consumption and heat dissipation for this solution should be lower than that of competing units. Quantance claims that battery current savings might be as high as 30-40% and temperature reduction could be as high as 50%. I would suspect that these savings are highly dependant on the peak-to-average ratio (PAR) of the incoming signal, as such, the numbers above don’t really tell the whole story. Additionally, the company claims that their new approach will allow for a 20-30% increase in coverage range without an increase in battery drain.

These promised power-savings are likely music to the ears of mobile device developers, who are tired of consumers complaining about the shorter than expected battery-life of the current batch of 3G devices – which is probably why the company was able to score some more funding. Quantance’s first chip, the Q1000, began sampling in Q2 of 2008 and is expected to enter full fledged production in Q2 of 2009. As mentioned before, initially the company will target 3G mobile devices, but eventually plans to address other potential markets such a long term evolution (LTE) and WiMAX devices.