Lattice Brings Reuse from Smartphones to General Embedded

Three new Lattice FPGA reference designs show how smartphone components such as cameras or SoCs with MIPI interfaces can be “bridged” into general embedded applications like automotive.

Lattice today announced three complete reference designs–based upon their FPGAs, of course–that show how users can reuse/repurpose smartphone designs in non-smartphone applications.  If the component, such as the SoC application processor, camera sensor, or even RF front-end, uses the industry-standard MIPI interface (Mobile Industry Processor Interface), Lattice has a way to bridge that device to other components.

Lattice FPGAs in three new reference designs bridge smartphone any embedded design.

Lattice FPGAs in three new reference designs bridge smartphone MIPI…to any embedded design.

Automobiles, for instance, now incorporate forward-looking cameras for lane departure, and driver-facing cameras to gauge alertness. Neither of these applications are smartphones but with Lattice’s help could utilize low-power smartphone CMOS image sensors.

Subaru's EyeSight system uses twin forward-facing cameras for lane departure and other adaptive safety features. (Courtesy: Subaru of America.)

Subaru’s EyeSight system uses twin forward-facing cameras for lane departure and other adaptive safety features. (Courtesy: Subaru of America.)

It’s an interesting concept that’s not only useful, but brings Lattice back into focus as a niche supplier of specialty FPGAs that aren’t behemoth power processors like those from Altera or Xilinx.  I say hat’s off to Lattice. We’ll watch this evolve and keep you updated.



USB Charging Becoming More Common than a 110V Socket?

USB is everywhere, and increasingly, it is the “socket” through which we charge all of our stuff. Microchip wants to be inside all those “sockets”.

[UPDATE: 8/13/13 11:45pm PDT. Changed from 12V to 12W; corrected sentence accordingly. C2]

The common 110VAC wall socket. (Courtesy: .)

The common 110VAC wall socket. (Courtesy: .)

I noticed that Home Depot now carries replacement 110VAC sockets with USB ports built in. What a handy way to avoid fumbling around for the elusive USB charger “wall wart”, especially since most non-Apple portable devices now sport micro USB charging ports and accept cheap $1 USB cables.

As USB proliferates to nearly every battery-operated consumer device, I’m wondering when the number of USB ports (2.0 and 3.0 SuperSpeed) will exceed the number of 110VAC wall sockets. The number of 110V sockets in the typical home is static (remodeling and extension cords notwithstanding), but the number of USB-charged digital cameras, Kindles, Samsung Galaxy IV smartphones, tablets, hubs and so on is increasing.

The common USB 2.0 connector. (Courtesy: .)

The common USB 2.0 connector. (Courtesy: .)

With this USB onslaught around us, it’s no wonder that more IC suppliers are focusing on USB silicon. And not just for USB 2.0 480 Mbits/s connectivity.

Case in point: Microchip’s latest USB port power controllers–the UCS1001-3 and -4, and the UCS1002-2. The new family not only supports active cables such as Apple’s irritatingly proprietary Lightning connector, but also 12W charging, enabling higher current up to 2.5A, plus priority charging capabilities.


Microchip aims to charge any USB device up to 12W. (Courtesy: Microchip .)

Microchip aims to charge any USB device up to 12W. (Courtesy: Microchip.)

A built-in sensor can report the amount of current being sourced so the port’s host system will know the difference between say, a Nexus 7 or an iPad (which require over 2A charging), and a digital camera battery or child’s toy. The family can also support future USB charging profiles, since USB is essentially becoming not only a 3.0 SuperSpeed connection channel (currently 5 Gbits/s), but an intelligent battery charger. Microchip is so bullish on USB, that they’ve even created two Eval Boards,

One of two eval boards, this one allows designers to explore programmable charging profiles. (Courtesy: Microchip.)

One of two eval boards, this one allows designers to explore programmable charging profiles. (Courtesy: Microchip.)

one of which emphasizes USB’s programmable charging capabilities.

If USB does become more ubiquitous than the common 110VAC wall socket, Microchip hopes they’ll be the equivalent of “Intel Inside” every USB host device. Yeah, I think it could happen.

End of an embedded era: Emerson De-”Mots” Motorola Embedded

As Emerson Network Power gets sold off to Platinum Equity, Motorola Computer Group, Force Computer, Artesyn, and more names may disappear into the history books soon.

8/7/13 UPDATE: Several people have commented that the napkin analysis below neglects to account for the “power” side of Emerson Network Power. ENP was also partly assembled via acquisition including: Astec, Liebert, and others. A comment also was sent to me saying “The embedded power unit has been on the market for a buyer for quite some time…”  Finally, there are some questions raised about the size of the open standard ATCA/xTCA markets, with one person agreeing with my statement that the telcos are successfully using the standards to build their own hardware. This would reduce the TAM for non-captive vendors like Emerson Network Power. Thank you to all who corresponded with me privately.  C2


Emerson today announced plans to sell 51 percent of Emerson Network Power to Platinum Equity for $300 million. It’s a shame, for sure. But what’s equally interesting are the embedded technologies and their creators leaving the Emerson camp, and how we got to this place.

Embedded Consolidation by Acquisition

Emerson Network Power became the $1.4 billion business it is today partly by acquiring Motorola Embedded Communications Computing in 2007 for $350 million, when ECC’s turnover was about $520 million (2006). The sale closed in 2008.

Perhaps a bargain for Emerson at the time, in the interest of buying “embedded computing products and services to equipment manufacturers in telecommunications, medical imaging, defense and aerospace and industrial automation,” wrote the St. Louis Business Journal at the time. Motorola’s $520 million in sales was added to Artesyn’s $100 million embedded computing business, acquired by Emerson Network Power the year prior, adding up to over $600 million in revenue 2007.

Just three years prior, Motorola was then called “Motorola Computer Group” (MCG) and had acquired the then-heavyweight Force Computers from board-stuffer Solectron. The terms of the agreement were not immediately disclosed, but I was able to ferret the price of $121 million from a footnote on page 47 of Moto’s 2004 10K here. Interestingly, it was slightly prior to this when Motorola spun off their semiconductor operations to Freescale Semiconductor, a separate financial entity at the time. The combined MCG and Force division became known as Motorola Embedded Communications Computing and was all about standards-based telecom and military products like VME, AdvancedTCA, and so on. But mostly about the telcom-focused AdvancedTCA (ATCA).

If you’re following the math, the cumulative total of acquisitions for these embedded technologies was about $721 million to this point. As I recall, Force didn’t belong to Solectron for very long; less than two years, I think. MCG + Force = Moto ECC added up to about 1,500 employees in August 2004, said the press release at the time. The division’s corporate vice president, Wendy Vittori (previously of Dell Computer if memory serves), said at the time: “We will be able to provide solutions for a wider range of customer application needs, supported by a broader portfolio of boards, systems, and services.”

Moto was number one in VME, although they’d ceded the rugged mil/aero market to the likes of Dy4 Systems (later Curtiss-Wright), Radstone, and SBS (both later part of GE Intelligent Platforms) in the late 1990s. Motorola lead the non-mil market with Motorola’s/Freescale’s own PowerPC-based single board computers, whereas Force had leadership in Intel-based SBCs and broader networking products. Wendy was right: it was a pretty decent technology fit, and Motorola was at that time already parlaying their embedded products into the data center and telecom. A year prior, in 2003, Motorola acquired NetPlane Systems, a telecom provider with data and control plane products…and captive customers.

When the Emerson/Motorola deal closed in 2008, an Emerson press release quoted several analysts praising the acquisition. It also said “A significant trend in the embedded computing industry is the adoption of industry standards, including ATCA, MicroTCA and AdvancedMC (AMC/xTCA)…currently more than 40 percent of network equipment providers are shipping ATCA-based systems.”

Present Tense

So far so good. In fact, I’ve followed the industry closely and agree that wired and wireless infrastructure build-outs continue to favor these embedded open standards-based products, and ATCA et al have replaced proprietary telecom equipment. Emerson Network Power’s VME business, I suspect, never recovered since the market for VME (and now the VXS and VPX variants) is almost entirely in defense. (Recall that Motorola walked away from that business ten years ago.) That leaves ATCA, xTCA targeting the telecom markets.

As recently as two months ago (May 2013), the head of the PICMG standards group responsible for ATCA, xTCA and AMC told me how well the telecom markets were growing. You can read my interview with Joe Pavlat here, where Joe estimated the market for ATCA at somewhere between $1.5 billion and 2.5 billion per year.

What happened?

In February 2013 Emerson’s CEO David Farr went on record with Fortune magazine as saying he wants to “double down in businesses that help manufacturers produce their wares” and to focus on cooling products (like air conditioners and chillers for data centers).

This might explain why Emerson would opt to leave this business along with Emerson’s pre-Motorola power business. The press release issued today cites the group’s revenue at $1.4 billion in 2012, probably less than the cumulative total of the price in real dollars of all those acquisitions if you linearize them from 2008. In fact, the group should probably be selling over $2 billion to achieve the correct ROI on all of those acquisitions, but that bumps up against the ATCA TAM cited above by Joe Pavlat of PIGMG. Did Emerson run out of ATCA runway?

That possibly explains the $300 million purchase price for 51 percent, making the overall sale roughly 50 cents on the dollar of last year’s gross sales. That also puts years’ worth of leading-edge VME, control plane, data plane, networking IP, ATCA, xTCA and other embedded technology up for sale by Platinum Equity. Or maybe not.

Sell it, or Keep it?

Who might want this technology? If you assume that no Emerson Network Power customers will be lost in the process (CapEx equipment is not quickly designed out), Emerson’s competitors like Radisys, Kontron, IBM, Dell, and HP already have their own (open-standard) hardware. My bet is that the key value will be any proprietary IP owned by Emerson plus customer relationships (read: backlog). Yet I can not think of a single open-market company that would want to buy this technology that doesn’t already have the core technology. So why buy it?

But Platinum may own a core company that needs Emerson’s technology for themselves: perhaps a telco or wireless provider who wants to produce their own ATCA equipment and not buy it on the open market. This certainly is a viable strategy for a mere $300 million (to start) to buy a multi-billion dollar telecommunications outfit. When asked to comment on this story, PICMG’s Joe Pavlat said: “Platinum Equity is extremely well regarded and has several other significant telecom investments that, at first glance, appear to be very complementary to the Emerson offerings.” Bingo.

So it may be the end of an era–when companies like Motorola, Force, Artesyn, NetPlane–created and implemented open standards-based embedded computers for the telecommunications industry. Hopefully these names and their creations will live on at another recognizable open standards company. But I’m not hopeful; I suspect they’re gone forever and de-Mot’ed  to the history books.