Your browser is not supported. Please update it.

6 October 2011

Quantenna comes out fighting on the power consumption front

Quantenna called this week seeking redress for comments made by Celeno about its power consumption, during an interview with us at IBC. Sam Heidari CEO of Quantenna gives off the tired air of someone who has heard it all before, and who knows he has to be patient, and ensure that his explanations are absorbed. He is fighting a war not so much on video network performance, but of perception. WiFi in video has a bad name and his company single-handedly has to change it.

The criticism was that Celeno uses 3 radios and therefore 3 power amplifiers, and because it has 6 antennas and can switch the radios between them, it claimed that its chips used less power due to the lower amplifier numbers, but the same performance.

The power issue was dealt with by Heidari swiftly, ‘Yes we have 4 power amplifiers because we have 4 radios, but in each region, Europe, the US, we all take the regulator’s power limits and when using a 4 x 4 MIMO we divide them by 4 and set our radio power usage. If you make a 3 x 3 you divide by 3, so each radio in a 3 x 3 is working closer to its saturation limit, which means it is working less efficiently and gives off more noise. There are codes of conduct on power in both the US and Europe and we operate easily within them.’

All of which puts paid to the suggestion that Celeno fed us at IBC, and gives us an opportunity to quiz the Californian company about its new deal with Samsung,

‘We know that the press release was a bit ambiguous. It says both that Samsung will use our chips in their products and that we will work together to continue to develop high-performance wireless solutions. I can’t say which products Samsung will use Quantenna chips in and I can’t really tell you the ways in which I see the partnership developing new products, but they are definitely there. It should be more obvious in about 6 months,’ none of which takes us forward so we have to speculate a bit here.

We don’t see why you would need a 4 x 4 transceiver chip in a SmarTV or a Blu-ray box, so we would hazard a guess that Quantenna chips will be used to merge Samsung set tops at the very least into multi-room DVRs or an entire home gateway. In terms of developing the technology Samsung might use the Quantenna know-how to get early into the IEEE 802.11ac chip market, but surely that would make it a competitor to Quantenna, so more likely it could put the Quantenna designs into SoCs. It might also develop transceiver chips which are far cheaper, designed specifically as receivers for its TV and Blu-ray sets, which work with Quantenna transmitters. Of course the relationship could be a lot closer and this could be Samsung making its core WiFi know how available to a company that it later plans to acquire.

Another area where Samsung may find it profitable to work with Quantenna might be in creating a two layer fully wireless home network ‘ for instance a Quantenna wireless WiFi backbone around the home, coupled with using 60GHz for delivery across a room, perhaps using WiGig or WirelessHD?

But whichever way you see it, the big rival in this market right now is Celeno and the small Israelli start-up is the incumbent in the Horizon device, announced in detail at IBC, which is a device put together by Samsung (although it is under the direction of both the prime contractor NDs and the operator UPC). So Quantenna seems to be moving aggressively into Europe and trying to ensure that Celeno has less and less oxygen on which to survive, by taking away one of the design partners it already has. Hence Celeno’s desire to make it appear that Quantenna chips might not work here. As Heidari says, he is fight a war of perception.

It would take a real world test, of both electrical power used and of performance, or someone with a PhD in electrical engineering to tell the difference between the two approaches so we won’t try. Celeno does only have 3 radios, but it can marry up any radio with any of 6 antennas, so its beam forming capabilities may well match up to Quantenna’s. Or they may not.

If we measure the two companies on shipments, Celeno simply doesn’t give out numbers and will only say that it has been shipping chip samples since 2009, while Quantenna talks about being on its third generation chip, having already shipped in the hundreds of thousands.

‘We will ramp production in late 2011, early 2012,’ Heidari said and agreed that Faultline’s estimate of in excess of 5 million chips by the end of 2012 was ‘not a million miles from its own figures,’ along with a potential break even financially.

But this is the problem right there. No market of less than 5 million chips is worth the effort for the big WiFi monsters of Broadcom and Atheros ‘ and once the market for video capable WiFi chips reaches this type of size, they will both enter it through acquisition or R&D investment and drive the price to the floor. Right now they offer both 2 x 2 and 3 x 3 MIMO WiFi chips, but have not done the extra investment needed for dynamic beamforming on 4 x 4. That’s not to say they couldn’t, but companies of this size are intrinsically less efficient and need to see a big payback before they agree to enter a market.

The first commercial operator to take the Quantenna chips was Swisscom, which went on to be an investor in the company in a deal uncovered at the beginning of this year.

Quantenna claims a further 6 or 7 unnamed operator customers are lined up and in some cases already in field trials, and it has said all along that it hopes to catch the eye of the important consumer electronics majors. including Samsung, trying to tear them from their interest in other wireless delivery formats to attach TVs including WirelessHD, WHDI and the emerging WiGig ultra-fast Wi-Fi effort in 60 GHGz.

The Swisscom field trials were carried out with a retail product which the company announced with Netgear in a Wireless video bridge. Heidari says, ‘these chips will be built into set tops, DVRs and home gateways. Our 3rd generation chip is an order of magnitude better in performance and its price will be based on mass deployment.’

When we asked about costs Heidari tried to help but wouldn’t go the whole hog, ‘The difference in costs between a Quantenna chip and another chip using less radios, is less than $5 in the total Bill of Materials for a device,’ which is as close as he would go. Heidari is fond of claiming that Quantenna chips operate at 5 times the reach and twice the throughput, so $5 more in a set top for that type of improvement, would almost certainly be worth it in pay TV, a market where errors at the customer end are toxic to repeat business.

‘The Swisscom deal was based on achieving 95% home coverage in the field trial and our chips beat that comfortably. The real world performance is pretty much 100 Mbps throughout a home. I know we cite 600 Mbps, and we can tick that box with two devices right next to each other. But when we are getting 300 Mbps, other chips are getting 50 Mbps, and when we are getting 150 Mbps, other chips can’t connect.’

So when will Quantenna or anyone else’s chips be ready to take on the fixed wire technologies such as MoCA and G.hn? ‘This is just a perception problem and slowly it will go away. We can already be used without any fixed line backbone. Look, Verizon chose MoCA because although it was more complicated and costly in terms of installation than anything else at the time, it led to less support calls and less customer issues all round. Operators will pay extra for the technology if it gives them that.’

By implication Heidari is saying that if an operator gets less fuss, but also gets a less complicated installation, then of course they will buy into it. You can almost hear him shrugging at the other end of the phone, inviting you to make you own conclusions in favor of Quantenna 4 x 4 WiFi. ‘By the end of 2012 and certainly by 2013 that perception will start to change in our favor ‘ because there will be a lot more pay TV services using only WiFi connections for video networks.’