Cheaper – not better – batteries will rule the EV market

In 2010, Panasonic announced they had developed a 4.0Ah silicon anode battery that would go into production before the end of March 2013. At the end of 2013 the battery is nowhere to be found.

It’s easy to chalk this up to product development delays. I don’t doubt this is hard work – science is hard. If it were easy we’d have fusion and flying cars by now.

Rather, I think that for large battery manufacturers like Panasonic and LG, they’ve redeployed resources from making better batteries at the same price, to making their existing batteries cost less. Reducing the cost per kWh of battery capacity is now job one. The disruptive threat from some start-up coming out with a novel process to make a 4x energy capacity battery is mitigated by the fact that the large mainstream suppliers will be able to undercut them on both price and manufacturing capacity. Their novel process might work for niche applications in low volumes, but it won’t matter much because the cells will, by comparison, be in short supply and too expensive for a mainstream EV.

The change in priorities wouldn’t be a surprising one, given the principle issue with EVs today – cost. If you don’t bring down the cost of the battery packs in the EVs today, there won’t be broad-based demand for EVs tomorrow no matter what the range of the vehicle is. When EVs take the place of a second or third car in a household, the range issue isn’t nearly the problem it is made out to be. And the market for second and third cars in households that can afford suburban homes to recharge them in is sufficiently large this early in the electric vehicle adoption curve.

Battery costs are not generally published, but we can estimate that as of mid 2013 are around $350 per kWh at the finished pack level. This makes the price around $30,000 for the Tesla 85kWh battery, $5800 for the Volt, and $8500 for the Nissan Leaf pack. By the end of 2015, prices could be around $250 per kWh, and $150 per kWh by the end of 2017. At the second long term price goal of $150, it becomes possible to sell a 60 kWh/220 mile range EV for $35,000 because the pack is only $9,000 (or 30% of the bill of materials), roughly in line with the Nissan Leaf. As Tesla has shown, a purpose-built EV can accommodate the amount of cells necessary for this battery pack size. Existing battery technologies (NMC, LiFePO4) will continue to improve marginally each year, providing more energy capacity per unit volume and per unit weight.

Once the price issue is resolved, battery makers can focus exclusively on incorporating new technologies to make EV range no longer an issue without spiking the price. While emerging technologies like lithium-air technologies may become practical after 2020, when the cost can come down enough to put them in affordable EVs for the driving public is the larger question.

Apple October 22 Predictions

Apple will give a release date for Mavericks (October 25) and price ($20, $50 for server)

Apple announces new MacBook Pros (retina and regular) featuring Haswell CPUs for better battery life – discrete GPUs don’t go away, but they’re only available on the 15″ models due to PCB size constraints (as has been the case for many years now) but integrated Iris Pro replaces GF GT 650-level option

Apple mentions Mac Pro will ship at the end of November, with pricing and pre-orders to come some time next month – brief mention its Made In America, including gratuitous shots of manufaturing/assembly line

Apple announces new 9.7″ iPads in thinner design with same battery life and A7X CPU at the same 499 and above price points

Apple announces retina iPad mini for $379 with an A6X CPU, and regular (current gen) iPad mini for $279

Nothing groundbreaking, nothing crazy – no Apple TV (as in actual TV sets), no iWatch. And since its between the time the quarter closes but before the earnings statement on October 28, not a lot of detailed information will be revealed about sales numbers or material things like that.

Routes for I-11 through and north of Las Vegas

Nevada’s Department of Transportation is working on figuring out where to put I-11 through and north of Las Vegas. The route between Las Vegas and Phoenix is pretty well established – along the current US 93 alignment from Las Vegas, with a bypass south of Boulder City, over the Hoover Dam Bypass Bridge, until US 93 hits Wickenburg, AZ, about an hour outside of Phoenix, where it branches southwest through the Hassayampa Valley down to I-10 west of Phoenix. A full listing of all the possibilities are available here (20MB PDF).

The route through Vegas is a difficult one – it would cost at least a billion, likely two billion, to upgrade US 95 to be wide enough to carry the extra traffic – the portion of US 95 east of the I-15 interchange would need a billion-plus makeover to fix the 1970s-style viaducts, and recently widened western half would need to be widened again, with some additional interchanges constructed. Two options are to use the existing 215 three-quarters beltway around the city (one around the city clockwise, which would come within 2 miles of my house, and one counter-clockwise) – however the lack of an eastern leg of the beltway means that a new leg would need to be created, and the alternatives indicate building it behind Sunrise Mountain through the Lake Mead National Recreation Area and connecting it to the I-15 north of Nellis AFB. After reviewing all the alternatives, I think the best route is to build the leg near Lake Mead, and connect it to the existing 215 highway (Alternate QQ). Work would need to be done ASAP* to secure additional right-of-way along the highway to allow it to expand up to 5 or 6 lanes in each direction. The clockwise usage of the 215 would require some additional right-of-way in certain areas, but would overall be much cheaper than expanding US 95 in place, as well as allowing for the highway to be up to 5 lanes in each direction, whereas portions of the existing US 95 are already that width, and would need to expand an additional two lanes to 7 lanes.

I’m sad to see the I-11 committee has removed my preferred route north of Las Vegas – Alternate GG through central Nevada. I would have liked to see I-11 run through central Nevada, with a spur line (I-511) connecting it from where it turns north from US-95, on to I-80 near Reno. I think this would have provided the travelling public with a much faster way to get north to places like Idaho, Oregon, and Washington, rather than going all the way over to Reno, then all the way back west to Boise. If the route continues north to I-84 in Oregon, then going through Reno would be a better choice. I expect Alternate SS to win out – following the US 95 corridor northwest to I-80 near Reno, and from there up US 395 to I-84 near the Columbia River.

But to me, its putting the cart before the horse – lets focus on getting I-11 from Las Vegas to Phoenix located, funded, and designed (glaring at Nevada’s past and present Governors and NDOT officials who have spent half a billion dollars for a highway to connect Reno to Carson City, but don’t want to fund the Boulder City Bypass with state dollars, instead suggesting a toll road), and from there we can map out where its going to go next. The Boulder City Bypass is job #1 at this point, with the Kingman Interchange next, along with all the improvements between the two to make it a full fledged interstate highway, at least from Vegas to Kingman.

From there, its on to US 93 between I-40 and Wickenburg, which has had substantial improvements over the last 10 years or so, with only three projects left to complete to make it four lanes between I-40 and the Santa Maria River – I’m guessing AZDOT won’t do any work south of the Santa Maria River until the I-11 corridor is finalized so they know where to expand it to interstate standards and where not to spend the money.

On that front – Arizona DOT has selected a preferred corridor and preliminary design for the Kingman interchange (PDF), with an estimated cost of only $86M. That pales in comparison to the $300-400M estimates for the Boulder City Bypass, but then again the BCB is a lot more work digging through the Eldorado mountains. However, Arizona would also need to spend more money on interchanges along the I-11 highway to remove at-grade crossings (I’m guessing 5 between Hoover Dam and Kingman).

* Sadly, this cant happen, as the PISTOL amendment to the constitution has a five year window for governments to use the property for the stated purpose (in this case, highway expansion) – if it doesn’t the original land owner can reacquire the property by paying the government back. In this case, dumb limitations hamstring government to do its job in a cost-effective manner.

Brief 2c on the Government Shutdown

I believe the root cause of the shutdown isn’t the budget deficit or Obamacare, but rather Gerrymandering – and two specific symptoms of gerrymandering – uncompetitive districts evolving into extremist candidates, as well as illegitimate majorities stemming from packing and splitting.

There have been a few articles, most pointedly this one (from a conservative paper no less), that illustrate how the shutdown is being driving by a minority of the majority party. To paraphrase the article, there 30 GOP members who believe that Obamacare is in fact very harmful and will do anything to stop it. There are another 30 that are willing to compromise, but worry about being primaried from the right. It is my estimation that if those 30 were primaried, their successor would join the 30 true believers.

The emphasis in the above paragraph illustrates the problem – uncompetitive (from a partisan angle) congressional districts drawn so that they are solid for one party or another. Because one party has a lock on the district, its likely that more and more extremist (left or right) individuals will be elected – up to a point at which there are enough defectors from the majority party to elect the other candidate. This almost happened in 2012 with Michele Bachmann – she is in a solid conservative district, but has become extreme enough where there were enough defectors (or non-voters) in both her party and independent voters that she nearly lost the race (she won by 1% of the vote).

If all congressional races were competitive from a partisan angle, I believe the extremist element in each party would be rendered unelectable because independent voters (whom upon most elections ride) will choose the less extreme candidate, and that in order to win political parties would choose the more centrist candidate instead of the extreme.

So how do we stop gerrymandering?

California, after substantial bipartisan pro-incumbent gerrymandering in 2000, chose an non-partisan board to draw the districts. Other states have had similar approaches by using a non-partisan group to draw the lines.

Can we fundamentally change the rules of the system? Do we have all representatives represent the entire state they’re from like Senators do? Do we use topological rules to divide districts in a more algorithmic manner? Use an alternate voting method (Fair Majority Voting for example) to ensure that a state that is split 35:65 has a 1:2 representation ratio? Beyond an appropriate partisan representation ratio, how do you choose representatives from each party that are aligned with the voters wishes (e.g. you not voting for a party, you’re voting for a person).

The voters weren’t represented in the 2012 congressional elections – totaling up all congressional races, Democrats got more votes than Republicans by about 500,000 votes (slightly less than half a percent), however failed to gain majority control of the House because of gerrymandering, in fact they are at a 30 seat deficit (233-202). This was due to the Republican successes in 2010 getting control of state legislatures such that they could control the redistricting process, gaming the system.

The sooner we deal with the issue of Gerrymandering the better. Then maybe we can start removing some of the dysfunction from Congress and move towards a legislative body that can accurately represent the people of America, both in terms of majority/minority parties as well as fewer extremists.

The Importance of Infrastructure – alternately, Bootstrapping Civilization on Another Planet

I thought of this question when I read the news about a SuperEarth that’s only 22 light years away. What would you need to take to another planet to bootstrap civilization? To build up civilization enough such that it could survive on its own without any assistance from Earth? Initially, you’d probably send some autonomous robots and satellites to the planet to scan for information to make sure it’s hospitable to humans. From there, you’d want to build up infrastructure.

Infrastructure is the parts of civilization that give you the ability to grow food, get clean water, and have a peaceful and prosperous society. It starts with an energy source, but you’ll bring that with you (probably nuclear or fusion-based). You might want to build a primitive navigation system (something like GPS) and a geographic database so you could have maps to know where resources are. You’d also want some sort of weather satellite to know when you’re going to get hit with hurricanes/typhoons, tornados, etc. The atmospheric data would be useful for growing crops. The first few generations would be rough – there wouldn’t be much in the way of creature comforts like sports and entertainment, unless you want to watch the World Series from 22 years ago.

Eventually you have to be able to stop using what you brought with you and transition to using what is on the planet. In other words, you have to get to sustainability. Eventually, the nuclear fuel you use will run out, the satellites orbiting the planet will run out of propellant and no longer be able to control their orbit. The good news is that we have knowledge! This new civilization wont need to go through stages like the Stone Age, Bronze Age, and the Iron Age. We know how to smelt aluminum, what Portland Cement is made of and how to make large batches of it, and we know what it takes to get things into orbit.

But we still need the manpower to make all this stuff and to collect the raw materials. Its the other end of the economies of scale – building out a small civilization with only several thousand people means you have to pick and choose what you do – you don’t have an entire planet’s population to draw on for a diverse set of resources. Picking the right things to specialize in might make the difference between survival or colony collapse. This means government intervention – the people in charge will have a plan (likely drawn up before they left Earth) and want to stick to it to ensure success.

Lets say we’ve picked an area near a large river, and we want to provide a source of power, control the annual floods as well as create a lake to use as a reservoir for our potable water supply (and sometime down the road, recreation). So we want to build a large hydroelectric dam. We need an incredibly large amount of steel and cement, along with the turbines and power distribution system. What goes in to Portland Cement you might ask… well we need to mine a large amount of limestone and gypsum for the basis of the cement, plus some other minerals. From there, we need to build a cement mill to grind the raw materials into the powder needed to create the cement mix. How do you build a cement mill? Beyond that, we need the people to make all that stuff, and build the dam! The average amount of people working on the Hoover Dam in the 1930s was 5,000 people. Even with technological advances in terms of automation, will you even have that many able-bodied men and women in the colony to complete the project?

In modern society we take so much for granted in terms of infrastructure. But when none of that exists, not even foundries to create steel for building other factories with, you start to have to get creative on how to rebuild that infrastructure or what can you take with you from Earth to the new colony.

Do we take raw materials? Can we mine minerals from asteroids on the way there? How does the new civilization launch replacement satellites, or get back into space if a few generations down the road they figure out that the planet just isn’t hospitable, or to study and explore nearby planets and moons? They know how rockets work, but do they have the incredible industrial machinery necessary to build one? Do they have the manpower necessary to operate the supply chain from raw materials to finished rocket and replacement satellite?

With all this work to do, the last question is how do you keep the colony going and growing? How do they find time to do all of these things with such few people, and still manage to raise sufficiently large families (5-6 kids each for many generations) necessary to populate the planet and ensure a viable future? How do they educate them? Who runs the schools and universities? What about technological advancements and research and development?

This is the “soft infrastructure” side of development – the government, the institutions necessary to operate in a civilized society? Does it operate as a military-style dictatorship for the first 50 years, then evolve into a democracy? If you choose democracy out of the gate, how do you keep it from making bad decisions based on self-interest that could imperil the entire colony?

In short, infrastructure (hard and soft) is incredibly important. And it turns out that colonizing another planet might be equal parts developing the technology to make the journey, building the colony when you finally arrive, and successfully governing the colony’s early years.

iPhone 5S is a great upgrade from the 4S, but…

As a nerd who follows Apple, I have to maintain a delicate zen-balancing of wanting the latest and greatest, with the best specs and fastest everything, with the understanding Apple isn’t about the specs.

But the iPhone 5S is mostly a miss for this iPhone 5 owner. Its a great upgrade over a 4S model – users who upgrade from the 4S will see amazing speed improvements in the CPU, GPU, cellular connectivity (LTE), and WiFi reception (using the 5GHz band instead of the ultra-congested 2.4GHz band).

But for the iPhone 5S, the areas in which Apple showed the biggest improvement (CPU, GPU) were already fast enough me, and iOS was always responsive even with the iOS 7 betas. I remember many times on my 4S while I was waiting for something to download and process, but I don’t have that on the iPhone 5 – between the A6 CPU and LTE, I’m quite content.

The areas that I did want to see improvement – LTE-Advanced radios, 802.11ac WiFi, and battery life went missing or didn’t get improved that much (battery life got a tiny boost, but not as much as I was hoping for if they had used IGZO screens).

I’m almost tempted to hold out and wait another year for the iPhone 6, particularly because I pay the premium to buy the 64GB model. I cant hold out because I already promised my phone to a family member when I bought it and used their upgrade, but I think that if I could, I would hold out for the 6.

Its a change in the way the cell phone market exists – phones only need to be compelling in two respects 1) against current competitors and 2) upgrade-worthy over the two year ago model once you’re locked into an ecosystem. All other concerns for Apple reside in the “ease of use” and “margins and profit” categories.

Net metering for solar power isn’t sustainable, so it will end

Recently, utilities in two states have started the process of ending the net metering incentives for solar power installed on homes. With the advent of solar power financing companies like SolarCity that will finance the cost of the solar power system and take the payments from the savings on your power bill, the number of solar power installations on homes is proliferating. But when you play out the scenario of most single-family homes having solar power on the roof, and net metering providing them with a bill of around $10-20 each month, you discover that net metering isn’t sustainable for the balance sheets of utilities – so here we are at the end of net metering.

The underlying economics is why net metering isn’t sustainable for the grid or the utility that runs it. The 11c per kWh (national average) you pay is really three component parts – the electricity generation (including the power plant and the feedstock used as fuel), the transmission and distribution infrastructure costs (power lines from the plant to your city, high-voltage distribution throughout your city, lower voltage distribution to your neighborhood, and finally 240V distribution from you local transformer to your house), and overhead and profit (the costs of running the company).

The costs of each are about 5-7c for generation and fuel, 3c for transmission and distribution and the remainder for the overhead/profit. Beyond these cost estimates, the price of electricity varies throughout the day (on- and off-peak pricing) as well as different times of year due to supply (from hydroelectric generation stations in the spring) and demand (air conditioners in the summer and electric heaters in the winter).

When you’re generating solar power to use within your own house, you’re eliminating the demand for the first two items in that list – generation and transmission/distribution. But you haven’t eliminated the third, while the grid just sits underutilized but still must be paid for (capital costs and maintenance). Beyond that, when you’re generating more than you can use within your own home, you’re increasing the cost of the generated fuel from 5-7c to the net-metered cost (11c), and using the local distribution network (at a fractional rate of the whole), so the cost of electricity to the power company goes from 11c to about 17c per kWh. This was acceptable to most utilities when the on-peak summer prices for energy were 15c or more per kWh. But current spot market prices hover anywhere between 5 to 8c per kWh, which means they could buy extra power cheaper on the spot market than a customer selling their excess solar back to the grid under a net metering program. Bottom line is that the utility is forcing customers without solar power to subsidize those with solar power.

I’m still bullish on solar power – but the rate structures will have to change, and likely become more complex, and more in line with how commercial and industrial customers are charged. The structure would like be both consumption and capacity based. You would be charged for power consumption at 6c/kWh, plus demand charges based on the maximum power you’ve drawn from the grid over the 12 months (e.g. 6.3kW on July 15). Using solar would cut your consumption and likely decrease your maximum power, but the free ride is over.

Beyond net metering, grid-battery integration will have to become cheaper and more feasible so that those who generate solar power can store their own over-generation and use it at later times to bring down their costs.

Apple iPhone 5S/5C Predictions

So we’re a little less than a month away from the September 10th unveiling of the new iPhones.

iPhone 5S:

  • A7 processor – still 32nm Samsung fabbed. An incremental upgrade over the A6 processor. Likely just improving the SWIFT core, accomplishing all the things they wanted to do with the first generation SWIFT (A6) but didn’t get time to. For the CPU, I expect a little IPC performance improvement, but stays dual core and increases the max clocks. For the GPU, I expect a non-trivial performance increase with a switch to the “Series6” Rouge GPU from PowerVR, probably in the neighborhood of 50%, with a much smaller increase in the die area.
  • Third-gen Qualcomm (MSM9625) baseband and WTR1625L radio front-end. This supports LTE-Advanced (carrier aggregation) and UE Cat4 speeds (150Mbps, which you’ll start to see as Verizon and T-Mobile roll out 20MHz FDD networks).
  • IGZO screen to save energy to make up for carrier aggregation (running two radios at once).
  • Support for 802.11ac (single stream up to 80MHz channels, similar to other smart phones).
  • Same storage and price points as the current iPhone 5. Apple’s margins are shrinking enough on the iPhone as-is, so they don’t want to let users pay $200 for a 32GB phone when they could pay $300. Small chance of seeing a 128GB iPhone (again, to boost margins).

BOM: ~$180

iPhone 5C (the C is for COLORS!)

  • Either an A6 CPU downclocked (parts that couldn’t make QA for the iPhone 5) from the iPhone 5 version, or a modified A6 CPU that is missing some features.
  • Second-gen Qualcomm LTE 9615 paired with WTR1605L radio. This will enable TD-SCDMA and TD-LTE for China Mobile.
  • Rest of parts similar to the iPhone 5, except for that the case is much much cheaper to make (cost savings of around $15/phone alone).

BOM: $135

OXIS Energy to scale up production of Li-S batteries in 2014

OXIS Energy is planning on commercial production of their Lithium Sulfur batteries in 2014. The cells are expected to be around 200Wh/kg (low for Li-S but its still early) and achieve a little more than 1500 cycles to 80% capacity. They have been tested to be very safe (a nail puncture test resulted in a 1.4C rise in temperature and no expansion or pouch rupture).

These batteries are suited well for EVs and marginally for EREVs.

For EVs, the energy density is about double of what was in first generation EVs (Nissan Leaf). This means that replacing the pack with the new cells would provide almost double the range, from 75 miles to 130-150 miles. The cycle life of 1,500 cycles would provide for about 195,000-200,000 miles to 80% capacity (a nominal range of 104-120 miles). This would be a big boost for EVs.

For EREVs, the weight and size of the pack could be reduced by 1/2 over the 2010 baseline with the same electric range, or by 1/3 to achieve a 25-30% range increase. For something like the Volt, this would mean a return of the 5th seat and a boost in electric range from 38 miles to about 45 miles in a 19kWh pack using 13kWh of energy. Because cycle life is non-linear in Lithium batteries (well, it is for Li-Ion, I’m assuming that it is also that way for Li-S), by using only 70% of the battery we extend the cycle life by almost 60%, increasing the cycle life from 1500 to 2400 to 80%, which would be good for 108,000 electric miles – just barely exceeding the 8-year, 100,000 mile warranty supplied by auto manufacturers on EREVs and EVs (notwithstanding any gas-powered miles that also count under the warranty). This just-clearing-the-hurdle approach however doesn’t leave a lot of margin for environmental factors (heat, cold) and time-based degradation of the cells. Cycle life would need to be extended more (from 1500 to 2000 cycles), or cell energy density would need to be improved (more miles per battery cycle) to remedy this.

Ultimately, the biggest issue facing OXIS Energy isn’t the performance of the cell in 2014, rather what is promised by other companies for 2015 and 2016 – 300 and 400Wh/kg energy densities that will revolutionize EVs. If they can keep their Li-S chemistry competitive (and outrun Li-Ion in the mid- and long-term race), or if their competitors fail to deliver on their promises, they will be successful.

Grid Storage set for prime time?

Its been a long while since I published anything renewable energy related – I had become burnt out by extravagant promises and the green bubble bursting in 2010. But things are starting to look up – solar leasing is a huge success, more and more projects are being built using renewable energy, and even thrifty casinos are getting in on the game.

So what happens when we start to scale up to even more renewable energy generation? The intermittent nature of renewables both in the immediate (sun goes behind a cloud) and daily dispatch (sun goes down at night) requires that there be some sort of grid stabilization and storage capacity.

A slide deck from EOS Energy Storage shows off their energy storage product – a 6MWh, 1MW capacity grid-scale battery storage system. The most impressive piece here isn’t the technology or scale, its the price they’re promising. Now I don’t know if that’s their off-the-line price when they start to ship in volume down the road, or the price they are debuting in 2014 when they start to ship their product for its first field deployment, but its a very compelling price for grid storage.

However I’m not entirely ready to jump on board yet, especially with the 2014 release date since they released a slide deck about a year earlier, where they said that they would ship product in 2013 (and they haven’t). They have acquired a customer since then, which is a promising sign.


Their quote price of $160/kWh and $1000/kW and promise of 10,000 cycles and 75% round-trip efficiency translates to a cost of 2.1 cents per kWh – which is more than the typical spread between off- and on-peak wholesale (spot market) prices along the west coast, which is around 1 to 1.5 cents per kWh. However, a net cost increase of half a cent per kWh to be able to include more renewable energy into the grid is a rather small price to pay. As solar and wind prices continue to decline, renewable energy will become more of a player, and storage and buffering of that power source will be needed. It wont take much storage – maybe 3-5% of total plant capacity (250MW solar would need 7.5MW of generation from battery storage) to smooth out bumps in the grid, and around 20% (50MW) to be able to store power for after-hours usage.