The lead time on manufacturing chips is fairly large. It takes 12-16 weeks to fabricate a chip from silicon wafer to end product packaged and ready to be soldered onto a PCB. So 3-4 months. If you wanted them mid-December, you’d need to start production mid-September. If Apple wants a million chips, Qualcomm would need to get going now.

The biggest question now is not when is it released, but when is it announced. This is a calculated decision – more than even the decision to make a CDMA iPhone (which more or less falls into the DUH category given how Android is doing on Verizon and Apple doesn’t want to cede a perfectly viable piece of the market).

So how do they decide when to announce a Verizon iPhone?

1. Speculative Momentum. Every time a rumor comes out, it generates headlines. Announcing its going to be out for sure kills this cycle. You’ve only got so much to announce after you’ve made the initial announcement – things like the prices of data plans, any other terms and conditions, visual voicemail support, etc.

2. The Holiday Season. You probably want to announce it before December 1 for a corresponding January/February launch. People generally only get to update their phones every 2 years. If you announce 2-3 months prior to the release date, people will hold off long enough to get a suitable demand at launch. If you assume people get a new phone every 2 years, and Verizon has over 90M customers, that’s 3.75M customers every month that get a new phone. Let Christmas pass without an announcement, you’re likely to have some people frustrated that they just got a new phone and have to wait so long to get a Verizon iPhone. The counter-argument is that Apple is likely to be supply constrained for a while (first 3 months) and they’ll still sell every unit they make, so pumping up demand isn’t necessary.

3. FCC Certification. This used to be an issue, but isn’t as much anymore. Apple seems to have few problems these days with submitting devices to the FCC and requesting confidentiality. The only minor slip-up was the internals of the iPad ending up on a website the day before the launch in April, which isn’t that big of a deal since they would have been discovered the next day anyways. Assuming a device takes 2 months (maybe more around the holidays) then it would be submitted in late November for a late January launch.

4. An actual, factual deal – handshake and signatures. This is somewhat obvious, but they’ll need to actually come to terms and agree on things like phone price, feature set (from a phone/network perspective) and other things like what Verizon expects Apple to filter out of the App Store (network issues).

There is a lot of talk about unveiling it at CES since the CEO of Verizon has the keynote. I think that’s incredibly stupid speculation. It would be very un-Apple like for them to let a partner announce the phone. Even if Steve showed up, Apple would want to hold its own event. And January is probably too late – after the holidays and many purchasers are stuck for another year or two on other phones.

I’m inclined to pick a mid-November announcement. I think a September announcement with the refreshed iPods is possible, especially in light of Apple’s September 30th self-imposed deadline of figuring out what to do with the iPhone antenna issue. Apple could announce the iPhone 4-and-a-half in September for a January release with a physical fix, but who knows.

Bonus: If Verizon wanted to get a leg up on AT&T they’d do the WiFi hotspot thing.

I’ve been using iOS 4,0 for two weeks now, and I realize I forgot how much I dislike iPhone OS iOS x.0 releases. I’m especially disappointed that a bug from the 2.0 version of the OS is back.

Issues with Mail Part I

First issue is that iOS (since 3.0) doesn’t play well with POP3 accounts. iOS does not close its POP3 connection to the server, leaving it open and letting the server timeout the connection after a minimum of 10 minutes (per the RFC that governs POP3 mail servers). How do I know this? I run my own mail server and look at the logs myself – iOS on the iPhone or iPad does not close its POP3 connection when you close the Mail app on the phone. It just leaves it open and lets it timeout. On the other device (or the desktop) the system will report an error about being unable to log into the POP3 account.

This is an issue when you have multiple devices checking a POP3 account every 15 minutes – an iPad, an iPhone, a desktop PC and a laptop – and half of them don’t play nice with each other. For the uninitiated, you can only have one client open a POP3 mailbox at once – so until that 10 minute timeout passes you cant log in with a different email client.

Issues with Mail Part II – Electric Boogaloo

The “I’m not checking mail automatically” bug from iOS 2.0 has returned with a vengeance. It only took about 12 hours after the initial install of iOS 4.0 before I noticed that my iPad was getting new emails but the iPhone wasn’t. I played around with the phone and realized, yup, the same problem that I had with iOS 2.0 has returned and there isn’t much I can do about it other that spend a weekend switching from POP3 to IMAP (not what I was looking forward to) and hope that it solves my issues (it should since IMAP allows multiple simultaneous connections).

For reference, if anyone at Apple ever reads this and cares enough to fix it – I have 4 POP3 accounts and an Exchange (Google) account on both an iPad an iPhone set to 15 minute intervals. Switching to IMAP fixed the POP3 issues.

Still no useful AVRCP support

AVRCP is a remote control protocol over Bluetooth. The iPhone supports a very limited set of commands – play/pause/stop. However other devices, specifically my Ford Sync system in my car, support more advanced AVRCP implementations that support next/previous, searching, etc.

Whats that noise in my pocket?

Twice in three days, when I’ve disconnected my iPhone from the USB port in my car (Ford Sync system), the iPhone continues to play music out of its speaker instead of stopping (which is what it used to do, and what it should do). If I don’t notice it immediately I’ll put the phone in my pocket and then hear this background noise…

Battery Life

When the iPhone was checking mail regularly, it ate through battery life quicker than with OS 3.0. Nothing streaming or going on in the background, just treating my phone like I did before. I turned on the battery percentage indicator to get a better feel for it, but around that same time the mail app stopped checking mail automatically.

Multitasking

Not much multitasking to speak of – the apps that support it haven’t been released to the App Store yet. There is a neat animation to illustrate when you’re switching between apps so you know the app is still running in the background.

Charging $15/mo for data plans? Moving up six months (or possibly more) worth of upgrade eligibility to now?

AT&T knows they’re losing exclusivity. The question is when. Based on the six month window they provided, I’d venture to say we’ll see the announcement in September (along with the other iPod upgrades and a rumored iPad upgrade), with release in early November – just in time for the Christmas shopping season.

Previously, in January 2010 at CES in Las Vegas, AT&T announced a network improvement plan. This plan was supposed to be completed by the end of the year. At the D8 conference last week, Jobs stated that things should be fixed by the end of summer (September). Moving up their time frame 3-4 months (20-25% of the project time frame) is a huge change from a project management standpoint. Something significant must have made AT&T change their plan.

AT&T does have the advantage over a Verizon or Sprint iPhone 4 – namely that HSPA+ (7.2 or 14.4Mbps) is faster than CDMA can ever be. Its why Verizon is going to LTE before AT&T is – Verizon needs LTE to get user speeds past 1.4Mbps. AT&T is about that right now with HSPA 3.6 and users could see 2Mbps with HSPA 7.2, possibly more if AT&T’s backhaul can handle it (in Canada, users have benchmarked speeds up to 3.5Mbps on HSPA 7.2 using the same GSM frequencies AT&T uses in the US). The fly in this ointment is that T-Mobile could end up with a faster 3G network than AT&T (they’re moving straight to 21Mbps HSPA+, however the iPhone 4 cant handle speeds that fast).

Whats Known

We know of the new design, the rear and front-facing cameras. Of the slightly larger battery. The possibly radio-transparent back of the case. The internals powered by the A4 CPU that Apple has worked on for a while now. The higher resolution screen was recently confirmed.

What It Means

We’re certainly in for video conferencing. What remains to be seen is how it is structured. Does it work over WiFi only, or will it work on 3G? Does it go over the cellular phone network or the internet? Does AT&T heap another $20/mo per phone to actually use it? How does it work with non-iPhone devices? Or is it just there for use in non-phone  applications like Skype or iChat?

The radio-transparent glass back of the case is one of the more intriguing pieces for me (the sort-of electrical engineer). If Apple has devised a material which is radio transparent, they’re looking to improve battery life. Since Infineon hasn’t progressed much in their 3G baseband chip (still fabricated on a somewhat old RFCMOS 130nm platform), Apple has to step up and figure out ways to reduce the battery usage of the rest of the device. Reducing radio transmission power will help battery life (the RF transmission is usually between 50-100mW depending on the strength of the signal needed to get to the nearest tower). Likewise, the ultra power efficient A4 chip will help in consuming power of the device. All these enhancements combined with the larger 1400 mAh battery (5Wh) its possible we see 3G call times and 3G internet usage approach 10 hours, WiFi internet usage up to 12 hours and video playback times reach 15 hours. The sales of Mophie Juice Packs has probably convinced Apple they need better battery life – to go all day at work and all night out having fun.

But What Else?

I haven’t seen anyone talk about the Infineon 3G baseband chip. The chip currently used is presumably the same Infineon Smarti UE chip, Infineon hasn’t released a new one since 2007. It supports up to 14.4Mb/s HSDPA (downstream) and 5Mb/s HSUPA (upstream). This is twice as fast as the current iPhone supports (7.2Mb/s, real world speeds around 1.5-2Mb/s on adequate networks – not AT&T, 14.4Mb/s would probably provide real world speeds of 3Mb/s).

Other networks are possible. Just like how Apple announced their original iPhone in 2007 six month ahead of time (for several reasons), Apple would likely pre-announce the phone coming to other networks. I would expect that any pre-announcement would be about 4 months ahead of time, and Apple would try to get the phones out in October or November in time for Christmas (were this to take place, I’d put AAPL at around $350 by the end of 2010). The two main reasons to pre-announce are to prime subscribers of those networks (to keep them from upgrading their current phone to a new one) and to avoid it being announced through either the FCC or through leaks of phone company employees due to the necessary engineering work that would need to go in to supporting the iPhone (visual voicemail, possible video chat, etc). I expect Apple will need to support all four major US cell phone carriers within the next 9 month because of the gains the Andriod platform is making on them because the iPhone isn’t available through every carrier. In the desktop model Apple is happy to keep their margins at 40-50%, the phone however has much larger margins – an iPhone 3GS cost around $190 to manufacture (in mid-2009) and probably more for licensing costs but sells to AT&T for $599, a $400 subsidy. Would carriers be willing to subsidize the phone as much if its available on all four networks? Verizon sells the HTC Incredible for $200 (with $325 ETF) subsidized and $599 unsubsidized. We’ll see…

I certainly do not see support for LTE or WiMax in this generation of phones, mostly due to the fact that there are no networks ready. Apple held off on 3G for a year because the AT&T network wasn’t sufficient. I can see them holding off until 2012 when LTE for voice and data are fully supported on the 3 main LTE networks (AT&T, Verizon, T-Mobile). What I do see happening, coupled with the unveiling of iPhones for other carriers is AT&T going live with 7.2Mb/s or even 14.4Mb/s data networks before the end of the year in select cities (top 25 markets, plus select smaller ones). By providing the user with faster data access, AT&T looks to set themselves apart from Verizon. T-Mobile can also offer this, but their status as a “mid-major” carrier probably doesn’t threaten AT&T.

We’ll find out in just about a week. It’ll be a long seven days.

So since 4.08kg of pure (elemental) lithium is equivalent to about 21.75kg of LCE (5.33kg of LCE is 1kg of Lithium). This means that the 24kWh battery contains 21.75kg of LCE, or about 0.9kg of LCE per kWh. This is 50% larger than the 0.6kg I had seen others cite as a figure for LCE per kWh, which would diminish the amount of batteries my in my estimates by 33%. But it seems there still will be plenty of LCE to go around since the numbers before were absurdly high.

WLC, in a recent press release, also stated that their target for their stage I lens production is 27,700 metric tonnes per year of Lithium Carbonate (LCE, or chemically Li2CO3). The math to turn that into the number of EVs is easy – 27,700 metric tonnes is 27,700,000 kg of LCE. In 1kWh of a Lithium-Ion battery there is 0.6kg 0.9kg of LCE (figures in this article have been updated). This means that 27.7M kg of LCE per year is about 30.8M kWh of batteries that can be produced. They have an expected 18 years at this rate of supply to mine (approximately 500,000 metric tonnes LCE total).

In a pure EV (like the Nissan Leaf) the battery is 24kWh, so from 27,700 tonnes of LCE comes 1.28M Nissan Leaf battery packs per year. In a EREV like the Chevy Volt, its battery pack is 16kWh, so 1.9M battery packs would be able to be manufactured for the Volt.

To put these numbers in perspective, in 2009 there were a total of 10.4M cars sold in the US, and in 2008 approximately 13M cars sold. So this single lithium mine could power up to 15% of all the US EVs and EREVs sold, if the automakers could build and sell that many (which they wont, at least initially).

So the question is, how many tonnes of LCE would it take to make every car sold in America a plug-in? From a small two-mode system that would allow for 8-12kWh batteries for 10-15 miles at speeds below 60MPH, all the way up to pure EVs with 50kWh batteries. If we assume that 70% of cars sold are two-mode at 10kW, 20% are EREV (18kWh) and the last 10% are pure EVs (35kWh avg), the total kWh for a year of 14M cars is 197.4M kWh, or 177.3M kg of LCE. So in order to produce enough LCE, we would need to produce about 180,000 metric tonnes of LCE, or about 6.5x the amount of stage I.

The stage II lens has approximately 1.365M tonnes of LCE, and at 180,000 metric tonnes per year, it would be exhausted after 7.5 years, assuming the production rate could be sustained.

Seven and a half years might not be a long time, however there are still several other stages to this mine area (stages three and four), plus there are other lithium mines in the Nevada and the US. It appears that Lithium supplies wont be a blockade on the road to electric cars. While Li-Ion batteries can also contain other precious metals that might be scarce, Lithium shouldn’t be an issue.

[Update 5/27 - updated Lithium math based on Leaf's use of 0.9kg/kWh]

Nissan this week announced that their LEAF EV will be priced at $32,780 in the US before a $7,500 tax credit for electric vehicles. The 220V charging dock (required if you’re traveling more than ~50 miles per day due to recharge times from 110V/8A circuits) is an additional $2,200 (also eligible for a 50% tax credit). The total price (vehicle + 220V charger) after all tax credits is $26,380, and possibly as low as $21,380 is California, Georgia and Oregon due to $5,000 state income tax credits.

Nissan Japan has set a retail price of approximately $40,700 including VAT. After government incentives cost around $33,000.

The most interesting prospect is a $349/mo lease, plus $2,000 down. Further terms and conditions of the lease haven’t been specified (mileage, etc), but considering that electricity is pennies compared to gasoline, even for a 30MPG vehicle, people could get used to paying $30/mo more on their electric bill every month in exchange for no more gas stations. Though it makes me wonder what Nissan would do with all those lease turn-ins in 2014 with a comparatively out-of-date battery and larger EVs.

Mitsubishi responded by cutting the price of their iMiEV electric car by $7,000 in Japan in response to the lower price of the LEAF.

The low price will put pressure on other auto makers to push the price of their electric car downwards. Suddenly, an all electric sedan doesn’t look too appetizing if it is priced over $30,000 (after credit). Specifically, the price that GM will set for the Chevy Volt will likely be pushed down slightly from what they might have been expecting to sell it at. Even though the Volt is essentially an unlimited range electric vehicle due to the gas take and electric generator, GM’s difficulty will be conveying how the Volt works and explaining its benefits over EVs to the general public. An old political axiom applies – if you’re explaining, you’re losing.

It does look like a world of hurt for niche EV makers, Tesla might survive due to its luxury status, but more… exotic cars like the Aptera are likely to see harder times ahead due to price ceilings. The real question is battery supplies. How do the cost of batteries change when mass-manufacturing hits, and the demand for automotive batteries starts to greatly outstrip supply. Do we see prices go up, or can sufficient quantities be made with the existing supply chain?

The Nissan LEAF EV is expected to start selling in limited markets at the end of 2010, and nationwide sometime in 2011.

The raw numbers are very interesting – a theoretical maximum of over 2,000 Wh/kg, with initial numbers (though not available for many years) around 700Wh/kg. This figure is a huge leap above current technologies. Given “Musk’s Law” of 8% per year, we wouldn’t hit 700Wh/kg until about 2030. As long as they can deliver sometime in the next 5 years it would provide for a new level of energy storage capabilities. Batteries would reduce in volume by 50%. Size and weight would no longer be the issue (how to recharge such batteries quickly would still be).

The batteries also offer superior thermal behavior – down to -60C and up to +160C. This would be beneficial to the automotive applications where cold weather performance can be an issue.

The real question is how fast can they make a product and make it manufacturable. Contour hasn’t even bothered to announce any target dates for product delivery on their website or available documents. They have the talent and are attracting venture capital, so we’ll see how quickly they can turn their lab work into factories putting out large volumes of their unique chemistry.

You’re not Apple. Stop trying to be. You haven’t earned 1/10th of the mind-share Apple has earned for creating their simple, easy to use products. So cut the ridiculous hype and earn it the hard way – by providing a simple user experience that still delivers the important features while leaving behind the 10% of features that are used less than 1% of the time.

Hyperbolic invites and talk about how you’re going to change the world when you’re just releasing a new revision of an existing product make you look stupid. If you aren’t even going to bother entering or creating a new product category, you should just frame it in more humble terms (our best X ever!). It looks like the lesson of the Segway has been forgotten or willfully ignored. Even the short-lived hype around the recent Bloom Energy generator should be instructive on how not to unveil products, or at least how to do it in a reasonable manner without promising to change the world and fall short.

Elon Musk (Tesla CEO) postulated a “weak Moore’s law” for Li-Ion batteries, that the price/performance ratio will increase by 8% per year, or 9 years to double. The price/performance ratio is the ratio between the price per kWh of the battery pack and the amount of energy the battery can store. If current batteries can store 140Wh/kg and cost $500/kWh, an 8% improvement means either the storage goes up to 150Wh/kg, the price goes down to $460/kWh, or somewhere in between (145Wh/kg and $480/kWh). A Tesla battery pack would go from $35,000 (53kWh at $650/kWh in 2009) to $24,000 ($400/kWh in 2014), a reduction of about 10% of the entire price of the car over approximately 5 years. Combined with other cost saving methods, the next stage of the Tesla evaluation - the Model S – starts to look feasible. Its still not going to be the most affordable car, however significant progress is being made.

The cost per battery pack can be broken into two parts – the batteries themselves and the pack. The pack costs can be trimmed considerably with mass-manufacturing. Instead of hand assembling each battery pack and set of battery modules (a series of cells), semi-automated assembly can increase the throughput of the teams assembling dramatically while keeping the same number of people around, reducing the amount of employee-hours spent per battery pack.

The cell costs don’t come down as easily. This is the decidedly slower part of the electrification of vehicles. Following the 8% rule, automotive battery packs due in 2009 cost approximately $650/kWh. In 2014 this cost is about $430, and by 2017, the cost is $330/kWh, and by 2020 $260/kWh. Following the more agressive price decreases noted above, prices in 2017 would be $235/kWh, and by 2020 $172/kWh.

So by 2020, a Volt-style battery would cost $4,200, or about the cost of a new engine (a rebuilt one can be had for less). This assumes that other battery performance parameters do not improve – rather the Volt still requires a 16kWh battery and only uses 8.8kWh of the battery pack. If the current estimates of what battery specifications will be by 2020 (2,500W/kg, 250Wh/kg, 2,000 cycles and 4,000 recharges at 70%DoD) the Volt would be able to have its pack size reduced to 12.5kWh (50kg, 110kW), thus reducing costs further to $3,250 for the battery pack, and the total price premium of the E-REV system would be approximately $5,500. Factoring that cost over 5 years is $1,100 per year in savings needed over gasoline, which is achievable when factoring in savings in electricity costs over gasoline (approximately 9c or 11c/mile savings depending on cost of electricity), reduced maintenance costs ($150/yr for oil changes, etc) and reduced variability of fuel costs – my electric company needs a regulatory body’s approval to change the price of energy, the local gas station chain can add 10 or 15c to the price of gas over a holiday weekend because they feel like sticking it to us.

By 2030, barring any new technology that would leapfrog Li-Ion on price and performance, battery prices would reach $110/kWh, and total costs would be equivalent to a Prius premium today.

Over the long term, E-REVs are workable from a consumer finance standpoint. Initially, subsidies, longer warranties and extended payback periods will be needed to entice the consumer to buy in to the electrification of vehicles. If we can manage to stick with it for the next 5-7 years, it will take off and the nation can start to wave good-bye to oil and petroleum for their in-city commutes, and we’ll all breathe easier with less smog.