GBatteries and their BatteryOS and BatteryBox

Today, I’m going to look at GBatteries, a startup incubated at YCombinator in 2014 that is now based in Ottawa. Their first consumer product, BatteryBox, was launched on Kickstarter last year, and as of January 2015, they were shipping. BatteryBox is a 60Wh (~14,000 mAh @ 3.7v) backup battery pack for Apple MacBooks and USB powered devices.

Last summer, I decided to dig into the subject of lithium ion battery reuse. Tomorrow’s used batteries are today’s new batteries, so I’ve been learning what I can about what today and tomorrow hold for new batteries. Its really too much for one person to go into in great depth, but I’m still trying to get a better lay of the land, and part of that effort is looking at battery-related startups like GBatteries.

If you visit the GBatteries site, the first thing you see isn’t the BatteryBox, instead you see a plug for “BatteryOS” their new battery management system that “increases battery capacity and cycle count.” It goes on to say:

BatteryOS is a new way of controlling rechargeable lithium-ion batteries that enables any normal battery to have higher capacity and not degrade over time.

It’s clear then that BatteryBox is an application of their new technology, a way to prove the value of BatteryOS value while bringing in some revenue, with he longer term goal of licensing technology and/or selling components that device makers will incorporate into their own products.

Digging in a little deeper, we find that GBatteries gets more specific about the claims about BatteryOS:

 10-40% more capacity (varies between specific type of li-ion chemistry), have a 4x higher cycle count, and not degrade in capacity over time.

In considering these claims, it helps to understand a bit about industry standard approaches to lithium ion battery design and management. To start, lets define key factors:

  • Energy capacity / density
  • Power delivery
  • Lifetime
  • Safety
  • Cost

These factors are all, to some degree, interdependent, which allows designers to trade them off aginst each other. For portable consumer electronics devices, like phones and laptops, the working range fall into a relatively narrow space.

Within that space, one tradeoff designers make is energy capacity vs cell lifetime. A simple choice offered to designers is to sacrifice ~10% of the overall energy capacity of a battery in exchange for a doubling of its lifetime. This is usually accomplished by limiting the maximum voltage allows when charging, and the minimum voltage when discharging.

With this background, we start to understand how BatteryOS might deliver on its promises:

  • 10-40% more capacity (varies between specific type of li-ion chemistry)
  • 4x higher cycle count
  • not degrade in capacity over time

What they seem to be saying is that a device designer can have their cake and eat it to. Rather than giving up capacity for lifespan, or trading lifespan for capacity, they can have it all, and more.

Gbatteries claims that they have patents pending, but despite starting the company in 2012, and participating in YCombinator a year ago, I can’t find any relevant patents for father and son co-founders Tim and Nick Sherstyuk, so it is hard for me to evaluate their claims in more detail.

The BatteryBox would seem to be a chance to test these claims. They’ve been shipping them out to Kickstarter supporters for the past few months, but so far, I haven’t been able to find any independent tests and I’m not in a hurry to buy one and reverse engineer it.

We can, however, evaluate the published claims about BatteryBox BatteryOS, and other products and see if everything makes sense.

Salient details about BatteryBox:

  • $219 price ($129 to kickstarter backers)
  • 60Wh from 5 pouch cells
  • 49x95x63mm
  • 404g

Lets start with the price. There are a number of external battery packs available for laptops, but only a handful that work with the MagSafe connector on Mac laptops.

Hyperjuice sells a 360g, 60Wh battery pack that works with Macs for $199 that includes a DIY kit to adapt your existing MagSafe cable, for $249, you can get one with an already modified MagSafe adapter.

Voltaic Systems sells a 72Wh pack for $129, and a modified MagSafe cable for $20. It weighs 558g.

Both HyperJuice and Voltaic are bundling a MagSafe connector with rebranded external battery packs designed and manufactured by someone else, and similar packs are available from other brands as well, such as this $99 Anker-branded version of the same 72Wh pack Voltaic is selling.

Adapting MagSafe cables isn’t officially sanctioned by Apple, but clearly, HyperJuice and Voltaic have been willing to take the risk, and seem to have achieved some success over the years. For some reason though, Gbatteries decided to go another route with BatteryBox, developing a non-magnetic, MagSafe compatible connector that clips on to the side of the laptop.

I don’t know why Gbatteries decided to take this route. It may be the risk-intolerance of their funders, or perhaps they want to avoid antagonizing Apple because they hope to sign them up as a BatteryOS customer in the future.  The advantages to consumers over a $20 MagSafe cable from Ebay aren’t apparent.

What is obvious is that for 55% the price of the BatteryBox, one can get a $20 cable and a $99 Anker pack with 20% more initial capacity.   The Anker pack is also 20% heavier, but its thin form-factor makes it easier to tuck into a laptop bag than the BatteryBox.

From a cost, capacity and density point of view then, BatteryBox’s use of BatteryOS doesn’t provide obvious benefits over the status quo at time of purchase. Of course, that’s not everything it promises.

The other benefit BatteryOS is supposed to provide BatteryBox users is a 4x longer lifespan. Lithium ion batteries wear out due to a combination of time and use. That wear generally shows itself in a few different ways: increased internal resistance, decreased charge/discharge rates, and decreased capacity. In ordinary use, the decline is gradual and continuous, every charge discharge cycle results in a tiny bit of wear.

GBatteries is rather inconsistent in the way it talks about the battery lifetime benefits of BatteryOS. In one sentence, they claim both 4x the cycle count AND that the capacity won’t degrade over  time. And yet, they also show a graph of runtime vs use that shows that the capacity of cells in BatteryBox decline noticeably over time, though at a lower rate than cells managed by traditional means.  Its also worth noting that the graphs they show don’t start the Y axis at 0, which exaggerates the differences between the alternative management techniques.

In the long run though, it seems that the 20% capacity advantage and the almost 2x cost advantage of the non-battery-box solution goes a long way to offsetting the purported advantages BatteryOS brings to the BatteryBox.

Looking closely BatteryBox does help clarify some of the claims about BatteryOS. GBatteries claims significant capacity AND cycle-life benefits from BatteryOS. BatteryBox makes claims about significant cycle-life benefits, but the capacity tracks closely with that of other devices with the same weight, suggesting that the claims for BatteryOS don’t make the remaining tradeoffs clear.

While BatteryOS isn’t enough to make BatteryBox a compelling product, and while it may not provide all the benefits claimed, at least not at the same time, it might still be an important improvement to battery management, and provide a compelling benefit when incorporated into new versions of existing products.

Even so, I think GBatteries and BatteryOS face an uphill battle, because it isn’t clear to me that they bring anything new or unique to the table.

The lithium ion battery market was long dominated by the needs of consumer electronics devices like laptops, cameras, and mobile phones. The overall pace of progress in these product areas has been rapid, which drives relatively short upgrade cycles (2-5 years). When lithium ion batteries became durable enough to go mainstream, device manufacturers expected them to be replaced once or twice over the lifetime of the product. Since then, capacity improvements have been given at least equal weight with durability advances. Durability has improved enough that they are now expected to last the lifetime of the product they power, shifting even more emphasis to capacity improvements.

In this era, anyone needing greater durability often found it made more sense to work with cells from the consumer electronics market trading a bit of their superior capacity and cost for better durability.

This is exactly what Tesla did when designing the Tesla Roadster, which used 18650 batteries designed for laptops. Along the way, they developed their own techniques for better extending battery lifespan. In addition to the well-worn technique of limiting charging and discharge voltages to trade some capacity for durability, they also use a new charging profile. The traditional charging profile for LiIon batteries is to use a constant current (often the current required to fill the battery in 2 hours) to charge the cell to the voltage limit (typically 4.2-4.35v, depending on cell chemistry), and then hold the voltage constant until the current has drops to some predetermined threshold (often 1/10th the initial charge rate).

Tesla’s approach is to add another charging stage at the beginning of charge. This charging phase uses constant power, up to a voltage level below the final charging voltage target. This results in a decline in charging current as cell voltage climbs. This helps reduce heat generation at higher cell voltages, avoiding the combination of conditions that lead to the fastest cell wear. Once a threshold voltage is reached, charging switches to constant current to the final cell voltage which is then held until the current declines to its cutoff threshold. I expect that other companies have developed their own novel battery management schemes.

Of course, it doesn’t mean that GBatteries doesn’t bring something more to the table, but its also worth keeping in mind that some of the advantages they bring might be either offset or reduced by improvements in other areas. The most obvious is improvements in that batteries themselves. As I explained above, Tesla started by using cells from the consumer electronics industry for the Roadster.

For the Model S, they worked with Panasonic, their battery supplier, to design a cell for automotive use. It should be obvious that automobiles have longer lifetimes than consumer electronics by a, and that electric cars need to prove that their batteries will meet expectations set by existing automobiles. One obvious improvement for vehicle batteries then, is durability. In addition, cost and capacity will remain important considerations. The result I think, is that applications that need more durability than available from consumer electronics cells will now have the option of working with cells developed for the automotive market.

In the end though, the market for Lithium Ion batteries is so large already, and will quickly grow even larger, that an innovation that can offer a 10% improvement to 10% of the market could be very valuable. Whether or not BatteryBox was a good first step is in doubt. It targets a small portion of the overall market. It isn’t a compelling product in its own right, which makes it less compelling in demonstrating the value of BatteryOS.  On the other hand, the experience they gained in developing and selling BatteryBox will probably help them better understand the issues that device makers, the potential customers for BatteryOS, deal with.

 

 

 

Notes on JB Straubel’s Keynote at the 2014 SEEDZ Energy Storage Symposium

I thought I’d share my crude notes on a very interesting talk about energy storage from Tesla CTO JB Straubel. The slides are available as a PDF.

  • JB Straubel Keynote
    • CTO, Tesla
    • Previously CTO & co-founder of Volcom, specializing in high-altitude electric aircraft platforms
    • Board Member, Solar City
  • Why is Tesla Deeply invovled in Energy Storage
    • EV Battery History
      • Status quo was led-acid in 1995
        • Stagnant performance
      • LiIon made possible Tesla and the rebirth of EVs
        • 4x gravimetric energy density
        • 6x volumentric energy density
        • 2x cycle life
      • Tesla was first to do Li-ion R&D for vehicles in 2003
    • Tesla Roadster 2008
      • Sought to challenge internal combustion vehicles on more than just fuel economy
      • ~2,500 sold
      • Largest battery pack in a vehicle
        • 50kWh +
      • Optimism based on progress of LiIon cells
        • ~2x improvement in energy storage capability of batteries in the decade between EV1 era and the Tesla Roadster
          • 300Wh/litr to 600Wh/litre
          • Accompanying weight reduction as well
        • In talking to companies involved in 2005, they saw no reason for progress to plateau
        • Ten years later, they expect progress to continue for the foreseeable future.
        • 40% improvement between introduction of the Roadster and the Model S
          • ~300 miles range
          • 85kWh
          • Smaller pack than the roadster
        • See this trend continuing for the next 10-20 years
        • Only 10-20% away from being able to compete broadly with the internal combustion engine.
    • Long term goal is to sell millions of cars
      • Energy storage is the biggest factor influencing cost, and therefore, volume.
    • Current offerings
      • Roadster
      • Model S
        • Main focus
      • OEM
        • Toyota RAV 4
        • Mercedes-Benz
          • BClass
    • Model S Battery Pack
      • Scale & Scope have enabled a cost point that is competitive in other markets outside cars
      • Utilities
      • Energy density is a key path to lower cost
        • Automobiles
          • Not intuitive
          • Some automobile manufacturers have tried to cut energy to cut costs.
        • 200Wh/Kg
        • Benefits for stationary storage
          • smaller footprint
          • easier retrofit
      • Current costs and projections are on track
      • Manufacturing volume is important too.
        • Goal to build 500,000 Gen3 vehicles/year
          • using capacity in existing Bay Area Plant
        • Energy storage required to meet that goal is another story
        • In 2013 ~34 GWh of lithium ion batteries were manufactured, but from ~21 GWh in 2010.
        • Tesla’s projects to use 3-4GWh in 2014
          • Approximately 10% of WW volume
        • 500,000 Gen3 vehicles/year will require 10x current, or ~35GWh of batteries in 2020.
        • Gigafactory sized to meet that demand
          • Would still only supply 0.5% of WW new car market.
          • Doubling WW capacity
          • Reengineering supply chain with major opportunity for cost cutting
          • 30% cost reduction by 2017 ramp-up
        • Thinks that the market for non-automotive storage will grow even faster than the electric car market.
          • Grid storage is slow to mature, but once you cross certain price thresholds, it becomes a commodity market that sells on cost savings
        • Approach to cost reductions at the gigafactory
          • Looking at everything, including the sourcing of the raw materials.
          • Materials are a meaningful portion of the cost of cells
          • Plan to use their purchasing power to use demand as leverage for sustainable practices throughout the supply chain (labor practices, energy, etc)
        • Gigafactory
          • 35GWh of cells
          • 50GWh of packs
            • 15GWh for stationary packs
              • 10x of the California mandate
  • Tesla and Stationary Storage
    • ~2012 adapted vehicle pack architecture for residential energy storage packs
      • 5kW/10kWh
      • Originally targeted at people with solar and teslas
      • Slow to scale
        • ~1000 systems
      • People who wanted to play with energy in their house
        • Peak demand management
        • “Islanding” to isolate their system from grid and allow operation in the event of a power outage
      • Experiment with aggregation of systems
        • Coordinated across households
        • Business models are a challenge
    • New experiment is a larger-scale system
      • Module
        • Based on model S pack architecture
          • Different arrangement of modules
          •  Same cooling and management
        •  Two hour-rate pack.
          • 200kW / 400 kWh+
        • “Single skid” unit
        • Roughly the size of a shipping pallet
        • 800Kg
      • Pilot deployment
        • SuperCharger location
          • Big enough to be worth the time
          • Very peaky load
            • Up to 250KWh peak demand from cars
            • <100KW peak from meter
      • Thinks that utility side opportunities are significant
        • Backup power
        • Demand management
        • Scale up of renewables
          • This will be the major driver for storage!
        • Major adjustments needed in rate structures/incentives
        • es- I wonder where trends in peak generation costs vs storage costs cross
      • Large Pilot (2013)
        • “One of the bigger installations”
          • Tesla’s Fremont plant
          • 1MW / 5 modules
          • Manages ~10% of peak demand
          • 100 MW substation with 100kV tie.
          • 1-2% of installed grid-attatched storage
          • In the process of doubling capacity
  • Thinks of Tesla is an energy innovation company, more than a car company.
  • Cautionary note about building for economic sustainability, avoiding dependency on incentive programs
  • Innovation needed in storage management technology and relationship to grid and generation
  • “The stone age came to an end not for lack of stones, and the oil age will come to an end not for lack of oil”
  • Thinks the economics of renewables are already compelling and now constrained by storage
  • Q&A
    • Can the world support the scale up from a primary resource point of view
      • Li-ion are primarily graphite/carbon for anode, nickle for the cathode (for their chemistry). Steel for can, polymers for separators, organics for electrolytes
      • Predictions of shortages of lithium aren’t well founded
      • Thinks there is a bubble in lithium prices, but Li-ion costs aren’t the major contributor to lithium cell costs
    • Have you thought about the policy changes needed to drive this change?
      • They have been building relationships with CPUCs and utities and even state and fed regulators, but it isn’t a big focus.
      • Tesla is an innovation company that focuses on creating products that drive the need for regulatory changes.
      • Utilities are conservative. Automotive experience is demonstrating that this technology is ready for utility use.
    • Recyclability of lithium ion batteries
      • Recycling is being built into the Gigafactory. With enough volume there are great opportunities for materials reuse.
      • Challenge now is that growth is so steep that the volume of cells coming back is so much lower than current production
    • What is the commercial lifespan of Lithium Ion batteries?
      • Lithium ion batteries for the next 5-10 years
      • Not waiting for the next big thing…
      • Lithium ion cells make business sense now, which is why they investing.
      • Looking historically, improvements aren’t necessarily revolutionary.
        • Often it is through incremental improvements to components of the system that can leverage a lot of existing infrastructure for building cells and packs
          • Better cathodes
          • Better electrolyte
          • Better separators
        • Thinks we can double energy densities with better anode and cathode materials over the next 10 years.
    • 18650 vs other formats
      • Thinks people are hung up on form-factor for no good reason
      • At the scale they are dealing in what matters is whats inside.
      • What matters
        • Cost of materials
        • Thermal / Safety parameters
      • Their conclusion is that relatively smaller form factor are a reasonable optimization tradeoff between safety, thermal and cost.
        • Slightly larger than an 18650
        • Cylindrical makes sense from a cost perspective
      • Some applications will have different characteristics
      • Large cells just move complexity inside the cell
    • Deployment plan and pricing metrics for fixed storage
      • Depends on market
        • Residential
          • Leased to customer, bundled with larger offering
        • Mid-market commercial customers
          • Many prefer to buy outright
          • Others want Tesla to provide savings, take care of utility interconnect, etc
          • Some people don’t want to arbitrage their energy
            • He did it at his for a while. Got bored with it
          • Thinks at the end of day, Tesla or a 3rd-party financing company will end up owning most of the packs and be responsible for management
        • Utilities are another matter, given the scale and dollars involved.
    • Have you developed a center to deal with dispatch of modules?
      • They are building it.
    • Failure rate and warranty for modules
      • Quite reliable
      • Automotive application is much harsher
      • Maybe over-engineered for fixed application
      • Warranty rates are great
    • Degradation profile
      • Similar to cars
      • Optimized for 10 year life
      • Avoid paying too much for a product that outlasts the product, or the pack.
    • What are your perspectives on the power converter, are you building it yourself?
      • System design and integration are their major focus, just as with automobiles.
      • The power electronics for Tesla Model S is a very capable starting point for these fixed storage modules.
      • Costs for the power electronics are dropping quickly, will see less than $0.10W very quickly.

Apple’s Batteries, Apple Cars?

Earlier this year, a series of rumors and reports emerged that Apple may (or may not) be making serious investments in developing an electric vehicle. Taken as a whole, the rumors and reports are intriguing, if not convincing, but there are also plenty of alternative explanations for the circumstantial evidence.

This isn’t a site about Apple though, or cars, or technology in general, it is about batteries, and my purpose with this post is to note some interesting tidbits about Apple and batteries that have emerged in the attention to the Apple Car.

First, Apple has been recruiting people from battery companies. They are being sued by A123 for hiring some of their employees. A123 is a battery company specializing in Lithium Iron Phosphate (AKA LiFePO) cells. The employees in question were A123’s former CTO, Mujeeb Ijaz, and various engineers responsible for developing and testing lithium ion batteries. A123’s earliest design wins were in batteries for tool packs, but there current focus is squarely on providing batteries for the growing electrical vehicle market. They also faced bankruptcy recently, so it is not surprising that their employees were open to be recruited to working on battery technology at Apple, whether or not those batteries will be used in electric cars.

In the complaint, A123 claims that Apple has hired employees from other li-ion battery companies A123 has ties to, including SiNode Systems, Samsung, Toshiba, LG, and Panasonic.

Apple is clearly interested in advancing the state of the art for battieries in service of its publicly known product lines. The design of the upcoming Apple Watch seems strongly influenced by the issue of battery life and the new MacBook goes to extraordinary lengths to maximize the volume of the device occupied by the battery.

This isn’t anything new for Apple, but they’ve gone to extraordinary lengths with the new MacBook.  Apple used to use battery packs built around the cylindrical, metal encased, 18650 cell, but about 10 years ago, they were pioneers in the use of pouch cells, which use a plastic pouch, rather than a metal can, and are shallow rectangles, rather than cylindrical. Apple’s move to pouch cells allowed thinner systems with less space wasted in the baps between cells. The followed up by forgoing field-replaceable batteries, for battery packs installed in the main cavity of the computer. The space saved was used to increase battery capacity. The Retina Macbook Pro went even further, gluing the cells directly to the laptop case. It also used a mix of cell sizes.

The latest MacBook goes even further. The logic board is tiny compared both to past Apple laptops, and the batteries, which occupy most of the volume and floor-plan of the computer, are unprecedented. The cells aren’t simple rectangles, they have rounded and notched edges.

Screen Shot 2015-03-17 at 3.36.43 PM

For even more battery density, they are stacked in a terraced fashion.

Screen Shot 2015-03-17 at 3.36.51 PM

Whether an Apple car makes it to market, if it is being developed at all, I am certain that Apple wants to stay on the cutting edge of battery development, whether in packaging, chemistry, management or manufacturing. Whether it is Apple, or its suppliers in the lead, Apple needs people who know all about the development, manufacture, and use of lithium ion batteries. It is no surprise they’ve been hiring them.

 

Used Tesla 444 cell 6s74p Modules for Sale

Last night, a new member on the EEVBlog forums posted that he had  ten battery modules from a Tesla Model S for sale. The asking price? $1,900 each, shipped within the US.

According to the poster, it is a rev b pack from a car with ~4,000 miles on it. Each module has 444 18650 cells, configured as six in series, 74 in parallel (6S74P). Capacity is 233Ah and 5.3 kWh. Each module has its own Tesla BMS (battery management system), and plumbing for heating/cooling.

I also found another thread about the same modules in in a DIY electric vehicle forum. There is more discussion there, including some testing results by one of his collaborators. And here is a thread on another EV forum where the collaborator posted earlier in the month about hunting down the source pack and break down some of the modules into cells for sale.

Doing a little math, this works out to $4.28/cell, $0.12/Ah and $2.78/Wh. Back in November, I came across a Hack-a-Day interview with the first person I know of to publish a Tesla Model S pack teardown. I noted that he’d paid ~$20 for his pack, which worked out to about $3/cell, so, this is ~40% more expensive/unit, but with the advantage of 1/10th the initial outlay.

The cells in these Tesla packs are a variant of the Panasonic NCR18650 cells. The exact variant isn’t known, and is probably Tesla-specific, but there are NCR18650 cells with similar capacity on the market. Like all 18650 Li-ion batteries, any NCR18650 cells available retail, or in smaller wholesale quantities have passed through a few middle-men, making the wholesale prices hard to estimate. What I do know though is that it is hard to find such cells for less than $6.50. Laptop packs with 9 similar cells are over $100 new, though you can generally find surplus packs for $50 or $5.50/cell. With only 4,000 miles on the battery pack, and given that Tesla treats them pretty gently, the cells in these packs are going to be pretty close to new condition.

Given all this, these modules seem like a reasonable price if you can use the entire module intact.

On the other hand, using cells from these packs individually probably doesn’t make sense. In addition to the effort required to disassemble the module, these cells may need to be wrapped. It is also quite possible that these cells don’t have some of the safety features people expect with 18650 cells, since the pack has other provisions for dealing with cell shorts and overheating.

If you buy any of these modules, I’d be interested to hear about your plans for them. If you’ve seen other Tesla modules for sale, I’d appreciate a link, or information about the pricing.