[ausev] anyone heard of these guys?

[Chris Robison]

Brian,

Your conclusions are different from what we've seen in our research of 
lithium options so far. I'll try to run some numbers again below, and 
try to approach mathematically what we've seen on a larger scale. 
Hopefully my math and assumptions aren't too far off.

(This one is really terrible to read through. I repeated the numbers I 
got in a summary at the end.)

Brian Lasseter wrote:
> So lets compare it to a Trojan T-145...
> http://www.trojan-battery.com/Products/ProductSpec.aspx?Name=T-145
>
> The Trojan is 10.38" x 7.13" x 11.5", at that size I think I can cram
> 56 of those Lifebatt cells in the same size box (I'm presuming they do
> not need the connector tabs in a box).  That's 7 x 4 x 2 = 56 of the
> LifeBatt cells.
>   

I have not seen the inside of LifeBatt's modules. However the modules 
and custom arrays I've seen from other manufacturers and private 
projects (including the Valence U-charge module we dismantled, and 
others in pictures I've seen) do not pack the cells in a square grid, 
but rather in a tight hexagonal grid for maximum density. The better 
higher-power designs I've seen make use of the resulting gaps -- they 
have air holes in the side plates to pass cooling air through them.


> If we connect the 56 Lifebatt cells in 2 parallel strings of 26 cells,
> we get a 6.4V battery with 260Ah of energy.  The T145 has a voltage
> just over 6V, the same, with a capacity  of 244Ah.  The Lifebatt has a
> little more capacity than the T145 in the same volume.
>
> 56 Lifebatt cells will weigh 20kg, and the Trojan T-145 weighs 33kg.
>   

Although I understand there's a significant difference in Peukert 
constants even between models from the same manufacturer including 
Trojan, we'll assume the most optimistic multiplier I calculated in my 
last excruciating post on this subject which was 0.71.  So 244Ah (20hr 
capacity according to Uve's Battery Page and most sources online) x .71 
= 173.24Ah of usable capacity at a 1-hour rate. You may discharge these 
faster than a 1-hour rate in practice, and the actual constant may be 
higher than the 1.1 we're using, so you'll probably get a little less 
than this in reality. But we'll go with it. The Peukert capacity loss 
for lithium is pretty close to nil at nominal (not racing) discharge. 
Plus, in a 100-mile lithium configuration, individual cell discharge in 
terms of C is about half of that from a 40-mile pack typical with lead.

To better compare energy capacities apples-to-apples, we'll use units of 
energy rather than charge. The energy capacity for the T145 is 173.24 * 
6 = 1.039kWh. At 33kg, specific energy is therefore approximately 
31.5Wh/kg. LifeBatt claims 80Wh/kg.

With the dimensions given by Trojan, the volume of the T145 is 
26.4*18.1*29.5 = 14096.28 cubic cm, or about 14L. So, the Trojan's 
volumetric energy density is 74Wh/L. LifeBatt claims 170Wh/L which I'll 
assume is for each cell (and hopefully includes the cell can, or 
otherwise they're a bunch of slimy bastards). But we can't use that in 
the real world, because we've got a bunch of cylindrical cells. Darn 
geometry. So....

The density ratio of an optimum hexagonal circle lattice is 
approximately .906 (from sqrt(3)pi/6), meaning the rest is lost to the 
gaps between cells. From that we can get the real density, about 
154Wh/L. The module container will reduce this a bit further due to the 
flat sides but of course the amount of loss decreases with increasing 
module size. I can't predict that without knowing the size and I'm not 
smart enough to calculate it anyway, so I'll be lazy and leave it out 
and we can just assume that the results will differ from reality by a 
small margin.


> In the final analysis, you get 6.5% more Ah, you save 40% of the
> weight, and you pay 3000% more for the lifebatt.  The T145 costs about
> $100, while the Lifebatt will cost you $3080.  And that is not
> including the extra cost for their chargers, and battery management
> system, or your cost to assemble 6V car size batteries from the cells.
>   

By the numbers I've come up with above, specific energy of the LifeBatt 
cells is approximately 230% (2.3 times) that of the Trojans, meaning a 
130% improvement. Volumetric energy density of the LifeBatt cells is 
double, improved by 108%. These numbers are not quite as good as I 
recall seeing from other suppliers (memory is really fuzzy), but are 
realistic.

I don't have LifeBatt's pricing, so all I can give are the wholesale 
prices I've seen from other manufacturers, which are typically around 
$0.80 per watt-hour, excluding the unobtanium high-end (A123) and 
unapologetically fraudulent low-end (ThunderSky) offerings.

$100 is pretty optimistic for a T145 at today's prices. A quick google 
shows that not many vendors sell them online or provide prices, but I 
found these with prices well over $200:

http://ebatteriestogo.com/merchant2/merchant.mvc?Screen=PROD&Store_Code=EBTG&Product_Code=CBT145
http://www.batterysales.com/trojan.cfm

I called our dealer, who has the lowest prices we've found on Trojan 
batteries (Comal Golf and Battery, they give a good price for AustinEV 
members and our company), and he quoted me $164 each for 20 pieces, with 
prices set to go up significantly in January. You may have gotten your 
price from a reliable source, so we'll use that too (and I'll be very 
interested in who you talked to!)  :o)  Again, the T145 contains 1039 
usable watt hours. At $100 each you get about 9.6 cents per watt hour. 
So you pay about 8.3 times more for lithium, and this is in the 
neighborhood of what we've found so far. It would be more like 10 times 
more if comparing against less expensive (e.g. USBattery) lead-acid.  
With the price I was quoted, lithium is only (heh, "only") 5 times more 
expensive at 15.7 cents/Wh. Lead prices continue to rise, and increased 
manufacturing scale and competition continues to push the price of 
lithium slowly downward.

I haven't gotten good figures yet from anybody about the price 
difference for using configured lithium modules instead of raw cells. We 
signed an NDA yesterday to get some near-future data from one 
manufacturer I'm hoping we can do business with. (unfortunately, I will 
therefore not be able to share it without permission.) But, the battery 
management system doesn't necessarily have to add a lot more expense. 
For example, Valence charges $150 for their BMS unit, which covers up to 
a 450V series string. A good portion of the management circuitry is in 
their battery modules of course, but that isn't necessarily the case 
with a custom design or cells assembled yourself. The entire BMS in that 
case could be a truly one-time cost, perhaps minus some embedded thermal 
sensors.

> So... Even if the Lifebatt cells last 3x longer than the T145 battery,
> which their warranty assures, the cost-to-benefit ratio is still way
> out of wack.  I'll stick to Lead-Acid.
>   
It's a little more reasonable than that, but it's still way out there. 
It's certainly not a money-saving proposition, yet.

However, there are other benefits besides energy density and specific 
energy that some folks even today are willing to pay for. First, no one 
really cares about energy capacity -- what they care about is range. 
You'll use less energy in stop-and-go driving due to the lighter pack, 
so you'll actually get more range in the city from the same quantity of 
energy from lithium. Second, lithiums have a very flat discharge curve, 
so performance won't deteriorate as your pack discharges, the way it 
does with lead. You get the same performance until it goes off the cliff 
at the end of the curve, and it's time to recharge. Third, lithiums do 
much better in the cold than lead-acid batteries do, so you don't lose a 
large chunk of your usable capacity in the winter months (not that 
that's going to matter much anywhere south of Kansas a decade or two 
from now). Fourth, it's the only practical option for those who really 
need a driving range beyond 40-50 miles.

And finally, the real-world longevity of LiFePO4 is pretty much unknown 
at this point. Cycle life testing suggests between 10-15 years of 
typical commuter use, but of course calendar life will most likely cut 
that short, and the sophistication of your BMS algorithm will have a big 
effect too. Conservative estimates of 5 years have been thrown around, 
but the reality is that we just don't know, and no one wants to 
overpromise. We'll see in the coming years as the early adopters blaze 
the trail.

Summary:

Specific energy:
Trojan T145: 31.5Wh/kg
LifeBatt lithium: 80Wh/kg
---
LifeBatt lithium holds 2.3 times the energy per unit mass


Volumetric energy density:
T145: 74Wh/L
LifeBatt: 170Wh/L (bare cell) 154Wh/L (continuous packing)
---
LifeBatt lithium is twice as energy dense by volume


"Monetary" energy density:  :o)  (excluding one-time costs)
T145: $.096/Wh ($100/batt),  $.157/Wh ($164/batt)
LifeBatt: 1.25Wh/$
---
Lead is 10x to as little as 5x more expensive, depending on price of 
lead battery


Again, hope I didn't screw any of this up too bad. Check my work...

  --chris