The LiFePO4 Headway cell 38120P TEST REPORT inside

[safe]

It's interesting data, but it doesn't do anything for my balancing "masters thesis". :? In order to make a mathematical argument for the "Smart Battery" I need to see the data sliced across all the cells at the same time and ideally I need to see a cross section of the cells as they discharge. The idea is to be able to prove that a group of cells that are controlled independently can produce more power if optimized than the losses that MOSFET resistance adds to the equation.

My "mathematical intuition" suggests that balancing is philosophically flawed and can be improved upon by "Smart Battery" theory. (note: that's the "On State" or "Off State" idea)

Well... I guess my Ph.D will have to wait... :lol:

 

[Doctorbass]

to prove that a group of cells that are controlled independently can produce more power

That's a fact.. my idea of independently charging all parallel 1s group and then to protect cells with a LVC circuit like Bob and Gary did is similar i guess.. no control on the cells.. just chargefully charge each time and also watch each cell to detect the first one to drop under a given voltage to stop the load...By that way, the cells will get their max cycle, be protected and will be equalized.. ee oupss.. fully charged! each time they are used.

Ph. Doc

 

[safe]

That's a fact.. my idea of independently charging all parallel 1s group and then to protect cells with a LVC circuit like Bob and Gary did is similar i guess.. no control on the cells.. just chargefully charge each time and also watch each cell to detect the first one to drop under a given voltage to stop the load...By that way, the cells will get their max cycle, be protected and will be equalized.. ee oupss.. fully charged! each time they are used.

I want to be able to prove that you can go beyond the LVC and go the extra distance to where the very last cell gives up it's last bit of energy. If one cell is, say, 10% weaker than the rest that means that the WHOLE pack needs to stop at a point that is at least 10% from the true end. If one cell goes to 20% weaker (as it ages) then the whole pack drops to the 20% loss. And what makes matters worse is the "runt" cell gets worked the hardest because it can less and less carry it's load so it drags the whole system down even more. By eliminating the controller based low voltage limit and instead letting the cells cut out one by one you could also sense that the pack was running low. (the end of charge sag would be very noticeable)

I just want to be able to present the math in a format that is undeniable...

Many people might be surprised that using the LVC and balancing does not do as good of a job as they think it is doing. I'm seeing that you could get as little as a 1% loss of capacity using MOSFET's (if you do it right) so the argument for a full blown "Smart Battery" design might be worth it from a performance standpoint. (and the older the pack gets the better the "Smart Battery" design would be)

That's what I want to prove to myself and others... however, I'm already becoming fascinated with the ultracap solution that might be coming down the road... so much change happening, so fast...

 

[oatnet]

I want to be able to prove that you can go beyond the LVC and go the extra distance to where the very last cell gives up it's last bit of energy.

While that might be interesting and all to get every last watt, does it have any practical application? Running a cell to lvc - and below - kills cycle life. Why strive for 100% DOD and get 400 cycles when you can size your pack to meet your needs at 70%-80% DOD and get 2000+ cycles? It just seems wasteful of the battery and your time.

I never exceed 70% DOD, and to extend the life of my pack I seek out opportunity charges to reduce DOD even further. If a cell is out of balance and hits 75% DOD, the impact is negligable. For me, LVC is there only for a cell that gets a bad charge.

-JD

 

[safe]

While that might be interesting and all to get every last watt, does it have any practical application? Running a cell to lvc - and below - kills cycle life....

You are behind the curve in understanding how the "Smart Battery" concept works. (so your assumption is wrong) The idea is that when a cell drops to it's cutoff point (that point were damage does NOT occur) you simply put the cell into an "Off State" condition and that cell alone is now offline while the rest of the pack keeps going. You might have 20% of your pack left and this way you can get that without adding wear and tear. If you wanted to extend the life more you COULD set your lower limit higher... like if the absolute danger zone is 2.5 volts you could set the MOSFET's to trigger an "Off State" condition at 2.7 volts and that would guarantee protection for that cell. Maybe all the other cells at that point are above 3.1 volts and have a long way to go... so you get to keep going until they are equally (proportionally) drained.

People seem to be having a hard time understanding the concept of the "Smart Battery".

The core concept is that cells that are outside of their safe operating range are functionally disabled, while the strong cells are allowed to keep going. This way you get the added benefit that you get longer life on average because all the cells are proportionally aged at the same rate.

This is why I want the data, because I'm sure I can present the mathematical case that MOSFET's and the "Smart Battery" are actually better in the long run than LVC and balancing.

As for the idea of buying more battery... sure, that's always good, but that has to be subtracted from the economic "life cycle" cost of the system. If you buy 1000 Wh of battery you want to get the full usage out of it... and don't forget that the clock is ticking all the time too, so the "Shelf Life" is killing your pack while it's sitting there doing nothing.

I might add that even if you don't hit your LVC you are still working the weaker cells harder than the stronger ones because the stronger ones never have to work as hard to give the same result. Even if you go down 70% from full the weak cells are still being worked harder for a proportionally longer time relative to their capacity. Depending on how bad the cells go out of balance you will see a decline in direct proportion to the weakest cell...

 

[Dalecv]

Looks like a lot of work to demonstrate the ambient air temperature fluctuations in your lab.

Just another parameter to report on, keep up the good work

 

[rf]

I want to be able to prove that you can go beyond the LVC and go the extra distance to where the very last cell gives up it's last bit of energy.

While that might be interesting and all to get every last watt, does it have any practical application? Running a cell to lvc - and below - kills cycle life. Why strive for 100% DOD and get 400 cycles when you can size your pack to meet your needs at 70%-80% DOD and get 2000+ cycles? It just seems wasteful of the battery and your time.

I never exceed 70% DOD, and to extend the life of my pack I seek out opportunity charges to reduce DOD even further. If a cell is out of balance and hits 75% DOD, the impact is negligable. For me, LVC is there only for a cell that gets a bad charge.

-JD

We admittedly don't have all the life-cycle information we'd like on the various new batteries. But you're making it sound like we're still back in the lead acid stone age. The evidence for LiFePO4 cells so far says strongly that manufacturers cutoff voltages are conservative and that protection from over-discharge is the one key element to provide for. R/C enthusiasts have abused A123 cells to great lengths with almost no ill effects. Depth of discharge seems largely unimportant IF you have an effective low voltage cutoff (LVC). Conservative discharge and other lead acid type concerns don't apply. Except where you're still testing your circuitry and range and things like that.

If you have real data that says something to the contrary I'm sure we'd all like to see it.

Limiting depth of discharge to 70% or even keeping batteries religiously topped off is good for having reserve power when you want it. It shouldn't be necessary to make this chemistry last noticeably longer. At least that's what the current collective experience seems to be saying.


Richard

 

[safe]

Capacity and Wear

There are two issues, Capacity and Wear.

Capacity - Looking at the chart below it's pretty obvious that if cells have differing capacities then when you cut off the whole pack prematurely because of one cell that you lose all the unused capacity. If you balance the cells as you discharge you could get around this, but most balancers can only balance a small trickle current which would not be anywhere near enough to equal the difference.

Wear - We know that wear is not an "On - Off" affair. With Lithium there is what is known as the "storage charge" which is a charge level that is right in the middle (something like 3.0 volts usually) which is the condition on which the battery degrades at the slowest rate. The Chevy Volt is rumored to be looking at a 30% to 80% range for their A123 LiFePO4 packs for this reason so that they can achieve the hoped for 10 year shelf life. As the weaker cells (the one's that are forced to cut off early at their 2.5 volt level) get weaker they accellerate in their relative degradation. It's pretty obvious... just look at the chart and think about it... the cells that have higher capacity barely do any work, while what I call the "runt" cell has to work really hard just to survive.

"The Rich Get Richer, The Poor Get Poorer"

...so the LVC / Balancing technique has it's flaws.

 

[safe]

It shouldn't be necessary to make this chemistry last noticeably longer.

I'll take that as a joke. :lol:

The fact that people have to spend hundreds of dollars on cells that only last a few years is a crime. To say that "shelf life" and "cycle life" aren't serious (and potentially future ending) flaws is just not true.

We want to find every way possible to preserve the life of these cells because they are as expensive as gold... (but end up disposed as garbage)

 

[wanders]

It shouldn't be necessary to make this chemistry last noticeably longer. At least that's what the current collective experience seems to be saying.

I sincerely mean no offense, but I seriously take issue with this statement (and I'm not trying to pile on). The good news is that with excellent prospects for Li-Ion anode improvements (e.g. the silicon nanowire design from Stanford) and hopefully similar improvements for the cathode, we are looking at 3-4X energy density improvements over conventional Li-Ion in a relatively short time, 4-5 years to market, and I'm not really sure what the power density implications are but they are probably good as well.

Keep in mind that there is no Moore's Law for batteries, and the recent nanotech improvements (e.g. LiFePo4 particle-based electrodes) are the leading edge of a revolution in a practical science which has had steady but small percentage growth for the decades since the electrochemical battery was invented.

Think of it this way - with any luck, if you buy a quality LiFePo4 system today and take care of it, by the time it wears out you will be able to replace it with something 3-4X better. Exciting times, I say, and none too soon.

Willie

P.S. For those who are into such things, I've attached a PDF (freely retrieved from the web) of the Yi Chi silicon nanowire battery paper. I have more of this stuff if you like it...

 

[recumbent]

Thanks "Wanders" for the "nano wire paper" you attached, and of course your right about the comment.

The futur has great batteries in development for us, couldn't happen at a better time.

 

[rf]

It shouldn't be necessary to make this chemistry last noticeably longer. At least that's what the current collective experience seems to be saying.

I sincerely mean no offense, but I seriously take issue with this statement (and I'm not trying to pile on).
[...]

I make some important points about LiFePOs and you guys get lost in this odd little, out of context quote. Wow. All those words say is current LiFePO is pretty good. Of course our lust for more power will never be quelled. Perhaps not even when someone perfects the Naquita generator or ZPM (http://en.wikipedia.org/wiki/Zero_Point_Module).

I was trying to say that you folks seem to be putting too many Lead Acid/Nicad/etc. attributes onto LiFePOs -- where they do not belong. No memory. No self-discharge to speak of. No cares about being left in a discharged state for weeks -- they really don't care. Just don't reverse their polarity -- that's all.

A thousand full cycles is pretty damn good. Two thousand, three thousand -- wow! Those are apparently doable with good cells. I don't know what the issues are up in that rarified air. That's beyond my reach. Beyond the reach of most of us, I think. We're still flying Sopwith Camels and you guys are debating the finer points of afterburners on jet engines. I say lets learn how to utilize our current good fortune -- we can finally fly! If they bring us jets tomorrow -- great! Otherwise, let's learn how to loop our Camels.

3000 cycles is 15 years worth of riding to work every day. That's approaching too much longevity. We need to figure out how to use it all. Poor us.

Richard

 

[safe]

3000 cycles is 15 years worth of riding to work every day. That's approaching too much longevity. We need to figure out how to use it all.

"Shelf Life?"

You seem to have forgotten that LiFePO4 still dies after 5 years whether you use them or not. There are two things:

"Shelf Life"

"Cycle Life"

You can get many cycles with LiFePO4 without doing much damage (if you keep it within the safe voltage range) but time will eat away at it no matter what you do.

However, if you use a "storage charge" rather than a full charge there might be some hope that you could extend the life beyond 5 years. The Chevy Volt guys are trying out this idea in order to stretch the life to their goal of 10 years.

15 years... that's just not going to happen...

(but I'm not against the idea that maybe in the future they can extend the shelf life so that one day that might be possible... but today, with current knowledge, it's not a reality)

 

[fechter]

You seem to have forgotten that LiFePO4 still dies after 5 years whether you use them or not. There are two things:

I think that's true for Li-Co cells, but I've heard up to 10 years for LiFePO4. I don't think any of them have been around long enough to really know...

 

[safe]

I think that's true for Li-Co cells, but I've heard up to 10 years for LiFePO4. I don't think any of them have been around long enough to really know...

That's true... it will take years to know for sure... right now I think people are making a lot of claims, but no one can be certain. We know more about cycle life wear because you can measure that through testing in a short period of time, but we can't know the shelf life for certain.

And there's still that chemical purity question...

Maybe the better LiFePO4 cells use more pure chemicals and that translates into longer life. So it just being "LiFePO4" might not be enough, it might need to be "99.99% pure" in order to have 10 years of shelf life.

Makes those Chinese cells suspicious...

 

[rf]

And there's still that chemical purity question...

Maybe the better LiFePO4 cells use more pure chemicals and that translates into longer life. So it just being "LiFePO4" might not be enough, it might need to be "99.99% pure" in order to have 10 years of shelf life.

Makes those Chinese cells suspicious...

Doesn't mean a darn thing. Now you're sounding like Don the network terrorist. Creating doubt where no evidence exists.

For all we know you can mix the lithium half and half with gopher dung without effecting performance. Add to the discussion, don't divide it. Please.

``Didn't I hear that there's a purity problem with those batteries?'' No you didn't!

``Didn't I hear there's a reason to believe the Chinese cells have problems?'' No you didn't!

Almost all LiFePO4 cells on the market come from China.

Richard

 

[safe]

For all we know you can mix the lithium half and half with gopher dung without effecting performance. Add to the discussion, don't divide it. Please.

We have to ask the simple question:

"What causes a LiFePO4 cell to decay?"

...I think you are going to find that there are differing processes that go on inside the liquid... and that's all it is... a liquid... that cause it to decay from one form to another.

My "guess" (and I stress that I'm just using the way most other things behave as a baseline) is that there are processes that take place all the time to contribute to decay that are not tightly bound to the cycling of the cell. So you will have a natural constant decay of the chemical just based on shelf life. Then on top of the natural decay rate you can add to it with hard cycling... the harder you stress the cell (and the deeper or higher the voltage) the more you exaggerate the decay.

All things die... all things decay... it's life...

However, some things like a rock made of granite will decay very, very slowly so it can seem almost like something lasts forever if you make it hard enough.

 

[Doctorbass]

UPDATE 13 april 2008 --- 50 cycles done.. end of cycles test for cell 1 ---

I'm pretty confident that these 30-50 cycles slope would more represent the real comportment of that cell.

I agree, the temp is much affecting this cell like any LiFePO4 cells could be.

But in this case that seems to affect more the charge/discharge efficiency than the damage that could happen to a cell about heat...

Remember.. that 8Ah cell no 1 charged at 10A... ( 1.2C ) for the needs of the cycling speed... to avoid taking a years to test... I agree that a cell that you charge slower will have a better cycle life.. so consider this case as the worst case you can get...

Next: cell 2.... we will be able to begin comparaisons!

Doc

 

[Link]

So it looks like they've gone down by about 150mAh. Not bad, especially considering that that is still over their rated capacity.

 

[Doctorbass]

Urgent Data Request

The group as a whole likes to see the charts of the cell performance, but I'm urgently requesting that you post the data for the charts so that folks like me can do analysis with them.

What I'm really looking for is a constant "C" rate discharge for each cell... that's one set of data per cell per discharge... so that I can compare how they differ between themselves. From this data I can evaluate the effectiveness of balancing theories.

If you can post the data as simple "Text & Commas" (the "*.csv" extension) then I don't have to worry about Excel format related problems...

SAFE, AS YOU REQUESTED
HERE ARE THE DISCHARGE DATA OF THE CELL 1 to 3

 

[PJD]

As far as shelf life, Valence advertises "excellent float life" - i.e life when held at 100% state of charge - for their LiFePO4 batteries. So, presumably they must mean at least as good as a well maintained, high quality PB-acid battery in float/standby use. How long would that be?

 

[TylerDurden]

As far as shelf life, Valence advertises "excellent float life" - i.e life when held at 100 state of charge - for their LiFePO4 batteries. So, presumably they must mean at least as good as a well maintained, high quality PB-acid battery in float/standby use. How long would that be?

20yrs.

 

[Link]

As far as shelf life, Valence advertises "excellent float life" - i.e life when held at 100 state of charge - for their LiFePO4 batteries. So, presumably they must mean at least as good as a well maintained, high quality PB-acid battery in float/standby use. How long would that be?

20yrs.

Your rebuttal, Safe? :wink:

 

[PJD]

One problem with resolving this issue is that this forum may be chock-full of EE whiz kids, but no chemists. (I'm a Civil Engineer - but not in the environmental field where the chemists congregate). But considering that the LiFEPO4 is supposedly a much more stable compound than LiCoO2, it would stand to reason that it should exhibit better shelf life.

 

[Doctorbass]

He he.. guys.. eee.. i would keep this post for the report of the headway cells :lol:

:wink:

Doc