Alternator Size with LiFePo4

[tanglewood]

Originally Posted by kmacdonald

Four or five years ago when I was shopping for 4 small capacity LiFePO4 cells I used google to search for them. I would click on the links found and call about pricing and availability. A common response was "we don't sell them any longer". Failure rate was often cited as the reason.

Well, to draw any conclusion from that you would need to know why they failed. Probably the single biggest take-away from all the discussion here on CF is that LFP (and other Li types) need to be managed differently in some key ways compared to LA. Anyone taking an LFP cell and using it like an LA or other cell will result in a really high failure rate. It's taken quite a while for people to grasp that and build systems that properly manage them.


My own take is that we are looking at a technology that is still in it's infancy and has yet to really take off. Not something that has been tried and failed.

[kmacdonald]

Originally Posted by tanglewood

Well, to draw any conclusion from that you would need to know why they failed. Probably the single biggest take-away from all the discussion here on CF is that LFP (and other Li types) need to be managed differently in some key ways compared to LA. Anyone taking an LFP cell and using it like an LA or other cell will result in a really high failure rate. It's taken quite a while for people to grasp that and build systems that properly manage them.


My own take is that we are looking at a technology that is still in it's infancy and has yet to really take off. Not something that has been tried and failed.

They have been around for 15 or 20 years I think. With all the benefits they offer you'd thing someone would have designed a system with enough fault tolerance to be market viable. That hasn't happened yet but it's not due to lack of trying. Boeing failed at their first attempt (with another lithium technology) and about burned up an aircraft in the process. Several others have failed. I agree, they aren't for everyone and require a knowledgeable user. That includes a VERY small percentage of the boating community. People just want to use their boats and are not looking for a science project. Trojan may be the last hope for a "drop in LiFePO4" designed for the masses.

By the way, the systems that manage them have been blamed for a lot of the failures. I don't know if it's an excuse or true but I suspect a little of both.

Fault tolerant systems are essential to the long distance cruisers and LiFePO4 has show a complete lack of it. I hope that changes because I'd like to take advantage of all it has to offer but I'm not convinced it's ready for prime time. I use a small LiFePO4 bank in my Capri 22 and have been for years. It works great and does exactly what I expect out of it but I wouldn't put a LiFePO4 bank on my Bristol.

[kmacdonald]

There is some really good info on some of the RV sites about LiFePO4. RV's are a much bigger market than boats and there are a lot of stories of the trials and tribulations of LiFePO4.

[evm1024]

Originally Posted by kmacdonald

No, you clearly misunderstood SOMETHING, I don't know what but SOMETHING. If a battery technology requires sequential serial numbered batteries to function correctly in a pack, it's just not going to work. Next time you order a set, specify sequential serial numbers and let us know what they say. Part of the beauty of mass production is that ALL the items produced come out the same within reason. If not something went wrong or the allowable tolerance is too large. We all know they won't be exact but if they aren't close enough to work in a pack, any pack with the same battery, then the tolerance is too great. Unfortunately that seems to be the case. Don't take my word for it, read what some of the pack builders that quit selling them said. My best guess is that most of the prismatic cells end users are able to buy are factory rejects.

Try to not have such an inflated opinion.

You used the term exactly which means exactly that. No differences at all.

So lets take some electronic components. How about resistors? Run a batch of resisters. OK so we sort them and bin them. These are greater than 10% out of spec of the design resistance. And those are within 10%, these are within 5% and lastly this much smaller batch are within 1% of the design.

So much for exactly.

Or how about the variation in speed between CPU die on a wafer. When I worked for Intel we were getting around 400 chips per die. We binned them and tossed the out the dead or too far out of spec chips. They were binned for speed.

So do you want to check your attitude and the door and dial back your Exactly comment?

[kmacdonald]

Originally Posted by evm1024

Try to not have such an inflated opinion.

You used the term exactly which means exactly that. No differences at all.

So lets take some electronic components. How about resistors? Run a batch of resisters. OK so we sort them and bin them. These are greater than 10% out of spec of the design resistance. And those are within 10%, these are within 5% and lastly this much smaller batch are within 1% of the design.

So much for exactly.

Or how about the variation in speed between CPU die on a wafer. When I worked for Intel we were getting around 400 chips per die. We binned them and tossed the out the dead or too far out of spec chips. They were binned for speed.

So do you want to check your attitude and the door and dial back your Exactly comment?

Do CPU's in a dual or quad processor computer have to be consecutive serial numbered? Of course not. You can even add a CPU at a later date if needed. I hope you can see the difference here.

I'd also like to add that 10%, 5%, and 1% tolerance resistors carry a different price. The greater the price the closer the tolerance. Are you buying your prismatic cells from an authorized distributor listed on the manufacturers web site? Or are they coming from the reject 20% tolerance pile of shi.........

[Delfin]

Originally Posted by kmacdonald

Do CPU's in a dual or quad processor computer have to be consecutive serial numbered? Of course not. You can even add a CPU at a later date if needed. I hope you can see the difference here.

They likely will be, and for the same reason builders of LFP packs want sequentially produced cells.

[evm1024]

Originally Posted by kmacdonald

Do CPU's in a dual or quad processor computer have to be consecutive serial numbered? Of course not. You can even add a CPU at a later date if needed. I hope you can see the difference here.

Dude, get serious. Do you know anything about how CPU are made?

400 plus die on a wafer back then, 42 wafers in a boat. The variations of the working die change with the position on the wafer and where the wafer is in the boat.

And don't forget that there are some 60 or more steps in the pipeline.

Now again tell me how the battery manufacture gets infinite precision in the various processes that go into making up a battery cell. And how they compensate for changes in weather. Oh, and of course this is China so you get the QC that you pay for.

How is it that a battery cell made in December by the night shift is exactly like the one made in June by the day shift....

Suggestion, tone it down.

[evm1024]

Oh I should add that part of my job was "Low Yield Analysis".

We did circuit analysis of die still on the wafer (under a microscope with a prober) to see if we could figure out why they were different or did not work at all.
Ans remember that these are for die on the same wafer, often side by side another die that worked fine.

Of course there are many, many parameters to measure with something as complex as a CPU die.

Relax and contribute

[kmacdonald]

Originally Posted by evm1024

Dude, get serious. Do you know anything about how CPU are made?

400 plus die on a wafer back then, 42 wafers in a boat. The variations of the working die change with the position on the wafer and where the wafer is in the boat.

And don't forget that there are some 60 or more steps in the pipeline.

Now again tell me how the battery manufacture gets infinite precision in the various processes that go into making up a battery cell. And how they compensate for changes in weather. Oh, and of course this is China so you get the QC that you pay for.

How is it that a battery cell made in December by the night shift is exactly like the one made in June by the day shift....

Suggestion, tone it down.

I have to admit I've never worked in manufacturing. We farm that out overseas to low budget labor. We prefer the more cerebral work in the States.

Like you say, there are tolerances and there are in fact rejects also. What's in your boat?

[kmacdonald]

Originally Posted by Maine Sail

A number of months to go... The last batch had some major variations in capacity changes between cells, all lost capacity, some quite badly, and I just wanted to repeat it to be sure, due to the wide variations. Some are elcheapo/no-name and others are premium K2, A123/LithiumWerks etc...This all takes time. This is not something I am planning to publish, just for my own curiosity, and done specifically because some cell makers claim their cells can be floated without diminishing capacity. Those claims are not what I have seen so far, and this is why I am repeating the test.

This is what I'm talking about. Tolerances so lose that it's hard to believe, so much so that the tests are being repeated. And those were premium batteries, not black market rejects.

MaineSail knows what he's doing and I'm sure he had reason to do the testing he did. He must have seen a trend. And you want to built fault tolerant banks with those? Are you kidding me?

[evm1024]

Originally Posted by kmacdonald

I have to admit I've never worked in manufacturing. We farm that out overseas to low budget labor. We prefer the more cerebral work in the States.

Like you say, there are tolerances and there are in fact rejects also. What's in your boat?

I've got 4 Winston 700 AH LifePO4 cells in my boat. I've been running them for 3 or 4 years now. Have to look it up to see how long for sure.

And how about you? What is in your boat?

[kmacdonald]

Originally Posted by evm1024

I've got 4 Winston 700 AH LifePO4 cells in my boat. I've been running them for 3 or 4 years now. Have to look it up to see how long for sure.

And how about you? What is in your boat?

I have a 20Ah LiFePO4 GBS bank charged only by solar. 30W panel with Genasun controller. LA in the Bristol.

[evm1024]

So in LiFePO4 construction there are films deposited on foils. These films have some variation in thickness and composition. All of the other processes that go into making a LiFePO4 cell have their own variations.

It is clear that there will be some differences between the foils made a month apart and equally clear that cells made from the first foil will be different from the one made from the other foil.

After assembling the cells they are tested. Any cell that fails the minimum parameters is rejected and if it works at all it is sold as a substandard cell.

The other cells that pass the minimum are then sold as fully functional cells.

It should be obvious that if you assemble a battery from 4 cells all from the same foil that you will get cells that are more closely matched in all parameters that a set of 4 cells from different foils.

Now if I'm just going to put 64 cells together to power my truck I may not care how closely matched they are.

But if I am only going to use 4 or 8 cells and push them closer to their limits I would want cells that are closely matched.

Oh, what vendor was testing cells to match the parameters and assembling batteries with matched sets?

[kmacdonald]

Originally Posted by evm1024

Oh, what vendor was testing cells to match the parameters and assembling batteries with matched sets?

I was referring to MainSails suggestion of using consecutive serial numbered cells. I don't know if vendors do that or not or if it's even possible to reliably get sequential cells. I do agree with you that they could be more closely matched that way.

[evm1024]

Originally Posted by kmacdonald

I was referring to MainSails suggestion of using consecutive serial numbered cells. I don't know if vendors do that or not or if it's even possible to reliably get sequential cells. I do agree with you that they could be more closely matched that way.

There you have it....

China is fully capable of making "things" of very high quality. But to get that quality you need a bunch of things. Some companies put one or more of their engineers over in the plant in China to ensure that they get the quality they want.

I would expect that any cell made to have the A123 label on them would have very good QC at the plant in China. YMMV for others cells.

And even with an engineer on site their could be shortcuts taken by omission or commission - Just think back to Rocna's move to China....