Alternator Size with LiFePo4

[john61ct]

Originally Posted by rgleason

Keeping the SOC range to 75% to 25% indicates that a 200ah bank will have about 100ah available, which would be ok.

Not at all necessary, IMO much more wasteful than simply avoiding the voltage shoulders. Assuming 4S, 12V nominal,

stopping charge at 13.8V is within ~10% of even the high voltage + absorb profile spec'd by the vendor, and likely equal to or within just a few percent of rated AH capacity.

And staying above 12V sacrifices hardly any storage capacity at all, compared to going lower into more dangerous territory.

Further details here http://www.cruisersforum.com/forums/...ml#post2788925

If you are looking to intentionally keep a 50% Reserve, fine, but for the sake of longevity during normal cycling, do it by drawing down to only 60% DoD, rather than stopping charging at 75% SoC, there's nothing to be gained by that.

[Q Xopa]

Some people are saying that stopping charging around 75% SOC maximizes Cycle life. So that could be considered a gain.

However as you say that does come at a 'cost' of leaving 50% capacity unused, wasted. So essentially you effectively end up with a small bank for your $.

As someone also correctly states this also comes down to specific definitions (of which there are many) of of what is considered to be 75% SOC.

I think it is useful to be aware of as many of these factors as possible but there is obviously a balance to be weighed with all of this. This probably comes down to personal design philosophy goals.

[rgleason]

I'd rather not pay double for LiFePo4 batteries, leaving 50% of capacity unused. What are some more appropriate SoC to use? Approximately what SoC is 13.8v and how deep can I draw down maintaining reasonable cycles and life? For me 10 years would be fine.

I see thar John61ct is suggesting I can use close to 100% of the capacity, call it 96%.

..."my C definition" of 100% is 96% of the A rating

By charging to 3.45vpc or 13.8v for C definition of SoC 100% full.

That would be great, because that would be about 2 days worth of capacity.

So knowing these things what size and type of alternator do I need? I am going to mount it forward and outboard as much as possible to provide good airflow and avoid the hot manifold. I will duct fresh air to the alt. I need about 60-80 rock solid contunuous amps. I will use J10 pulleys.

Would a 165a alternator with built in rectifiers but dual fan provide this capacity for long term performance and life?

[rgleason]

The other thing I'd like from our new alternator is good performance at low rpm, 60a minimum at 1000 engine rpm (preferably 750 engine rpm).

[john61ct]

Originally Posted by Q Xopa

Some people are saying that stopping charging around 75% SOC maximizes Cycle life.

They're just not measuring SoC changes accurately.

That requires load test comparisons between different "stop charging" profiles.

Those counting AH going in, using any Absorb Time and especially once past 13.6V, only a tiny fraction of those AH are actually going into storage.

[john61ct]

Originally Posted by rgleason

The other thing I'd like from our new alternator is good performance at low rpm, 60a minimum at 1000 engine rpm (preferably 750 engine rpm).

the most expensive part of a custom alt install will be the custom ratio pulley(s) to accomplish that.

And that seems low to me, unless the alt's required torque is truly a tiny proportion of what the engine puts out.

Or maybe with a specialized alt like Eco-Tech, only needs 2000rpm to put out decent amps.

[john61ct]

Originally Posted by rgleason

I'd rather not pay double for LiFePo4 batteries, leaving 50% of capacity unused.

Exactly, I have never advocated that.

What are some more appropriate SoC to use? Approximately what SoC is 13.8v and how deep can I draw down maintaining reasonable cycles and life? For me 10 years would be fine.

No guarantee of that of course, but I've already given you **my** formulas, follow the math details in my linked post, cross reference to MS load test results.

Functionally, use voltage basis for the top end, SoC AH-meter only for drawing down.

> I see that John61ct is suggesting I can use close to 100% of the capacity, call it 96%.

> ..."my C definition" of 100% is 96% of the A rating

Only applies if your actual delivered capacity is higher than A the vendor AH rating.

If you try to save money and get "factory rejects" that won't be the case.

Also, more like 80-85% comparing to the mfg definition of Fully charged, B rating.

Do load testing on **your** setup to verify your own capacity numbers, choosing a given 100% Full definition to compare to your voltage-based usage profile.

> I need about 60-80 rock solid continuous amps.

The ratio between continuous output as compared to mfg rating will vary greatly, large frame vs small, expensive specialized for marine deep cycling vs stock vehicle type, and also by brand.

My approach if money were no object would be

Eco-Tech, followed by MS or Bruce for Mark Grasser

If cheaping out and I knew I could make room, a large frame 150+A Leece Neville fire truck alt.

[Q Xopa]

Originally Posted by john61ct

They're just not measuring SoC changes accurately.

That requires load test comparisons between different "stop charging" profiles.

Those counting AH going in, using any Absorb Time and especially once past 13.6V, only a tiny fraction of those AH are actually going into storage.

Yes, I agree, they are not measuring SOC correctly. As aren't most of us. This is, as I'm sure you agree, not so easy to do. Apart from the fact that there are so many methods for doing so, ie Voltages and Coulomb counting for a start. It will be good when SOC definition is decided and accepted a little more widely.

If we accept Coulomb counting as the preferred method, which I agree is preferable then we still need to try and get an accurate AH to SOC number out of our BMV with our Charge Efficiency factors etc tweaked accurately. The SG200 sounds like it's getting pretty close, when they are widely available.

All that said I think the concept is in line with what many others are saying.

It doesn't sound like you are so convinced, please feel free to elaborate so we can all learn.

[Q Xopa]

Originally Posted by rgleason

I'd rather not pay double for LiFePo4 batteries, leaving 50% of capacity unused. What are some more appropriate SoC to use? Approximately what SoC is 13.8v and how deep can I draw down maintaining reasonable cycles and life? For me 10 years would be fine.

I see thar John61ct is suggesting I can use close to 100% of the capacity, call it 96%.

..."my C definition" of 100% is 96% of the A rating

By charging to 3.45vpc or 13.8v for C definition of SoC 100% full.

That would be great, because that would be about 2 days worth of capacity.

So knowing these things what size and type of alternator do I need? I am going to mount it forward and outboard as much as possible to provide good airflow and avoid the hot manifold. I will duct fresh air to the alt. I need about 60-80 rock solid contunuous amps. I will use J10 pulleys.

Would a 165a alternator with built in rectifiers but dual fan provide this capacity for long term performance and life?

Smaller alternator pulley and larger crank pulley will give a better low eng RPM performance.

As for the Cycle life- SOC range issue. Well this is some peoples theory out there. It most likely has some element of truth. Time will hopefully tell how close their graphs are.

The next issue is how relevant it is to us. These are after all, like most of our systems just devices that will give us service, hopefully good service for a limited life, hopefully at a reasonable cost. Service life cost, is a good question, there will be pretty strong trends that emerge, but they are seldom absolutes.

All the things you are saying sounds like having a good shot at having it work as you would like, or at least certainly a better shot than most who dont put this thought into these considerations.

The trouble with most things is 'absolute' certainties is they are mostly not. Most of us are looking for the simple answer, but it rarely works out that way. Most people telling us they have the simple answer can talk a good story.

Most people that know more of the truth are reluctant to offer a simple answer. It's especially true with a newly developing technology like Lithium battery technology.

This isn't the answer most people want but all we can do is try and find the best information we can. There's plenty of info out there, some of it might even be true.

In the end it comes down to 'we pays our money, and we takes our chances'.

[evm1024]

I find the attached graph to be most instructive. \

All indications is that for LiFePO4 you charge at some (fractional) C rate and the battery voltage rises in proportion to SOC. Bulk or CC charging as it were.

Then when the battery voltage reaches a Max voltage (as shown in the graph) rather than carrying on a CV absorption phase for a long time (as needed for LA) you terminate charge after a short time (when I <= C/x, some small tail current).

The graph does not show how long CV was held but it does show that once you reach 3.4 VPC you are close to 100% charged and that there is little to no difference between charging to 3.5 VPC on up.

I'll have to try to find the original source for these curves to see how long they stayed in CV at the specified voltage. Anyone know?

The saying for FLA is that the last 20% takes 80% or the time (or some such).

We should start saying that the last 20% of LiFePO4 takes the same time as the first 20% (or some such).

Just to get us out of the extended CV mindset.

[rgleason]

The old Balmar 90-65 65amp is running at 74C or 165f and outputs 50a for a short period then drops to the 40's. I dump fresh air with a 3" fan onto the single front fan. Low rpm performance is poor.

I like the idea of using a 4" fan to draw hot alternator air down into a fabicated sideplate chamber that is put against the inside of the keel to enhance cooling (about 18" square).

If I can fit a large case alternator by mounting it higher and outboard of the manifold. I will fabricate the mounting brackets and possibly get the 200a CMI alternator for $600. In that case I will need one crankcase J10 pulley mounted to the front of the existing pulley. The water pump will just have its own V belt with a small idle wheel. I will need to modify the system cooling hose to relocate it forward. Here is the drawing
http://www.cruisersforum.com/forums/...ml#post2762002


The alternator belt setting will be put at b-10 which is 50% or 100a or at b-14 if that exists in Balmar's regulator....

Bruce Schwab has provided excellent full load graphs for MG alternators. It is too bad Balmar has not updated their graphs accordingly.

I'd like to thank everyone for contributing so much to this thread. Re reading it again I found many gems. Including this one from Paul, please read his well documented link of 29 page PDF. I know I will be referring to it in the future.

http://www.cruisersforum.com/forums/....php?p=2753595

Please keep adding good quantitative information and alternator experience with lifepo4 to this thread. Thank you


Looking at EVM1024 post just below, it looks like we could stop charging LiFePo4 at 3.4vpc.
EVM, I have that graph, let me see where I found it.

[john61ct]

Originally Posted by Q Xopa

Yes, I agree, they are not measuring SOC correctly. As aren't most of us. This is, as I'm sure you agree, not so easy to do. Apart from the fact that there are so many methods for doing so, ie Voltages and Coulomb counting for a start

There is no difference of opinion, on how to accurately measure total AH capacity, just different levels of correct knowledge.

Once you pick a given definition for 0% and 100%, then AH counters are just fine for measuring SoC on the way down, while **discharging**.

And let's say within those 0% and 100% points, normal cycling may be between 10% and 90%. If the owner-defined 100% rarely gets hit to reset the AH-counter, it's still good enough for guesstimating so the shoulders are avoided. As you point out Balmar's new SG200 may be more accurate at this than other meters, but for myself, I don't mind precisely hitting my 100%

for .2-.3C, 3.45Vpc, Absorb until endAmps @ .03C (3A per 100AH)

once a week or so.

However, as you get past 13.6-13.8V, depends on current level, an increasing amount of the AH being counted, is going into heat production, no longer actually increasing SoC.

As MS demonstrated with his documented test results, a load test is required to determine how many extra AH are **stored** by varying the two stop-charge parameters for a given charging rate, Voltage and Absorb endAmps (if any).

> It doesn't sound like you are so convinced

sorry, not convinced of what?

[john61ct]

Originally Posted by evm1024

when the battery voltage reaches a Max voltage (as shown in the graph) rather than carrying on a CV absorption phase for a long time (as needed for LA) you terminate charge after a short time (when I <= C/x, some small tail current)

For daily use cycling, best and simplest is to "just stop" charging when your end-point voltage is reached. As you point out, a long Absorb stage is holdover "lead thinking".

However when testing, or otherwise a precise repeatable benchmark is needed, stop-charging needs to be based on some higher than usual 100% SoC definition.

Mine is highlighted above.

MS demonstrated several others, including one where Absorb V is held until **zero** current is flowing. Which personally I would not use often on a given bank.

[evm1024]

I found the source as well. See:

https://www.powerstream.com/lithium-...ge-voltage.htm

They tested single cells and charged CV until the tail current dropped to 30 mA. Which looks to be about 0.013 C (9 amps for my 700 AH bank).

I note that they say that "in most cases" they CV until I = 0.013 C....

[john61ct]

Originally Posted by evm1024

The saying for FLA is that the last 20% takes 80% or the time (or some such).



We should start saying that the last 20% of LiFePO4 takes the same time as the first 20% (or some such).

I think an endAmps setpoint of .02 or .03C is precise enough for most benchmarking purposes, and at lower current levels (well below .3C) I find that rarely means holding Absorb / CV for more than a few minutes.

As I said "stop at setpoint V" is all that's needed for normal cycling usage, Absorb serves no useful function in that context.

And if SG200 really is that accurate without ever needing 100% resets, then the only times I can think to Absorb at all would be load testing, and (if using top) balancing. Both are very rare events in most cases.