Which batteries should I buy (rhetorical question)

[ghatikar]

2 hours ago, plonkster said:

I agree. I often see the sales material say things like you only need a third of the capacity with Lithium Ion, but that obviously makes no room for a cloudy day. With lead-acid, the 50% you (almost) never use doubles as both buffer for a longer life AND for the rainy days. I tend to go with a 50% rule rather than the one-third rule the marketers calculate so meticulously :-)

Hybrid idea... mmh let me think about that :-)

I will try to d that at some point in future. using Ac coupling - basically, feed the Victron Quattro  5KW ( lithium Ion) from my Axpert 5 x 2   KVA with 400 Ah battery bank on AC/IN 2

Potential losses 30 % along the way  DC ->AC and thenAC->DC.  Since the Axpert Feed only AC and water pumps I can afford to drain the Lead ACid and keep house lights on in Emergency 

 

And Yes I know it will be more practical just to install a manual change over to feed the house from Axpert. but this will be in the name of Science and will be easier to implement. Only two additional wires MCB and some clever assistant Logic 

[pilotfish]

I often read about how expensive it is to run off batteries, and that we should use the minimum possible DoD to extend battery life and therefore reduce cost of ownership - it all sounds sensible but I decided to sit down and take a look at the actual numbers.

Obviously I used my own batteries for this because how much I am paying is much more important to me than how much anyone else is paying

So the following is what I have come up with;

According to the above I am much better of at an 80% DOD than 30% DOD - am I doing something wrong?

[Chris Hobson]

I think if you look at the trends it is important. At roughly 2000 cycles and 100% DOD your batteries are poked. At 30% DOD you could get 6000 cycles and still have 40 of you original capacity.

[Guest]

Also to keep in the back of ones mind is that the batteries internal processes may have a set lifetime.

So trying to keep them to last long after that could be dodgy in that you never got to use all them cycles, they die of old age. Not so?

Got the thought form I think @PaulF007 who spoke to someone in the forklift industry who said their batteries last about +-5 years?

[pilotfish]

The thing is that the time that the batteries last is not the pertinent factor, cost per kWh should really be all that matters. From the above graph and spreadsheet it can be seen that;

• at 80% DOD my cost per kWh is R1.29 to R1.47
• at 30% DOD my cost per kWh is R1.53 to R1.72
• at 0% DOD my cost per kWh is infinity, because any number /0 = ....

So my point is that maybe we should be looking a bit closer and not making the blanket statement that the lower your DoD the better it is for your batteries.

Note that the above costs are calculated based on the fact that my batteries charge "solar only" so charging cost is effectively zero, the only cost I am allowing is the battery cost divided by kWh derived from the battery (until 80% capacity is reached, they could run for longer) - if you charge from grid or generator then you would have to add that cost which makes it far too complicated for me.

[ghatikar]

One important factor to consider is  the LOAD  or the power source the batteries replace and the cost involved - do you need to cycle the battery fully every day to meet your load requirements 

 

n terms of ROI ( return ON investment ) it makes sense to prolong the life of your investment ( CAPEX) since obviously you are taking the money for batteries  that came from somewhere even if it is your own funds they could be put to better use.

 

So generally accepted practice is to use the 50 % DOD and get average returns and life  

 

I have enclosed a sample simplified model for project IRR /NPV. you can put your numbers to see  which model is best to follow considering the cost of energy you will have to pay which the batteries replace ( Grid , Generator, net metering ) 

 

Or alternatively, use this website where terms are explained  in simplified versions http://www.free-online-calculator-use.com/npv-calculator.html

 

basically the higher the NPV ( Net Present value) the more desirable the solution 

Example NPV IRR.xlsx

[___]

I agree, cost per kWh is the metric. But the life is also an estimate and slight rounding or manufacturing differences and you may find that the couple of cents it's cheaper at 80% don't really exist. From the math I've done in the past it seems that there is essentially no difference in cost per kWh whether you go down to 50% or 80% DoD.

[PaulF007]

One thing that has me thinking the last couple of months , looking at the bmv , is that most of us use the soc to determine you DOD but I am more and more leaning towards your total AH available as a measure of your battery life. As an example:

If you have a 100 ah batt that will produce 1000 cycles at 50% ( Making it easy to do the maths )  in theory you should have 100 * 0.5 * 1000 = 50 000 ah in one batteries life cycle then the chemicals should be done for. As mentioned before batteries also have a "shelve life" but on the higher quality batteries it should not have an impact so lets leave that out. If you take the BMV it needs to Sync from time to time ( Many has noted that the BMV will "jump" to 100% near the end of the charge). If you keep track of the AH consumed you will find that over time you will "loose" some ah on the batts as it jumps 100 % and zero the ah consumed. 

The other thing that we tend to miss is that you have quite a loss on the conversion of battery power to ac and does not get factored in the costing and that could make quite a difference. 

[Sidewinder]

The analogy I always like to use is to compare a battery to a turbo.

A Turbo also has a finite useful life. The harder you work it (just like the battery), the shorter the life expectancy.

Squeezing that loud pedal and releasing it to create that nice turbo whisss...bang shortens the life, same a prolonged use at too high temps (boost).

Similar to the BMV for a battery, a EGT is in my view a must for any turbo car/bakkie etc. specially the diesels, to give one a early warning of abuse.

[___]

3 hours ago, Sidewinder said:

The analogy I always like to use is to compare a battery to a turbo.

I like it!

3 hours ago, Sidewinder said:

EGT

Oh, that might depend on the age of the vehicle. I believe the newer ones aim for low exhaust gas temperatures to reduce NOx anyway. And then there is my little family box that makes a pedestrian 93kw/300nm with a 2 liter mill... I bet you that turbo never really works hard, not compared to the newer stuff (saw a Volvo S90 the other day, they make 170kw+ with a 2-liter Diesel engine... that's some crazy power for an oil burner).

[___]

5 hours ago, Sidewinder said:

The harder you work it (just like the battery), the shorter the life expectancy.

To continue with that thought, the more interesting question is whether the relationship is linear, or in layman's terms: If I work it half as hard, does it last twice as long?

For batteries the answer seems to be a qualified "yes", the qualification being that it breaks down towards the ends of the scale. A battery that's taken down to 95% DoD every day will die in less than a tenth of the same time of the battery that is taken down to 50%, for example. On the other end of the scale, a battery that spends it's live between 0% and 10% DoD won't last 5 times longer than the one that is taken down to 50%.

In the middle of the scale, however, it appears that things are more or less linear between 80% DoD and 30% DoD, though the answers will differ between different batteries.

Edited September 28, 2017 by plonkster
Fixed a DoD vs SoC number.

[SilverNodashi]

On 9/26/2017 at 9:31 PM, pilotfish said:

I often read about how expensive it is to run off batteries, and that we should use the minimum possible DoD to extend battery life and therefore reduce cost of ownership - it all sounds sensible but I decided to sit down and take a look at the actual numbers.

Obviously I used my own batteries for this because how much I am paying is much more important to me than how much anyone else is paying

So the following is what I have come up with;

According to the above I am much better of at an 80% DOD than 30% DOD - am I doing something wrong?

Did you factor in your batteries' Peukert Factor? 

Basically, the more ampere you withdraw, the shorter the actual period of the cycle. i.e. at 50% a 1Kw cycle might last 32.4hours. But at 80% a 1Kw cycle won't last 36.3 hours. 

Look here: https://planetcalc.com/2283/

Lithium doesn't have this problem though. 

[pilotfish]

6 hours ago, SilverNodashi said:

Did you factor in your batteries' Peukert Factor? 

Yip I did. The batteries are rated at factory rated 490Ah @C10, the 540Ah at C24 that I have use in the spreadsheet is derived using Peukert formula.

[___]

Quick note on this, based on some info I found on the grapevine. There are companies out there who take a 3.1kwh LiFePO4 bank and sell it as a 4kwh bank. You can spot them by checking the spec sheet, they spec their battery with C10 and C20 discharge rates. Lithium cells are usually measured at 1C. Then there are others who oversize the bank by 15% to 20% to get a longer cycle life, which is also a kind of cheating... but a much better kind. Do your homework and buy batteries from the second kind of cheater. I'm not naming names for obvious reasons :-)

[pilotfish]

15 minutes ago, plonkster said:

There are companies out there who take a 3.1kwh LiFePO4 bank and sell it as a 4kwh bank.

I use between 10-15kWh of battery power every night so I would require at least 4x 4kWh banks if I went the LiFePO4 route - as a matter of interest about how much would that set me back? (another rhetorical question)

[Hannes7212]

9 minutes ago, plonkster said:

Quick note on this, based on some info I found on the grapevine. There are companies out there who take a 3.1kwh LiFePO4 bank and sell it as a 4kwh bank. You can spot them by checking the spec sheet, they spec their battery with C10 and C20 discharge rates. Lithium cells are usually measured at 1C. Then there are others who oversize the bank by 15% to 20% to get a longer cycle life, which is also a kind of cheating... but a much better kind. Do your homework and buy batteries from the second kind of cheater. I'm not naming names for obvious reasons :-)

I agree.  Have seen exactly (to the decimal) those same shenanigans.  Lets just say 16 x 60Ah cells is not a 4kWh battery.  As if it is not bad enough that the lead acid battery market is littered with junk claims.

But there is no excuse to be caught out (unless ignorance is an excuse!).  Insist on knowing the cell manufacturer and model that goes in the battery and check out the reviews.  Getting an opinion on the BMS used is also important.  The quality cell manufacturers are huge companies (LG, Pansonic,  Mercedes for petes sake, Samsung, BYD, CATL, Winston (mmm not sure how big), CALB etc etc and there is data on these guys.  

But I guess I am preaching to the converted as the people getting caught don't normally hang around these forums.

Those are my happy thoughts for now. cheers

[Hannes7212]

14 minutes ago, pilotfish said:

I use between 10-15kWh of battery power every night so I would require at least 4x 4kWh banks if I went the LiFePO4 route - as a matter of interest about how much would that set me back? (another rhetorical question)

Hi, about R8300 ex VAT per kWh for the brand I like, but you can do better if "crank amps" is not that important to you.  It gets less and less of a problem if you have a large bank relative to your inverter(s).

[ghatikar]

24 minutes ago, Hannes7212 said:

I agree.  Have seen exactly (to the decimal) those same shenanigans.  Lets just say 16 x 60Ah cells is not a 4kWh battery.  As if it is not bad enough that the lead acid battery market is littered with junk claims.

But there is no excuse to be caught out (unless ignorance is an excuse!).  Insist on knowing the cell manufacturer and model that goes in the battery and check out the reviews.  Getting an opinion on the BMS used is also important.  The quality cell manufacturers are huge companies (LG, Pansonic,  Mercedes for petes sake, Samsung, BYD, CATL, Winston (mmm not sure how big), CALB etc etc and there is data on these guys.  

But I guess I am preaching to the converted as the people getting caught don't normally hang around these forums.

Those are my happy thoughts for now. cheers

The Other kind of cheaters should be also avoided for home use in my opinion. I am not sure if you really want to invest in a battery with 5000 cycles.( or even 3200)  Who knows in 10 years what technologies will come to market. I found for off-grid use BYD was the best match in terms of energy density and no of cycles for  Blue inverters. For Axpert  the sky is the limit as you are not dependent on CANBUS BMS control.  But then also I found buying directly from the supplier is cheaper as you are actually not paying for the aftermarket integration. Also How sure we can be if the companies who are buying the known brands and putting it in steel enclosure with custom BMS will be around to support it in 10 years  time 

[___]

2 hours ago, Hannes7212 said:

Have seen exactly (to the decimal) those same shenanigans.

You and I are using the same source, that is, in as much as this is from "the horses mouth", I had a chat with the same horse.

I think the trick here is that if you use a lower discharge rate -- and technically that is not that far-out an assumption, for powering domestic loads you're not going to do 1C for extended periods of time -- then you do your work at a higher voltage and you can get a bit more power out of the batteries for the same amp-hour kitty. Now with that said, it still seems a terrible stretch. A LiFePO4 cell lives between 2.8V and 3.6V (those are the extremes), and if you work out the worst possible loss due to voltage efficiency, that would be 3.6/2.8 = 1.28, that is an extra 28%, and if you add that extra 28% to your 3.1kwh cell, then you ALMOST have 4kwh. And thus ends the lesson in marketing :-P

2 hours ago, Hannes7212 said:

LG, Pansonic,  Mercedes for petes sake, Samsung, BYD, CATL, Winston (mmm not sure how big), CALB

Victron uses Winston, so it can't be THAT bad :-)

Apparently CALB has a cell called the CAM72 that has amazing cycle life, at least it's marketed as such!

2 hours ago, ghatikar said:

But then also I found buying directly from the supplier is cheaper as you are actually not paying for the aftermarket integration

The simple BMS systems simply trip if there is a low/high cell, to protect the battery. They may also do passive balancing on the high side. When the BMS trips, the inverter turns off and your power goes out. The more complex setups allow the inverter to go into bypass mode, or perhaps even charge the batteries slowly to prevent damage.

For off-grid setups, a simple BMS might be perfectly suitable. Set the MPPTs properly, and the lack of full integration might not be a problem at all.

I have no idea how deep the integration goes with the mypower24/Axpert setups. I suspect it doesn't go to the same depths as the Victron setups, where the system is intimately aware of the charge current limits, voltages, shutting down and going into low-power mode or bypass, etc etc...

:-)

[ghatikar]

I think the crippled capacities are the main reason why many suppliers refuse to support off-grid use - Tesla, LG resu , Samsung SDi all currently do not allow off-grid use ( They are Li -On cells I think)  BYD, for example, recommends only one 2.5 KWH for 5 KVA inverter when used in on grid/grid Tie setup but 3 units of 2.5 KW for the same inverter when in off grid.

I found a supplier Herewin from China who are very upfront about the capacities and loading. 2000 cycles with 80 % DOD at 0.5C rate.- 10 KWH at 3000 USD 48 V .  I beleive all suppliers will be close to these figures as the chemistries cannot be so different  from different suppliers 

And the suppliers who claim to be drop-in replacement to Lead acid simply oversize the battery with clever BMS 

 

 

[___]

It's in, finally...

The RS485 cable to the BMS still runs shortcut-style through mid air because (as usual) I didn't have one of the right length :-)

[ghatikar]

Envious andSAD . I wish somebody sold Victron compatible batteries in Nigeria !

The kind of hoops I am jumping through before I can lay my hands on one

[pilotfish]

I see you keep the batteries in a treasure chest - any special reason for that

[___]

1 hour ago, pilotfish said:

I see you keep the batteries in a treasure chest - any special reason for that

I couldn't just stick them on the cold cement floor and just wire them up as I did with the lead acids, it didn't feel right. My first idea was to find a network cabinet, something that is about 9u or so high with a glass door, but those things are expensive and second hand I could only find small ones. Then I figured... nice wooden toy chest. Eventually that idea grew on me. I'm not that good with woodwork, so I eventually bought the chest at a local shop. So it cost a bit more than I planned, but considering the cost of the bank, that the BMS and fuse is nicely contained, and that wood has a nice warm look and feel to it, I'm happy with the solution.

These things are amazing. They just suck up energy like big sponges...

[___]

Oh man, I love these things. You know how lead acids taper off over 85%, still accepting a couple ampere but not really wanting to... well not these things. They recharge all the way up to 98% at all the current you can throw at it and then stop... like a plane against a mountain. A few more minutes and it drops down to 500mA... and then the BMV resets to 100% :-)

Edit: Look how the charge current drops of around 5PM...