LTO for SailBoat

mblowes

New member
Joined
Oct 18, 2018
Messages
27
I'm still doing duediligence on what lithium to put in a sailboat. Initially had thought LiFePO4s.. but now more likely LTO.

The larger form factor (66160) should mean less mechanical connections and work. My intention of this thread is to solicit responses and update my body of knowledge by linking elsewhere in this forum as I find relevant references.

The boat.

Existing
1. 300W solar
2. 4.5kWh diesel generator
3. 100A alternator on main enging
4. Currently 450AH (3x 150)lead acid (LA) house battery. I run this to just 30% (135AH) between charging to extend cycle life. Approx 100A start battery. Do not plan to replace start battery..
5. 12V System
6. Have 3 separate marine type ACs (Master berth 5kBTU, main cabin 16kBTU and rear cabins 12kBTU). I will be installing a soft start for the 5kBTU if required.
7. 2KW Charger Inverter. Will be switching this over to full Victron set up prior to new battery (3kVa Multiplus (2.4kW), Solar Charge Controller etc).
8. Boat is located in hot climate near equator.

Objective/s
1. Replace existing LA.
2. Significantly upgrade to approx 9.2kWhbattery to allow silent running of 5kBTU AC overnight.

There marine ACs are not so efficient.. Based on the measurements I've run all have a COP of about 1.3. Since all 3 are about the same I've concluded this is due to the high sea water temps 30-32 deg C.

I'm assuming running the 5kBTU AC overnight for 12 hrswith a 50% duty cycle (this confirmed by meter readings in marina).
So 5kBTU * 12 *.5 / 3.4 (BTU to kW cooling)/ 1.3 (COP)= 6.78 kWh ie 73% of target 9.2kWh battery.

So there are several things to be deducted.
Q1. Just what is the nature of these LTO cells available from China. Are they second hand , or discarded new.. and discarded why? because they cant be charged fast enough in bus / truck situations.

Q2. What is best configuration for pack/s. This needs to account for any bms / balancing required.. also the ability to manage the back .. a single 10kW pack is not viable (too heavy).
Q2.2 Is there are neat module size (series or parallel) that can be then expanded.

Q3. Just what is the safety of these cells vs others... presumably better than LTO>>LiFePO4>>liOn.

So far.

1. New electrics on order , still to arrive and be installed.
2. Have received 5 x 40AH 66160 LTOs from China. have configured them into 1S (12V).
3. Yet to receive a sample BMS and separate balancer only.


The 5 cells arrived with either 2.23 (x2) or 2.22 (x3) volts on them, open voltage. I thought that indicated similar SOC. Once I put them in series I connected them to a small 8A 12V smart lead acid charger.

All looked fine...charging away merrily for about 3 hours. The voltage steadily increasing from 11.1 to about 12.5. I looked away for about 5 mins and it shot up to 13.5. One of the cells hit 3.1. Not good (max from supplier is 2.9). I paralleled all the cells for about an hour and got back to 2.54V cells. Reconnecting in series and connecting the charger again.. and once again the two cells that had the highest initial reading once again started increasing in voltage faster.. I disconnected the charger once the hightest hit 2.8.. the second was at 2.7.. the other 3 all at 2.54.

Approximately 12 hours later the 2.8V unit is not 2.65. I have not tried to figure out if they are at their rated 40AH capacity but it suggest the high voltage reading (ie 2 cells) e are of lesser capacity than the others.

This now creates another question

Q4. Does this variability add significantly to the amount / type of BMS that needs to be used.
Q4.1. If this variability is manageable from a BMS point of view , do the life cycle projections (15000+) for these cells still hold, or is that for ideal.
Q4.2 BMS lifespan . Most only are 30000 hours , or 5 years when quoted. How do square that with cells that can last 30-40 years. eg planned BMS replacement every x years? External notification when BMS fault detected? BMS will now be important if cells are so variable.


So let me add some hypotheses to disprove.

1. LTO Cell variability can be managed, probably just with a beefy balancer.
2. BMS lifespan can be managed.. by putting in short term solution now, looking for longer term one.
3. Modular design (for weight management) for 12V system can be reached accounting for BMS needs, eddy currents.
 
Installed 80A Chinese BMS and connected 5s (12V 40AH) pack as house bank ie using existing lead boat electrics. When charging same all cells rose uniformly from about 20% SOC until 2.61 when the same 2 cells shot off up. They got to about 2.78 each.. the BMS balancing kicks in at 2.65 and seemed to struggle to bring them back (balancing current 40mA). The other 3 cells never rose above 2.55.. so did not get fully charged. BMS also ran super hot even when charger said it was only putting in < 50A..

Question out of this remains.. will this variability occur in future bulk order? Answer: most likely. Is it a problem? Possibly not.. currently its just a single series pack.. if it were a 5s3p or >p then presumably there would be some balancing along the parallel rows.

5S config may not work for when cells are at 2.3 (nominal) the total voltage is 11.5.. Freezer does not operate at 11.5 and fridge cuts off at 11.. there must be some kind of protection circuitry there. So this charger will need to be replaced with new one capable of required voltage..

I also got to properly measure the current battery bay.. Only enough space to get to 5.9kWh with LTO.. not the 8 - 10 kWh I need.
 
mblowes said:
I'm still doing duediligence on what lithium to put in a sailboat. Initially had thought LiFePO4s.. but now more likely LTO.

The larger form factor (66160) should mean less mechanical connections and work. My intention of this thread is to solicit responses and update my body of knowledge by linking elsewhere in this forum as I find relevant references.

The boat.
Wow , I did not think your sail boat was the size of a double wide trailer , floating in a hot bath tub of water :D

Q1A1) These new cells are sold as:

Grade A) They are to specifications listed by Yinlong , price 100.00$ plus each (the 30k cycles are not full 100% DOD and are at prime temperature , .2C discharge , .2C charge , and no vibration conditions with some poetic license )

Grade B ) These cells are sold by the resellers , higher resistance , at rated or slightly above capacity , some cosmetic defect ,
booger -ed threads , some minor shorts (parasitic drain )

Grade C ) These cells are sold by the resellers , higher resistance , at rated or slightly below capacity , some cosmetic defect ,
booger -ed threads , repackage (No Name) ,some minor shorts (more parasitic drain ) , possible used.

The resalers are the same mix of people any where on earth , some good , some bad , make more money , some not.

To get a good end result, Is either pay for grade A or invest some time and some money to get a cheaper end result.

New cells ,Grade B will not have scarring on at least one end from a nut tightened on it. The aluminum is "dead soft" it will scar up if used. One end only usually means that somebody put on a nut without any lube, it's that soft. Don't use the nuts they supply they are usually over sized .

The way to get the most out of batteries is to:

1) weight the battery -- 40ah is 1.24kg , 35ah is 1.2kg , 30ah 1.15kg , Yinlong is the only manufacture of this size and type.
This is the nominal weights for the sizes.

Rating system 1 best , lowest resistance , highest ah , lowest v drop ------------- 5-6 Worstresistance,highest v drop ,lowest ah

2) Measure for shorts (self discharge ) Charge all batteries in the group that your going to use to 2.7v .
Leave them unconnected to anything for at least a week. Then sort to the least amount to most of voltage loss and mark them. 5 categories for 5s setups or 6 cats for 6s setups. From the lowest to highest

3) Measure resistance of each cell at 2.7v , Mark each one , 5 ranges for 5s setups or 6 ranges for 6s setups From the lowest
to highest

4) Measure the amperage each cell at 2.7v to 1.9v, Mark each one , 5 ranges for 5s setups or 6 ranges for 6s setups From the lowest to highest, I have tested some 40ah and some at 44ah , marked 40ah


For assembly you need to know where most of the power range is 75 % is 1.9-2.0 to 2.4 volts .
5s1p power R 9.5v to 12v---- 6s1p 11.4v to 14.4 If both charged to 14.4v

A) 5s = @ 500 usable watts , slightly over charged 87% capacity 2.88v/ cell -- for ac, inverter
B) 6s =@ 400 usable watts , slightly under charged 72% capacity 2.40v /cell --for freezer, refrigerator

The imbalance of the cells of resistance and some ah capacity causes issues with the ramp voltage from above 2.4v, It does not take much power to get that spike.

You probably noticed that the two cells closest the end positive terminal had the spike and the one in the center was the lowest.

Using the ah numbers and resistance ratings

5s count AH numbers < Most Important Rule >
+#1cell -- #5ah
#2cell -- #4ah
#3cell -- #1ah
#4cell-- #2ah
-#5cell-- #3ah

5s count resistance numbers Use this formula only if you have a lot of difference
+#1cell -- #5ah
#2cell -- #4ah
#3cell -- #1ah
#4cell-- #2ah
-#5cell-- #3ah

Small shorts less than 10% 30 days , try to keep them at end of the string on the positive side, because being cycled 1plus a
day.

Using the ah rating , I was able to get away without using a voltage equalizer on some sets, because the end batteries tended
to fill up a little faster that the center ones.

On 5s2p or more are easier to mix and match , because a 30ah ( drastic example ) and a 40ah in parallel will work to get the combined watts as long as you draw out what the 30ah cell rate capable of, not to exceed the C rate of the 30ah cell

I'll get to the other questions later
 
Jim Jr.. you've packed a lot in here and will take me a bit to get through and digest. fantastic response thank you.

But just scanning.. everything I've experienced with these first 5 cells seems to line up with what you say, mind you just on a single series string of 5! :)

My guess is that mine are the Grade Bs. They have Yinlong logos. Interesting intel on the cell grades.. not sure how you built that up but excellent. $100 seems expensive.. but then if you really have a 30-40 year application the cost / kWh is still not bad at all. My cells came from Kylin (thus a reseller) and they've been most helpful. I asked if they do some capacity / resistance matching on my next (larger) order and they said they could. Will see how that goes. Interestingly they said it would be easier for them to match resistances closer rather than matching capacity. Based on your pack construction technique (ie capacity before resistance unless resistance is bad) that should fit quite nicely. In the final model the pack should be at least 5P so the capacity variation should be self solved a lot as you suggest. I'd already started to guess these 'B's would be okay for me to build my packs. You just need a management plan.

My 5 cells ranged in internal resistance from 11 mOhm to 12.9 mOhm.. if I did that measurement correct. And certainly the capacities seemed different based on the spiking that occured on a couple of cells at the end of charging.. (I didnt properly note the cells for that.). My QNBBN equalizer (5S) hasnt arrived for testing yet.. so keen to see if that solves charge spiking on this 5S1P config (ie without extra parallel cells).

Discharging was interesting.. as you noted the cell closest the positive connection seemed the most active.. on discharge it would show the lowest cell voltage.. with the cell furthest away the highest.. if I left it long enough then it was a linear decline along the pack. If I moved the positive and negative lead to other end.. after about 1 hour.. the gradient was reversed.. if I connected both to the middle cell the gradient was out and up to both ends.. I tried connecting to the positive to the first cell , but the negative lead to the negative on the furthest cell... Once again the cell with the positive lead attached was the lowest ie this effect is related to the positive connection only. My head is still spinning from this , especially since going into this I thought all the cells wouldbe at the same voltage on a parallel bus!! Take the load off and within about 30 mins to 1 hour the cell voltages did equalise. I wonder if i build a harness .. connecting at three points along the parallel to one point going to the load - if this gradient would be solved.

Based on the other bits I have .. discharging was as a parallel pack, charging is as as series pack.. so I dont think what I've found contradicts what you found on charging ... which I'm inferring as the first cell (cell with positive external connection) is the most active.

I will charge up tonight (and look for that charging pattern) and leave for a few weeks.. I need to sail elsewhere to get some other parts. When I get back I will check the parasitic readings.

So if I heard you correctly .. the basic plan of attack for series string building is :if internal resistances arent too bad , you should be able to balance 2P+ strings fairly well.. probably just throw out cells with really bad parasitic loss.

I need to read you battery stacking algorithm and get clear in my head too:

5s count AH numbers < Most Important Rule >
+#1cell -- #5ah
#2cell -- #4ah
#3cell -- #1ah
#4cell-- #2ah
-#5cell-- #3ah

This is referring to a 1x 5S string , right? And suggests putting the weakest cell with the positive terminal .. correct? I'm not sure how I square that with my discharge learning above.. from that I guessed I should be putting the most powerful cell next to the positive.. Anyway, as I said .. you've packing a lot I need to digest.

Other.

Not real happy with the 80A BMS.. Ran way to hot with just 40-50A.. dont want that on my boat. Assume that is since its using Mosfets or something to do the switching.. Dongwant to go down the Batrium route for the this pack , these cells seem pretty safe... but would like something. The one brainwave I had was potentially using one of these cheap monitors to drive a large external (say 200A) relay.. use the bms charge/loadcircuit to drive the coil on the relay. I'll draw that up in the next few weeks..but its working in my head after a few beers. Not cheap,, but Victron even appear to have a microprocessor relay that would not only fit quite nicely (230A) but would have a much lower parasitic draw.

One issue is one bank or two battery banks.. Most of what of read on lithium is go or 1. But boat has 2 banks.. and I would like to keep that.. its a redundancy when underway that I dont want to lose (ie sail on one bank.. if something goes wrong you have 2nd for Plan B). Unfortunately both the charger/inverter I have now, and the Victron I'm about to upgrade to have a single charge/load circuit. Having two banks means there could be an inrush if they are at different SOC when paralleled together. I've seen solutions for when the charge / discharge are different paths.. but not for common.. more searching to be done.

I'm also at the boat now. And able to measure the battery bay. Interestingly enough the density of LTO is significantly less than LiFePo4. I may not be able to get enough LTO kWh in the space I have. I may have to go with LiFePo4 (I can get 32650 LiFePO4 6AH for USD2.14 delivered to my door here) Not the end of the world.. the whole point of learning here was to do even larger (LTO) installs at a house and farm next year.
 
Jim Jr.. Dude.. you were right on the charging.. the middle cell is the lowes !. two on either end are about the same .. only an 8A charger so a ways to go.

Also.. what's the best/simplest way to do capacity.. at moment I've been doing it through a resistor.. that is not really constant current.. but if all I'm trying to do is rank them does that matter..? Answer is likely not.
 
mblowes said:
Jim Jr.. Dude.. you were right on the charging.. the middle cell is the lowes !. two on either end are about the same .. only an 8A charger so a ways to go.

Also.. what's the best/simplest way to do capacity.. at moment I've been doing it through a resistor.. that is not really constant current.. but if all I'm trying to do is rank them does that matter..? Answer is likely not.

Your are correct, as long as you test with the same load to the same voltage.
I use a load meter with a external power supply and a bank of 6v light bulbs at .5C and let them go to 0v.
Testing the bank, I use a 12v car ptc with fan attached to try to get "Usable watts "

https://www.ebay.com/itm/Portable-3...epid=2112328471&hash=item284753688b:rk:1:pf:0

Max wattage is @400 watt

As for the Chinese bms , they will run hot because of the chopping resistors and mosfets. I have a machine shop and made
heatsinks for my capacitor chopper boards, I use a 2.5v unit for only limiting the max volts a 6s bank.
The 80a unit is too small for what your doing, it's for like a 10-20ah bank 48hr cycle.
With these batteries you only need to limit max volts on charging.
 
mblowes said:
I'm still doing duediligence on what lithium to put in a sailboat. Initially had thought LiFePO4s.. but now more likely LTO.








Q2A2) The best configuration is what suits your needs, 48v systems and higher are efficient when you need 120v, But you need
need more of a bms system because of more cells in a series, and more cost.
12v systems are not as efficient to 110v because of higher amperage parts the heat generated. But you have less cell count to
manage .
But if you have systems already 12v there are less conversion losses.
As for keeping all your batteries in one spot is convenient but not , written in stone
the batteries can be in other areas as long as, not in living area, fuses or circuit breaker on both ends , and not looped .
Connecting in a loop can have a magnetic effect on a surge of power, AC circuits would be like being inside a large transformer
Q2.2A2.2) See above
Q3A3) Barring external fire ( not the battery) , external shorts , and being squished in a press, They have shone not to burn,
explode from over charge, puncture,dents to halve size of thickness or zero volts and over discharge.
The cells do not contain any lithium metal or carbon. The only hazard is putting in a fire , like a aerosol can in a fire , they
will go boom.
Q4A4) Yes
Q4.1A4.1) Those are perfect condition testing with grade A cells with some poetic license . Thur the mountain of data from
the different manufacturers and scientific papers that i have read are 12k to 20k cycles for storage cells and 30k for starting
performance cells aka" Toshiba 2.9ah and 10ah cells" on average to 80% and if used to 50% most people won't live that
long. The cell manufactures will not sell to individuals . Imagine a car that would last a lifetime , that's only one per person per lifetime.

Q4.2A4.2) I don't see a problem with 20-30 years, provide that they are kept relatively dry and free from major jolts and vibrations. the cells are sealed to keep the good stuff in and the bad stuff out.
The cell chemistry and material are copper, aluminum , anode -cathode ( oxides) materials and a lithium salt with poly-vinyl
alcohol (electrolyte) and plastic type separator. None of which is corrosive to each other to any extent.
 
Jim Jr: sir I'm starting to suspect you may be Edison incarnate. I'm getting a whiff of most of what you say from my own experiments but you are way ahead of me on everything.

Truly thankful for your input.
 
Today I toggled back to LTO.. I realised that though I can fit about 432 32650 LiFePO4 6AH cells in the space.. by the time I derate them 30% to get 3000 cycles I have the same capacity capacity as 60 40AH 66160 LTOS at 100% DOD.

Have decided I need 2 x 6s5p since not only is there an inverter , and air conditioner to drive but also a 12V water maker which needs a lot of herbs in the upper voltage range.

I've attached a basic idea of how I want to configure the 2 packs of 6s5p to fit the battery bay. There is an engineering shop here at the boat yard who can whip the plates (and some busbars) up. (I have not included the series connections in the graphic yet, or nuts or wires). Is there anything inherently wrong with this construction.. 1) is there anything wrong with having the cells vertical (is there any electrolyte to run down?)I figure aluminium plates will have a fair degree of mechanical strength. There is some movement , but its a boat not a car.

There is 2mm between each cell. Have not included cell fuses.. may add some to the pack overall.


image_hnahmf.jpg



Mind you , in parallel I put some feelers out to a few suppliers today in China and they are either nil stock or price has jumped from USD18-20 EXW to 25-40 EXW. Ouch.
 
All of the cells are recycled, none of them are new as YinLong do not sell new cells outside of BYD. Stay away from cells that do not have the original blue wrapper.

With 5s you can't use them with a typical engine alternator (which peaks at 14.2V : 2.84V per cell) as the alternator will deliver maximum current and your BMS has to be a custom job to cope with a 100A balancing bypass. Basically, 5s will not work economically on a 12V system with an engine involved and you will have a very large balancing problem to deal with.

12V charged to 14.2V from a typical engine alternator before it switches back to float is only 2.37V per cell on a 6s setup and here you are only at a 72% charge state.... 60Wh to play with per cell. Most 12V rated equipment for vehicles/boats will usually cope with 15V, which is a higher cell desulphate voltage level from some lead-acid, which ends up as 2.5V per cell on 6s and then your up to 82%.

Get rid of the starter battery if you are staying with a 14.2V upper charge level from the alternator because the LTO cells do not sag anywhere near as much as lead-acid with a high starter loading and your pack should then give the engine a higher turnover speed, better starting and prevent starter motor burnout with a comparable low charge level from a lead-acid. This can prove to be a big win if you have engine issues and using the starter a lot without the engine firing up....

The batery pack setup seems more designed to sit on a flat workshop more than in the unused spaces in the hull of a boat....
 
Thank you completelycharged.

The boat is now nearly 15 years old so, I'm upgrading the electrical system in advance of the batteries. This will be Victron gear (when it gets here!). It is a little pricey but seems purpose built for marine and gives me a little more flexibility when I'm learning. (And an old guy can write off just about any amount on his beloved boat).

The Victron charger has a number of profiles including for lithium, which can be customized. There is a max voltage of 16V (12V system) so I'm hoping I can tweak their LiFePO4 profile up to this to run the 6s LTO ( ie 2.67 which seems pretty close to full). I'm pretty sure the desulphate / equalize function on my current charger is north of 16V (they do whizz a little faster) so I'm relatively confident all the loads can handle the likely voltages.

I'm keeping lead acid for the start battery. The boat is in Thailand and this is both the readily available battery technology now and well into the future. The house (LTO) bank can have a problem , but the main engine needs to be able to be sorted locally and fast if there is ever an issue. And the cost of one of these starter batteries every 4-5 years is not consequential.

The Victron 12/200 BMS deals withlead acid start battery combined with lithium house. It charges the lead acid first , then the lithium and can self regulate how much charge is taking from the alternator . The alternator is 100 or 120 A.. and my other research told me that lithium could suck all that in and more so needed some kind of limiter to the standard alternator / regulator to avoid overheating.

Most of the nooks and crannies in the boat are already jammed full of stuff! So these batteries have to fit in the old battery bay. More and more I'm doing things in a way that should help make the boat more saleable down the track. For the right buyer , lifelong batteries and night time silent air con in the tropics should be a great feature.. and if they dont want to pay a premium I will simply pull the LTO out, drop in lead acid, and recycle the LTOs to an ESS somewhere.

I've been trying to figure out what's wrong with these batteries from the start.. they seemed too good to be true. This supply hiccup may indicate problem number 1. You're dependent on one supplier (YinLong) who can yank the supply(from the resellers) at any time. Kylin , who seem to have great customer service , jumped from 17.50 -> 30.00 in about 10 days. I'm really hoping this supply gap/price rise is as a result bunch of Americans buying in ahead of 25% tariffs in January rather than something else more structural. We will see.
 
Alternators regulate the upper current based on the engine rpm so just avoid running the engine at high rpm when charging or take the alternator 12V coil feed wire and switch it separately so you can then keep the alternator turned off if needed.

I would wire the LTO to the starter and keep the existing battery on float in that case and the reason being was from what difference I saw. Granted I only had a 20hp twin V but that was starting a lot better than the 65Ah lead acid with just 2x6s packs connected, I had an extra pack for capacity to run some computers on the vehicle while it was not moving. I had the starter running (on and off) for the equivalent of over a minute while I was sorting out getting the engine running again. Started up, 55A charge back into the cells. The voltage drop with the starter also knocked the computers out with the lead-acid as the voltage fell too far, but not with the LTO.

With LTO it would be ideal to replace part of the keel balast in a boat but for existing boats near on impossible ! I was thinking more along the lines of shaping the cells to match the shape of the hull so that they could fit each side or along the bottom and only add one cell width to the space used.


I'm surprised that at the momemtn YinLong are not doing what the vehicle manufacturers are doing and recyling their own cells into home or commercial powerwalls. That situation may change but the cell will still end up on the market, just like relatively new Tesla batery packs from crashed cars.

If I had a supplier who quoted 17.50 one week and then quoted 30.00 the next I would not necessarily call that great customer service, depends on their real reasoning for the hike.... special pricing just for you or market wide ?
 
Believe the pricing is market wide.. I hit them all and got nothing under $28.. and of course they then make up that in the freight anyway.

Rev 2 for the battery packs.. Will stick with aluminium (aluminum) bus bars to match the pole material. Needed to beef them up.. essentially anything serial has to take up to 400A. The load is fused at 300 so battery internals should withstand that and more. May power (inverter load) should be 200A (per battery. There will be two batteries.. and I plan to add a 225-250A circuit breaker to one to handle situation if I've run one and accidentally switch them in parallel ie different SOC.

For first build I wont be fusing cells. Is this necessary in a 5P? I assume if one cell internally shorts it can only take current from the adjacent parallel cells so in this case 4 other cells for a total of 40 * 2.3approx 370W.. if near fully charged. Interesting in the alternative LiFEPO4 4s54P is would be 53*3*6 W to dump into a shorted cel ie 1017 W ie 3 x the power into a volumetric mass 1/10 the size (32650/66160).


image_zcfllr.jpg
 
Add a breaker on each battery pack if your hoping to be at sea and have a cell issue, that way you can switch/trip out either battery pack. Plus, there may be a time when you want to discharge half the pack or keep half a pack fully charged as a reserve.

If an LTO cell in your configuration shorted and was capable of actually holding an internal short (which they are not) then the short circuit current per cell is around 800-1200A depending on the internal resistance age so you would see around 4,000A trying to short into the cell at somewhere around 1.5V, so about 6kW. This would melt the aluminium rather quickly. There is a video of someone with a clamp meter and shorting a cell with welding cables on youtube. The current and power capabilities of the cells is a lot lot higher than you have in your calculation.
 
completelycharged said:
Add a breaker on each battery pack if your hoping to be at sea and have a cell issue, that way you can switch/trip out either battery pack. Plus, there may be a time when you want to discharge half the pack or keep half a pack fully charged as a reserve.

If an LTO cell in your configuration shorted and was capable of actually holding an internal short (which they are not) then the short circuit current per cell is around 800-1200A depending on the internal resistance age so you would see around 4,000A trying to short into the cell at somewhere around 1.5V, so about 6kW. This would melt the aluminium rather quickly. There is a video of someone with a clamp meter and shorting a cell with welding cables on youtube. The current and power capabilities of the cells is a lot lot higher than you have in your calculation.

Thanks CC.. I've just received the cells and inverter/charger and getting the busbars machined so should be able to get set up in test in the next few weeks. I intend to run with just QNBBN equalizers to start as this seemed to be the most effective in the 5 unit test. The inverter charger has a max charge of 16V so with the equalizers on it should hold things together , at least into the medlum term.

My plan now is to randomly assemble the parallel packs ie not try to pre measure for capacity .. the logic being that they should randomly total out at 5 x 40 = 200A per parallel. As there will be 2 banks.. I will strip one down alternatively every 2-3 months , let sit for a bit.. and measure the parasitic loses on each cell to see if there are any duds.

Believe the equalizers are rated to at least 30000 hours. At this stage I will simply wholesale replace them every 4-5 years.

In that time I will give some thought the BMS I really want. For such long life cells I suspect something different is needed to current off shelf and Batrium seems like overkill. My guess is shunts in the balance lines and tracking the balancing current might be the trick.. Failure to balance (eg equalizer failure) or changes in relatively amount of balancing (indicating possible cell failure ) seems to make most sense. Also a need to alert externally and maybe some kind of proof-of-life feature is needed remotely eg a routine on a Linux server pinging the BMS every month or so to confirm it is functionally ,seems to make most sense to me where I'm at now.

I will be fitting breakers to each battery. For the reason you suggest and also in case I inadvertently connect them in parallel and they are are wildly different SOC.
 
Before assembling the packs, connect all the cells together and leave them for a few hours to balance out. If your cells are anything like a typical delivery they will be in a few different voltage groups from whatever state of charge they were at before they were taken out of whatever it is they came from. If the voltages are greater than around 0.5V difference you may want to ease the balancing current or have some rather warm cells and wires....

Having seen some test data I did not bother measuring before assembling packs and have been running my cells for 6mth now without a BMS at all and they are still very close (cell to cell voltage) to when they were put in. The QNBBN units should last a long, long while and you only need them connected once per week for a few hours..... the more they work the more you have something wrong in your system. Some form of over/under charge protection is a must. My over charge protection is in the MPPT charge controllers and under charge is the inverter cut-off. Again no separate all in one BMS style.

Batrium is good for some as it provides a lot of information, pricing in my view, it is very expensive for what it does for the specs if you have a high series cell count which you get with LTO on 48V. That's my opinion and will probably get flamed for it.

The breakers are must in my mind, I have seen and caused enough shorts in my life and dare to imagine some of them without fusing to reduce the peak current before the weakes point vapourises, melts and burns a hole through the bottom of the boat....
 
So put the first of 2 batteries together 6s5p this evening. The engineering shop did a good job on the busbars (4mm Al) so all fitted quite nicely.

Made up all the 5p packs but connected them in || overnight to equalize per CC suggestion.

The only issue here where the aluminium nuts. I'd bought in 150 from China. I stuffed up my first 5P assembly by 'forcing' with a wrench the nut.. it burred into the cell pole.. ugg. After this I finger tightened all of them first. those that would not finger tighten into the sin bin. I had one 'golden' black carbon nut and was able to verify that not a single Yinlong cell (out of 60) had a problem on the thread... ie very clean. All the problems were with the nuts with 22 failing..on the way to geting the 120 I needed: they probably can be easily cleaned up ... but you really have to be super careful working with aluminium! (Spelling is Australian).

After the first 5P disaster, the next 11 x 5Ps went fine (after applying the finger tighten test to the nuts) so lesson learned. Tomorrow I'll sort out that dud 5p and || the second pack.. any suggestions for getting these frozen Al nuts off?

Inspected the new inverter.. looks like its been banged a bit in transit.. .fingers crossed. it will work.. that test likely in about 2 days.

Have to say I'm pretty pleased with the bus bars.. put a lot of effort into the design of those and it paid off - assembly was remarkably easy and each 5P group seems quite robust. One issue.. lock washers.. Based on the compatibility issues alerted by Jim Jr I've been trying to keep everything Al to match the cell poles. So busbars and nuts are also Al. The metal is so soft (kind of locking the nut) that I have no problems with testing and even some production use as is.. but there is some vibration when sailing.. For the life of me I have not found Al lock washers anywhere... does anyone know a source? I could double up on nuts (but thats about USD60 landed.. which is a bit there).. or even dab a little loctite in the end.. I dont want to revert to stainless ones (which seem to be everywhere) as I'm all to well aware of the galvanic issues that will create.

I may not fit the equalizers for a while .based on CCs recently and Jim Jr some time earlier.. I will be all over the cells for a while constantly checking and learning.. ..

As for high / low voltage protection. . For high side the inverter /charger max is 16V. I figure one of the advantages of the 6S config is that 16 / 6 is 2.66 V.. just a tad (0.04) under full charge for these LTO.. I'm hoping that's close to perfect (esp when they are balanced). For the low the inverter has some adjustable shutdown settings and I will likely set this to 12v. I believe in the box of stuff I ordered there is a Victron Battery protect device which can also de-coulple at a selected voltage.. so this will ultimately be what I will use.

I need to give some more thought to comments of the weekly or monthly need for balancing ie only connect equalizers then... this kind of adds complexity to the balance wiring (ie switching somehow)..... I have solar panels that are and can can be connected even when away form the boat.. (about 900W). I assume the problem with leaving the equalizers connected all the time is that they chase parasitic loses around the pack until they flatten it. I reckon if I leave the solar always connected (as it is now for the lead acid) , and have the equalizers connected this should become a non problem. Comments on this logic welcome.
 
Here's the first battery (12V 6s5p) in the flesh...

image_ifhfzb.jpg


I'll recant on my earlier favourable busbar comment.. when I tried to put the heavy duty joiners in .. they did not fit.. the holes in all these 5p end plates were not quite centred.. will have to go back and enlarge holes in joiners to accommodate and work with the machinist on accuracy !

Separately.. went back and tested the first 5 cells after a month of sitting. They had all been stored at 2.56V.

Today , 1 month later and with the same multimeter, the voltages are now:

#1 2.55 -0.4%
#2 2.52 -1.6%
#3 2.53 -1.1%
#4 2.53
#5 2.56 0%

For an average loss of 0.8%. Need to relate that back to the specs and also note that its not a statistically valid sample size. But still interesting. Back to my last update on parasitic losses.. these presumably will chase themselves down in parallel packs and ultimately lead to different voltages between the series connections.

Since I do have a solar panel available I'm inclined to leave the balancers connected perpetually.
 
Back
Top