LTO Powerwall

completelycharged

Active member
Joined
Mar 7, 2018
Messages
1,083
Yep, not 18650 cells...
This is with the standard 66160 LTO cells that are in 30Ah to 40Ah capacity. The 30Ah cells are widely available as recycled units.


image_cnfhqh.jpg


The cells are arranged in an 11s2p pack with a small space between the cells for additional air circulation and tollerances.

Additional small holes in the joining plates (2mm Aluminium) are to screw the bars into an outer wooden frame. and for connecting BMS cables.

The LTO cells have a working range between 2.7V and 1.5V and clipping the top and bottom of the voltage ranges to where the capacity lies with an 11s arrangement this gives the closest proxy to a standard 24V lead-acid battery voltage range and makes it more compatible with a standard UPS unit.


image_omyyjw.jpg


Each pack will then be 24V and a capacity of 2kWh (using 40Ah cells).

External connection will be via two XT90 connectors so that either more than one connection can be made or two connections to a cental busbar.
 
The busbar setup for the system will be a 12mm (or 18mm) Ply board with plastic stand off mounts for an aluminium busbar.

The plastic stand off mounts are actually part of a PVC windowframe, which after taking appart for the glass some of the sections looked really useful. So, with a bit of cutting they are then the ideal material to mount the aluminium bar to, which then also prevents any issues with mounting it to wood and moistrue. With mounting directly to wood, at 12V your quite safe, however my system is going to be 72V and I'm starting to get cautious. This is comming from someone who used to play with 7kv transformers at the age of 12 seeing how big a jacobs ladder I could make, while blanking out the TV reception for around a quarter of a mile.... long time ago...


image_xgkngd.jpg


The busbars have earth blocks attached rather than drilling for lugs because it works out cheaper per connected wire and a simpler build process.... while allowing for future flexibility.

Using Aluminium also works out to be a lot less costly than copper and all you have to do is upsize by dividing your copper csa requirement by 0.65 and the aluminium resistance is then the approximate level of copper. Plus, there are not a lot of places selling copper bar at sensible prices in small quantities.


image_kmnuxq.jpg


The busbars are arranged so that the bottom bar is 0V, next bar up 24V, then 48V and the top bar 72V.

The middle 2 bars do not need to be anywhere near as thick as the outer bars because the terminals will be taking the majority of any current flow and only pack to pack balancing/imbalance will flow along the middle bars.

The terminal blocks at the ends (may revert to lugs for these connections) are then the offtake points for the load via a 300A shunt then to secondary MOS arrays on a second single aluminium bar mounted at the end of the board...

The idea is to then have leads from the bars and separate leads from the battery packs, if the pack is at the same bus voltage then just plug and switch on.. or if they are out of balance, try and switch on and see if the packs 100A breaker trips or melts :)
 
This is what the PVC window frames look like for the busbar supports and the thinner Aluminium bar (38mm x 3mm) that will be used. They are cut into 80mm lengths and then drilled for wood screws.

Free, fully insulated and strong busbar mounts... good for any conductors, copper pipe, coper wire, etc...

[diy]rkxosk[/diy]
 
i like the approach to go with LTO cells very much! since storage space in physical dimensions is secondary for powerwalls (in my case) these cells should be a lot more uncomplicated in handling as well as less dangerous when misused. and last: great compatibility with existing systems designed to deal with lead-acid batteries.
the only drawback is hard availability in my area (austria)

waiting to see more progress on this!
 
With almost all of the parts now arrived the actual build process is set to star.... Still waiting for some DB15 conectors for the BMS hot plug for the packs, which is that last critical item that would prevent the packs from being used due to the wiring offtake positions.

Have altered the front panel slightly so that the breaker is now mounted on a DIN rail as I may add a smaller 20-60A breaker(s) on some packs for XT60 connectors that would allow individual packs to be used with a lower than 100A breaker ready to make magic smoke from unwitting devices. The smaller battery isolator switch would control these.

The rotory isolators are not rated for much above 24V so can't be used on the packs for isolation when in series connection, because the first switch couldbe trying to isolate upto 90V for the 72V pack and potentially a lot higher if many packs are connected together.

With the DIN rail the breakers can be easily added later fo all packs if needed. The DIN rail also made mouting the breaker a little more robust and allowed it to be sunk into the panel more so that the trip arm is less likely to be accidentally caught. The brekers are not intended to be used as regulat isolators/switches because I'm not entirely sure they would last a lot of use and expect them to only last around 5 fault trips

[diy]djssmi[/diy]

The ESP array was the test setup for the BMS system and the first unit will be a hot plug box with 12 units in (11 : one for each cell and one for pack overall shunt or volts).

The large power supply is a used 2800W Cisco Catalyst power supply that can deliver 50V at 28A - 62V at 28A or 113A at 12V. There are another two units rated at 4200W that are intended to be standby charger units from the grid/generator. Very cheap high amp power supply option... plus the outputs are fully isolated....
 
Progressing with building the battery boxes (18 of them !) and cut just over half of the wood required for the build. Appart from the sheets all the other wood is from local skips throwing away a variety of good wood... still around 200 battons to cut and then final measuements for the width after the cells arrive.

9mm plywood is the main construction material, 3mm was too flexible for the weight and 12mm may create an issue with the bolts on the ends of the cells as there will be 3mm Aluminium cell connector + 12mm wood + 2mm washer + 8mm nut and I don't think the threads are long enough... reminds me I don't have enough washers.

[diy]rflngw[/diy]

The generator will be used when in a remote location as a backup for when there is not enough solar or wind. The 6kW charger until will be explained later in another post, recycled equipment costing less than 50 and will be CC CV with low upper V limit.

If it is not raining tommorrow (and sunny with solar power !) I will hopefully finish off cutting the battons, start fixing some of it together and start drilling out the holes for the cell mounts.
 
Bit more progress with the battery side supports for the boxes, which are now all drilled with 792 x M12 holes and now need the Aluminium cell joining plates, which is the next job to cut up.180 cell joining plates plus36 end connections and 1600 holes to drill.

[diy]aalskb[/diy]

The other 18 thin strips of wood in the image are for extra strengthening and going to be used to join the corners/edges together in the boxes and came out of two long floor boards that were hanging out of a skip I passed by.
 
Working on the pack balancing arrangement and on the front will be a DB15 connector with fused access to each cell within the pack. Axial glass fuses of 3A maximum (might de-rate to 2A) would provide internal pack protection from anything that is plugged in.

Small volt meter will be included in the pack and this will be swiitched with separate 1A fuse protection.



image_evfcsu.jpg


Connecting into the DB15 will be a connector that is separately fused and passes through a 0.27 Ohm resistor.

All of the connectors for a given parallel pack will then be connected together, i.e. in my instance 6 packs of one voltage will all be connected together.

The main balancing would then be done from the central connections, i.e. six of the above circuits would be connected together with the left hand DB15 connector. These connections would then be where the whole pack is balanced, current limited to a maximum 2A per parallel set. This is the diagram for just one of the 11 celll sets in the packs.


image_pdmycy.jpg


The reason for passing the connections through a nominal 0.27 Ohm resistor is to prevent the fuses from being easily blown on faults and excess over/under charge states. With a pack the cell to cell difference between two packs may be a maximum of 500mV. With a maximum cell to cell difference of 500mV passing through a 0.27 Ohm resistor the resulting current flow would be 1.85Amps or 0.93W. The aim is not to put all the cells in balance all of the time at every charge point, rather gently push or pull them into balance over days or weeks.

This interconnect arrangement will be a temporary solution ahead of the separate BMS wireless boards being plugged in once I finish them after the pack builds.
 
Bit of progress, cells arrived and have managed to mount most of them into the frames/boxes that I have been preparing. The packs are 22s2p and give a working range close to a couple of lead-acid batteries of 22.0V to 28.6V.

[diy]fbmxjl[/diy]

The cells appear to have been recovered from at least three different items as the voltages, labeling and consistent marking on the cells seemed to show three distinct groups of cells. The 66160H is the 40Ah cell that Yinlong produce and is th emost recent version of the cell, the 66160A is 30Ah and the first generation cell they made, the 66160F is second generation at 35Ah and I have not seen any of these in the used market so not sure where they were used if they were produced in any quantity.

[diy]aveeoz[/diy]

Change of design on the fly as I ended up putting the connecting bars on the outside of the boxes, which is not ideal because of the connection. The threadedbolt will carry the current and then back through a spring washer (compressed flat), which makes the compression washer the narrowest part of the conducting are along the path and is around 20mm2 steel. I'm not planning to pull more than 20A on a continuous basis per cell from any of the packs (40A per pack or approximately 1kW at 25V) so it should not cause any problem.


image_icymac.jpg


Working on the balancing leads and connections, the image is part of a test board to help wiring up the rest of the leads to each of the packs with 14 wires to connect per pack to provide cell voltages external to the packs on a plug-play basis and hopefully not a plug and bang basis :D

[diy]aojdli[/diy]

The heat shrink on the back of the connector is just as much to prevent the cable from touching anything should it break off from the conector as the heat shrink will hopefully still extend beyond the bare wire. The Connection is arranged so that the maximum voltage between any two adjacent pins is only 4.6V so any bent pins or shorting will not result in a big flashover. Fusing, still having a think about internal axial fuse or external fuse as I don't like anything that is not on a fuse.

While the sun has been out each of the packs has taken a turn in a slow 10A / 250W charge to bring the voltages up, check the balances and try a few experiments on cross connecting cells and seeing how many Amps flow. Total charge so far is around 4kWh, just need to get the first UPS connected.
 
where did you source the batteries? here in the region of austria/germany it is almost impossible to get those...
 
They were from OSN Power in China, shipped via sea freight as a single pallet and crossed fingers that they would actually arrive.

OSN may try to say that the cells are "new", they are most deffinately not, but they appear to be in a very good condition electrically and still have virtually the full life remaining.

All of the European LTO manufacturers will not sell direct and seems the same in the US, I have tried contacting all of them and spoke for a long while with one of them but still no prospect of external sales for at least 2-3 years. At the moment any LTO cells in the marketplace are recycled or manufactured in China in a small factory with limited quality control and would not touch them with a bardge pole (old English saying...)
 
First pack almost fully wired up withconnecting up the balancing cables after building a small test box to display the voltages for the cells.

[diy]fadcgv[/diy]

Slight issue I realised after and too late was the volt meter would not alter the decimal so it stayed at 0.1V resolution...

[diy]cxmwpe[/diy]

The last cell reading is double because of the way I had to wire the volt meters as they will not power up from less than 3V so the power is stepped with the last cell reading the cell 10 and 11 combined. Will change this in future with a boost converter just for the last cell because each volt meter needs it's own -ve reference and so I could not supply all the volt meters from a single source.

Wiring is held in final position with hot glue.

[diy]pboeil[/diy]

Reverse side of the front panel, DIN rail fix with the two volt meters and switches still to install. The DB15 is flush with the front and the back filled with hot glue to hold it in place without any screws and supports the cables better.
 
First 9 boxes wired, charged and powering my computer at the moment. Wooo, sun power with no sun outside...

[diy]pnqkak[/diy]

This was an earlier test with 3 of the boxe - yes, no separate breaker - testing :D Now with 63A breaker...

[diy]mnnplj[/diy]

The MGE Unit pulls hardly any power when nothing is plugged in, while the Borri units (T3 2U 3KVA) seem to pull around 70W even with nothing attached when the UPS is active. At the moment with 14% load the Borri unit is pulling about 6.5A at 81.1V so a little bit of losses in the UPS, at least it will save on a heater in the winter... :p

[diy]xoizna[/diy]


The hot-plug unit I made to check the voltages works ok, if the displays actually worked properly and moved the decimal so I had 2 decimal place accuracy, but I guess that's all you get for cheap China diaplays.. Will be making a second one with better displays and a boost for the last volt meter (powered across 2 cells and measures one for each voltmeter). Worked well enough while wiring the leads up.

[diy]uwiscw[/diy]

Downside is that I did not connect any of the voltmeters into the battery boxes because of the accuracy and a realisation that the 30V units will have slight issues if the bus side voltage is reading 48V, which will occur with 2 pack breakers on and the third one in an off state.

Next up, a bit of soldering for the balance leads..... all 225 of them.

[diy]ykxhqd[/diy]
 
Running the UPS for a day now and have the battery packs unit on charge at 1300W (78.3V and 16.7A with 1A from the UPS while it is pluggeg in and 15.7A from the charger).

The Borri units I have (TU U2 3KVA) appear to have a lower cut-off voltage around 10V of lead-acid or 60V so that provides for a cut-off per cell around 1.8V and the upper 100% full reading is 13.75V lead-acid or 82.5V total or 2.5V, while the charger may go to around 14.2V lead-acid or 85.2V during the 3 step charge for the desulfation phase which would then be around 2.6V per cell.

This is why a pack size of 11s works for Lithium Titanate cells if your using it with anything that was made to work with lead-acid batteries and gives a resulting working range for the cells of 1.8V to 2.6V. The cells will work between 1.5V and 2.7V but this is similar to all Lithium cells in that you may see some issues running 0% to 100%.

[diy]jaftxs[/diy]

Balance lead boards are in progress with fuses and resistors - still adding resistors and fuses to the 15 boards...
 
Our cells arrive two days ago, brand new...atthe moment we are testing the same cells, Yinlong 2.3V40Ah, our plan is to build 14P24S 48V30kw. max output 5kw peak 10kw.
For .125C we running 2.05V -2.5V a straight line 39,7Ah... charging them higher to 2.8v brings nothing, discharging them lower than 2.0v also nothing.
Our Italian friend did the same test and with the same results and conclusions.
Max power we seeis about 42Ah 1.5V-2.8V.
We are interested in your progress with balancing as we noticed that we need more than 1.2A.

My plan is to use two aluminium sheets 3 or4mm and sandwich the cells in between.

Not sure how you build you pack, but is there really wood between the cells and the connection, or is there a nut in the wood..?
 
At 0.125C (5A) between 2.05V and 2.5V can you confirm that 39.7Ah (5hr) is correct because 42Ah full 2.8V to 1.5V less approx 12Ah due to 2.5V and 2.05V range makes 39.7Ah for an operational voltage range curious ?

Only wood and the 12mm thread from the cell. Good for 35A per cell without any warming.
 
Till now the plan is to 2.0V to 2.5V, beyond and above there is no possibility to balance the cells in this system.
Currently testing 2P5S 10Ah cells for a portable pack, with a simple balance board set at 2.7V.
For static charge and discharge there will be no problem balancing a bigger system, for a solar system it is a whole other story.
Your approach to build smaller parallel packs I had in mind when starting with LTO, but there is a lot of electronics and connections involved, that is the mean reason I will go for 14P and figure out what the best option is to balance.
If we can not find anything on the market, I wil make it myself.
Another important part will be battery protection/ and control settings for charger/inverter/generator, that wil be easiest part to build.
Main problem on this side, I did not design a pcb for over 15 years now, and to hook up the new software and dig in it cost a lot of time.
I wil start design/develop on plain pcb, and solve one step at a time.
It would be nice to find someone who is running the pcb software on daily base who can design a pcb for this system, as I am sure a lot of people sneaky reading our progress, with a almost unbeatable storage system...
 
After a lot of searching around for parts, inverters and charge controllers I am comming to the conclusion that 72V is currently the most unsupported system voltage. While EV and bike packs may be 72V there is a big lack of choice with solar charge controllers and inverters that will work for 72V packs. The critical issue is that there are no off grid wind charge controllers that work with 72V which are actually designed properly as wind charge controllers, not just PWM rubbish that will destroy a wind turbine or fixed MPPT voltage style.

The72V UPS units work well and gives out a 1A charge when plugged in (would take 17 days to charge 30kWh) and pulling the power cord turns it back to battery mode, which is ok for testing and running some computers and electronics.

Charging 72V is cheap and easy via the power supply I posted separately about and the system is currently charging at 1000W from one unit, while sorting out the BMS leads and parts before contemplating re-arranging the solar to charge the packs directly rather than going through the grid and back.

Changing to 48V arrangement is now looking very appealing but a little scary as the busbars will have feed in breakers for 900A (100A x 9) if I add in the last 3 packs to make 18 in total. The real issue for concern is actually during a fault if you have say 5000A flowing the resulting magnetic field pulse created may end up pulling the busbars off the mounts.
 
Hello,
i bought lto 66160 too, mines are written like "excell lithium titanate" 2.3v 30ah real 25ah and sadly 2.1v
they are labeled in 2 ways:
- YL14052400208
- 104251F28200124
if someone know how to read it would be nice to know.
i paid them 9 usd/pc , how much did you?

i used a bms of electric bike programmable and the bleeding balance current is 0.2A. would be nice to have 1 or 2 A my configuration is 24S4P links of copper and having problem with the nuts that are from stronger metal than aluminum and break the terminals even at 7 Nm.

im cycling one cell to know life of it and doesnt seems to be very good compared to lifepo4 maybe like someone said they might be strongly used.

nice to contact
 
Back
Top