My 38kw, 138v, 3.8kwh LTO Battery

Sunder

10 MW
Joined
Sep 6, 2011
Messages
3,054
Location
Sydney, Australia
Nothing to report yet, but grand adventures await.

Planning on connecting 57 or 60 of these batteries in series, for an electric motorscooter. I am intending to build 3 x 46v packs to be joined by solid copper bars. Makes them easier to work with both weight (Each cell is 900g, so the completed battery is going to be about 55kg) and electrical safety. It also means that if I ever decide I've had enough, I can take the batteries out to use for an off-grid system.

My plan (though this could change), is to 3D print assembly guides, and build custom acrylic boxes to contain these in.

As LTO is virtually indestructible, I am using individual cell protectors, no BMS. That was a conscious choice, as I've had the BMS cut power on me just as I finished overtaking a car. The driver was not happy, but I was very lucky.

Their energy density is fairly poor, so I'll be losing range. But their power density and cycle life are huge improvements, and 99% of my journeys can be done without a mid-trip recharge.

It will probably take me 2-3 weeks of opportunistic evening work to QA each of these cells, so updates will be slow, but I'll post them for anyone interested.

My Cells.jpg
 
where do you get LTO cell balance boards from?
 
They aren't LTO cell balancers. They're actually super capacitor protection circuits. As soon as it hits 2.7v, it starts bleeding up to 0.5A. Not perfect, but still easier than trying to balance 3 x 20S BMSes as we were discussing in my other thread (about Zenners).

A small update, I've been unusually tired due to cycling caffeine. So all I've done is unpack the 6 boxes worth of the stuff, and test 3 cells. First 3 cells all tested between 27,500mah, and 28,200mah, so close enough to spec. All three cells also tested to 0.9mOhm internal resistance, so again, within spec.

I'll post up pictures if I can.
 
Hi Parabellum,

These are OSN 28Ah prismatic cells:

28Ah
2.3v nominal
<1mOhm
$19USD per cell
$8USD per cell shipping by air to Australia.

I have verified all of those stats within 5% tolerance.

Related to the cell balancers rather than the cells themselves. I did some testing over the weekend:

At 2.68v, the cells use 0.05 milliamps. So there is definitely a parasitic draw:

20180609_175158.jpg

At 2.70v the cell balancers use 8 milliamps. Not quite the 500ma the ads claimed they were speced for:

20180609_175234.jpg

Even at 3.00v the cell balancers use 165 millamps:

20180609_175345.jpg

I avoided using Zenner diodes because of the so called "soft knee". There's definitely a sharp knee at 2.7v here, being 160x increase in current over 0.02v, but at 8 milliamps when my charger puts out 4300 milliamps, I am somewhat sceptical these are up to the task.

I will be trying the same test on Zenners next to see the result. Some time this week, I gotta learn CAD to draw up the assembly guide. Due to the size of the tabs (70mm x 30mm or so), the chance of accidental shorts without a non-conductive guide to pin things to are pretty high.
 
Good point. I tested only two units which behaved identically (well, the second one bled marginally more current.

I didn't think they had LEDs, but on close inspection, it seems like one of the tiny surface mount somethings, have been replaced by a surface mount LED:

20180611_110917.jpg

Will try a few more.
 
Those are actively bleeding when LED is on, I have no idea what they do when damaged, but it is interesting to know. What I did is just replacing units that did not glow up with >2.8V and they where cold, so no bleeding.
Be careful, I have no way to develop statics, tile floor, glass tables, no socks at 30C, hope you did no damage to the boards just by touching them.
BTW. they look same to those I got, 20% DOA.
 
I thought my 1kW battery was huge. That's one beast of a battery :) Have you considered how you are going to charge that thing? Many of them can be charged at 5C without diminishing the battery life much. Looks like you'll be able to fill up almost as fast as those gas guzzlers, given the right (30kW!!) charger.
 
I think it might have been a trick of the flash on the camera. I reran the tests, and all of them lit up at around 2.7-2.8v. I actually do recall the one on 3.0v getting warm as well.

@thorlancaster328 - I'll be charging with an ultra powerful, really fast... 600w Meanwell :p The budget didn't extend to the 6.6kw charger that I wanted. (Actually wanted 2 x 3.3kw chargers, so I could charge at home @ 15A, and at the shopping centre at 30A.). Unfortunately as they say in my industry... "That's out of scope".
 
i dont think the power outlets at those locations would like it as well...
dragging along somehting like a chademo isnt really "cool" on a bike...
 
Well, I educated myself on 3D modelling. If you can call it educating. I won't be drawing anything more complicated than a plate with holes for a while.

I knocked this up in about 20 minutes, after spending about an hour getting familiar with the controls.

First Guide.jpg

To help you make sense of it, the 20 slots are for the tabs to feed through. The gaps between the slots are 8.75mm, even though the tabs are 25mm wide. I'll either need to fold the tabs, or cut them.

The four 5mm circles are to fasten a solid copper bar to the final tabs. This should reduce the likelihood of the tabs tearing from movement between each 20S pack.

The + and - signs are purely cosmetic, but also will help me quickly identify which end of the battery is which. I am slightly tempted to put the endless sphere logo on there as well ;)

This design isn't yet finished. I need somewhere to mount the cell protectors. My options are:

1. Put a vertical riser between the two slots (18mm wide), then print M5 studs protruding from either side. It's gonna have to be narrow. I don't know that I could get a lock nut onto it to fasten. It'd have to be a very thin nut.

2. Widen the plate, and place studs on alternating sides. As each cell protector is 35mm wide, and each cell is 8.75mm separated, again, I'm going to need to do some odd staggering. I think Parabellum had the same issue with his placement - nothing could be lined up linearly.

Any opinions on how I should mount the cell protectors?
 
What about fitting them horizontally in 4 lines?
Printed studs could get to hot in worst case scenario, vicat temp of PETG (I used to print in) is ~80C, boards get over 90C (with heat sinking to terminal) hotter if free standing.
What type of connectors will you use?
Are there some bolts already you could work with or extend on?
 
Not quite sure what you mean by "Horizontally in 4 lines". Could you explain?

Getting more confident with the program, I am thinking of making hollow dividers/risers between the cells which should be insulated, and then printing an M5 stud on it. Once I get it right, I'll upload that diagram too.

Looking at the cost of small volume printing, I figured it was cheaper to just buy my own printer. So I have one coming in 5-7 days. It prints ABS, which should have a Vicat temperature well above anything short of a battery fire.
 
I mean horizontally on top of the plate like I did but in 4 straight lines on some raisers like you say. ABS vicat is 105C and board will probably get there without any heat sink, or you remove 1 resistor and it probably will be OK then.
 
I have bought a few spares, so I might do some destructive testing on one or two. I have a contactless thermometer. I might just hook it up at different voltages until the temperature remains stable for 10 mins, then crank it up another 10th of a volt each time until it destroys itself.

A small AC fan hooked up to the charger might also be a good idea. Although if a battery is really unbalanced it will keep bleeding after the charger is turned off. Hopefully regular balancing and keeping it near balanced by never going too low will avoid this situation.
 
put a piece of painters tape on the cell, IR meters dont like reflective surfaces so the temperature wont read right. reading the tape helps a lot.
 
flippy said:
put a piece of painters tape on the cell, IR meters dont like reflective surfaces so the temperature wont read right. reading the tape helps a lot.
Black permanent marker is thinner layer, no trapped air inside, temp changes response and accuracy is higher.
 
Minor update.

In between ensuring that all the cells are good, I've been designing up battery stays.

This is the first one:

First LTO.jpg

There's 20 cells there (not yet soldered together). The ends are held together by 6mm acrylic, and the black stays are 3D printed to within a few 10ths of a millimetre (To the point where 3 strokes of a file went from not fitting to fitting perfectly), so there's a possibility that they will crack if the cells expand a little (as LTO are known to do during charge/discharge), but we'll see. I'll cycle it a few times before I install it.

It's a slow project, but it is also the end of the Australian financial year, and that's a tough time for me, especially with a working wife and two kids. But I do what I can to keep my hobbies going.

I really want these in before Winter is over. I notice how bad the sag is on the old batteries these 3-4*C mornings.
 
You pretty much have to, I once left a Cell open for a year. I wanted to do a capacity test, but while charging the voltage started jumping around as i was pressing down on the slightly formed bubble :shock:

It was a lithium Manganese Cell though.
 
My understanding is that you only need to prevent expansion, rather than actively compress.

So, what I did was measure the width of the batteries sitting under its own weight (imagine them tipped sideways from that photo). I measured the stack to be 183mm. I then compressed them using G-Clamps. At 181mm, it wasn't possible to compress them any further by hand. So I printed the stays to 182mm.

The cells won't fit if I try to assemble by hand, and just fit if I use the G-Clamps, so I'm satisfied that there's a very small amount of compression on them. The only question is if the 3mm thick stays will crack when they expand from heat/charging.
 
Back
Top