Honda Insight Forum banner

101 - 119 of 119 Posts

·
Registered
Joined
·
897 Posts
You may want to do this with the board attached to an LTO board that is not attached to a pack (as opposed to not having the "balancing" board connected to anything and relying only on the bench supply's voltage readout) since the LTO board probably provides greater voltage resolution than the variable power supply. Unless of course you also have a precision voltmeter you can parallel to the variable supply's output.
 

·
Administrator
Joined
·
12,441 Posts
Discussion Starter #102
I'm not bothered about the psu output as long as it is sufficient to charge the pack above the balancing voltage.
Mine is fine and tapers off at 31V which is about 2.58V per cell. This isn't a test of the psu etc.

I'm testing if the boards work on a live pack, and how long it takes to bring the voltage down from say 2.58V until the balancing shuts off at around 2.45V

I managed to find a time lapse action camera and it is watching the setup as we speak.
Voltage is down around 100mv this morning to about 2.48V.
I'm expecting some balancing loads to turn off in the next few hours.
 

·
Administrator
Joined
·
12,441 Posts
Discussion Starter #103 (Edited)
Something's happening! Now three loads have turned off. :eek: :coffee:

Once this block/board has finished I will check all the individual cell voltages with my decent multimeter.
I'll then do another block with a different board to check variation/consistency between them.

For safety you really have to remove the LTO BMS from the battery before installing the little adapter board and sil socket etc. That's a bit of a PITA as the tap screws are really tight.

Dropped screws and uninsulated screwdrivers are a definite danger to your terminals.

Block 20C Charging.
Block 20B Balancing.

Time lapse camera watching closely. (No idea if it is working yet!)

I cut little windows in the covers so I can see the LEDS and let balancing heat out.

I'm not great at dremel work, any great ideas to cut neater more regular holes in the covers?

A thinner slot just to view the leds would probably also let out enough heat.

87913
 

·
Registered
Joined
·
217 Posts
I cut little windows in the covers so I can see the LEDS and let balancing heat out.

I'm not great at dremel work, any great ideas to cut neater more regular holes in the covers?
You're looking for a hand operated nibbler, like this:
87914


There are pneumatic versions but this is fine for working with small plastic bits.
 
  • Like
Reactions: Natalya

·
Administrator
Joined
·
12,441 Posts
Discussion Starter #107 (Edited)
My balancing time lapse photography effort was complete rubbish.
I need to practise and maybe get a better camera. :rolleyes:

Anyway here is a first balancer report with them set at 2.45V.

1) The first block 48 hrs since charging the pack to 2.6V per cell are now back down to between 2.485 & 2.446V as measured with my decent meter.
A 39mv difference between the highest and lowest.

There is still a bit of intermittent activity with the balancer led's on a few channels so I shall just leave it connected and see how it goes over the next few days.

2) The second block 36 hrs after charging to 2.6V per cell are now down to 2.501 to 2.449V.
A 52mv difference between the highest and lowest.

Again there is still some balancer activity.

I shall just monitor the blocks for the next few days and check all the voltages again after all activity has finished.

I shall repeat the charge and balance process on both blocks and measure how much capacity is required to be input for the cells to again reach 2.6V this will gives us a figure for the amount of energy burnt off between 2.6 and 2.45V~.

I will then induce some deliberate imbalance on several cells and see what happens..

I'm not jumping to any firm conclusions at this stage more testing is def required.
 

·
Administrator
Joined
·
12,441 Posts
Discussion Starter #108 (Edited)
After another 3 days all balancer activity has finished.

One block has a maximum inter cell differential of 40mv and the other 55mv.
How significant that is who knows..

It looks like it would certainly pull separate blocks with different voltages into some sort of line around 2.46V if given enough time.
As it would be working 24/7 it will be interesting to see a full pack fitted with balancers versus one without in terms of general day to day balance in use.

Time to start charging and testing some other boards..

Code:
    $20B         $20C
    2.456        2.454
    2.46         2.472
    2.478        2.46
    2.475        2.468
    2.484        2.446
    2.444        2.466
    2.455        2.47
    2.476        2.466
    2.468        2.497
    2.456        2.501
    2.457        2.468
    2.456        2.459

Total    29.565        29.627
      
Average    2.463        2.468

Difference    0.04        0.055
 

·
Registered
Joined
·
897 Posts
I looked at similar numbers yesterday.

The difference between 0.04 and 0.055 is pretty large. The difference may be greater if a really mismatched dozen of devices are on the same board.

50 mV tolerance at 2.45 volts works out to be close to an amp-hour of difference in charge, which is about 4-5% of the pack capacity, based on the plot I posted in another thread.

Hopefully the pack average will halve the deviation from the center of the pack half this amount, but in the worst case, one cell could be significantly leading or lagging the pack. Someone running the pack to the knee and looking only at the total pack voltage may be in for an unpleasant surprise.

However, it would seem that one will see more equivalent full cycles if one makes many smaller cycles vs fewer larger cycles. (In other words, if one cycle consumes and restores X % of the pack and another cycle consumes and restores X/2 % of the pack, the pack will last more than twice as long with the smaller, more frequent cycles.

Given the wide tolerance, a greater concern might be reliability. Subjecting them to heat and cold and seeing if they drift or fail would not be a bad idea.
 
  • Like
Reactions: Natalya and Adria

·
Administrator
Joined
·
12,441 Posts
Discussion Starter #110
I'm testing again as we speak. Recharged the blocks to 2.6V so the balancers activated.

On one of the balancer boards I removed all the 0.1uf ceramic filter capacitors we added.
It is not shown in the manufacturers device pdf example circuit.
We/I added it as it's usual practise with IC's.
Maybe adding it adversely affects performance in some way.

Seeing if they fail or drift etc will have to be done in a working car.
I simply don't have time to do elaborate bench testing regimes on everything.

It remains to be seen at present if packs with no balancers will go out of balance increasingly as they age and cycle etc. A lot will be down to individual cells and chemistry etc. We might end up in a year or two with packs operating at 50% capacity, who knows it's too early to tell.

If adding balancers is guaranteed to keep a pack within 1ah that might not be too bad for long term use.
It's a bad idea anyway to use the top or bottom 10% of the SOC anyway.
 

·
Administrator
Joined
·
12,441 Posts
Discussion Starter #111

·
Premium Member
Joined
·
995 Posts
I'm surprised you guys are working for a balancing solution. The most important thing to have is a cell monitor for the LTO packs. Unless you have something connected to individual cells, you really should never see an out of balance cell.

You are much better off having the OBD2C&C monitor the voltages than an actual active balancer. The reason is that, I found that when you have a bad cell, it goes bad very rapidly and should be bypassed by cutting the cell out and soldering over to the next cell or replacing the whole pack.

The reason these cells do not go out of balance is that they are in a very sturdy enclosure, and they do not deform which I found is a large reason for pouch cells to not stay in balance. 18650 cylindrical cells rarely go out of balance also because they do not deform easily.
 

·
Premium Member
Joined
·
995 Posts
Let me add, I've found 2 bad cells in the LTO packs I bought, surprisingly and maybe luckily, both were on the same 12 cell pack.

Both cells had a bad smell to them, if you smell closely near the cell vent it has a bad chemical smell, almost like a magic marker smell. This is something you probably don't want in your car and if you had it in your car you probably would smell it. I store this bad cell out in my garage because I believe I can slightly smell it in my basement.

I believe, and I should do testing on the 2nd cell, the 1st bad cell only took a few cycles for it to really start getting out of balance. With a few charge cycles I noticed that it went from like 1 amp hour to 3 amp hours out of balance, it deteriorated very rapidly.

The rest of my over 100+ LTO cells I have, they are all in perfect balance, even the ones in my car after numerous cycles.

My 6 cell 12 volt replacement LTO pack, I question if that will stay in balance because I cut the pack and could have slightly damaged the last cells in the pack. My initial testing did not notice anything but time will tell on that pack.
 

·
Administrator
Joined
·
12,441 Posts
Discussion Starter #114
I'm surprised you guys are working for a balancing solution. The most important thing to have is a cell monitor for the LTO packs. Unless you have something connected to individual cells, you really should never see an out of balance cell.
We already have the cell monitor but that can't balance the cells.

You are correct in that you should rarely see an out of balance cell.
But it's a definite real world possibility, in fact I've seen it myself to varying degrees.
These LTO cells aren't perfect, and as they age are more likely to start wandering.

Why did Honda include balancing on the LTO BMS if it isn't needed?
Clearly they think it is/was needed, I agree, even if it's just belt and braces and rarely used.

Doing LTO conversions for other people is challenging. (Some are not tech savy)
I can't/don't want to ask them to unbolt the IPU and whip the covers off a few blocks and start clipping on loads or chargers to individual cells etc. That way madness, big sparks and danger lies. :eek:

Ideally we need some sort of automated system that just gets on with it. :)
 

·
Administrator
Joined
·
12,441 Posts
Discussion Starter #116
Hmm that's an interesting POV.. :unsure:

I think the tiny balancing currents would be unlikely to mask anything for very long in the real world.

I don't know all the answers, just pondering and tinkering as usual.
 

·
Registered
Joined
·
7,189 Posts
2 cents: For DIYers, enthusiasts, people who take more of a hands-on manual approach to things, balancing probably a hindrance, at minimum unnecessary. For everyone else, like making a system for a typical driver, balancing probably necessary.
 

·
Registered
Joined
·
897 Posts
With a few charge cycles I noticed that it went from like 1 amp hour to 3 amp hours out of balance, it deteriorated very rapidly.
Sounds like it got overcharged. This study* set a voltage cutoff at 3.6V to create an overcharge event which lead to an immediate 15% loss of capacity and degradation thereafter.

*Overcharge Study in Li4Ti5O12 Based Lithium-Ion Pouch Cell: I. Quantitative Diagnosis of Degradation Modes
 
  • Like
Reactions: retepsnikrep
101 - 119 of 119 Posts
Top