Honda Insight Forum banner

1 - 20 of 49 Posts

·
Engine-Off-Coast
Joined
·
1,766 Posts
Discussion Starter #1
So over the years I've built 3 dumb grid chargers -- one exactly following Mike Dabrowski's original plans, and two using more modern constant current LED drivers.

But all of them had in common the amperage was quite low -- about 350ma. The LTO installs have 2x or 3x (or more!) the capacity of the original IMA, and charging them can be done PHEV style.

I want to build a charger that will charge my LTO pack faster than a traditional 350ma grid charger.

I've started looking at power supplies with higher current.

A few model numbers:
ELG-200-C1050A -- $45, 95 to 195V, 1A
HLG-185H-C1050A -- $51, 95 to 195V, 1A
HLG-320H-C1750A -- $69, 91 to 183V, 1.75A
HVGC-320-1750A -- $100, 187V, 1.75A
HVGC-480-M-AB -- $170, 92 to 228.5V, 2.1A

I'm leaning towards the HLG-320H-C1750A (datasheet) with the 1.75 amps, but its max voltage is only a little bit higher than a 72 cell LTO max of 180V. I'm wondering if it's therefore not going to be good at charging past 175 volts, or that I might burn it out prematurely or something because it's not powerful enough?

Next up in the dollars is the HVGC-320-1750A. In the datasheet it also lists an operating range of 92 to 183 volts, which again has me wondering the same thing.

The two 1-amp models have a higher max voltage, so I'm sure they would work, but the extra current of the 1.75A models was appealing.

The most expensive supply seems to have a high enough voltage, and it outputs 2.1 amps, but it's twice the price. So I think it'll work, but if I can save $70 to $100 for only about 1/3 of an amp less then I'm sorta inclined to go with the cheaper options if they'll work okay.

Any thoughts on the maximum voltage issue? I've only done 72 cell LTO installs, and I don't think I'll ever do an 84 or 96, so I don't NEED a super high voltage supply.
 

·
Registered
Joined
·
824 Posts
Any thoughts on the maximum voltage issue?
I'd like to invite @Eli, @eq1 and @jime into this discussion.

I worry about whether concept of NiMH grid charging can be applied to an LTO pack.

If you only are looking at pack voltage, what happens if one or more LTO cells, that happen to have a higher SoC than the rest of the pack, reach the "do not exceed" voltage before the rest of the pack? Will they just keep increasing in voltage until they reach a point where there is permanent cell damage while the pack voltage is still below the charger's maximum?

In other words, is trying to grid charge a LTO pack of cells in series, without automatic monitoring every cell's voltage, and having a way to automatically terminate the charge when the first cell reaches its do-not-exceed voltage, exposing the pack to the potential damage should one cell come to lead the pack sufficiently?

If automation is used to stop charging when the first cell reaches maximum, doesn't that leave the pack as balanced or unbalanced as when the charge started? In other words, the stated benefit of NiMH grid charging, "pack balancing" is not acheived?

Furthermore, can one afford not to have automatic per-cell voltage monitoring, and a way to stop charging when the first cell reaches its maximum voltage, in actual in-car operation? for example, in a high current regen situation, could that not push one cell deep into overcharge voltages at high currents? What happens to an LTO cell that's pushed hard into overcharge voltages?

Conversely, in-vehicle, is there a similar problem for the lowest voltage cells? Can a cell damaged from overcharge become the first to be depleted during assist, and get pushed hard into reversal? What happens to an LTO cell that's reversed hard?

I think of the two reports of cells that had thermal failures, but unfortunately there was no log to help identify the cause.

Do LTO installs need to be have automation watching all cell voltages? Should those measurements be recorded so that failures can be analyzed?

Are the only ways to rebalance a pack that whose cells' SoCs are drifting apart, to manually charge or drain one cell at a time (with the inherent danger of potentially deadly high voltages and large currents), or install a BMS that can perform balancing?
 

·
Engine-Off-Coast
Joined
·
1,766 Posts
Discussion Starter #3
If you only are looking at pack voltage, what happens if one or more LTO cells, that happen to have a higher SoC than the rest of the pack, reach the "do not exceed" voltage before the rest of the pack? Will they just keep increasing in voltage until they reach a point where there is permanent cell damage while the pack voltage is still below the charger's maximum?

In other words, is trying to grid charge a LTO pack of cells in series, without automatic monitoring every cell's voltage, and having a way to automatically terminate the charge when the first cell reaches its do-not-exceed voltage, exposing the pack to the potential damage should one cell come to lead the pack sufficiently?
When driving the car can throw over 60 amps into the pack, I think that's way more to worry about than 1 or 1.75 amps.

If automation is used to stop charging when the first cell reaches maximum, doesn't that leave the pack as balanced or unbalanced as when the charge started? In other words, the stated benefit of NiMH grid charging, "pack balancing" is not acheived?
My intent isn't to balance the pack like a bad NiMH. I'm trying to recharge it most of the way over a few hours so that I'll have a full charge before I drive somewhere. Think of how a PHEV operates, except on a smaller scale.


Furthermore, can one afford not to have automatic per-cell voltage monitoring, and a way to stop charging when the first cell reaches its maximum voltage, in actual in-car operation? for example, in a high current regen situation, could that not push one cell deep into overcharge voltages at high currents? What happens to an LTO cell that's pushed hard into overcharge voltages?
If there was a major defect in a cell you might get the magic marker smell, but as far as I know, only one person here has had that happen with these.


Conversely, in-vehicle, is there a similar problem for the lowest voltage cells? Can a cell damaged from overcharge become the first to be depleted during assist, and get pushed hard into reversal? What happens to an LTO cell that's reversed hard?
We don't really know because no one driving real-world with these batteries on this forum has experienced that sort of failure yet.

I think of the two reports of cells that had thermal failures, but unfortunately there was no log to help identify the cause.

Do LTO installs need to be have automation watching all cell voltages? Should those measurements be recorded so that failures can be analyzed?
Some would say yes. Peter has been working on developing this capability.


Are the only ways to rebalance a pack that whose cells' SoCs are drifting apart, to manually charge or drain one cell at a time (with the inherent danger of potentially deadly high voltages and large currents), or install a BMS that can perform balancing?
We don't have much experience with cell drift. In my car I measured all 72 cells about 7 months after installing, after letting the car sit for 3 days unused, and the difference between the highest cell and the lowest cell was 0.02 volts.

I guess with the CAN system to monitor cells we could get it to throw P1449 if there was too great a cell voltage deviation??
 

·
Registered
Joined
·
652 Posts
Alibaba (and similar sites *gasp even wish) have low amperage lithium chargers of assorted voltage ratings for as little as $10’s of dollars

I myself have a $68 - 48-58 volt 15 amp “lithium” charger attached to my lead acid pack, sadly it’s “precision voltage control” is an internal pot and drifts randomly by large amounts
but for lead I just max it out and it’s ran Reliably for years this way.

String a few of these with a voltage kill circuit and it might be the cheapest (albeit slowest) option
(No recommendation or affiliation implied)

 

·
Registered
Joined
·
824 Posts
natalya wrote...
My intent isn't to balance the pack like a bad NiMH. I'm trying to recharge it most of the way over a few hours so that I'll have a full charge before I drive somewhere.
Awesome!

So yeah, then if your charger turns off when the first cell reaches the top, awesome.

After hitting send I left for dinner and I started thinking that I went overboard with my series of questions and feared they could be interpreted as calling into question your build - which was NOT intended - thanks for posting about them and thanks for answering all my questions.
 

·
Premium Member
Chicago & Detroit
Joined
·
887 Posts
I like using the OE LTO BMS boards because I believe the readings to be accurate.
Just to get people on the same 'definition' page, a single board monitors 12 cells called a Block. A 72-cell pack is ~180v and has six blocks.
With a total pack voltage of 176.8V I noticed one of my blocks had a considerably lower voltage than the other five. The delta between hi/lo voltage for the cells of a 'good' block was 4mV; the delta for the wondering block was 42mV.
I decided to use my 180V grid charger, Peter's simple Cycler, and monitored total LTO pack voltage and stopped at 183.7V., I figured an average 2.55V/cell was ok.
What I found was the block total voltages raised pretty evenly but did not level or balance the individual cells. The delta between hi/lo cell voltages for 5 blocks, raised, in mV, 4, 3, 3, 1, 6. The delta between hi/lo cell voltages for the 'wondering' block raised 61 mV.
So that attempt to level or balance didn't go as I'd hoped. Clearly I have an issue with one block.
To try to level/balance all blocks I used my 'Level' tool consisting of high power resistors of 25 ohm (2 each) and one 500 ohm resistor. Minor4326 LTO Conversion Post #218. I have very fine voltage-reducing control with the 500 ohm and gross control with a 25 ohm resistor, and medium-gross with 50 ohms. Today I reduced all blocks to within 10mV, monitoring using my DMM. Tomorrow I'll get more info after I read the BMS blocks.
I'm considering modifying Peter's Simple Cycler to a 30V charger using a Mean Well LPF-16-30 (cc/cv) because I believe a cv of 30V is necessary.
So for now I'm attempting to level my packs using my 'tool'. Hopefully in the future I can figure out how the Mean Well 16-30 can be used.
 

·
Super Moderator
Joined
·
8,231 Posts
I'm leaning towards the HLG-320H-C1750A (datasheet) with the 1.75 amps, but its max voltage is only a little bit higher than a 72 cell LTO max of 180V. I'm wondering if it's therefore not going to be good at charging past 175 volts........
Remembering that the "normal" charger for a Lithium pack is a cell balanced CC/CV setup, I think some care and understanding of how the charger terminates will be important. Even if the power supply under consideration is specified to charge to 183V, a safe level, the pack will experience some "rebound." (The pack voltage will fall when the charger terminates.) This rebound represents an actual misreading of SOC, i.e. you don't actually reach the SOC you wanted.

I don't remember how much these LTO packs rebound on termination. Your friend Adria has a Genesis charger which will execute a CC charge and then terminate at a designated voltage. You could do a simple experiment by having her charge her pack to 180V (2.5V/cell), and then measure the rebound voltage after an hour with a good voltmeter. (Check the offset between the Genesis and your voltmeter as the Genesis is charging.)

You will want a somewhat higher termination voltage to get to a stable 180V. I just don't remember what I saw for rebound when I did this. The think 183V may be somewhat short for a resting voltage of 180V

Of course, the process you are contemplating DOES NO BALANCING. That is an entirely different discussion. This process is relatively safe since the termination voltage is well short of the maximum spec voltage for the cells. It is especially safe if you occasionally check your cell balance by some reliable means.

The entire question of measurement accuracy, resolution and repeatability is going to come up shortly. The question gets very complicated and difficult to understand. As indicated, even a Genesis and a good Fluke are likely to differ a bit. It is best IMO to rely on a single meter when making measurements, if consistency is require as in this case. Cheap Harbor Freight meters are just an invitation to be mislead.
 

·
Administrator
Joined
·
11,778 Posts
I will be implementing a high power charger into my 96 cell setup shortly.

I'll detail it on the 96 cell thread so as not to derail this one.

It will certainly use the LTO BMS to signal via the OBDIIC&C if a cell is over V etc and shut down.
 

·
Super Moderator
Joined
·
8,231 Posts
I worry about whether concept of NiMH grid charging can be applied to an LTO pack.
As well you should. They are two different cell chemistries and they respond entirely differently.

If you only are looking at pack voltage, what happens if one or more LTO cells, that happen to have a higher SoC than the rest of the pack, reach the "do not exceed" voltage before the rest of the pack? Will they just keep increasing in voltage until they reach a point where there is permanent cell damage while the pack voltage is still below the charger's maximum?
In a word - YES. But it depends considerably on understanding and conservative measures. NiMH, the Insight OEM material, is difficult to overcharge. Within reason, NiMH just burns off excess charge as heat. Lithium, LTO in this case, reacts differently to overcharge.

LTO has a flat charge curve in the region of 2.0-2.5V. Outside this region, things start to go wrong IF charging and discharging is by non-BMS mode. (Even 2.5V might be too high if there is a preexisting bad unbalance.) The upper part of the LTO charge curve become very sloped or rapid. Just a little charge can dramatically change cell voltages. When bulk charging, one won't know what is happening to individual cells unless one uses Peter's monitor, OR one has an effective way to manually meter the individual cells.

Conversely, in-vehicle, is there a similar problem for the lowest voltage cells? Can a cell damaged from overcharge become the first to be depleted during assist, and get pushed hard into reversal? What happens to an LTO cell that's reversed hard?
Absolutely a concern! Lithium should have a good BMS. Given that most applications won't, there are a few simple measures which help considerably:
1. Measure your balance manually at frequent intervals to stay informed of cell-to-cell balance.
2. Stay well away from cell maximums and minimums.

Peter's OBDII cell monitoring/control setup makes real time monitoring of cell balance automated and easy. Not only can one check static balance, but one can trigger shutdown if a weak cell falls too low under load, or if a strong cell gets to high under high charge/high regen conditions. It won't balance, but it will protect the pack from the ravages of imbalance.

Just a few thoughts on the "as delivered" balance of the Greentech Fit LTO packs. The packs have not seen a balancing operation under BMS control in a long time, perhaps 3 years now. A lot seems to be made of the sterling balance that insightbuyer reported. LTO does seem to have quite low self discharge, so the packs do tend to stay balanced, but one never knows when there will be a cell which has a bit higher self discharge. Member insight buyer bought his packs at perhaps 1 year off BMS managed balancing in the car and he went to some pains to get equal block capacities. He may have also been somewhat luck. We are now buying packs at about 3 years off BMS managed balancing, so it is only natural that an occasional cell or block(subpack) will have a slight imbalance. iirc, Peter recently reported a 20mV block imbalance and even somewhat larger imbalances would be reasonable to expect 3 years after balancing.
 

·
Super Moderator
Joined
·
8,231 Posts
It will certainly use the LTO BMS to signal via the OBDIIC&C if a cell is over V etc and shut down.
Great, but still not cell balancing. Correct?
 

·
Super Moderator
Joined
·
8,231 Posts
What I found was the block total voltages raised pretty evenly but did not level or balance the individual cells.
And why should it? The charger is a bulk charger with no cell level balancing mechanism. The charger was designed for a completely different application.
 

·
Super Moderator
Joined
·
8,231 Posts
IMHO and from reading the comments in this thread, there is a serious lack of understanding about how Lithium packs function as series strings of multiple cells.

I also labored with this problem for a year or so and I certain don't claim to be an expert, but I got smarter. I searched out books and bought one that was immensely helpful.

Most of the books on Lithium cells, packs and multicell pack management (i.e. BMS) are a bit outdated in a rapidly changing environment, but the general principals don't change very much.

I looked for a book which treated Lithium packs at a systems level and wasn't overly bogged down in cell chemistry and in long equations, i.e. a readable reference. I found a very helpful aid in this book:


The book is a bit dated. It doesn't cover LTO or the Orion specifically, but it has some extremely helpful graphical material on battery balance, cell capacity and resulting pack capacity. It also has excellent section on BMSs. I could find no other worthwhile book with the same level of readability. It is available used on eBay but still not cheap.
 

·
Administrator
Joined
·
11,778 Posts
I would say don't get hung up on actual daily balancing or balancing with every charge.

1) These rugged LTO blocks seem to keep in line quite well if initially carefully balanced when assembling.

2) Minor overcharge isn't going to kill them or set the car on fire.

3) Using the LTO BMS to get the cell voltages, and then using that via the OBDIIC&C or some other simple control interface to terminate and or control charging means it's easy to prevent gross overcharging.

4) Adjusting your chargers own internal control so that it can't output more than the maximum pack voltage also gives you a second level of protection. (You can add in maximum charging time as another safety factor.)

5) As you gain knowledge about your pack and which cells are weaker etc, you can manually intervene (balance) perhaps once a year or something depending on how much they drift in use.

My LTO strategy will be.. A then B

A) Bulk charge with high power / constant current 8-10A or so until the first cell hits a target voltage say 2.5V
We have 240V AC at 13A maximum generally available in the UK
2.5kw is about the maximum charge power you can draw from domestic power outlets without melting things.

B) Lower current and or voltage until say about 350ma - 1A is flowing, then charge until first cell reaches 2.55V
 

·
Premium Member
Chicago & Detroit
Joined
·
887 Posts
here's my experience so far.
The original reason I wanted to charge/'balance' the pack was Block 1 was 60mV lower than the rest and I wanted to charge it up. The data below does not include Block 0 & 1 because they will soon get replaced.

A weird thing happened after I connected the batteries and started the car to check voltages. I reset everything so I had no SOC display, but the 3 charge bars were on as it high-idled. Before I started the car my total pack voltage was 183.6; after several minutes of internal charging it raised to 187.8, then after several minutes the voltage went to 189.3. That's when I shut down. That's pretty phenomenal. Didn’t want it to go that high, ever. So now I’m doing a discharge to 30.2V/block or 181.2V total.

My chart includes my original data before I started any charging; after a charge; and after a discharge. The overall results shows some improvement but not like a true cell-by-cell balance.
EDIT: I had removed the BMS Interceptor. After discharging I'll but it back into the circuit.
1594135122878.png
 

·
Registered
Joined
·
81 Posts
Remembering that the "normal" charger for a Lithium pack is a cell balanced CC/CV setup, I think some care and understanding of how the charger terminates will be important. Even if the power supply under consideration is specified to charge to 183V, a safe level, the pack will experience some "rebound." (The pack voltage will fall when the charger terminates.) This rebound represents an actual misreading of SOC, i.e. you don't actually reach the SOC you wanted.

I don't remember how much these LTO packs rebound on termination. Your friend Adria has a Genesis charger which will execute a CC charge and then terminate at a designated voltage. You could do a simple experiment by having her charge her pack to 180V (2.5V/cell), and then measure the rebound voltage after an hour with a good voltmeter. (Check the offset between the Genesis and your voltmeter as the Genesis is charging.)

You will want a somewhat higher termination voltage to get to a stable 180V. I just don't remember what I saw for rebound when I did this. The think 183V may be somewhat short for a resting voltage of 180V

Of course, the process you are contemplating DOES NO BALANCING. That is an entirely different discussion. This process is relatively safe since the termination voltage is well short of the maximum spec voltage for the cells. It is especially safe if you occasionally check your cell balance by some reliable means.

The entire question of measurement accuracy, resolution and repeatability is going to come up shortly. The question gets very complicated and difficult to understand. As indicated, even a Genesis and a good Fluke are likely to differ a bit. It is best IMO to rely on a single meter when making measurements, if consistency is require as in this case. Cheap Harbor Freight meters are just an invitation to be mislead.
A 24-cell pack for me usually has about a 1V rebound for me from when I terminate the charge to me checking it the next day if charged in a non CCCV manner. Never done 72 cells but your 183V seems about the right cut off but some trial and error should be done.

In general with non CCCV chargers and lithium the main sticking point is the charge termination. If the charger is just time based or dV/dT termination based it should not be used. Other than a true CCCV charger the only other termination method that would be ok is a simple voltage level termination. Because that voltage wouldn't have been held until the current level dropped to almost nothing is why there would be a rebound back the next day. Also since there is no balancing being done stay with the 2.5V/cell (30V for 12, 60V for 24, or 180V for 72).
 

·
Registered
Joined
·
81 Posts
@greggtwep16, could you explain what you mean by " voltage level termination "
Terminating the charge when the battery reaches a certain voltage. Constant Current chargers don't usually do this unless it happens to line up with the max voltage range of the charger (it is after all varying the voltage put out to keep the charge current constant) but nowadays there are all types of chargers.

This is different than CCCV (Constant current then Constant Voltage). Really CCCV is just a combination that switches from Constant Current (the faster charging part) to Constant Voltage at a battery voltage threshold. That CV part makes the amps flowing slowly reduce from the CC amps to almost 0, because the last part of the charge was almost 0 amps there will be almost no rebound when removed from the charger. If you are using a CC charger that does terminate the charge at a certain voltage (either because it's designed to do that or if it happens to but up against the max voltage of the charger) then you have the CC part just not the CV. So when you take the battery off the charger (or when the charger terminates the charge) it will read at that voltage. Then if you watch the battery voltage it will be be going down and will eventually settle on a lower number (this is not the same as NiCD or NiMH if you slightly overcharge there is no balance that it wants to settle on it's just rebounding because it was just charged). The amount it rebounds down will be related to how many amps you were charging at (more amps more rebound). The same is true in reverse for loads so if you hooked up a large load to it ran the voltage down and then stopped you'd see the voltage level slowly rise after you stopped the load.
 

·
Premium Member
Chicago & Detroit
Joined
·
887 Posts
so will a cccv supply like Mean Well LPF-16-30 that is rated to 30V actually reach that level and will it be a good supply to use for a 12-cell pack since it's right at the nominal voltage limit?
 

·
Super Moderator
Joined
·
8,231 Posts
According to my text rebound is a combination of internal resistance effects and chemistry effects. Internal resistance accounts for a small immediate rebound, current related, when charging ceases while chemistry rebound is usually larger and takes some period of time. Internal resistance is very small for LTO so that portion is relatively small compared to the chemistry component.

Internal resistance is a poorly modeled factor among battery manufacturers. It is frequently stated as AC impedance, which is easy to measure but does very little to represent what happens with the DC demand of a hybrid or electric vehicle. Ian has spoken before on this problem.

Watching high current DC discharge on NiMH and on LTO is instructional. On NiMH, there is an immediate drop in terminal voltage due to internal resistance followed by a short exponential drop related to chemistry. On LTO, there is a very small immediate drop, due to lower internal resistance, followed again by a small but exponential chemistry drop, related to chemistry. The same takes place when charging ceases, particularly if the charging was at significant current.

As several noted above, the rebound doesn't have to be much of a factor with slight overcharging near 2.5V.
 
  • Like
Reactions: minor4326

·
Registered
Joined
·
81 Posts
so will a cccv supply like Mean Well LPF-16-30 that is rated to 30V actually reach that level and will it be a good supply to use for a 12-cell pack since it's right at the nominal voltage limit?
Can you post a link to which one? For Meanwell power supplies the first number is usually watts not amps, so if those numbers are correct it's about .5 amps which isn't exactly speedy which I thought was your goal.

Also some are just CV on the output side not CCCV but if you send me the link it'll have more info since they make a lot of different types.
 
1 - 20 of 49 Posts
Top