Honda Insight Forum banner

141 - 160 of 173 Posts

·
Administrator
Joined
·
11,459 Posts

·
Registered
Joined
·
6,660 Posts
^ Well, I don't know, with these LTOs I don't think it's super critical. You can actually float charge these cells at 2.7V, for instance, something you can't do with regular ole lithium cobalt cells. Plus the risk of catastrophic damage, like explosion or fire, would seem minimal given what we've seen with Toshiba's abuse tests (for instance, they overcharged to something like 4.30V with little more than high temps)... A cell-level BMS would be nice. But, I think just keeping away from the top would probably be enough -- like if the charge in these graphs terminated at say 2.60V rather than 2.77V and 2.80V. That would leave a decent amount of headroom, particularly if you're not driving the cells to the very bottom, too...

These cells are slated to go into my car as a 12V battery, but the lowest I can adjust the DCDC output voltage is 13.85V, i.e. that 2.77V per cell. I might try it anyway, but I've also thought about doing some kind of BMS. The problem with the 12V usage, at 13.85V, is that that's not a once-in-a-while thing; rather, the battery will be floated at that voltage for much of the time...
 

·
Premium Member
Joined
·
2,856 Posts
Discussion Starter #143
3rd criteria ... Self discharge.
Good condition cells should have a very low/slow self discharge rate .. like single digit % per year .. if someone is connected to a good and working cell level BMS that does cell level balancing , well as long as that is working correctly they don't have to worry about this at all .. for anyone running without such a BMS , you have to go back and check on your pack and tweak/fix any variations that creep in .. greater the self discharge rates the more often one needs to go back and check it, or a larger buffer between checks.
Here are results from 12 months completed 24 cell self discharge testing.

Between ... 2.23 mAh/Day < SelfDischarge < 4.37 mAh/Day .. would probably be 'reasonable' in most conditions to expect around the middle of that Min/Max range , or around ~1.2Ah/Yr .. from these ~20Ah cells that is only around ~6% SoC self discharge lost per year of room temperature storage.

Test Method
1> I did cycle efficiency tests of each individual cell (PL8) .. rest.
2> Each cell was initially discharged to 2.000v (PL8 1A CC-CV) .. rest.
3> I charge into each a specific known Ah (5Ah , 10Ah).. each 6 cell group charges in series (Instek GPD-2303S).
4> Waited 6 Months at room temperature .. hotter faster , colder slower.
5> non-auto-balancing 1A CC discharge to 2.000v (PL8) .. 6cell5Ah groupA, 6cell10Ah groupB.
6> Waited 6 more months room temprature to reach 12 total months
7> non-auto-balancing 1A CC discharge to 2.000v (PL8) .. 6cell5Ah groupC , 6cell10Ah groupD.
8> All cells displayed (PL8) as balanced to 1mV.

If previous cell cycle efficiency not accounted for .. raw input vs output.
6 Month .. GroupA and B ..averaged 3.687 mAh/Day .. Standard Deviation 0.691 mAh/Day
12 Months .. GroupsC and D .. averaged 2.433 mAh/Day .. Standard Deviation 0.210 mAh/Day
5Ah .. GroupA and C .. averaged 2.629 mAh/Day .. Standard Deviation 0.414 mAh/Day
10Ah .. GroupB and D .. averaged 3.491 mAh/Day .. Standard Deviation 0.895 mAh/Day
Minimum cell was 2.232 mAh/Day
Maximum cell was 4.348 mAh.Day

Adjusting for the previously tested cell cycle efficiency.
6 Months .. GroupA and B .. Averaged 3.274 mAh/Day .. Standard Deviation 0.535 mAh/Day
12 Months .. GroupC and D.. Averages 3.298 mAh/Day .. Standard Deviation 0.976 mAh/Day
5Ah .. GroupA and C .. Averaged 3.497 mAh/Day .. Standard Deviation 0.767 mAh/Day
10Ah .. GroupB and D .. Averaged 3.075 mAh/Day .. Standard Deviation 0.744 mAh/Day
Minimum cell was 2.261 mAh/Day
Maximum cell was 4.360 mAh/Day
 

·
Super Moderator
Joined
·
7,989 Posts
Just received 4 more packs from Greentec and now into testing second pack. I am reminded of several observations which I think might be beneficial to other buyers.
1. As reflected in Ian's very thorough testing just above, the self discharge rate is very, very low, at least for "good" cells, therefore the batteries are still arriving well balanced and well above minimum cell voltages.
2. An anecdotal observation by me and others would indicate that the packs are arriving essentially cell balanced, I have confirmed this by removing covers and making pretesting measurements of individual cells. (In some 20 packs I have personally tested, I have not seen one with poorly matched cell voltages.)
3. My recent deliveries all had well matched overall pack terminal voltages. They came with voltage notation, so Greentec is apparently measuring this parameter before the packs are shipped.
4. The very tight pretesting cell measurements convince me that the Fit used a top balancing BMS approach. This would be expected, seems to me, and is fortunate for users who do not implement some sort of BMS. It means that the performance nature of the packs can be understood and somewhat managed without benefit of a BMS - as most are doing.
5. In testing, I find some variation in cell capacities. If one top balances with the PL8, then the discharge tails of the cells can vary a significant amount, up to 400 mV in my experience. Since the PL8 is currently set to a discharge shut-off at 1.5V, the low voltage tails become obvious on the PL8 discharge graph. I don't doubt that there is a pack somewhere which has a divergence of at least .5V. This would strongly indicate IMO that the owners should set a low end mental "stop" at 2V (i.e. 2V x 72cells= 144V). This is not a heavy price to pay for safety. There is very little capacity even in a "perfect" pack when one gets below 2V.

More testing results coming perhaps.

edited for clean-up
 

·
Registered
Joined
·
34 Posts
Just received 4 more packs from Greentec and now into testing second pack. I am reminded of several observations which I think might be beneficial to other buyers.
1. As reflected in Ian's very thorough testing just above, the self discharge rate is very, very low, at least for "good" cells, therefore the batteries are still arriving well balanced and well above minimum cell voltages.
2. An anecdotal observation by me and others would indicate that the packs are arriving essentially cell balanced, I have confirmed this by removing covers and making pretesting measurements of individual cells. (In some 20 packs I have personally tested, I have not seen one with poorly matched cell voltages.)
3. My recent deliveries all had well matched overall pack terminal voltages. They came with voltage notation, so Greentec is apparently measuring this parameter before the packs are shipped.
4. The very tight pretesting cell measurements convince me that the Fit used a top balancing BMS approach. This would be expected, seems to me, and is fortunate for users who do not implement some sort of BMS. It means that the performance nature of the packs can be understood and somewhat managed without benefit of a BMS - as most are doing.
5. In testing, I find some variation in cell capacities. If one top balances with the PL8, then the discharge tails of the cells can vary a significant amount, up to 400 mV in my experience. Since the PL8 is currently set to a discharge shut-off at 1.5V, the low voltage tails become obvious on the PL8 discharge graph. I don't doubt that there is a pack somewhere which has a divergence of at least .5V. This would strongly indicate IMO that the owners should set a low end mental "stop" at 2V (i.e. 2V x 72cells= 144V). This is not a heavy price to pay for safety. There is very little capacity even in a "perfect" pack when one gets below 2V.

More testing results coming perhaps.

edited for clean-up
My last 4 modules were cell balanced as well but were generally a little under what was marked on the tape from Greentec. To me this indicates that they were measuring this but likely not right before shipping. More likely they measured it awhile ago when they separated the 18 modules of the Fit EV.
 

·
Super Moderator
Joined
·
7,989 Posts
My last 4 modules were cell balanced as well but were generally a little under what was marked on the tape from Greentec. To me this indicates that they were measuring this but likely not right before shipping. More likely they measured it awhile ago when they separated the 18 modules of the Fit EV.
There might also be a little variation between your voltmeter and theirs. Are either calibrated reliability? Is mine? We don't know. The important thing is that the terminal voltages are within a few milli-volts of each other.

As Ian showed, self discharge is extremely low:)
 

·
Registered
Joined
·
6,660 Posts
5. In testing, I find some variation in cell capacities. If one top balances with the PL8, then the discharge tails of the cells can vary a significant amount, up to 400 mV in my experience. Since the PL8 is currently set to a discharge shut-off at 1.5V, the low voltage tails become obvious on the PL8 discharge graph. I don't doubt that there is a pack somewhere which has a divergence of at least .5V. This would strongly indicate IMO that the owners should set a low end mental "stop" at 2V (i.e. 2V x 72cells= 144V). This is not a heavy price to pay for safety. There is very little capacity even in a "perfect" pack when one gets below 2V.
In the interest of heading-off any potential future confusion, it's probably important to clarify that the difference in cell voltages at the 'tails' - at the point when discharge is terminated because one cell has reached the user-programmed low voltage threshold before others - is NOT indicative of any major variation in capacities.

If the low voltage cutoff threshold is 1.5V and the maximum voltage variation is only 0.4V, that means the cell with the largest capacity was at 1.9V at termination. The amount you'd discharge from that 1.9V cell down to a termination voltage of 1.5V is probably tiny, especially considering that these are 20Ah cells. Even a 0.5V variation - 2V for highest cell at termination, would be trivial - in terms of actual capacity difference.

Your conclusion makes sense - best to use something like a 2V cutoff rather than trying to run your cells down to 1.5V. Same logic applies to the top-end. At both the top and bottom, voltages rapidly increase and decrease, when a cell is full or empty, so over only a handful of seconds one cell's voltage can quickly deviate from the others.
 

·
Registered
Joined
·
34 Posts
There might also be a little variation between your voltmeter and theirs. Are either calibrated reliability? Is mine? We don't know. The important thing is that the terminal voltages are within a few milli-volts of each other.

As Ian showed, self discharge is extremely low:)
For sure self discharge is very low, I've had some previous packs long enough to tell that. The 4 latest weren't all the same variance from the writing they put on the pack of 12 (some a tenth of a volt, some two tenths, some four tenths, etc). Yet in the couple weeks they've been here have since been stable (so no really bad cell causing a pack to self discharge super fast). So even if the two voltmeter's are off it wouldn't explain that they aren't all off by the same amount if they were all done right before shipping. I would think they were measured/balanced some time ago by greentec (probably not all at the same time since they have hundreds of packs).
 

·
Super Moderator
Joined
·
7,989 Posts
@eq1, yes, mostly agree.

The cells as delivered by Greentec seem to be well top balanced. I have a slight concern that "some" packs may drift out of top balance over so significant time - whatever that means. After all, they are used packs.

I think that it is prudent to drill small holes in the covers, as I showed in my build thread, so that the balance of the packs can be checked occasionally. Balance is currently looking pretty stable, but if imbalance reached 80-100mV, I'd be looking for some way to rebalance - otherwise, "my" proposed operating range of 2-2.5V would be endangered by outlier cells. Just my opinion:)
 

·
Registered
Joined
·
6,660 Posts
Has anyone bothered to ask Greentec what, if anything, they do to the packs prior to selling? My gut tells me they probably at most charge them to some level - in bulk - but don't do a balancing charge. If that were the case then how well the cells remained balanced when purchased used would simply be a function of how well they were balanced in the car and of the consistency of self discharge from cell to cell...

edit: I sent them a message: "Do you charge these modules before selling them? And if so, what kind of charge - bulk, balance-charge, to what voltage? In general, after they were pulled from the cars (Honda Fit EV), what happened to them, if anything, before being sold?"

I'll report back if I hear anything.
 

·
Registered
Joined
·
34 Posts
@eq1, yes, mostly agree.

The cells as delivered by Greentec seem to be well top balanced. I have a slight concern that "some" packs may drift out of top balance over so significant time - whatever that means. After all, they are used packs.

I think that it is prudent to drill small holes in the covers, as I showed in my build thread, so that the balance of the packs can be checked occasionally. Balance is currently looking pretty stable, but if imbalance reached 80-100mV, I'd be looking for some way to rebalance - otherwise, "my" proposed operating range of 2-2.5V would be endangered by outlier cells. Just my opinion:)
Not sure if you live in an area where it gets cold but in winter the bottom 2.0v may be a bit high (only in winter). I'm also unsure of the typical C rates in the insight if it's only .2C or so 2.0v would be fine. If it's a higher C rate like in my EV then in winter setting the bottom at 2.0v is leaving a lot on the table and voltages don't nosedive at that voltage the way they do when it's warm outside.
 

·
Registered
Joined
·
6,660 Posts
^ 2V is just a general target, or concept, right? Like you guys would be doing something to compensate for high discharge current, rather than just having a blanket 2V cutoff at all discharge rates? For example, the OEM system allows 1V (120V total) with the NiMH cells, for short durations, but hitting that voltage doesn't mean the pack is empty -- voltage is only low because it's a high current discharge. At resting voltage the empty voltage is more like 1.2V, 144V total...

On the other hand, the Toshiba cells are so efficient, low IR, etc. that even at high current I doubt you'd be far off from empty if you did use a blanket 2V cutoff...

What you choose would probably depend in part on whether you're using the pack in a plug-in like situation or just a NiMH replacement with similar functionality, where plug-in you'd probably be trying to get maximum energy out of the pack...

Oh, also, on C rate, I have a figure scrawled on my desk that says "106V at 121 amps" -- that was one of the higher discharges Peter P. did with his 48 cell LTO setup. So that's 6C at average 2.21V per cell... And as I recall reading his results, at various output levels, even higher than above, the LTO voltages barely sagged at all...
 

·
Super Moderator
Joined
·
7,989 Posts
^ I never intended to be specific on usage scenarios. I was merely reporting what I had seen in testing with the PL8, and forming in my own mind some seemingly conservative PERSONAL guidelines. Whatever others choose to do with their packs is of course their own business and their own risk.

Just a little more information on my testing. The discharge current used in the above tests was 40A, or about 2C. The test temperature was room temperature or about 20C. Perhaps significant is that the internal resistance for these used packs varies widely, if the PL8 is to be trusted. At 20A charging rate and mid-range SOC (2.3V on charge), the IR seems to be vary from .4 to 1.2 milli Ohm. It seems to me that the fairly wide IR of the cells will tend to cause imbalance over time, but I don't currently have a view of the degree. Performance parameters will vary with temperature and load.

The stock MCM seems to constrain assist current to short bursts of 70A and then falling to 40A, at moderate temperatures. I think these currents are probably ok within the derated 2-2.5V range, at least for me personally. Those are the conservative limits under which I personally operate my conversion at this time. I may extend these ranges later after including a "light" current hack and a purpose built BMS. Any extension I make will just be a bit on the high voltage end, perhaps to 2.6V.

Current hacks change all this and I have no qualifications to judge their impact. I'm relatively comfortable with the lighter hacks because the LTO chemistry is tolerant of high burst current, but Peter is a much better judge of this aspect than I, and he has much more experience.

I hate to sound like a lawyer, but these are exactly some of the technical driven reasons why no one is willing to build a conversion for pay. Just too many open questions at the moment, and a total lack of control over how such a product would be operated or how it might be modified.

Correction: above stated internal resistance.
 

·
Super Moderator
Joined
·
7,989 Posts
Has anyone bothered to ask Greentec what, if anything, they do to the packs prior to selling? My gut tells me they probably at most charge them to some level - in bulk - but don't do a balancing charge. If that were the case then how well the cells remained balanced when purchased used would simply be a function of how well they were balanced in the car and of the consistency of self discharge from cell to cell...
I talked some with my contact at Greentec on some of this. They are currently doing two things before shipping, they are measuring the terminal voltage as a quality check, and they are reading the BMS board ID. They are marking both on the individual packs. They are doing no "cover off" stuff, so they are doing no cell level measurements. Of course, they have a generous replacement policy if you have a bad on arrival pack.
 

·
Super Moderator
Joined
·
7,989 Posts
^ I need to correct my post #154 above. I rechecked my internal resistance test data and I see that the IR was never as low as .2 milli Ohm. The approximate range has been about .4-1.2 milli Ohm. I have corrected the post.

Full range testing seems to indicate that the packs were originally the 20 Ah packs.
 

·
Registered
Joined
·
6,660 Posts
Has anyone bothered to ask Greentec what, if anything, they do to the packs prior to selling?.... edit: I sent them a message: "Do you charge these modules before selling them? And if so, what kind of charge - bulk, balance-charge, to what voltage? In general, after they were pulled from the cars (Honda Fit EV), what happened to them, if anything, before being sold?"
Here's Greentec's response:

"I do a voltage check of every cell before shipping.
We also do a capacity test of one module from the pack to verify capacity.
We don't manipulate the charge level too much with these. Usually ship them at what they were at when we pulled them off the shelf."
 

·
Super Moderator
Joined
·
7,989 Posts
^ Sorry. Urban dictionary says it is hummm. I guess they would be the authority:)
 
141 - 160 of 173 Posts
Top