Li battery vs. NiMH battery

The cells I am using are GR18650C Li-ion cylinderical cells. They have a higher energy density than LiFePo4 and you can fit upto 120Ah into the Vectrix battery space. The 3.0Ah of NiMh cells are to constantly balance the pack. Pack balance +or- 0.02v per Li-ion cell from 3.0 to 4.2 volts. There are 36 Li-ion cells in series. I do not ride the bike to below 120 volts and this is safe with a balanced pack. No cells go flat and if a Li-ion cell was to go flat it won't heatup. They develop an internal short and won't recharge. No electronic BMS has been fitted. I have ridden this bike for 10 months and it still has the same range as converted.

I have converted my Vectrix to Li-ion 10 months ago. The battery has 90 Ah Li-ion and 3 Ah Ni-mh. The Nimh cells balance the Li-ion cells and no additional electronics has been fitted to my bike. I fitted a volt meter to monitor the battery voltage and prevent overdischarge. I have a range of 280 km at 60 km/hr or 210 km at 80 km/hr. I don't have to worry if I have sufficent range to get home on my trip out for the day. The weight of the battery is 85 kg and it fits in the standard battery space. I have removed the fans. The pack runs at a higher voltage, 120 - 145 volts with slight improvment in acceleration.
The disavantages is that the standard vectrix charger won't fully recharge the battery if you travel more than 105 km. If you do a long trip, it will take several charges to bring the battery upto full charge. I have fitted a socket for a trickle charger that I occasionally use if I want to fully recharge a flat battery faster.

Sounds interesting!

Have you described this conversion in more detail somewhere already? If not, could you please give more details?

Are the NiMH in parallel with the Li-ion cells?

The disavantages is that the standard vectrix charger won't fully recharge the battery if you travel more than 105 km. If you do a long trip, it will take several charges to bring the battery upto full charge.

The Ni-mh cells are wired 3 cellls in series and them in parallel with each 90Ah Li-ion cell. They form a series string of 108 Ni-mh cells with every third inteconnected to the 36 Li-ion cells. The Ni-mh cells act as a shunt regulator for each Li-ion cell. When the Li-ion cell reaches 4.05v the the Ni-mh cells are shunting 150ma. When the Li-ion cell reaches 4.20v the Ni-mh cells are shunting 400ma. The pack is being slowly and continuously balanced by the Ni-mh cells and stays in balance. 3.0 mah of Ni-mh cells is sufficient to balnce 90Ah of Li-ion cells that are in good condition. The high voltage cutout of 152 volts on the Vectrix charger prevents overcharging. Using the characteristics of these batteries this way meant that I did not need any additional battery management. The Li-ion cells need to match and be in good condition. The Ni-mh cells need to be in good condition to keep the pack balanced. This system will only work with Li-Ion or LiPo batteries. LiFePo4 cells cannot be used with this technique as their voltage and charge curves do not match.
I balanced the cells before fitting to my Vectrix. 1 month later I checked the voltage of each li-ion cell and found 2 Li-ion cells that were 0.26 and 0.38 volts lower that the other cells. checking the Ni-mh cells in parallel I found I had 2 faulty Ni-mh cells in the new cells I had used. I replaced these and have had no other problems.
I have used this pack in my bike for 10 months with no degradation of performance. The longest trip I have made was from Croydon to Traralgon in Victoria Australia on a single charge.

The disavantages is that the standard vectrix charger won't fully recharge the battery if you travel more than 105 km. If you do a long trip, it will take several charges to bring the battery upto full charge.

I think the Laird is capable of making a firmware matching the real capacity of your battery.

The upper limit so far is 65Ah, but you can use a timer to reset the charger an unlimited number of times

Battery size is no problem

Matt

The cells I am using are GR18650C Li-ion cylinderical cells. They have a higher energy density than LiFePo4 and you can fit upto 120Ah into the Vectrix battery space. The 3.0Ah of NiMh cells are to constantly balance the pack. Pack balance +or- 0.02v per Li-ion cell from 3.0 to 4.2 volts. There are 36 Li-ion cells in series. I do not ride the bike to below 120 volts and this is safe with a balanced pack. No cells go flat and if a Li-ion cell was to go flat it won't heatup. They develop an internal short and won't recharge. No electronic BMS has been fitted. I have ridden this bike for 10 months and it still has the same range as converted.

Okay, so let me see if I can understand this. I looked up a GR18650C Li-ion cell that had 2.2AH capacity and 300 cycles life expectancy. Please let me know if this spec is way off. So it would take 41 cells to equal 90AH. Or 41x36 cells for pack; 1476 Li-ion cells in 36s41p configuration? Is that a close guess? Then for each 41p group, you have 3s Ni-Mh in parallel. Sounds doable, but I would hate to replace a cell or two in a parallel group.

What do you expect for a life cycle expectancy on this pack? Can you post a pic of each parallel group?

This is a better cell, dimensions make for an easier fit, its higher capacity and when shipped it comes with a listing of each cells actual capacity and internal resistance
http://www.ev-power.eu/CALB-40Ah-400Ah/SE70AHA-Lithium-Cell-LiFePO4-3-2V-70Ah.html
If you prefer 60Ah, use TS or WB, don't use Sinopoly as it is the wrong shape

Hi Matt,
The 70 Ah cell appears to be no longer available, I can only find 40 and 60Ah (The 60Ah appears to give 70 Ah capacity at low discharge rate)
Your comment confuses me - the 60/70Ah cell appears to be 14 cm wide, how does that fit in the 23 cm wide battery box?
Only the 40Ah cell looks like it is made for the Vectrix at 115mm width and there both CALB and Winston/TS 40Ah seem to have the same dimensions.
For Sinopoly 60Ah cells, the only option is to put 6 sideways (6x 34mm) and 5 rows of 14 to get 30 cells in the bottom layer, then lay two rows of 6 sideway on top for a total height of at least 37 cm (22, interconnects and 14). Of course, BMS wiring will need to be adapted for such a setup.

sinopoly make 60Ah cells in two different shapes

one in the old TS 60Ah shape and another in the 14cm wide shape

looks like GWL no longer stocks CALB cells.
CALB do still make the 70Ah 115mm wide cells though

Matt

The upper limit so far is 65Ah, but you can use a timer to reset the charger an unlimited number of times

How much time, at a minimum, is required between power-down and power-up for
the on-board charger to 'reset' and start a new charge cycle?

Since this is a going discussion, I'll post this question here instead of in the official Vectrix conversion thread. I am wondering why Vectrix includes a new charger in the conversion. Does anyone have any ideas (other than bald greed) why they would consider this necessary? Are the charging characteristics of the LiFePo batteries that different that a whole new design is desirable? Or perhaps they are taking this opportunity to replace a part they expect would have problems in the future so they are reducing liability under the new warranty by charging the customer for a new charger now? I find this odd.

The ESD charger output cannot get down to the low currents needed for equalizing the Lithium cells. The ESD charger will destroy itself if you try to draw less than an couple amps. So in order to switch to the Li, the Runke (EVPS) charger is needed.

Thanks X Vectrix. So, if the scooter already has the newer (Runke) charger, there is no need to switch, correct? I'm sure I have the original charger (still running the 62 mph software). This is my biggest question about a diy conversion, and from this answer, it seems to matter. The old charger will not work with Lifepo batteries.

That makes me wonder how antiscab, oob and others are dealing with charging in their Li conversions. Has anyone done a diy charger install that would work with Lifepo and make nice in communicating with the rest of the system?

I hope this isn't too off topic.

For my own clarity, are you saying that the charger upgrade will be to the same charger which they have been using ever since the switch to Runke from the original?

Thanks X Vectrix. So, if the scooter already has the newer (Runke) charger, there is no need to switch, correct? I'm sure I have the original charger (still running the 62 mph software). This is my biggest question about a diy conversion, and from this answer, it seems to matter. The old charger will not work with Lifepo batteries.

That makes me wonder how antiscab, oob and others are dealing with charging in their Li conversions. Has anyone done a diy charger install that would work with Lifepo and make nice in communicating with the rest of the system?

I hope this isn't too off topic.

The 62mph SW is on the motor controller, but you are correct, the old silver cast aluminum ESD charger will not work with Li batteries. In order to properly charge the Li cells the current has to be controllable down to 300 mA. The ESD cannot do this.
However, if you have one of the original Runke chargers, it too may be swapped out for a newer version Runke. The newer version has many fixes and upgrades in it.

Thanks X Vectrix. So, if the scooter already has the newer (Runke) charger, there is no need to switch, correct? I'm sure I have the original charger (still running the 62 mph software). This is my biggest question about a diy conversion, and from this answer, it seems to matter. The old charger will not work with Lifepo batteries.

That makes me wonder how antiscab, oob and others are dealing with charging in their Li conversions. Has anyone done a diy charger install that would work with Lifepo and make nice in communicating with the rest of the system?

I hope this isn't too off topic.

The 62mph SW is on the motor controller, but you are correct, the old silver cast aluminum ESD charger will not work with Li batteries. In order to properly charge the Li cells the current has to be controllable down to 300 mA. The ESD cannot do this.
However, if you have one of the original Runke chargers, it too may be swapped out for a newer version Runke. The newer version has many fixes and upgrades in it.

Hi the old Charger might work if you use shunt-regulators which can deal with 3A.
Not effective but useable.
Greetings Mike

Hi the old Charger might work if you use shunt-regulators which can deal with 3A.
Not effective but useable.

Hello. Thanks for the idea. Do you mean "not effective" only in the sense of wasteful of energy?

I'm sure I have the original charger (still running the 62 mph software). This is my biggest question about a diy conversion, and from this answer, it seems to matter. The old charger will not work with Lifepo batteries.

If doing a conversion with my kit, the ESD charger is preferred, as it can be reprogrammed, whereas with the runke we do not know how

due to the minimum charge current/voltage curve on the commodity LiFePO4 cells we use, the charge current shouldn't ever be below 2A.

In practice this means the charger puts out 3A, and the cells higher than 3.6v get shunted

It is best practice not to hold a cell at 3.4v or above for long periods of time, so balancing does not happen all in one go with the kit.
This is what allows use to use the original charger at a minimum charge current of 3A, even though it can go down to 1A

I presently use a TC Charger 2000W 15A on my own Vectrix, however, thats only because I did my conversion before The Laird was able to remove the Ah charge limitation on the ESD charger

All my new conversions retain the original ESD charger

Matt

Ah, Matt, very reassuring. As usual, you have an answer. I've been worrying about this all day--been wondering whether I could install a new Vectrix charger myself, how much I could get one for, etc.

In practice this means the charger puts out 3A, and the cells higher than 3.6v get shunted

Does "get shunted" mean automatically by the BMS?

So, one more time slowly, let me see if I'm following this. X Vectrix tells us the new charger can equalize using small currents which the ESD charger cannot do. Antiscab uses a different method of equalization which never requires fewer than 3A out of the charger which is 2A to the battery. I would like to put my understanding of the equalization up for critique. Some cells will reach the critical value of 3.6 V while others still want further charge. The charging is stopped here, the battery loaded for some period of time, and the charge is applied again. During the usage of the battery, the higher cells are contributing more energy and thus draining relatively faster, bringing the cells more into equal values. Then on the subsequent charge, the lower voltage cells are taking more charge relative to the higher voltage ones. Over several charge/discharge cycles, the pack will gradually equalize naturally without force-feeding small currents to the cells which are lagging in voltage. This all assumes healthy cells and relatively equal charge/discharge behavior among the cells.

It is best practice not to hold a cell at 3.4v or above for long periods of time

Does this recommend, as has been stated elsewhere, charging immediately before riding, especially until good equalization? Seems that operating below full capacity even more than usual during initial use would be good practice here--reminds me of the old cars which had to be broken in at slower speeds for a certain number of miles before being pushed.

due to the minimum charge current/voltage curve on the commodity LiFePO4 cells we use, the charge current shouldn't ever be below 2A.

Meaning that the repeated charge/discharge method of equalization is actually better for these batteries than the low voltage equalization done by the new Vectrix charger? (Not that it matters that much--I'm just curious about this.)

Last week I was speaking with a mechanic who has a souped up Gem. He replaced the gel batteries with Calb (sp?) LiFePo4. He swore to me that after they were equalized, equalization has never been a problem--popped off the seat (I wish we had such easy access) and put a meter on them for me to see. He is convinced that he does not need a BMS. I'm not sure if this is a characteristic of the battery type, his particular collection of batteries, his system in general, or even an accurate statement. It is reassuring, though.

My 4000Li has 9074 miles on it, and still running very strong. No BMS. Do not take this as scientific proof, but the last time I pulled my cycle apart and measured all 20 LifePO4 cells, the voltages varied at full charge from 3.52 to 4.08. But under load, and after running for several miles, the voltages were exactly identical, at 3.28 for all twenty cells. I do not have an explanation for this, but the balance was nice to see. It appeared the charger does indeed overcharge some cells, and I am now limiting the charging cycles to 3 hours (double green). But how they end up in perfect balance after driving a few miles is beyond me. But they do. I have 9074 miles that, in my mind, states BMS may not be worth the complexity. My top speed is right around 53mph and my range is still at 40 miles, which are both very close to my specs when my 4000li was brand new.

Pete, your cells are unbalanced. You are over-charging (damaging) the cells at 4+ Volts.
Reason they all get to the same voltage after some discharge is that LiFePO4 has such a
flat discharge curve that they stay at the same voltage essentially from 20% to 80% charge level,
but when imbalanced, some will be pushed over the "knee" and skyrocket in voltage while other
are still close to the "flat" range.
You must use either a BMS or regularly manually balance (so *you* are the BMS)
and never charge to the cells to 100% full because there will always be some that
have drifted and will get damaged before the rest catches up...

Thank you Cor! When my cycle was new I would leave it plugged in overnight and the charger would go into pulse mode, pushing the cells from 70 volts to eventually 79.9. I no longer leave it charging for more than 3 hours, which usually cuts off the charge at around 71 or 72 volts, so I am pretty sure none of my cells exceed 4.0 at that point. I have 9 months left on my battery warranty, and should be close to 15,000 miles at that point. In the meantime, I am adding a 3 hour timer to my charger, and considering adding a 21st cell when I do finally take the bike apart to check the cells for warranty replacement. Both of those should reduce the charge per cell somewhat. Complexity of BMS?? Are they putting BMS on the 5000li now?

...I would leave it plugged in overnight and the charger would go into pulse mode, pushing the cells from 70 volts to eventually 79.9.

Aiiieee! Even for perfect balanced cells they were pushed to 4.0 Volts on each charge !?!?!
That is certainly damaging to your cells and anything holding LiFePO4 cells at low current over 3.6V means over-charging.

I no longer leave it charging for more than 3 hours, which usually cuts off the charge at around 71 or 72 volts

Good, if the cells are well balanced then this would hit them with the max 3.6V that the charging profile of LiFePO4 should get: charge to 3.6 until current drops to 0.05C (2A for a 40Ah cell) then stop charging. So it is best if you have a BMS that detects any cell reaching 3.6V and simply switch off the charger (I believe Antiscab's Li upgrade does this).

... and considering adding a 21st cell when I do finally take the bike apart to check the cells for warranty replacement. Both of those should reduce the charge per cell somewhat. Complexity of BMS?? Are they putting BMS on the 5000li now?

I don't know, I have only some experience with Vectrix Li upgrade.

PlzPete, at least go with the "Poor Man's" BMS: the Cell-log 8M - costs about $13 out on HobbyKing.com. It will give you both LVC and HVC. You connect balance wires to each of your 20 cells and then run those wires to Cell-log mating connectors. So you would end up with three 9-pin connectors (8+8+4 cell bundles). Test which cells are the weakest cells: first to hit 3.65V and first to hit 2.5V, by moving the Cell-Log between the three connectors. Once you know your weak cells, put them next to each other so their signals are going out on one Cell-Log connector. Then when charging or while running the bike, connect the Cell-log to the bike so it can monitor your weaker cells. Doesn't get simpler or cheaper than that. The Cell-Log does take some current so don't leave it connect when not riding the bike.

The Cell-log displays the voltage of 8 cells at a time, so periodically you can move it around the three connectors to see how all your cells are doing. Much easier than tearing down your bike to measure each cell by hand.

Or if you want to get analytical, you could buy the Cell-log 8S for about $25. It has a USB port and data logging function. With it, you could go for a ride, and then plug the Cell-log 8S into your laptop and view exactly how your cells performed during the ride. Cheap solution either way for folks w/o a BMS.

That makes me wonder how antiscab, oob and others are dealing with charging in their Li conversions. Has anyone done a diy charger install that would work with Lifepo and make nice in communicating with the rest of the system?

With Litium, if I need to put in a new charger, I just wire it straight in.

There is no CAN bus, and the fuel guage does not work
I use a cycle analyst as the fuel guage

It's not really a problem now, as the ESD charger can be made to work with the larger Lithium batteries

Matt

In practice this means the charger puts out 3A, and the cells higher than 3.6v get shunted

Does "get shunted" mean automatically by the BMS?

yes it does, by the circuit board that is on each cell
the circuit board only takes ~0.7A (cell voltage dependent)

that way as the charger puts out 3A, the highest cells don't get less than 2A even when balancing
if things are bad enough that a full cell reaches 4V, the charger is shut down.

In the case of the TC Charger, it restarts when the highest cell voltage is below 3.8v

if using the original charger, the charger stays off until you come and reset it.

It is best practice not to hold a cell at 3.4v or above for long periods of time

Does this recommend, as has been stated elsewhere, charging immediately before riding, especially until good equalization? Seems that operating below full capacity even more than usual during initial use would be good practice here--reminds me of the old cars which had to be broken in at slower speeds for a certain number of miles before being pushed.

It is best to keep the bike fully charged, but I normally wait until night-time off peak utility rates before charging

The reason I suggest charging immediately before riding for nimh is to help deal with self discharge, of which lithium does not really have a problem.

I do suggest not using the full capacity until you know the battery is top balanced, however, the battery is usually shipped from the manufacturer already balanced (all cells at 50%)

due to the minimum charge current/voltage curve on the commodity LiFePO4 cells we use, the charge current shouldn't ever be below 2A.

Meaning that the repeated charge/discharge method of equalization is actually better for these batteries than the low voltage equalization done by the new Vectrix charger? (Not that it matters that much--I'm just curious about this.)

yes I believe so, however, the chemistry of their cells may not be exactly the same as the commodity large format LiFePO4, which may mean it doesn't matter.

On my old scooter, I got 22'000 km out of 16 x 40Ah cells before capacity fell to 50% of original.
however, I used to use a charger that would float, at an average of 3.6v
I think that was the cause of the continuous capacity loss

my present battery has done 30'000km, with no capacity loss

Last week I was speaking with a mechanic who has a souped up Gem. He replaced the gel batteries with Calb (sp?) LiFePo4. He swore to me that after they were equalized, equalization has never been a problem--popped off the seat (I wish we had such easy access) and put a meter on them for me to see. He is convinced that he does not need a BMS. I'm not sure if this is a characteristic of the battery type, his particular collection of batteries, his system in general, or even an accurate statement. It is reassuring, though.

You can get away with running no BMS while the pack is new, all cells are nearly the same capacity and it starts out balanced.

The only issue is if you drive a little to far one day and reverse a cell, you will not know unless you get under the hood with a multimeter and check.

the devil in the detail is that a cell that is overcharged might not go short circuit, but rather develop self discharge instead. meaning even if you check just after a drive, it will become dead later on.

This can all be avoided if you absolutely do not use more than 80% of rated capacity.
That de-rating should be increased if the temperature is very low, and the battery is small relative to the car size.

If you charge a battery with a reversed cell, the charge voltage may be too high, which could overcharge another cell causing capacity loss in that cell.

next time you drive that cell also gets reversed, meaning now you have multiple dead cells.

next time you charge, the voltage the charger is aiming for is never reached as there are too many cells that are short circuit. when this happens the charger never shuts down, not only damaging every cell in the pack, but also potentially causing a situation that can result in fire.

Using a per cell BMS is not the only way to avoid this outcome.

Lee Harts batt bridge will tell you if you have a dead cell even while you are driving, for instance.

Matt

How much time, at a minimum, is required between power-down and power-up for
the on-board charger to 'reset' and start a new charge cycle?

Not sure, somewhere less than 10 seconds

My own timer can't go less than 15 minute increment, so I've never bothered to check

He swore to me that after they were equalized, equalization has never been a problem--popped off the seat (I wish we had such easy access) and put a meter on them for me to see. He is convinced that he does not need a BMS. I'm not sure if this is a characteristic of the battery type, his particular collection of batteries, his system in general, or even an accurate statement. It is reassuring, though.

As has been already mentioned, unless the cells are just off charge, you won't be able to tell how balanced they are due to being in the flat part of the voltage curve

LiFePO4 cells have *very* low self discharge when new.
The only way they get self discharge is by overdischarge, or something conductive ends up between the terminals (like iron ore dust)

running without BMS works ok if the cells start out balanced, until the first time a cell is overdischarged and develops self discharge. At which point it really all falls apart.

Matt

the old silver cast aluminum ESD charger will not work with Li batteries

Hi X Vectrix,

Would it be possible to buy the old ESD chargers that have been removed from official Lithium conversions?

I would be happy to buy on an as-is basis with no warranty

to everyone else,
If you are going to get an official lithium conversion which requires a change of charger, see if you can keep your old one :)

cheers,
Matt

He swore to me that after they were equalized, equalization has never been a problem--popped off the seat (I wish we had such easy access) and put a meter on them for me to see. He is convinced that he does not need a BMS. I'm not sure if this is a characteristic of the battery type, his particular collection of batteries, his system in general, or even an accurate statement. It is reassuring, though.

As has been already mentioned, unless the cells are just off charge, you won't be able to tell how balanced they are due to being in the flat part of the voltage curve

LiFePO4 cells have *very* low self discharge when new.
The only way they get self discharge is by overdischarge, or something conductive ends up between the terminals (like iron ore dust)

running without BMS works ok if the cells start out balanced, until the first time a cell is overdischarged and develops self discharge. At which point it really all falls apart.

Matt

Don't mean to get picky but I did not author the information in the above quote....
mk

oops, sorry, its was in fact Geomoo

I seem to have lost the ability to edit my posts, how odd

Matt