DIY stage 2: received Li Ping LifePo4 battery [photos]

[daniel.weck]

Box received, no Hanna Montana cuteness I'm afraid

http://www.pedelecs.co.uk/forum/electric-bicycles/4769-diy-stage-5-prepping-fitting-electricals-photos.html#post60345

 

[dan]

I guess no (nobody asked me to pay anything extra). How would they charge me ? Would the delivery man require an extra payment, or would I be charged on my credit card automatically ??

Very lucky Daniel, if customs had checked your parcel you would not have received it....lithium batteries are classed as dangerous goods and have to have special labels on the parcel...clearly your parcel was wrongly labeled as containing a power supply, and was undervalued. It is normal for the Royal Mail or any other shipping company to collect import duty and VAT on goods they carry

 

[daniel.weck]

Very lucky Daniel, if customs had checked your parcel you would not have received it....lithium batteries are classed as dangerous goods and have to have special labels on the parcel...clearly your parcel was wrongly labeled as containing a power supply, and was undervalued. It is normal for the Royal Mail or any other shipping company to collect import duty and VAT on goods they carry

I see. Thank you for the information. It seems common practice from Chinese seller though...

 

[dan]

very common practice...but would your supplier refund you if it was confescated

 

[flecc]

Custom checks are extremely random and infrequent in nature. After large numbers have bought the chargers
I recommend for two bike types from Powerstream in the USA over the last two years, not one had ever been
caught for any duty etc.

Then the other day for the first time one very unlucky lady was, the parcel held back by the Post Office who
demanded the customs charge of £9.38 for VAT plus £13.50 as the Post Office clearance fee. Fortunately no
import duty was levied.
.
​

 

[daniel.weck]

very common practice...but would your supplier refund you if it was confescated

Or if it was damaged / lost in transit: no guarantee. Yep, i've been lucky indeed !

 

[daniel.weck]

Ah, one red LED is now constantly off, even after several attempts to give the built-in BMS some time to balance the prismatic pouch cells. Normally, each cell reaches full capacity (ceiling voltage threshold) at the end of the first phase of the charge cycle, all the red LEDs go red, and the BMS subsequently blinks like a christmas tree whilst it balances the current across the cells to even out the final voltage (Li Ping LiFePO4 v2.5). This can actually take several hours, especially when the battery is new-ish.

So, I checked the individual cell voltages via the provided tabs connector (after removing the plastic shrink-wrap that protects the BMS circuit board) => 12 cells in total, or more precisely: one "cell" is made of 2 actual units, wired in parallel (for twice the amps / capacity), and the 12 pairs are then wired in series (3.3V nominal voltage for each pair, 10Ah capacity in total). One of the "cells" is indeed clearly under full capacity. It could either be a BMS problem (although Li Ping has a good reliability track record) or a faulty cell.

...time to take the pack apart!







...and now to "manually" charge the weak cell directly, using my Turnigy charger/balancer (I first tried at 1A for a few minutes, now charging at 2A).



We'll see if this cures the problem (there's a second cell in the pack that is a bit low, I'll give it some as well).

Cheers, Daniel

 

[daniel.weck]

Okay, so out of the 12 LiFePo4 cell pairs, 10 of them settled at 3.6V after a full charge (allowing sufficient time for the internal chemistry to stabilise: from 3.7/3.8V immediately after end-of-charge, to 3.6V after ~30mn or less).

The 2 remaining cells settled at ~3.3V, i.e. slightly above nominal voltage (which is reassuring). These 2 cells are therefore not dead, but there is however a clear discrepancy in SoC (State of Charge). From what I read, this can happen, and it has an adverse effect only on mileage (no damage to the pack). The ride autonomy is reduced because the BMS shuts down at the LVC (Low Voltage Cutout) of 2.1V for any given individual cell when under load (discharge). Because one of my cells was notably imbalanced (way below median capacity), this was enough to shut down the whole pack despite the other cells still having plenty of juice left in them). The Cycle Analyst (or any voltmeter) was hinting at a healthy pack, because the low capacity of the culprit cell was only detectable during a significant discharge rate (voltage sag under load).

Note that the charging voltage ceiling threshold for any given cell is 3.9V (with the Li Ping v2.5 BMS, when charging the entire pack as a whole), so it may well be that the great imbalance with 1 or 2 cells forced the charging cycle to end "prematurely", in the sense that the remaining 1 or 2 cells were left under maximum capacity.

When I first individually-charged the weak cell (after a full charge cycle with the official charger/BMS combo), it took about 40 minutes, so clearly it needed a proper top-up (which the built-in BMS somewhat did not achieve, probably due to the great imbalance). Same remark about the second weak cell, albeit not to the same extent (it needed less of a "booster" charge via the Turnigy). Now that I have forced a top-up of the individual weak cells, I am expecting a much better mileage (i.e. LVC at 2.1V should occur later).

So, I'm now going to leave the official Li Ping charger attached to the BMS for a few days, to see if the pack balances better. That's with a green charger LED instead of red, which indicates that the main charging phase has ended, and the charger is actually pushing virtually no amps (unit is barely warm), producing the full 45V. Apparently the Li Ping BMS must be connected to a power source for balancing to occur. I wrongly assumed that the blinking red LEDs on the BMS (whilst the charger is disconnected) are a indication of active balancing, but it seems that's not quite the case.

Daniel

 

[Deleted member 4366]

I had exactly the same as you. First, the LED not coming on unless I left it on charge for a very long time. Then the slightly lower voltage on one cell, then I notice and measured a reduction in capacity. My battery tester showed 15aH instead of 20aH. I openedit up and thisis what I found. A it disappointing really after only about 50 charge cycles and two years old.

 

[daniel.weck]

Interesting, thanks for sharing!
I don't have any swollen pouch / prismatic cells in my pack though. I hope it is just a typical LiFePO4 cell discrepancy.

I have not used this pack a lot actually, but I have owned it for some time. whilst in storage, I did not float-charge the cells frequently (trickle), I only regularly plugged in the supplied charger (mostly low depth of discharge, apart from a few long distance rides).

Unfortunately I do not have a 12s discharge rig to test the effective battery capacity, so instead I will ride my bike and will collect the CycleAnalyst statistics (my local hills should quickly exhaust the battery pack ).

Dan

I had exactly the same as you. First, the LED not coming on unless I left it on charge for a very long time. Then the slightly lower voltage on one cell, then I notice and measured a reduction in capacity. My battery tester showed 15aH instead of 20aH. I openedit up and thisis what I found. A it disappointing really after only about 50 charge cycles and two years old.

 

[cwah]

You need proper compression on prismatic cells to avoid this type of issue

 

[daniel.weck]

It looks like the trick consisting of topping-up the 2x individual weak cells in my LiFePO4 pack has worked its magic (3.6V normal charge voltage + 1C charge current ~2.3A ... although I tried 5A as well on one occasion).

The procedure basically boils down to balancing the pack manually (therefore bypassing the integrated Li-Ping v2.5 BMS) immediately after the Constant Current phase of the charging cycle, in place of the Constant Voltage phase normally handled by the provided charger.

All 12 red LEDs on the BMS now reach full brightness, the voltmeter indicates >3.6V on each pairs of cell tabs (even after several minutes of chemistry stabilisation, but of course the cells go down to nominal voltage ~3.3V when discharging under load). The BMS itself now does a better job at balancing the cells after it has finished charging the pack.

For now I am keeping the pack open so I can get to the cell tabs (I have taped a "flap lid" at the top, for easy access), because there's one particular cell that still has some discrepancy in its Depth Of Discharge (I am still testing the 100% State Of Charge, hoping that the cell will align with its siblings after a few cycles).

I have also changed the packaging from the rigid aluminium box (see earlier photos in this thread) to a more discrete soft-padded bag which fits straight inside the front-mounted Brompton bag. I might post photos of that later My motivation for this change is that I ride the Brompton offroad from time to time: the vibrations and the harsh bumps are pretty nasty on the prismatic pouch cells, when standing on their edge rather than lying on their side.

Dan

Okay, so out of the 12 LiFePo4 cell pairs, 10 of them settled at 3.6V after a full charge (allowing sufficient time for the internal chemistry to stabilise: from 3.7/3.8V immediately after end-of-charge, to 3.6V after ~30mn or less).

The 2 remaining cells settled at ~3.3V, i.e. slightly above nominal voltage (which is reassuring). These 2 cells are therefore not dead, but there is however a clear discrepancy in SoC (State of Charge). From what I read, this can happen, and it has an adverse effect only on mileage (no damage to the pack). The ride autonomy is reduced because the BMS shuts down at the LVC (Low Voltage Cutout) of 2.1V for any given individual cell when under load (discharge). Because one of my cells was notably imbalanced (way below median capacity), this was enough to shut down the whole pack despite the other cells still having plenty of juice left in them). The Cycle Analyst (or any voltmeter) was hinting at a healthy pack, because the low capacity of the culprit cell was only detectable during a significant discharge rate (voltage sag under load).

Note that the charging voltage ceiling threshold for any given cell is 3.9V (with the Li Ping v2.5 BMS, when charging the entire pack as a whole), so it may well be that the great imbalance with 1 or 2 cells forced the charging cycle to end "prematurely", in the sense that the remaining 1 or 2 cells were left under maximum capacity.

When I first individually-charged the weak cell (after a full charge cycle with the official charger/BMS combo), it took about 40 minutes, so clearly it needed a proper top-up (which the built-in BMS somewhat did not achieve, probably due to the great imbalance). Same remark about the second weak cell, albeit not to the same extent (it needed less of a "booster" charge via the Turnigy). Now that I have forced a top-up of the individual weak cells, I am expecting a much better mileage (i.e. LVC at 2.1V should occur later).

So, I'm now going to leave the official Li Ping charger attached to the BMS for a few days, to see if the pack balances better. That's with a green charger LED instead of red, which indicates that the main charging phase has ended, and the charger is actually pushing virtually no amps (unit is barely warm), producing the full 45V. Apparently the Li Ping BMS must be connected to a power source for balancing to occur. I wrongly assumed that the blinking red LEDs on the BMS (whilst the charger is disconnected) are a indication of active balancing, but it seems that's not quite the case.

Daniel