Notifications
Clear all

Recelling a 18650 based battery pack

Page 4 / 6
Kevin
(@kevin)
Estimable Member Customer
Posted by: @djibattworkshop

By the way what is model of new cells planned to use, is it BAK as you mentioned before?

And what cells were originally installed?

Also check presently preprogrammed chemistry profile by running these commands

WW 00 0008

RW 00

 

Original cells:

Brand: Samsung
Model: ICR18650-22F
Capacity: 2200mAh Rated
Voltage: 3.60V Nominal
Charging: 4.20V Maximum
Charging current (standard): 1100mA
Charging current (max): 2200mA
Discharging: 2.75V Cutoff
Discharging current (standard): 440mA
Discharging current (max): 4400mA

New cells:

Brand: BAK
Model: H18650CQ
Capacity: 2550mAh Rated
Voltage: 3.60V Nominal
Charging: 4.20V Maximum
Charging current (standard): 1250mA
Charging current (max): 2500mA
Discharging: 2.75V Cutoff
Discharging current (standard): 2500mA
Discharging current (max): 7500mA

ChemID result:

WW 00 0008
RW 00 0218

Thanks

ReplyQuote
Topic starter Posted : 10/09/2021 3:30 pm
Kevin
(@kevin)
Estimable Member Customer
Posted by: @torture

@kevin 

The Cycle Count must also be reset after battery recelling.

At the moment it's 1141.

Also not a bad idea to change the manufacture date since the battery will be considered as new.

Both values can be found on the SBS Configuration tab of BQEvsw.

Yes that's the plan but unfortunately, enerprof.de have not shipped my cells yet.

 

Thanks

ReplyQuote
Topic starter Posted : 10/09/2021 3:32 pm
Kevin
(@kevin)
Estimable Member Customer
Posted by: @djibattworkshop

So FCC just changed from 1504 to 2416mAh. This is because impedance tables in chip eeprom left from previous worn cells. They should be reset to reflect impedance of new cells.

There are three ways to do it.

1. Make relearning cycle according to TI requirements. Full Charge -> Relax 2 hrs -> Full Discharge -> Relax 5 hrs -> Recharge to Full or to Storage Level. Chip should recalculate impedances after this.

2. Updating the Chemistry tables using bqEasy plugin. It will rewrite impedance tables, OCV curves tables, temp correction tables and OCV flat range to your specific cells. But it require v1.04 support. Using v1.05 could break something in dataflash if any of updated parameters have different layout.

3. Updating Ra-tables manually using Hex-editor or SMBUS commands.

 

 

I am trying to perform method one to calibrate the battery but the maximum allowed Relax time before charging is 4 hours (240 minutes).

If the relax time is 4 instead of 5 hours will it be acceptable?

 

Thanks

ReplyQuote
Topic starter Posted : 19/09/2021 4:49 pm
DjiBattWorkshop
(@djibattworkshop)
Member Moderator
Posted by: @kevin

If the relax time is 4 instead of 5 hours will it be acceptable?

Minimum relax interval after discharge is 5 hours according to TI requirements for z-Gauges. This is to wait until voltage is building up and become stable to take proper DoD(OCV) to update Qmaxes.  Less interval is acceptable only if voltage is stabilized faster then in 5 hours period. Chip check this condition by measuring dV/dt. It should be less then 4uV/s otherwise Learning Cycle is disqualified. You could monitor these events using reqisters bitmap in bqEVSW software.

There is also LED on front panel of NLBA which draws about 10mA and could prevent early OCV stabilization event. You could check workaround made by @torture to switch it off for relearning of z-Gauges.

https://www.laptopu.ro/community/postid/1990/

P.s.

@azzido

Could we make appropriate limits for relaxation periods to follow TI recommendation?

It looks z-Gauges cant be relearned at all with present setup and front LED continuously attached as a load.

 

ReplyQuote
Posted : 19/09/2021 7:27 pm
Kevin liked
Nicusor
(@azzido)
Member Admin

@djibattworkshop 

Is it sufficient 1 cycle to learn and update the parameters correctly ?

NLBA Laptop Battery Analyzer with Chip Reset & Repair Option

ReplyQuote
Posted : 19/09/2021 7:36 pm
DjiBattWorkshop
(@djibattworkshop)
Member Moderator

If this one cycle is Qualified by chip then yes. For this TI rules of Qualified Discharge should be followed.

 

ReplyQuote
Posted : 19/09/2021 7:44 pm
Kevin liked
Kevin
(@kevin)
Estimable Member Customer
Posted by: @djibattworkshop
Posted by: @kevin

If the relax time is 4 instead of 5 hours will it be acceptable?

Minimum relax interval after discharge is 5 hours according to TI requirements for z-Gauges. This is to wait until voltage is building up and become stable to take proper DoD(OCV) to update Qmaxes.  Less interval is acceptable only if voltage is stabilized faster then in 5 hours period. Chip check this condition by measuring dV/dt. It should be less then 4uV/s otherwise Learning Cycle is disqualified. You could monitor these events using reqisters bitmap in bqEVSW software.

There is also LED on front panel of NLBA which draws about 10mA and could prevent early OCV stabilization event. You could check workaround made by @torture to switch it off for relearning of z-Gauges.

https://www.laptopu.ro/community/postid/1990/

P.s.

@azzido

Could we make appropriate limits for relaxation periods to follow TI recommendation?

It looks z-Gauges cant be relearned at all with present setup and front LED continuously attached as a load.

 

If I do it manually like this, would the calibration be valid?

 

1. Charge the battery from the Charge tab with "Ignore SoC" enabled.

2. When finished, disconnect the positive/red wire from the battery.

3. Wait at least 2 hours and reconnect the red wire.

4. Discharge the battery from the discharge tab with "Ignore SoC" enabled.

5. When finished, disconnect the positive/red wire from the battery.

6. Wait at least 5 hours and reconnect the red wire.

7. Charge the battery to 100%.

 

If this is OK, is the last step necessary? i.e. will the FCC update on completion of step 6?

 

Thanks

 

 

 

ReplyQuote
Topic starter Posted : 19/09/2021 7:55 pm
Kevin
(@kevin)
Estimable Member Customer

@djibattworkshop 

 

By the way, the LED seems to be drawing 12mA and that means the capacity of the battery will be reduced by 24mAh in two hours.

ReplyQuote
Topic starter Posted : 19/09/2021 8:03 pm
DjiBattWorkshop
(@djibattworkshop)
Member Moderator
Posted by: @kevin

1. Charge the battery from the Charge tab with "Ignore SoC" enabled

TI rule says "Charge the battery until FC bit is set". Branded TI chips like you working with detects this event very well so Ignore SoC 100% is not required. It's only useful for fake chips with poor gauging.

Posted by: @kevin

7. Charge the battery to 100%.

This step is not necessary. You could charge it then to any level. Qmaxes and FCC should be updated after step 6.

To be sure, export GG-profile before and after Relearning Cycle through bqEVSW DF pane, File Export feature. Then compare Qmaxes and Ra-tables using text editor. Me recommend NPP with FileCompare plugin. It will definitely show you if Relearning was Ok.

 

ReplyQuote
Posted : 19/09/2021 8:19 pm
Kevin liked
DjiBattWorkshop
(@djibattworkshop)
Member Moderator
Posted by: @kevin

By the way, the LED seems to be drawing 12mA

Yes, it depends on 2s, 3s, 4s configuration and common voltage.

Posted by: @kevin

and that means the capacity of the battery will be reduced by 24mAh in two hours.

Will check TI limit for passed charge disqualification rule.

 

ReplyQuote
Posted : 19/09/2021 8:23 pm
Kevin liked
DjiBattWorkshop
(@djibattworkshop)
Member Moderator

So TI disqualification criteria for passed charge during relax period (they call it Offset Error) is 1% of Design Capacity.

For your battery it is 4300*0.01=43mAh.

So 24mAh is Ok for this rule.

The question is dV/dt with LED attached.

It could be calculated on well relaxed battery by taking two Voltage readings on predefined period, say 1 or 3 or 5 hours. And then dividing it on time interval in seconds. For example for 1 hour interval Voltage should not drop more then 14.4mV under LED load.

 

 

ReplyQuote
Posted : 19/09/2021 9:30 pm
Kevin liked
Kevin
(@kevin)
Estimable Member Customer
Posted by: @djibattworkshop

So TI disqualification criteria for passed charge during relax period (they call it Offset Error) is 1% of Design Capcity.

For your battery it is 4300*0.01=43mAh.

So 24mAh is Ok for this rule.

The question is dV/dt with LED attached.

It could be calculated on well relaxed battery by taking two Voltage readings on predefined period, say 1 or 3 or 5 hours. And then dividing it on time interval in seconds. For example for 1 hour interval Voltage should not drop more then 14.4mV under LED load.

 

 

Thanks for that! I will try to do that in the morning.

ReplyQuote
Topic starter Posted : 19/09/2021 9:36 pm
Kevin
(@kevin)
Estimable Member Customer

I have calibrated the battery and these are the results:

image

Now it is time to check the capacity drop caused by the LED. I will repost after one hour.

 

Thanks

ReplyQuote
Topic starter Posted : 20/09/2021 7:06 am
Kevin
(@kevin)
Estimable Member Customer

Looks like the LED voltage drop for 1 hour is 6mV

image
ReplyQuote
Topic starter Posted : 20/09/2021 8:09 am
DjiBattWorkshop
(@djibattworkshop)
Member Moderator

6mV drop should be Ok if it could reach at the end of 4 hours period. After discharge voltage build-up has exponential form and dV/dt is quite fast at the beginning.

By the way have noticed FCC is quite small for this battery after relearning.

Have you analyzed GG-profiles before and after?

 

ReplyQuote
Posted : 20/09/2021 11:28 am
Kevin liked
Page 4 / 6
Share: