BQ34Z100-G1: SOC falls back to zero at 90%

Dig Kleppe86

Prodigy 150 points

Part Number: BQ34Z100-G1
Other Parts Discussed in Thread: GPCCHEM, EV2400

case: Lead Acid battery 36V 9000mAhr; 2438.logs.zip

Cell term voltage 1.6x18 = 28.4V

The hardware of my board misses the filter at the current input. R sens : 0.01E

Discharge @2A: OK

Discharge @4A: at Remcap exactly 8000mAhr : SOC 90->0% , Full ChargeCap 9000->1001 . The same happens with a constant Voltage instead of the battery.

It happens with all my boards.

Must be some setting?

When recharging the battery the Remcap goes from 0-> 1001, then steps to 9000.

Log from Bq studio:

Sample	DateTime	ElapsedTime	Control	CtrlStatus	StateofCharge	MaxError	RemCap	FullChgCap	Voltage	Current	AvgCurrent	Temperature	AverageTimeToEmpty	AverageTimeToFull	AvailableEnergy	AveragePower	InternalTemperature	CycleCnt	StateofHealth	ChargeVoltage	ChargeCurrent	PackConfig	Flags	Flags B	GridNumber	LearnedStatus	DoDatEOC	Qstart	TrueRC	TrueFCC	StateTime	DOD0PassedQ	DOD0Time	DOD0	QmaxDOD0	QmaxPassedQ	QmaxTime	LogRowTime(ms)	LogStatus
1	2018-12-19 11:03:49	4034	0x0203	0x0203	100	100	8979	9000	42534	-4180	-2966	25.6	182	65535	17685	-9877	27.3	0	100	43200	4500	0x09D9	0x1291	0x0400	1	4	0	0	8977	9000	45	23	0	0	0	23	0	886	SUCCESS
135	2018-12-19 11:12:45	540091	0x0203	0x0203	93	100	8358	9000	42552	-4165	-4167	27.0	120	65535	17538	-9846	28.5	0	100	43200	4500	0x09D9	0x1191	0x0400	1	4	0	0	8357	9000	579	643	3	0	0	643	3	887	SUCCESS
182	2018-12-19 11:15:53	728112	0x0203	0x0203	91	100	8142	9000	42552	-4164	-4164	27.1	117	65535	17487	-9844	28.6	0	100	43200	4500	0x09D9	0x1191	0x0400	1	4	0	0	8141	9000	766	859	4	0	0	859	4	887	SUCCESS
210	2018-12-19 11:17:45	840125	0x0203	0x0203	90	100	8013	9000	42552	-4162	-4162	27.2	116	65535	17456	-9837	28.7	0	100	43200	4500	0x09D9	0x1191	0x0400	1	4	0	0	8012	9000	878	988	4	0	0	988	4	887	SUCCESS
211	2018-12-19 11:17:49	844118	0x0203	0x0203	89	100	8009	9000	42552	-4162	-4162	27.1	115	65535	17455	-9839	28.7	0	100	43200	4500	0x09D9	0x1191	0x0400	1	4	0	0	8007	9000	882	993	4	0	0	993	4	888	SUCCESS
212	2018-12-19 11:17:53	848128	0x0203	0x0203	89	100	8004	9000	42552	-4162	-4162	27.2	115	65535	17454	-9839	28.7	0	100	43200	4500	0x09D9	0x1191	0x0400	1	4	0	0	8002	9000	886	998	4	0	0	998	4	888	SUCCESS
213	2018-12-19 11:17:57	852122	0x0203	0x0203	89	100	7999	9000	42552	-4162	-4162	27.2	115	65535	17453	-9839	28.7	0	100	43200	4500	0x09D9	0x1197	0x8400	2	4	0	0	-1	1001	890	1002	4	0	0	1002	4	887	SUCCESS
214	2018-12-19 11:18:01	856133	0x0203	0x0203	0	100	0	1001	42552	-4163	-4162	27.2	0	65535	0	-9841	28.7	0	100	43200	4500	0x09D9	0x1197	0x0400	2	4	0	0	-6	1001	894	1007	4	0	0	1007	4	886	SUCCESS
215	2018-12-19 11:18:05	860125	0x0203	0x0203	0	100	0	1001	42552	-4163	-4162	27.1	0	65535	0	-9841	28.7	0	100	43200	4500	0x09D9	0x1197	0x0400	2	4	0	0	-11	1001	898	1012	4	0	0	1012	4	887	SUCCESS
216	2018-12-19 11:18:09	864135	0x0203	0x0203	0	100	0	1001	42552	-4163	-4162	27.2	0	65535	0	-9841	28.7	0	100	43200	4500	0x09D9	0x1197	0x0400	2	4	0	0	-15	1001	902	1016	4	0	0	1016	4	886	SUCCESS

settings and log are in included zip

Thanks in advance

Dig

over 5 years ago

0 Batt over 5 years ago

TI__Mastermind 42535 points

Have you characterized your PbA cell using GPCCHEM? If not, you can see inconsistent performance.

0 Dig Kleppe86 over 5 years ago in reply to Batt

Prodigy 150 points

The cells were characterized by TI (801?).

6 HCNF J- 7.2 Betta batteries.

0 Bryan Kahler over 5 years ago in reply to Dig Kleppe86

TI__Mastermind 25955 points

Hi Dig,

Please let me know if the steps taken in this thread help resolve the issue: e2e.ti.com/.../2052274

Sincerely,
Bryan Kahler

0 Dig Kleppe86 over 5 years ago in reply to Bryan Kahler

Prodigy 150 points

Hi Bryan

Before i jump into this:

I suppose these steps will give me a more accurate gauge. But i dont think this will solve my problem: at 2A ( and 3A) discharche the gauge works perfect.

Which parameter causes the SOC jump to zero at exact 8000mAHr at 3.7A? , even at a constant voltage (40V) far afbove the battery-empty point (28.5V) ?

The Ri is 0.36E maximum.

My evaluationboard does exactly the same.

Dig

0 Bryan Kahler over 5 years ago in reply to Dig Kleppe86

TI__Mastermind 25955 points

Hi Dig,

Please send me the SREC from your device so I may investigate further. A log of the event in 1 sec intervals would also be helpful for debug.

Sincerely,
Bryan Kahler

0 Dig Kleppe86 over 5 years ago in reply to Bryan Kahler

Prodigy 150 points

Hi Bryan

Here are the files.

DigP18364-1.zip

0 Batt over 5 years ago in reply to Dig Kleppe86

TI__Mastermind 42535 points

Thanks, we will look into this and get back to you soon.

0 Bryan Kahler over 5 years ago in reply to Batt

TI__Mastermind 25955 points

Hi Dig,

After running a few simulations with this data I do see the jump but not a trigger from voltage. Could you please send me:

1) The cell spec sheet
2) The input files used to the GPCCHEM tool
3) The output files from the GPCCHEM tool
4) Another log at intervals of 1 second, but for a longer period of time. Please include a rest period at the beginning of the log where the cell is relaxed prior to discharge. This log should be on the order of hours. Please ensure the discharge portion of the log is greater than 600 seconds so I can monitor qualified discharge of the gauge in emulation.
5) The SREC prior to running the test above in step 4, and the SREC after running the test above in step 4.

Also, please use a learned pack in the subsequent test of step 4. The RA tables have not been learned with this device. A successful learning cycle is critical for gauging accuracy. The learned status of the pack should be a 0x06, not a 0x04.

Sincerely,
Bryan Kahler

0 Dig Kleppe86 over 5 years ago in reply to Bryan Kahler

Prodigy 150 points

Hi Bryan,

We don't have GPCCHEM files.

Included is the datasheet of the battery.

for your question 4: The logfiles I send you where from a Voltage source, not from the battery. ( the results were the same). So relaxing of the cell is not appliable.

Do you need logs of the real battery?

The accuracy of the gauging process is Ok as long as we stay below 3.7A , so the RA tables are not so important now.

Regards

Dig.

6-hcnf-7.2.pdf BATTID-report (1).zip

0 Bryan Kahler over 5 years ago in reply to Dig Kleppe86

TI__Mastermind 25955 points

Hi Dig,

In this case since we are trying to run an emulation to replicate the issue, the relaxing of the cell is extremely helpful in ensuring that state machine between the emulator and the device are aligned. Also, a longer time period is critical for analysis.

If you could send a log file with a real battery over a longer time scale (for example a learning cycle), it would be helpful.

Sincerely,
Bryan Kahler

0 Dig Kleppe86 over 5 years ago in reply to Bryan Kahler

Prodigy 150 points

logs2019.zipHi Bryan

I had blown my EV2400, took a while to repair.

Here are logfiles ,charge , decharge at 3A and 4A

0 Bryan Kahler over 5 years ago in reply to Dig Kleppe86

TI__Mastermind 25955 points

Hi Dig,

Thank you for the logs. I will update this post by Monday.

Sincerely,
Bryan Kahler

0 Bryan Kahler over 5 years ago in reply to Bryan Kahler

TI__Mastermind 25955 points

Hi Dig,

Still working to replicate this issue with your logs. Will update again tomorrow.

0 Bryan Kahler over 5 years ago in reply to Bryan Kahler

TI__Mastermind 25955 points

Hi Dig,

Still working to replicate this issue with your logs. Will update again tomorrow.

0 Bryan Kahler over 5 years ago in reply to Bryan Kahler

TI__Mastermind 25955 points

Hi Dig,

Still working through reproduction of the issue.

With the updated logs2019 file, do you have SREC files immediately prior to each test that correspond to the gauge settings utilized in the device prior to each test?

Sincerely,
Bryan Kahler

0 Dig Kleppe86 over 5 years ago in reply to Bryan Kahler

Prodigy 150 points

Hi Bryan

I don't have srec files prior to each test.

I have created one now.

310119-bq34z100G1.zip

0 Bryan Kahler over 5 years ago in reply to Dig Kleppe86

TI__Mastermind 25955 points

Hi Dig,

Thank you for the files - will try to reproduce the issue with this SREC. I will update this thread by Wednesday (2019-02-06).

Sincerely,
Bryan Kahler

0 Bryan Kahler over 5 years ago in reply to Bryan Kahler

TI__Mastermind 25955 points

Hi Dig,

I have put the log files together, but its not of much help because they are out of order and the state is not preserved between test/simulation

When logging, please use one log file. In that log please provide:

(1) 3 hour rest
(2) event of interest

The rest is so we capture an OCV point for the simulator. The event of interest to to compare results against (Remcap goes from 0-> 1001, then steps to 9000.)

Sincerely,
Bryan Kahler

0 Dig Kleppe86 over 5 years ago in reply to Bryan Kahler

Prodigy 150 points

Hello Bryan,

Attached the new log file.

Best regards,

Dig

080219.log

0 Bryan Kahler over 5 years ago in reply to Dig Kleppe86

TI__Mastermind 25955 points

Hi Dig,

In the log the current is 0 for the first 36-40 seconds, not 3 hours. Was the gauge allowed to rest for 3 hours prior to the start of the log?

Sincerely,
Bryan Kahler

0 Dig Kleppe86 over 5 years ago in reply to Bryan Kahler

Prodigy 150 points

Hi Bryan

The unit has been in rest for a whole night.

Dig

0 Dig Kleppe86 over 5 years ago in reply to Bryan Kahler

Prodigy 150 points

Hi Bryan
The unit has been resting for a night.
Dig

0 Bryan Kahler over 5 years ago in reply to Dig Kleppe86

TI__Mastermind 25955 points

Hi Dig,

Thank you for the update - will try again with the simulator and provide an update tomorrow.

Sincerely,
Bryan Kahler

0 Bryan Kahler over 5 years ago in reply to Bryan Kahler

TI__Mastermind 25955 points

Hi Dig,

Good news - have the new dll for the simulator. Simulation is now in progress - will be able to report the results on Monday.

Sincerely,
Bryan Kahler

0 Bryan Kahler over 5 years ago in reply to Bryan Kahler

TI__Mastermind 25955 points

Hi Dig,

Not getting the expected results with the simulator. Will update Tomorrow (Tuesday).

Sincerely,
Bryan Kahler

0 Bryan Kahler over 5 years ago in reply to Bryan Kahler

TI__Mastermind 25955 points

Hi Dig,

The Ra tables and Qmax values for the device have not been updated.

Please run through a learning cycle.

Update status needs to be updated from 0x04 to 0x06.

The gauge is simulating off of unlearned Ra tables - this is the core reason for the result.

Sincerely,
Bryan Kahler

0 Dig Kleppe86 over 5 years ago in reply to Bryan Kahler

Prodigy 150 points

Hi Bryan

I have run a number of cycles with a discharge current of 3A, and long rests between charge and discharge,

The learnes status stays 4.

Am I missing something?

Dig

0 Bryan Kahler over 5 years ago in reply to Dig Kleppe86

TI__Mastermind 25955 points

Hi Dig,

Here is the generic learning cycle guide with respect to lithium ion: www.ti.com/.../slua903.pdf

Since you are using lead acid, the learning will occur on the discharge, not on charge and discharge, so two full learning cycles will need to be run.

For the OCV update, please ensure that the cell fully relaxes. This relaxation time allows for a valid OCV reading to be taken and stored for the Qmax update. The
valid OCV reading occurs when the dV/dt of the battery is < 4 µV/s. The voltages do not need to be monitored, the gauge monitors for this condition and takes the OCV reading once met.

Section 2 and 3 discuss most gauge settings. If the improper chemID is used, learning may not complete.

1. Set charging current to at least C/5 rate but never more than C/2 rate
2. Charge the cell to full. For charge termination the current taper method is used,
a. During two consecutive periods of Current Taper Window, the AverageCurrent() is less
than Taper Current AND
b. During the same periods, the accumulated change in capacity > 0.25 mAh /Taper
Current Window AND
c. Voltage() is > Charging Voltage – Charging Taper Voltage. When this occurs, the
[CHG] bit of Flags() is cleared. Also, if the [RMFCC] bit of Pack Configuration is set, and
RemainingCapacity() is set equal to FullChargeCapacity().
4. Relax until the first DOD is taken
5. Set discharge current to at least C/5 rate but never more than C/2 rate.
6. Discharge the PbA cell until it is at least 10mV below terminate voltage
7. Relax again until the DOD is taken, this is indicated by the VOK flag being cleared.
8. At this point Qmax for the cell will be updated
9. Repeat the cycle again to get a subsequent Qmax and Ra table update

Sincerely,
Bryan Kahler

0 Dig Kleppe86 over 5 years ago in reply to Bryan Kahler

Prodigy 150 points

Hi Brian

My system refuses to learn. Learned status stays 4.

We have done more than 5 cycles following the steps in the guide, at different discharge currents, and more than 4 hours relaxing.

We noticed that in step 3c the RMFCC bit was not set, the other registers were updated. The Qmax is updated.

Can you explain step 4: relax until the first DOD is taken?

Is it OK to wait longer between step 4 and 5?

Do we use the right chemID? our battery was not in the standard list, the ID was advised by TI.

It is still unclear to me how a wrong Ra could cause too early zeroing SOC. If the RA is too high it would happen to late instead of too early.

Negative Ra? As you see in the last log , discharge at 2A the event happens at the right voltage of 28.8V.

When discharged at 4A this voltage is 42V.

Regards, Diglogs20-27-2.zip

0 Batt over 5 years ago in reply to Dig Kleppe86

TI__Mastermind 42535 points

Thanks for the logs. We will get back to you by next Wed.

0 Bryan Kahler over 5 years ago in reply to Batt

TI__Mastermind 25955 points

Hi Dig,

It is okay to wait longer between step 4 and 5. The cell is fully relaxed when the voltage stops increasing and has reached steady-state. We also refer to this as the unloaded voltage or the open circuit voltage (OCV). The DoD will be taken when a qualified OCV measurement is made. The valid OCV reading occurs when the dV/dt of the battery is < 4 µV/s

Please ensure the battery is fully charged, allowed to rest, then fully discharged and then allowed to rest.

These errors may be caused by an incorrect chemID. Please send me the information on your cell including datasheet as well as the archive you sent to the gpcchem tool for the chemId match.

Sincerely,
Bryan Kahler

0 Dig Kleppe86 over 5 years ago in reply to Bryan Kahler

Prodigy 150 points

Hi Bryan

You already have the datasheet of the battery, see my post from jan 8.

We have no gpcchem data available. The cells where characterized by TI.

Regards,

Dig

0 Bryan Kahler over 5 years ago in reply to Dig Kleppe86

TI__Mastermind 25955 points

Hi Dig,

Which chemID are you currently using?

The chemID given in the batt report listed above was 0801 with a 13.88% deviation. Is this correct?

Sincerely,
Bryan Kahler

0 Dig Kleppe86 over 5 years ago in reply to Bryan Kahler

Prodigy 150 points

Hi Bryan

It is 801.

Dig

0 Kang Kang over 5 years ago in reply to Dig Kleppe86

TI__Mastermind 24870 points

Hello Dig,

Why does the voltage not change for this cell during a discharge? It is really flat.

Thanks

0 Dig Kleppe86 over 5 years ago in reply to Kang Kang

Prodigy 150 points

Hi all

Because we don't get any result after 3 months we decided to stop with this device.

0 Bryan Kahler over 5 years ago in reply to Dig Kleppe86

TI__Mastermind 25955 points

Thank you for the update.

Power management

Power management forum

BQ34Z100-G1: SOC falls back to zero at 90%