BQ24278 CHG LED Odd Behavior

A new charge cycle starts when input power is removed and then replaced.  Are we certain that the battery voltage as measured at the IC's BAT pin does not indicate a need for charging for those 20 minutes?  What is the charge current and related termination current?  If high, then it is possible that charging starts again for those 20 minutes.

The charging current is set to 1.8A and the termination current is set to 0.15 A.  The battery is known to have full charge based off the CHG LED output.  Can a new charge cycle keep the CHG LED on even though the charging current is at 0.0A? 

No.  If charge is done, the CHG LED should go off.

Hi,

I am trying to use BQ24278 as a bay charger (standalone). I use your eval board to verify the function.

The board is supplied from a 5V supply. Nothing (except as on th eeval board) is connected to the System output.

 I get odd behaviour when I remove the battery.  The data sheet claims that whenever the battery is removed (along with the NTC resistor) the CHG output should go high impedance. This is sometimes true and when it happens the BAT output toggles between 0 and 3.5V.

My problem is that most of the time the CHG output does not go high impedance, it is still set to low (charging), when I remove the battery. And now the BAT output is set to approx 1V.

If I set the TS input to 2.0V all the time the CHG always goes high impedance when I remove the battery. In this case I do not get temperature surveillance of th eBattery, which I want.

Any Ideas on how to solve this?

How are you connecting the battery's thermistor?   You will need to remove JP4 and connect the thermistor to J7.  You will also need to change the EVM's R4 and R5 resistors for a 10kohm thermistor as the resistors on the board are not for 10kohm thermistor. 

Then, the /CHG pin should go low during the new charge cycle (following battery attach or VSUPPLY POR) and will stay low if TS indicates a battery temp fault even if the battery is removed.   However, if the battery is removed during a recharge cycle and TS is indicating a battery temp fault, /CHG will stay high impedance. 

I am connecting a 103AT thermistor between TS and ground. Resistor between DRV and TS is set to 5k6 and resistor between TS and ground is set to 12k.

The thermistor is included in the battery pack and is disconnected when the battery is removed. VSUPPLY is normally always connected. I thought that the battery detection circuitry described on page 12 would set /CHG pin high impedance regardless of TS pin voltage.

Is your second part about line two in Table 2 "Charging suspended by thermal loop"? It kind of complicates things for me. It should be overruled by "No battery present"

A complicated way to come around this issue would be to measure the BAT voltage and when it is not toggling between 0 and 3.5V (stays at 1V) I connect a resistor between TS and ground  (8k2) to keep TS voltage in "correct" temperature range.

Or should I select a different (more expensive) device such as BQ24104?

By the way VDPM pin is missing i the block diagram on page 6.

Did you ever get an answer to your question on TS behavior.  I have exactly your set up for the TS resistors (this allows 10K external NTC) and see this behavior

1) When VSUPPLY is connected with NO battery, CHG pin is LO (Should be tri-stated)

2) When the battry is removed during charge cycle CHG pin remains LO (should be tri-stated)

Also, I see ~0.6V on the battery pin when NO battery is connected.

Regards,

If termination is enabled, when no battery is present, there should be a square-ish wave at the BAT pin. 

The CHG pin never tri-states.  It is either lo impedance or high impedance.  When there is no battery attached to the BAT pin, the CHG pin should go high impedance if termination is enabled and there is no dc load attached to the BAT pin that is greater than the termination current.

Hi Jeff,

Thanks so much for your replay.  With this device there is NO explicit Termination Enable/Disable pin.  Termination condition is automatically detected.

This is only 1 case where I see CHG pin not behaving properly.  in ALL other cases the device seems to work exactly as specified in my application.  It charges, terminates, and both LEDS function accordingly.

The ONLY one case that CHG has unexpected behavior:

NO battery connected in the system, both CD and CE pins are pulled down with 10K to GND, both CHG and PG are pulled up via 10K resistor, and TS pin is configured for a 10K Thermistor type NTC (5.6K to DRV pin and 11K to GND).  

I power the system up via 5V at IN pins.  PG goes low Z and CHG pin goes Low z.  I see approximately 0.6V at BAT pin and NO square wave for battery detection.

NOW, if I first connect the battery and then power up the IN pins to 5V I see PG Low-Z and CHG low-Z .  The battery is still not fully charged and CHG is low-z as expected.  If I REMOVE the battery at this point PG stays low-z, CHG goes Hi-Z AND I see the battery detection square wave at the BAT pin.  Works great!

I am wondering if I should toggle CE and/or CD pin at power up to properly initiate internal state machines and this could be the reason for this behavior?

regards, Sadry

This part was designed to have the battery installed first and then apply/remove input power.  If you install power first, you need to toggle the CD pin (enter/exit HiZ mode) to reset the state machine.

Thanks Jeff.

Unfortunately my CPU is powered by the SYS pin and I cannot toggle CD but I can toggle CE.  I tried this but it did not resolve my power up CHG=LOW-Z with NO battery connected.

Any other remedies you can think of?

Best,

Sadry

I have seen this ODD behavior with the CHG pin too in this way.

1) connect a charging voltage to IN pin

2) Lipo Battery is connected and needs to be charged.

3) Load is requesting current from SYS pin at the same time

The charger provides the requested total current (way below ISET and ILIM) and charges the battery at the same time while providing the load current.  If I leave the set up like this for a long time, then the battery is fully charged (CHG goes hi-impedance) and the load current continues to be provided from the IN source.  All works great!

NOW at this point if:

1) If I turn off the load and then remove the power supply from IN and reconnect the supply, the CHG pin is LOW-Z FOREVER although battery is FULLY charged.

2) They only way I can get the charger working properly again is to remove the charger from the system, use the battery to supply current to the load for a short time and then reconnect the charger.  It will charge the battery and CHG will go Hi-z again after a while.

Looks to me if the charger is charging the battery while providing current to load, it may slightly overcharge the battery and then it subsequently gets confused when a new cycle is started with this fully charged battery until the battery voltage is reduced.

Any ideas?

thanks,

Below is a scope plot with the bq24278 EVM tied to a 20 ohm load, 5VIN and a sourcemeter simulating a battery.

When I repeat your steps above, I see the /CHG pin properly change stage.  If the battery reaches full charge and then bleeds off and enters a recharge cycle (within VIN being removed and replaced in between), the /CHG does not change state because it is designed not to change state for recharge cycles.

Hi Jeff,

Thanks for quick reply!

Just to make sure I understand the picture.  Yellow is Voltage applied to IN pin.  what is the green label? Is the charge voltage or current?

1) 1st low pulse on CHG - Why does CHG go LO-Z for the duration Vin is removed? 

2) 2nd and 3rd low pulses on CHG - Again CHG goes lo-z when IN voltage is removed why (2nd pulse)?  I understand the 3rd CHG pulse as it seems like we are doing some charging for a short time.

I fully understand the CHG does not go lo-z while re-charging but the problem I am trying to explain is different.

The case I am explaining (not necessarily repeatable every time) is this.  You fully charge the battery and CHG goes HI-Z.  Then you remove source on IN pin and apply again.  The CHG pin remains LO-Z forever and the current into IN pin is almost 0.  This is sort of what the thread starter also saw but in my case the CHG get stuck in lo-z

It happens that I saw this condition where I had a load running while I left the charging voltage running over night.  The charger indicated fully charged in the morning (CHG Hi-Z).  Then I disconnected and reconnect the voltage source at IN pin and CHG pin went to LO-Z and stayed there indefinitely.  I have also seen this very exact behavior when I have FULLY charged the battery with another CC/CV LI-PO charger and then connected to BQ24278 charger.  I see almost 0 charge current and CHG stuck at LO-Z.  Another time was when I fully charged the battery and then attempted to charge (top off) a couple of more times but removing the source at IN and re-applying.

The only way out of this is to drain the battery for even a short time via load and then try a charge cycle.  It seems like maybe there is a case where when the battery is fully charged (or maybe if OC BAT voltage a bit too high) the charger thinks it needs more charging (CHG goes lo-z) but there is actually no charge current provided.  This always happens when the battery is fully charged first.

thanks, Sadry

Yellow is V(IN).  Green is input current into V(IN).  

1.  On the EVM, /CHG is pulled up to DRV.  DRV goes away when V(IN) is removed.

2.  Same as above.

To what voltage do you have /CHG pulled up?  IN or BAT?

What you describe could happen if termination is disabled, which only happens if one of the other control loops (VINDPM, DPPM, thermal) are in control.  Given the low charge current at termination, DPPM and thermal are unlikely.  What is your voltage at the IC's VIN pin when this happens?

How quickly are you removing and replacing the power supply on IN?  The battery detect circuit needs to complete one full cycle (250ms) in order to determine that the battery is fully charged.

BTW also the datasheet says the following.  I assume this discharging is done by BQ24278.  Also, during such condition will CHG pin be lo-z?  Even if that is the case, my situation is a bit different such that the CHG pin never goes to Hi-Z until I actually take some current out of the battery and then try to charge again

Beat regards,

"If the battery voltage is ever greater than VBAT(REG), the PWM converter is turned off and the battery is discharged to VBAT(REG). This prevents further overcharging the battery and allows the battery to discharge to safe operating levels."

No, the discharging is expected to come from the system load. 

Is your load attached to SYS or BAT?

OK I understand the behavior in the picture now.

"To what voltage do you have /CHG pulled up?  IN or BAT?"

>>In my setup the CHG pin is pulled up to SYS pins so they can be driven by either the battery or the charger's regulator output.

" What is your voltage at the IC's VIN pin when this happens?"

>> ~5.4v.  and VINDPM is set for 4.7V and I agree that the thermal loop is not active here.  This is on a test setup on the bench and nothing is even warm.

"How quickly are you removing and replacing the power supply on IN?"

>> It is done manually by hand as it is actually a micro usb connector.  so say 5 seconds or more

The load is connected to SYS pins and it is actually a boost converter so it is only running when converter is on.  So, what happens if the OC battery voltage is a bit too high?  what does BQ24278 do wrsp to CHG pin?

I cannot actually measure this right now as I am not in the lab but i left the battery on the charger to see if the CHG stops going lo-z.  I can check the battery voltage later.

Anything you want me to check?

Best regards,

What is the voltage on TS during your testing?

Good point!

It is ~2.3V with battery attached.  Is that an issue?  

I did notice that it does NOT have the actual values that are calculated based on my actual NTC (10k at 25c, Rhot=2.88k, Rcold=29.57k).  Looking at the datesheet and the EVM more closely.  I notice this.

from datasheet

Vcold = 60% VDRV (5.2v typ) = 3.12v

Vhot = 30% VDRV = 1.56V

From Figure 4 of DS I looks like charging is DISABLED if the voltage at TS is GREATER than Vhot and LESS than Vcold.  so disabled if TS voltage is between 1.56 and 3.12 volts

1) In my test case the Vntc is 2.3V and it looks like it should disable and elongate the charge cycle, correct?

2) Looking at EVM they use RHI=1.87K and RLO=4.12K (it says for 10K NTC).  This will result in a TS voltage of 3.16V typical.  This is just 40mV above the Vcold threshold.  Is this correct? looks too close to me

3) My NTC datasheet will calculate RHI=5.34K, RLO=10.92K.  With battery at 25C (NTC=10K) this will result in 2.56V and will create a problem based on what I see in the datasheet figure 4

Am I missing something in the calculations?

thanks, Sadry

V(TS) = 2.3V should be okay.  The EVM resistor values are not for a 10k thermistor.

Thanks Jeff,

Can you please look at my analysis in previous post.  I looks to me that TS voltages between 1.56 and 3.12V should delay charging and the values in EVM will fall in this range with 1.87K up and 4.12K down and a 10K NTC @ 25c.

Best regards,

Charging is disabled if V(TS) < VHOT = 30%*V(DRV) or > VCOLD = 60%*V(DRV)

Charging is reduce by 1/2 and termination is disabled if  VCOLD>V(TS)>VCOOL where VCOOL is 56.5%V(DRV).

VBATREG is reduced to 4.06V if VHOT<V(TS)<VWARM where VWARM =38.3%*V(DRV).

Thanks again!  

Please follow below and tell me what is wrong

"Charging is disabled if V(TS) < VHOT = 30%*V(DRV) or > VCOLD = 60%*V(DRV)"

So VDRV=5.2V typical.

VHOT = 1.56V

VCOLD = 3.12V

1) From the below figure from DS, charging is DISABLED when V(TS) > 30% V(DRV) or V(TS) < 60%V(DRV) if that is an OR gate with logical one for DISABLE function.  That is the opposite of your statement.

2) In any case, the EVM has 1.87K/4.12K resistor divider for a 10K NTC battery resistor.  This ends up being a 3.16V TS voltage when V(DRV)=5.2V typical.  TOO close to COLD threshold and by your definition above charging will be disabled at 25C ?!

Regarding 1, the figure is incorrect. The text below it is correct. 

Regarding 2, the EVM resistors are not for a 10k thermistor and will result in V(TS) being in the COOL to COLD region.  This means termination will be disabled and /CHG will not go HiZ even if the battery voltage reaches regulation.

AHah NOW it all makes sense!!!!

BTW, the block diagram in page 6 of 2012 Datasheet is also then incorrect.

Also, wrsp to 2nd comment above, are you saying that the EVM voltage divider BY ITSELF will result in V(TS) of COLD to COOL region or in conjunction with a 10K external NTC resistor?

By themselves they have a V(TS) = 3.57V and with an external 10K thermistor they will yield 3.16V.  Both of these values are above both VCOOL and VCOLD.  The one with Thermistor is just above VCOLD.

If you attach a 10k thermistor/resistor to the EVM, you will see 3.16V which is outside the COOL/COLD region.  You can change the shunt on the jumper to connect the on EVM potentiometer and set V(TS) to ~ 2.5V which corresponds to 25C operation.

OK Jeff, I was mainly commenting on your statement a couple post a go.

This is NOW clear for me and thank you for your support.  Although, I did not find out why it seems like the charge cycle never completes with a fully charged battery, I now better understand the thermal loop control in BQ24278.

PLEASE have some sort of errata issued regarding the sense of the thermal loop detection pictures so others like me don't get confused as to the functionality.

Best,

Sadry

I have a pressing issue in production with BQ24278.  Basically, when you connected a FULLY charged Li-Polymer battery to the charger, the CHG pin remains low for many hours (perhaps until the timer kicks in.  Have not verified but could be 6 hours) before going Hi-Z.  Here the info.

1) Fully charged battery voltage at IC pins ~4.17v

2) Current entering the IN pin ~15mA.  The voltage is 5.02v.

3) Battery has NTC resistor and the voltage at TS is ~2.54v with resistor dividers.  TA=25

4) The CD = LOW and CE/ = LOW pins are not toggled and hard connected

I REALLY need to understand what is happening here since this was an issue seen by this thread's starter and I also saw it during development in 2014.

Thanks for your help,

Sadry

Thanks.  The problem occurs like this.

1) Battery is connected and never removed and is fully charged

2) CD and CE/ are pulled down.

3) power is being applied or removed from the IN and CHG pin stays low

If I take a bit of current from the battery by turning on load and then try to charge again everything works fine until battery if full again.  At this point if I remove power from IN pin (with full battery) and apply again I can get the CHG pin asserted low for very long time.  This might not occur the first time while battery is full and may take a couple of tries.

I cannot assert/deassert DC since I have no idea when this condition can occur in real product application.

Hi Jeff,

Do you have any other thoughts here?

1) EXACTLY what are the conditions that can result in PG and CHG staying LOW on a fully charged battery from a fresh charge cycle (via external power being applied to IN) until the 6 hr timer expires.  

I only see ~15mA of current going into the IN signal (5v supply) and there is NO load connected to battery.  

2) The battery voltage does not exceed 4.17 voltage while charging, NTC is connected and the voltage is 2.54v.

Thanks!

Thanks.  That is exactly how I reproduce as well.  

Depending on the 1-cell Li-Po battery I use I can make it re-occur repeatedly.  For example it happens with a fully charged 6000mAh battery all the time where it does not with a fully charged 1100mAh battery.  I have measured the OC battery voltage of both and they seem to be the same within 20mV.

Thanks for following up on this.  We have 2K boards made with BQ24278 and need to have some sort of a solution (SW based?) to detect end of charge properly.

A few questions

1) The condition can occur when VBAT > VBATREG-VRCH.  Isn't this the case whenever a battery is fully charged using the charger?  If yes, then this device has a fundamental operational flaw, NO?

2) The spec is confusing.  Why two values for TA=25c

3) You mention IC needs to be power cycled when CHG latches.  Does removing power from IN pin and reapplying again with battery connected constitute a "power Cycle"

4) Can this be fixed by toggling the CD or CE pin?

5) Will the CHG pin unlatch after 6 hour timer expires?

6) Can you think of a way to be able to intelligently charge the battery and determine a full charge without relying on CHG pin.  I have the PG, CHG, CD, CE pins going to a micro controller that drives the charge LED.

Thanks!

Hi Jeff,

I appreciate it if you could reply to the 6 questions I asked.  

Based on your comment on the CHG latch-up, it seems like the condition can occur even during a normal charge cycle when the IC is trying to charge the battery to VBATREG (as soon as the battery voltage is greater than VBATREG-VRCH)?