This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

bq51050b faulty transition from fast charge to taper mode

Other Parts Discussed in Thread: BQ51050B

Hi I have recently received some sample BQ51050b wireless power rxr ics. I have built a couple of prototype boards.

I am using a couple of bought in QI compliant power transmitter chargers to test the boards. The boards operate properly in fast charge mode and in current regulation mode. There however seems to be an issue when transiting from the fast charge mode to the taper mode. The power output oscillates and causes the transmitter chargers to beep.

Please see the following screen shots:-

 2766.FaultyTransitionFromFastChargeToTaper.tiff

CH1 shows Vrect, CH2 shows VBat

The same is shown on a longer 500ms time base here :-

1780.FaultyTransitionFromFastChargeToTaper_TB500ms.tiff

I also checked the TS/CTRL line :-

5417.TS_CTRL op during FaultyFastChargeToTaper_TB500ms.tiff

The fast charge current is set to around 200mA and was measured as 218mA. The current termination setting is set to approx. 10% and during testing this was 28mA.

So every thing seems to be fine apart from the transition from fast charge to taper mode.

The circuit is shown below

Any advice on what might be the problem, would gratefully be received

Best Regards,

Gareth 

  • Sorry circuit image missing from original post.

  • Gareth,

    I have a few basic questions to start.  Can you tell me the values of:  L1, C1, C2, resistance from ILIM to FOD, resistance from TERM to ground and any information you can about R6?

    Are you using a battery or battery emulator to measure the current?  Any information on that would be helpful.  Typically, I use a source meter to emulate the battery and change the source meter voltage to control the transition between modes.

    Can you get the plots of the battery voltage and battery current during the shutdown points?  Or, at least the voltage and current levels before the trip and during the 500ms when the RECT is low.

    Regards,

    Dick

  • Hello Dick,

    Thank you for your response.

    The L1 value is 12uH,  C1 is 190nF,  C2 is 2nF

    Resistance from ILim to FOD is 1.245 kohms

    Resistance from Term to ground is 2.42 kohms 

    For R6 I am using a 10k NTC thermistor10k, 3435K Beta Value . Please see following link for further details

    http://www.rapidonline.com/electronic-components/10-0k-precision-ntc-thermistor-61-0515

    I am not using the thermistor in 'anger' at this point. That is, it is not next to anything that gets warm/hot and is in free air.

    The issue happened when charging real batteries so I built a battery emulator based on the one shown in one of your data sheets, to test further. The plots given in the initial post were obtained using the battery emulator.

    The sense resistor on the battery emulator was measured at 0.1ohms. Here is the plot of IBat and Vbat at the point the charge oscillates on and off.

    8407.IBat_FaultyFastChargeToTaper.tiff

    CH1 shows Ibat (volts across 0.1ohm), CH2 shows VBat.

    Please let me know if you need any more info.

    Best Regards,

    Gareth

  • Gareth,

    Thank you for the information and plot.  All the numbers agree with your calculations (as expected).

    The current is quite erratic.  Is there anything except the battery emulator on the output?

    Take a look at the electrical characteristics, top of page 6, spec IBAT in the Synchronous Rectifier section.  There is discussion related to this spec on the top of page 22 under Synchronous Rectifier.  I'm concerned that what may be happening due to the noise on the BAT pin that the system is switching between half and full synchronous rectifier modes.  

    Any chance of sending a JPG of the board layout?

    Can you send a scope shot of the system operating in normal fast charge mode?

    Regards,

    Dick

  • Hello Dick,

    Here is the Fast charge current plot showing IBat (CH1) and VBat (CH2).

    0410.IBat_FastCharge.tiff

    The circuit is laid out on vero-board. The BQ51050 is on an adapter board that plugs into the vero board. I know vero board is rubbish but unfortunately I have to demonstrate something working before we are allowed to go to pcb.

    I think your concerns about the noise on the IBat plots and the synchronous rectifier modes has pinpointed the issue I am having. So as long as the noise on IBAT, when on a proper board, is below 25mA (synchronous rectifier mode hysteresis) the on off cycling we are seeing should disappear?

    Many thanks for your help in this matter,

    Gareth

     

  • Gareth,

    Wow, that is really noisy for just an emulator.

    Something you might try to prove the issue is increasing the charge current with your emulator.  If you move the charge current to 400mA and verify the transition to taper that should confirm our belief.  You could still very termination as a percent of the 400mA.  

    Based on the vero board, I would also be concerned about the communication from the TX to the RX.  Take a look at the signal on AC1 on the scope.  The communication occurs about every 250ms.  It will show up on the AC1 line.  If the TX can't interpret it correctly, it will terminate charge and retry.

    Regards,

    Dick

  •  Hello Dick,

    I have tried the fast charge at 400mA. Please see attached plot. It seems to show the charging is cycling on and off before getting to the synchronous rectification mode transition point.

    6685.IBat_OnOffCycles_400mAFastCharge.tiff

    With your mention of the communications being a possible source of failure I have monitored this as well during the on off cycling. I did this my attaching a diode envelope detector /data slicer cct to the comm1 pin and then bit banged the output to obtain the data packet. Some of the detected packets are shown below

     04 30 34
     03 00 03
     03 01 02
     03 00 03
     03 FB F8
     03 F8 FB
     03 F8 FB
     03 F6 F5
     04 2A 2E
     03 F7 F4
     03 F6 F5
     03 F6 F5
     03 F5 F6
     01 58 59
     71 17 00 10 00 32 EA 09 A1
     51 0A 00 00 43 00 18
     03 16 15
     03 12 11
     03 0F OC
     03 0C 0F
     03 09 0A

    I hope this info is of use. I think the first few packets were received whilst the charge is stable and then the proceeding packets show the lead up to the first off/on cycle.

    Regards ,

    Gareth

  • Dick,

    Further to your suggestion about monitoring the AC1 line, I have attached a few plots at various stages.

    For when rxr is attached to the battery emulator but the battery voltage is turned off:-

    4130.AC1commsNoBatteryVoltage.tiff

    1732.AC1commsNoBatteryVoltage_TB250ms.tiff

    For when the rxr is in fast charge mode (~400mA):-

    4336.AC1commsFastCharge400mA.tiff

    For when the rxr is cycling on and off:-

    2061.AC1commsDuringFailure.tiff

    I think this shows that there is a lot of noise on the comms when charging. The  spike every 2.5ms during the comms packet seems to be what is causing the charging to fail. Do you have any idea what could be causing the 2.5ms  pulse? They only seem to show up during the comms packet

    Best Regards,

    Gareth

  • Gareth,

    I will have to decode the packets you included in the previous post.  Generally, I use a "Qi Sniffer" from AVID-Tech which decodes the communication.  

    If there are 64 packets missed, the TX will recycle the power transmission - and that could cause what you're seeing.

    Capturing readable comm packets is difficult,  My hat is off to you for the results.

    The 250ms/division plot shows the packets being sent very nicely.  Pretty regularly at 250ms intervals - wtih a few thrown in quick possibly responding to current changes.  

    I will do some more checking to see where the 2.5ms periodic signal comes from.

    Regards,

    Dick

  • Dick,

    I realised that I did not look at the AC1 line comms on a 250ms timebase yesterday when the charging was failing. Have done that today and I have attached the plot.

    Interestingly it shows that the comms is sent every 60ms or so. Taking on board your mention of the tx recycling power if 64packets are missed, this gives a recycle period of 0.060 *64 = 3.84secs which is close to the sort of period of charge failure shown in the earlier plots.

    I think that this is very good evidence that the problem is noisy comms. Most likely due to building the cct on vero board.

    With regards to the data packets, It does show a power recycle with the packets starting with

    01 58 59
    71 17 00 10 00 32 EA 09 A1
    51 0A 00 00 43 00

    This is the signal strength packet followed by an Identification packet and then a configuration packet.

    Regards,

    Gareth

  • Gareth,

    We have not communicated in some time.  I am check in to see how your wireless power project is coming along.  Let me know if I can assist.

    Regards,

    Dick

  • Hi Dick,

    Sorry for delay in responding. I am waiting for a proper pcb to be populated with components, (rather than breadboard). Unfortunately although we have the parts I am finding it difficult to get it allocated some time to put it all together.

    I am sure with a proper pcb, the issue mentioned above will disappear.

    Best Regards,

    Gareth

  • Gareth,

    Thanks for the update.  Let me know if I can be of any assistance in the future.

    Regards,

    Dick

  • Hello Dick,

    We have finally got a board made up and there are no issues at all. So essentially my prototype boards were too noisy, which caused the issue. The circuit built on a proper pcb works really well.

    Thank you for your help all that time ago.

    Best Regards,
    Gareth

  • Gareth,

    Congratulations on the great news! Thank you for following up with the resolution.

    Regards,

    Dick