DC2039A demo kit for LTC4015 - fire at the LTC4015 with 24V 72AH and 5ohms load

Hello,

 I have a trouble with the DC2039A demo board.

Really I made my PCB cancelling the USB support and the on board MCU and literally copying the gerbers coming with the Demo Kit.

 The load is made of a 24V 72AH SLA battery + a 5Ohm/200W resistor.

The battery is almost fully charged (by an external battery charger -  I am mean not charged in advance by the demo board).

I burned one board, the LTC4015 fired.

I left the chemistry switches opened and I set the CL0 1 2  to H.

Everything is OK when I power the unit by an external power supply; the battery is sinking a little (100-200mA) and the resistor is absorbing the right current (close 5A).

When I powered off the external Power Supply to try the auto-switch to the battery, not immediately, but within a couple of minutes the LT4115 literally fire close the pins IntVCC and DrVCC (top right corner).

Before the fire the current sink by the resistor was correct, almost the same when powered from the external power supply.

I am far away from the limits of the demo board (max 10A SYS current + max 8A BAT charge current, I am 5A and 200mA respectively) .

I read that it is better to power the DrVCC by an external dedicated LDO but, because I am far away from the absolute maximum ratings, I scare about a mistake somewhere else.

The MOSFETs were always cool, no any hot part but the LTC4015.

Any suggestion? Where is my mistake?

  • 0
    •  Analog Employees 
    on Oct 16, 2018 9:28 PM

    Hello,

    What event exactly caused this to happen? Was it upon connecting the battery? Or upon connecting the power supply? Or steady-state operation?

    I'm curious where exactly the damage to the IC is. Can you provide a photo?

    Inrush issues can occur with this circuit if a battery or power supply is hot-swapped due to the large amount of ceramic capacitance on the SYS node. I attached a guide on this thread that may help you if this is indeed the issue. If so, we have fixes for that.

    Regards,

    Zack

  • Hello Zack, thank you to help me.

    Attached the picture about the trouble I experimented.

    I confirm everything happened when I hot swapped a fully charged battery, 24V 36AH.

    The inrush current should be really the origin of the problem.

    Why don't use a NTC thermistor inrush current limiter?

    I use Ameterm NTC to limit the inrush current in big toroids rated 1000VA, the inrush is sometimes higher than 150-200A.

    Is there an hidden problem with the NTC? The calculation of the right model is not "simple" because it involves with joules. 

    I don't want to burn another assembled board ... I would like to decrease the risk as much as possible.

    I have already burnt 2 boards, I have another 3 left to experiment.

    LTC4015 cooked 

    The scratches you see on the top of the chip are cause I tried to desolder without success. The board was assembled by machine and by RoHS chemical so now it is very hard to disassembly.

    99% of times the batteries, when installed for the first time, are charged at 70-80% . So the hot swap is the every day situation. I scare about the strenght of the chip. Can an inrush limiter really solve the problem?    

    By the IRF540N is really possibe to limit an inrush current of more than 100A?

    I have two 24V 36AH battery able of a very high initial current, higher than 100A.

  • 0
    •  Analog Employees 
    on Oct 17, 2018 9:19 PM in reply to revenac

    Hello,

    Thank you for the picture.

    So, I see two burn marks on the IC. The lower-right one is what I would expect from a battery hot-swapping issue: that can happen if you blow the diode between the CSN/CSP pins due to inrush exceeding the differential abs max of those two pins. Since this happened on the battery connection event, I'll assume for now that the top-right burn marks were a result of this damage to the IC.

    The limiting thermistor could work (I assume it is actually a PTC so that the resistance increases with a higher temperature). The down-side is that it is a significant power drain. At 300mOhms, you'll be burning 3W with 10A discharge current.

    To prevent burning another board: apply VIN first. This will charge the SYS caps higher than the battery voltage, so you will not have any inrush through that destructive path. Also, use short leads between your battery and the board during connection (long leads have parasitic inductance that may cause voltage transients upon hot-swapping). Do this until you are inrush-protected and ready to test the inrush - at which point you should put the circuit into worst-case inrush conditions: VIN is missing and SYS caps completely discharged, then hammer on the battery a bunch of times. There is more detail in that document that I linked.

    Hot swap is common in these apps, it just needs to be handled appropriately. The LTC4015 is very strong in its specified operating conditions, but inrush is a common way to violate the sense input abs max differential voltage and blow up the chip. So, if the inrush limiting is designed to keep the circuit within the right operating conditions, the LTC4015 will be just fine every time.

    What is your plan with IRF540N? Is it to be used as an inrush limiting FET (as shown in the document)? Note that if you limit the inrush, you will not need to worry about operation at 100A. If you are seeing spikes of 100A now, that peak will not be hit once the inrush is limited with a slow FET turn-on.

    Regards,

    Zack

  • Hello Zack,

    mounted 2 IRF540N inrush limiters. One to liitthe inrush from the power supply and one to limit the inrush from the battery.

    EXPERIMENT #1

    I applied first the power supply voltage and plaied with no any strange result.

    I connected and reconnected many times the battery - 24V 36AH - and the circuit worked properly; the LTC4015 charged the battery with 3A and supplied the load (only150mA at the load by a 2K resistor).

    Every time I connected the battery there was a spike at the connector. I played withthis spikes and the circuit worked properly without any problem.

    No Inrush problems in this condition.

    EXPERIMENT #2

    I disconnected the power supply and the battery; waited 5 minutes so to be sure about a zero condition to start with.

    I connected the BATTERY ONLY, leaving the power supply disconnected. BOOM. Instantly I fired another circuit.

    I attach a picture. The right bottom corner of the LTC4015 has a hole.

    Help help help.

    Mario