LT8390 - input related voltage drop for low loads in boost region

Hi all, 

im using the LT8390 in boost region (input around 20-30V) to generate a output voltage of 32.8V with 1.7A. Reason for that voltage is, that I use the IC as a battery charger. That states, that of course in constant current mode, the voltage is rising and in constant voltage charging mode, the current is decreasing. During the decreasing current phase, the voltage should be at its max setting of 32.8V.

Unfortunately, when the load current decreases, a voltage drop around 40-70mV appears and this behaviour is hardly dependend on the input voltage. Means, that when I use 24V, the voltage drop happens from 32800mV to 32750mV at a load current of 130mA. When I use 22 oder 26V, the drop happens at slighly other load currents and may also be 30mV or 60mV.

When it comes to load and input dependent voltage drops, I tried to clean the feedback path with small capacitances but that worked not well enough. 100pF up to 1nF were used and only had small effect. 

I also tried to increase the output capacitance of the boost IC, which also had a minor effect.

The main reason, why this is a problem, is that the charging current at this phase also drops to nearly zero and therefore I will not be able to charge properly. Id like the voltage to be stable down to small loads of 40-50mA.

Im using a circuit really similar to the reference design as mentioned in the datasheets. The Layout is also similar to the LT journal article https://www.analog.com/media/en/technical-documentation/tech-articles/lt-journal-article/LTJournal-V26N4-00-df-LT8390-DawsonHuang.pdf

Id like to know, where I can start with to get rid of this problem and help and ideas would be highly appreciated.

  • 0
    •  Analog Employees 
    on Oct 5, 2020 6:01 AM 2 months ago

    Hi 

    I think you are describing the natural response of LT8390 at light loads (page 16 of the data sheet). 

    Please confirm if this is indeed what you are seeing at light loads. 

    If possible please take a snap shot of the switch nodes to verify pulse skip operation

    Regards,

  • Hi, 

    thanks for your answer. It looks like you are right, the IC seems to be in the light load region. So it is well noted behaviour indeed. I can see to pulse skip operation when the load decreases below 140mA. This is exaclty the range, where I am able to see the voltage drop. 

    So these pulse skips are triggered by reverse voltage events (1mV over shunt resistance). Is it possible to shift the area of light load operation, since I have to get rid of voltage drops in that area. Light load operation starting at 50mA would be ok, but in a range of 130mA+ this works not well for my purpose. Since the light use are is defined by voltage over resistance it might be shifted but for now I was not able to do so.

    Thanks again and best regards, 

    Marc

  • 0
    •  Analog Employees 
    on Oct 12, 2020 6:52 AM 1 month ago in reply to Heinemarc

    Hi Marc,

    You may want to consider inductance value as this changes the inductor current ripple. Please use LTspice to check this. You may also want to consider battery charger products ideal for charging different battery chemistry.

  • Hi Fil, thanks for your quick answer.

    I already tried to simulate the behaviour, but due to complexity of the LT model, the simulation is very very small and more or less not useable in practice. 

    Battery charger products with my requirements are not available on the marked (SMBus communication according to SBS 1.1, 8s Li Ion, 100W, ...)

    Some boards of mine show these light load pulse skipping, while others dont (at 130mA load). Why is that the case? Is it because of the tolerance of the 1mV reverse current detection?

    How can I reduce the reverse current through the shunt to avoid the pulse skipping operation (besides the already mentioned inductance changes)? 

    Thank you in advance!

  • 0
    •  Analog Employees 
    on Oct 29, 2020 9:44 AM 1 month ago in reply to Heinemarc

    Hi Marc,

    At this point, you may need local FAE support to further look at your application. Its hard to debug with the minor information you got. You are using this as a batter charger. I'm not so sure if how is that going to work. If you can provide me with project info and estimate annual usage, I will be glad to connect you to one.

    It is also better that you have one of our demo boards to play around with.