LTC3330: LDO problem and workaround (but why?)

As part of our research project, our group was finally able to make our LTC3330 work (or so it seems). It took us a week or so. Let me explain:

This picture is from the datasheet, figure 7, with the second chip removed (an LTC3388) so it reflects our setup. With everything connected exactly as in the picture, Vout will generate 5V but LDO_OUT will generate a really nasty-chip-burning 4.2V or more with some weird waveform. After fiddling with everything that made sense we started to alter everything else until we changed the inductor between Vout and SW from 22uH to 100uH and, inexplicably, LDO_OUT started to generate something close to 3.3V. So I would like to know:

1 ) Does this make any sense? What is going on?

2) Will this workaround (or whatever the best denomination is) actually work and be stable? Is 100uH an acceptable value?

  • 0
    •  Analog Employees 
    on Jun 18, 2019 2:34 PM over 1 year ago

    Hello,

    1) Hard to say without plots, but I might guess VOUT was not stable. Did you try increasing the Vout cap before swapping out the inductor? Try increasing to 150uF with the original 22uH inductor. Does that help?

    2) 100uH should be fine. It will slow the switching and possibly limit your max current output. See the Inductor Selection section of the datasheet.

    Regards,

    Zack

  • Hi Zach, 

    Please excuse the crude illustration, I am not located where the setup is right now. This is what we were getting on the LDO_OUT pin, a wave of what seem to be unevenly-spaced constant-exponential-decay sawtooths with a 4.2V average or more with a duration in the order of one second or so. The peak and minimum values are also not repetitive.

    About 1) yes, we tried all sorts of capacitance on Vout and LDO_OUT and we also changed the SWA-SWB inductor (and selector bits)

    Regarding 2) OK, if that is all it will change/limit then the problem is solved, I will look into the tables and try to adjust the value to our needs.

    BIG THANKS ! bowzee

  • 0
    •  Analog Employees 
    on Jun 18, 2019 8:21 PM over 1 year ago in reply to bowzee

    Ok, great, though it would be good to understand why this was happening.

    Also, you may want to test with a 68uH inductor and see if that works. Again, there's no problem with 100uH, but perhaps you are on the edge of stable operation and a 100uH inductor with a +/- 20% tolerance may not always work - or it may - the only way to know is to test.

    I'm sending you a friend request, if you'd like to privately share your PCB layout with me, I would review it for you. Up to you.

    Regards,

    Zack

  • Hi Zack

    Again thanks for the helping hand! As of right now we are using this layout:

    which, impressively, is working... HAHAHA!!! We will develop it more from this starting point and make it look nicer (a lot nicer). But it is very kind of you to offer to help with the board, as soon as we run into another hurdle I will post or contact you.

  • 0
    •  Analog Employees 
    on Jun 19, 2019 6:22 PM over 1 year ago in reply to bowzee

    Haha oh wow. Yes, I'd guess that this issue will go away once you get this into a proper PCB layout. You can sort of get away with this for now due to how low-power this circuit is.

    Regards,

    Zack