• Resolved

LMG5200: Switches short when using 2 devices cascaded

Part Number: LMG5200

Hi,

I am using two LMG5200 cascaded (Q1+Q2 is one package and Q3+Q4 is the other package) as can be seen in the schematics below. Here Q1 and Q3 are switched at the same time, while Q2 and Q4 are also switched at the same time. The Q1+Q3 pair and Q2+Q4 pair are never switched at the same time.

For my test I was switching at 100kHz, with around 2µs between Q1+Q3 and Q2+Q4. However, I saw that the resonant tank (Cr+Lr) was being discharged when Q2+Q4 are switched on. It seems that Q2 and Q3 are conducting at the same time, while the control signal for Q2 is high and the control signal for Q3 is low.

Does anyone have any idea what might go wrong, why Q2 and Q3 seem to short the resonant tank?

I have used a 2-channel isolator for the HI and LI signals of the upper LMG5200 with a 5V supply with respect to the main input voltage.

Used control signals:

Thanks,

Wouter

  • Hi Wouter,

    there are a few things that I would need to understand to give you a good answer:

    1.Is there a load connected to Vout?

    2.what does the Vout voltage look like during the transitions?

    3.If you apply a voltage source directly across the tank with everything else unpowered how much leakage current to you get (I would expect a lot less than 1mA) - also check the integrity of the capacitor.

    4.can you check the off state Rdson for Q3 and Q3 these should be several 100k ohm.

    If all the above look good there might be something associate with the actual operation of the circuit and might need further information.

    If at any time you don't feel comfortable sharing information on the e2e we can carry the exchange privately.

    Thank you and best regards,

    Alberto

  • In reply to Alberto Doronzo:

    Hi Alberto,

    1. Yes, a 4.7kΩ resistor

    2. In blue the output voltage, in green the voltage over the resonant tank, in yellow the current through the tank and in red the control signal of Q1.

    3. At 5V over the tank I measured <1µA

    4. I measured the off-state Rdson for all switches at >1MΩ

  • In reply to Wouter Krikken:

    Hi Wouter,

    everything seems to be pointing to a health circuit.

    One extra test that you can do to ensure the the LMG5200 are operating correctly is verifying that all the FETs can turn ON properly. 

    Make all Hi and LI high-level (don't power the bus voltage!) and check that the chain resistance is tens of  milli-ohms.

    If that checks ok, then the components are fine, and there has to be something to do with the values or operation of the circuit.

    In that case to further assist you I would need a schematic to read the correct values of components.

    Thank you.

    Best regards,

    Alberto

  • In reply to Alberto Doronzo:

    Hi Alberto,


    I measured them and saw that Q1-Q3 work properly, but Q4 did not switched. I had not expected this, since the output voltage was at an expected level.

    I replaced the broken LMG5200, now all switches work. Moreover, the tank current shows a resonant shape as expected. This seems to have solved my issued.
    I only had time to do some preliminary measurements, but those look good.

    Thank you a lot for your time and suggestions.


    Kind regards,

    Wouter
  • In reply to Wouter Krikken:

    Hi Wouter,

    happy I could be of help and and I am glad that you managed to fix the circuit!

    Feel always free to reach out if you have questions.

    Best regards,

    Alberto