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.

LM5046: Can I use FSFB as step up instead of step down?

Part Number: LM5046

Hi,

I know most of the application out there were for step down. However, for my case I'm interested in stepping up from a 12V input supply to around 48V output ~50W. I know that this should be doable with the change of configuration of the transformer's turn ratio.

What I'm curious was based on the datasheet mentioned in the PIN FUNCTIONS table, "If the power sources of VIN is outside of the range, LM5046 can be biased directly at VCC by an external regulator". I had tried that both within the PSpice model and the actual unit by sourcing 12V to VCC & VIN directly but no luck for me. For actual unit, it seems to sink a lot of current which tripped my external power supply. Based on the functional block diagram provided within datasheet, as long as VIN is no lower than VCC, it shouldn't biased the internal diode to sink so much current.

Any advice will be much appreciated.

thanks.

  • Hi Wei,

    Can you check if your PSPICE simulation will run correctly if the you connect the Vin in the normal configuration. If it runs in that way then you can try reducing Vin until you see the issue.

    It could be that the controller is not running because of some other issue. If the feedback is closed, I suggest setting up the feedback to ensure the controller runs at max duty cycle and trying to get the controller running first.

    Regards

    Peter
  • Hi Peter,

    I had tried as suggested with the standard TINA model provided in TI website. Seems like it keeps on complaining operating point not found when I'm doing the transient analysis.

    May be you can take a look as attached?

    Thanks.

    /cfs-file/__key/communityserver-discussions-components-files/196/LM5046_5F00_modified.TSC

  • Hi Wei,

    I relaxed some of the simulation parameters and the simulation now runs but I dont think its operating correctly yet.

    Regards

    Peter

  • Hi Wei

    Please contact me by email petermeaney@ti.com as it will be easier to send you the changes.

    Regards

    Peter
  • Hi Peter,

    Glad to hear that you can at least make the simulation running. I had tried to reached you through your email posted above but haven't got any response from you so far.

    Looking forward for any prompt response from you.

    Thank you.

    Rgds,

    Wei Chun Liau

     

  • Hi Peter,

    I had tried running the simulation with your suggested parameter tweaks. The simulation result shown that the controller only operational when Vin is >= 14V which is contradict with the datasheet. I had forced the UVLO, OVP, COMP to a standard DC voltage and grounded CS which shouldn't result in the shutdown of the controller. That was pretty much the scenario I have seen with the physical chip.

    Any advice to make it operational in 12V for VCC & VIN will be appreciated.

    Modified file as attached in 12V for Vin.

    /cfs-file/__key/communityserver-discussions-components-files/196/LM5046.TSC

    Picture as shown below is when Vin is in 14V where at least the mosfets are switching.

    Picture as shown below is when Vin is in 12V where the mosfets are  not switching at all.

    Thanks.

    Rgds,

    weichun

  • Hi Weichun,

    From reading the datasheet it says that the device should run from Vcc, but Vcc must be supllied from a regulated 12V rail. In that case I suggest you leave Vin open circuit and connect a diode from Vcc to Vin to bypass the parasitic diode of the Vcc regulator.

    I agree the sim does not run if Vin is below 14V, but that may just be an issue with the model which I need to check.

    If the min input voltage in the application is down to 12V then I think you will need a separate bias supply to provide the 12V for the Vcc pin, maybe something like a low cost flyback could provide a regulates 12V to Vcc for the application input voltage range.

    Regards

    Peter

  • Hi Peter,

    Thanks for your suggestions. I will give it a try then.

    Rgds,

    weichun

  • Hi Peter,

    I assumed that your suggestion is as shown below.

    The exactly setup was configured as shown as above. Upon power up, okay it didn't trip my 12V external power supply this time. However, once I source the COMP pin to 3V for switching it went haywire again. It not only didn't switch but drawing a few 100mA from the supplies causing the IC to heat up around 60degC.

    Same setup procedures without diode & 12V but with 14V through Vin, this time all things worked as expected as per our simulation model.

    I think that it might not be the model went wrong but the datasheet!

    Appreciate if you can have your hardware team to verify on this. If it really cannot support a 12V VCC input as claimed in the datasheet, then I'm going to scrap this and move on with other solutions.

    Thanks.

    rgds,

    weichun

  • Hi Weichun

    I agree, we will try and test it on the bench by modifying the evm.

    Thanks

    Peter

  • Hi Weichun,

    We need to order an EVM to test the application this will take a week to arrive. Sorry for the delay, I understand if you need to explore another option.

    Regards

    Peter
  • Hi Weichun,

    My colleague tested the EVM, he disconnected the Vin from the main input positive rail, hthen connected the Vin pin to the Vcc pin. The EVM was then connected to 48V and then Vcc was supplied from a 12V power source and the EVM started-up and ran on the bench. Please note that Vcc is also supplied by an aux winding from the transformer so that may have caused you an issue during your test ?

    Regards

    Peter
  • Hi Peter,

    Thanks for the information above. I would say it wasn't the aux winding from the transformer that affect the test. I haven't modified my EVM kit to try on that but with the sample units I got from TI store on my own prototype board. I am curious that all the 5pcs sample units I received was exhibiting the same behavior as mentioned above which in a way was tally with the result from the simulation model.

    I had confirmed with the modified EVM and the same unit controller had been transferred to my own prototype board. It worked like a charm.I also had confirmed the bad sample units by loading one of them back to the EVM and the EVM became not working.

    I'm not sure if you guys would want the bad units back for FA. Since I got them from TI store, I wouldn't doubt that they are counterfeit units. All the bad sample units work fine if they were powered up with a 14V through Vin without directly power up through VCC just as shown in the previous simulation model.

    This might give me a impression that TI might want to disable some features of it so that they were tally with the simulation model.

    Do you have any update regarding the simulation model? Is it being intended to behave in such a way?

    The subsequent IC controller that I buy from the market would it behave somewhat like the sample units I got from TI store or those that were installed on the EVM?

    Thanks.

    Rgds,

    Weichun

  • Hi Weichun,

    The IC designere checked the IC and says that the VREF rail (and consequently the internal 5V rail) is released when the UVLO pin voltage is above 0.4V (see pin description on the datasheet at pin functions section).

    Have you verified the voltage on UVLO pin when the device start-up when powered from 12V on Vcc?

    Thanks

    Peter

     

  • Hi Weichun,

    I hope you have made some progress in the meantime, since there has been no further replies on this post I am going to close it. Please open a new post if you have more questions.

    Regards

    Peter