• State TI Thinks Resolved
  • Locked Locked
  • Replies 5 replies
  • Answers 2 answers
  • Subscribers 62 subscribers
  • Views 110 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

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.

LM51231-Q1: Vin=3.3V to Vout=20V with 60W design support

Joern Ihlenburg
Prodigy 30 points
Part Number: LM51231-Q1
Other Parts Discussed in Thread: LM51551-Q1, LM51551, LM5155

Tool/software:

Dear Sirs,

we are not sure if we can use LM51231-Q1 to design a rather unusual boost converter:

Vmin = 2.5V
Vnom = 3.3V
Vmax = 3.7V
Vout = 20V
Iout = 3A

It will be used for driving a 20V BLDC motor, so output ripple is not much of concern. Also the output will only be used for short periods of time (max. 2s). But we are very space constrained.

Any concerns from TI for such a setup?

Many thanks

Joern Ihlenburg

  • 0
    TI__Genius 15935 points

    Hello Joern,

    Thanks for reaching out to us via e2e.

    In general, it is possible to connect the supply of a boost converter to its own (boosted) output.
    So, for example, if the normal input voltage of a car is 12V and during a cold crank it may drop to 3V,
    the booster can keep generating 12V during that time and will be supplied by its own output voltage.

    But every controller needs a start-up voltage which has to be met or exceeded to bring it out of reset.
    Also, this voltage will be needed to drive the gates of the eternal FETs.

    The LM5123 has a start-up voltage of 3.8V. So even your max input voltageb is lower than that.
    None of our BOOST converters can start up with such a low voltage that you would require.

    To get around this, you might use a small charge-pump to basically kick-start the BOOST converter until it can generate its own supply.

    Best regards
    Harry

  • 0 in reply to Harry Hofner
    Prodigy 30 points

    Hello Harry,

    many thanks for your support here. We are looking into an alternative solution with a possibly smaller footprint:
    - TI LM51551-Q1
    - On-Semi NVLJWS5D0N03CL FET
    - Diode PMEG4030ER-Q
    - 0.33 uH Inductor (Coilcraft XGL4020)

    We have a regulated 3.3V power supply available and we are wondering if we could power the LM51551 with that 3.3V supply (through the BIAS pin?) in order to be able to work down to the 2.5V.

    Kind regards

    Joern

  • 0 in reply to Joern Ihlenburg
    TI__Genius 15935 points

    Hello Joern,

    I am confused now. Maybe there is a typo in your initial post.
    What is the maximum output current? What is the expected output power?
    If it is 3A@20V, how can that work with a diode with 3A maximum current?

    60W at 2.5V minimum input voltage means 24A average input current plus losses plus tolerances.
    60% ripple may not be enough in this case, so your peak inductor current may easily exceed 40A.

    With a regular BOOST topology, this current will also flow through the high side switch (a diode in the asynchronous case).
    This would mean 12 Watt losses - just in the diode (!)

    You may be able to get around this problem with a (non-isolated) flyback topology, where the diode only sees the peak output current.

    Anyway, if you connect both, BIAS and VCC to that 3.3V supply and keep the power stage input completely separate, the controller will be able to start-up and also to
    support an input voltage of the power stage down to 2.5V.

    But please be careful with the FET that you choose because the maximum gate voltage will be below these 3.3V.
    And also, the 3.3V supply will need to deliver enough current for the controller and the FET.

    All information in this correspondence and in any related correspondence is provided “AS IS” and “with all faults” and is subject to TI’s Important Notice (www.ti.com/.../important-notice.shtml).

    Best regards
    Harry

  • 0 in reply to Harry Hofner
    Prodigy 30 points

    Hi Harry,

    many thanks again for your support. We have 6 Li-Ion battery cells in a parallel configuration. Each cell can be discharged with up to 5A. We have a small booster to generate the 3.3V for the logic supply... the main application is to drive a 24V BLDC motor for very short bursts of time (2s max, around 10-20 times per 24h). 

    We could tweak the gearbox in front of the motor a little bit to get to a 24V / 2.65A for the BLDC motor. That would roughly translate to 3V/3.5A per battery cell. We could probably live with 3V for the minimum input voltage. The maximum input voltage into the boost converter would be 4V due to the maximum cell voltage.
    We are extremely space constrained so every additional square millimeter hurts a lot.

    Any suggestions for a TI part from your side? We are under NDA with our OEM customer and we could share a lot more information with you if we would have an NDA between TI and us.

    Many thanks and kind regards

    Joern

  • 0 in reply to Joern Ihlenburg
    TI__Genius 15935 points

    Hello Joern,

    First I want to correct my previous statement.
    The peak inductor current may easily exceed 40A and in a regular asynchronous Boost topology, the diode has to be able to handle that peak current.
    Nevertheless. the (thermal) loss is calculated with the average load current. But this would still mean 2.4 W of thermal dissipation or more.
    So, a diode in a very big package is required plus some cooling concept (also if the load is only connected for two seconds).

    I also want to set your expectation right:
    The e2e forum is only meant to discuss technical issues with our devices.
    We cannot help you with any NDA related questions, nor do we have the bandwidth to create a design for you.

    Here are some ideas on the issues that you are facing.

    - 3.3V as a "kickstart"-voltage is still too low.
    Even if the internal logic of some of our controllers can start-up with that voltage, it will be extremely difficult to find a FET which can fully turn on below 3V.
    If you want to supply the booster with its own (boosted) output voltage, you will also need a diode from the Auxiliary voltage to the supply of our controller which will reduce the gate voltage even further.

    I spoke to a colleague, and he can also not recommend any FET which will be able to deal with this limitation.
    Please contact the FET manufacturers directly for some proposals.

    Long story short:
    You will need some kind of voltage doubler to generate something like 5V or more to properly kickstart the Boost stage.
    If the 3.3V is generated by a switch mode converter, you could for example use the frequency of the switch node of that converter and add a discrete charge pump, built from some capacitors and diodes.

    - There is a duty cycle limitation. A boost factor of Vin : Vout = 1 : 8 is pretty much at the edge
    The LM5123 will not really be able to handle that, so I would not want to recommend this as a solution.
    The LM5155 would be more suitable because it is possible to tweak the slope compensation with an external resistor.
    A (non-isolated) flyback topology, based on the LM5155 would be even better, because the transfer ratio of the transformer will help to bring the duty cycle down and also reduce the peak current through the diode.

    - I assume that there is some kind of controller which will turn on the load every hour or so.
    So, the majority of the time, the converter would operate with no load.
    Or are you planning to enable the Boost stage together with the load?

    We would propose that the controller will turn on the Boost stage without any load.
    Then, when the output capacitors are charged, turn on the load/motor.
    Either some Microcontroller does the sequencing or you could use some R/C delay stage.
    Or you could use the PowerGood signal from the LM5155 to turn on the load which will save you from adding any extra delay.

    Please have a look at this design on our website, which is a similar concept (but with different voltages, current and timing requirement):
    https://www.ti.com/tool/PMP30790 


    All information in this correspondence and in any related correspondence is provided “AS IS” and “with all faults” and is subject to TI’s Important Notice (www.ti.com/.../important-notice.shtml).

    Best regards
    Harry