• State
  • Locked Locked
  • Replies 3 replies
  • Subscribers 62 subscribers
  • Views 419 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.

LM5576 compensation network

user1738755
Prodigy 20 points
Other Parts Discussed in Thread: LM5576

I am using the LM5576 to convert 27....40V to +24V as in the webench circuit below.

As 1.2A is a peak output current  only with an average output current of 250mA I made the circuit more lightweight using the Quick start guide to recalculate values. See schematic below

L1 became 94uH (2x 47uH in series)

Rt 43k

CRamp  1nF

Rramp 100kOhm

Css 22nF

Rfb2 27k8

Rfb1 1k5

Cout 225uF

Ccomp 6.8nF Rcomp 30k0 Ccomp2 not populated

This configuration works OK at higher loads, but fails during light load conditions.

Simulating the circuit with tina shows that there is an overshoot at startup, which leads to some ms of inactivity, which potentially leads to undervoltage lockout caused by losing the bootstrap voltage.

Reducing the Output Cap to 20uF together with using an modified compensation network solves this problem in simulation, but is far away from calculated values:

Cout 20uF

Ccomp 22nF Rcomp 82k0 Ccomp2 22pF

Simulating the circuit gives good results between 6mA and 1.2A load.

Can I safely use these values, or are they outside the defined LM5576 range?

B.R. Ralph.

  • 0
    TI__Guru 57555 points

    I would be very carefull about changing the values.  You might go back to Webench with your new values and see if it accepts the design.

     

  • 0 in reply to Frank De Stasi
    Prodigy 20 points
    Hi Frank
    Thanks for your answer.
    To make it short: I inherited the LM7576 from a previous design state. It performed good, and in addition can be used to generate -24V too.

    The design performs satisfactorily in lab devices, now I would like to proof that it is good for series too.

    Using webench does not allow me to do so: It does not expects the 7576 to operate below 150mA minimum load, at least it does not let me simulate startup with lower load, or load transitions going lower than 120mA.

    I do not have the opportunity to add minimum load resistors burning 3.6W on my board (24V*150mA). 38mA is the maximum.

    Simulating the webench design in tina shows that compensation dimensioning leads to startup overshoot at light load. If this overshoot persists longer than 1-2ms it seems that the bootstrap voltage decreases and a lockout inside the IC prevents it from restarting before the output voltage has dropped to almost 0 Volt. (repeated infitely)

    Modifying the compensation network let me find values that simulate fine in transient mode, and perform quite good in real world. If there is no overshoot there is no problem with light load.

    I try to find an alternative way of proofing that the design is good for series production with alternative compensation network dimensioning.

    I need a specialist to judge whether this fits series production requirements. Can yo have a closer look at it?

    B.R.

    Ralph
  • 0 in reply to user1738755
    TI__Guru 57555 points

    The best way to verify the design is to run load transients.  Preferably over temperature and input voltage.

    If you can not do that then you will have to rely on the simulation.