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.

LM46002: frequency foldback

Genius 3870 points
Part Number: LM46002
Other Parts Discussed in Thread: LMR36506

Hi there


We use LM46002 to convert Vin=19V~12.6V to Vout=11.7V/1A, use ext. freq. 800khz (also connect RT=53.6Kohm at the same time).


When vin=12.6V/no-load, the SW node waveform indicated that converter works in decreased switching freq. ~401KHZ, and it further decreased to 258KHZ when load is 1A. Everything else (like Vout ripple and loop stability) seems remain ok.

We checkd with datasheet to confirm this "frequency-decreasing" behavior when Vin is near Vout. Section 8.3.8 may give the explanation that under 800KHZ freq, we will hit the Dmax = 84% limit. So we can not get 11.7V from 12.6Vin with 800KHZ because of "frequency foldback". It seems to be reasonable behavior of LM46002.

Could you please tell how this "frequency foldback" works? We have to ensure f_switching=800KHZ for all DC/DC in our system, and "12.6V to 11.7V conversion" is the system spec. Is there anything we can do to solve this problem ? Also , changing to another buck IC is the last option.

Any suggestion is welcome. Thanks!

  • Please see figure 34 of datasheet to understand how switching frequency varies with load current. Switching frequency is dropped at light loads to improve efficiency.

    https://www.ti.com/lit/ds/symlink/lm46002.pdf

    Switching frequency will decrease at high VIN if min,ton spec of the device is violated.

    See section 6.7, Switching characteristics table to find ton-min spec. 

    1. 800kHz -> tp = 1.25uS

    2. D = 11.7/19 = 0.61

    3. 0.61*1.25 = 0.76uS -> ton,min is not violated.

    Frequency will only foldback in light loads unless the device is synced used sync pin.

    If this is not an option, you will need to design in alternative device with FPWM feature.

    Two devices that come to mind that can achieve this are:

    LMR36006FSC

    LMR36506

  • Hi Yi,

    Yes, the frequency foldback is built in into the IC, what you see is normal behavior in order to maintain regulation at the output.

    Once the VIN is lowered down close to VOUT, the IC will try to maintain output voltage by reducing its switching frequency.

    Otherwise if the same frequency is kept, the VOUT will start drooping. 

    Since the headroom on dropout depends on efficiency, therefore to improve dropout performance then you need to increase efficiency. 

    You can probably select an inductor that has a lower DCR (winding resistance) to improve the dropout performance but unfortunately,  I dont think we can turn off this frequency foldback feature. 

    Will you be ok with running at lower frequency? or changing the output voltage?

    Thanks

    -Arief

  • Hi Marshall, you may misunderstaood my point. We only have the problem in LOW Vin (12.6V), so I thinkd it is the t-off-min which triggered the frequency foldback, and we are trying to fix or avoid this problem. thanks. 

  • Hi Arief, thanks for confirm our guess. 

    We are required by system design that all dc/dc should work on sync clock 800khz. But it maybe allowed to decrease Vout base on bench-test ( decrease Vout ) and I will look into this.

    Since the freq. foldback feature is fixed into IC, seems that there is not much thing we can do on circuits design.

    By the way, why "an inductor that has a lower DCR (winding resistance)" can be used to improve the dropout performance, and how low the DCR should it be?

    Thanks.

  • Since the freq. foldback feature is fixed into IC, seems that there is not much thing we can do on circuits design.

    You can slow down switching frequency if you want to operate at continuous frequency.

    Lowest DCR is best to improve dropout performance.

    DCR is conduction loss.

    See equation 7,8 of app note. 

    http://www.ti.com/lit/an/slyt293/slyt293.pdf