LMR16010: Functional Modes - how do I achieve CCM mode with very low loads eg 10mA, 100mA

Hello all,

first of all the main question is, how does the LMR16010 switch internally from his very different modes, active mode: sleep, CCM, DCM.

Or to put it another way, I did want to design a very variable power supply, for example 24Vin to 3,3V out, load from 10mA to 100mA usually, but maybe sometimes up to 500mA or even 1A.

So I did come around the LMR16010, dont know exactly why I did choose this one, maybe some TI product salesman could step in and convince me that this was the right choice, or has some better ideas, from my understanding, all SOIC8 power products do have quite a different pinout, so I cannot switch them around as easily as Opamps for example.

But i am drifting apart form my issue here.

SO my first design, I wanted to go with 700kHz switching frequency , has a 330uH inductor in my design, thought that was a good starting point, since more inductance is better for the ripple i thought.

Now what I could measure, is a huge ripple on the output ( 3,3V that is ) with a frequency to about 7kHz. So I startet wondering, why this is the case.

Now with some further investigation I got to the point to realize the IC is probably very bored with my loads, so its not switching right, it simply is working in sleep mode.

Datasheet states:

sleep-mode is entered, when internal COMP voltage drops to 400mV

What exactly is meant by that, how can I know with what load exactly this will happen?

Then I started to play around with different inductors, used from 1uH to 100uH, the lower the L value, the faster the sleep mode switching frequency gets.

This makes sense, because I assume the rise time gets higher for smaller values, since the mathematics behind is:

i = (Vin-Vout)/L

But with a 10mA load on 3,3V 1uH is still not enought to exit from sleep mode

with 100mA 1u still is not sufficient, but with 12uH a strange DCM mode is achieved, strage in the sense, its kind of 4 times DCM mode, after that one time a little longer , and looks a bit like sleep mode for one longer cycle.

Maybe this ia the overcurrent protection kicking in, I have not gotten to that understanding yet.

UPDATE#1:

on page 12 Datasheet: sleep-mode: if any switching cycle is below 300mA. The sleep-mode current threshold is the peak switch current level corresponding to a nominal COMP voltage of 400mV.

So I try to have loads higher than 300mA next, to see if this is the magic number, when the device exits sleep-mode.

But it seems rather odd to me, to have a device capable ot 1A be held in sleep mode until 1/3 of its possible output, I think i am missing something here, like switch current is not the same as load current, since load current is ( depending on the duty-cylce ) halfway sourced by the inductance thruh the free-wheeling diode.

So to get to some final big questions:

What is meant by 400mV on the COMP?

is that the voltage on the Feedback resistor network, but this in my measurment is of course always around 750mV as supposed to reffering on the datasheet.

Maybe someone has a nice formula for calculating for CCM mode in extrem low loads, like I=10mA, because for me it doesnt work, also 100mA does not seem to work.

Next step for me will be to play around with the switching time, maybe some slower times could lead to CCM mode, since current rise will have mor etime to develope, but I will try that out soon.

PS: what LM and TPS on the TI products stand for, i get linear monolithic? and texas power supply?

Thanks for reading.

seb