# LTC3112 undefined behaviour when power supply reaches current limit

I am getting a behaviour of LTC3112 that I do not understand. I am using it very close to the schematic recommendations in the datasheet. It works perfectly fine until my power supply (TOE 7610-60) goes into current limiting (at 2.5 Amps). The current limit is only reached due to switching peaks in the current, but not actually used by the circuitry. Then the LT3112 seems to get stuck and constantly draws 2.5 Amps (and gets really hot).

I have done a simulation with LT Spice, that does not show this behaviour. Note the voltage source V2 (going from 0.5 V to 0V in steps) and the voltage source V (with current limiting for 2.5 Amps).

In short: the power supply reaches its current limit and even if I decrease the output current of LTC3112, the power supply keeps outputting 2.5 Amps.

• What does V2 do when it reaches current limit? – Andy aka Aug 7 '19 at 7:29
• V2 is just going from 0.5 V to 0 V in steps (it is in fact a DAC from a MCU) V is the source that reaches current limit at 2.5 Amps and gets stuck there. – Jana Aug 7 '19 at 7:34
• I have done a simulation with LT Spice, that does not show this behaviour. And that's to be expected as V2 can deliver unlimited amounts of current. You should really use a bench supply which can deliver the peak current without it going into current limiting mode. As a possible workaround you could add a high value capacitor (like 1000 uF) in parallel supply at the input of the DCDC converter. This large cap might be able to "take over" that peak current so that the supply doesn't reach current limiting. – Bimpelrekkie Aug 7 '19 at 7:36
• Thank you for your answer! I nevertheless do not understand why it never ever again exits the current limit, even if I for instance change the input from 5 V to 10 V. I have tried it with a supply that has a higher current limit and it works. Just really want to understand the behaviour. I cannot add a capacitor for the final application since the design of the PCB is final. – Jana Aug 7 '19 at 7:46
• From data sheet: Should the output become shorted, the average current limit is reduced to approximately one half of the normal operating current limit. – Andy aka Aug 7 '19 at 11:58

My load is about 10 $$\\Omega\$$ and I sweep the voltage from 1...12V out, but this happens if $$\V_{OUT}>V_{IN}\$$.
My best theory so far is that the $$\V_{cc}\$$ pin starves. There is a mention in the data sheet that a diode forwarding 5V from another source might improve boost functionality.
If you get $$\V_{in}\$$ dips it possibly also affect $$\V_{cc}\$$, so it might be something to try. I realize this question is old but it might help someone to be careful about the $$\V_{cc}\$$ pin on the LTC3112.