How to increase the undervoltage lockout hysteresis range of DC-DC converters?

[Faranight]

Hey, an advance question. So, I'm not sure, if this is going to be a problem or not, but I did have some bad experience in this regard with an older DC-DC circuit. I'm about to build a new board for a small outdoor day-time solar power supply (hobby project). It's going to be a small supercap charger with two DC-DC regulators. The first one will be a MPPT IC to charge a 12V supercap bank off of a small solar panel, and the other IC (TPS63060) will be a buck-boost converter that powers itself from the supercap and produces a regulated 5V output. This second IC has an Enable/Uvlo pin that causes the chip to power itself ON once the voltage on the pin is high enough, and shut itself OFF once the voltage drops below a certain value. There is some small hysteresis present, so the pin basically acts like a schmitt trigger for the IC. This range is unfortunately fixed according to the datasheet (page 6), and cannot be modified.

Back in the days I had an issue with a similar circuit where the input voltage of the DC-DC regulator was high-impedance. The issue was that this converter drew a very large current upon powerup, which caused the input voltage to drop. This in turn triggered the undervoltage lockout on the IC and shut it down. Once the input voltage recovered, the cycle would repead in an oscillation-like fault. Back to my current project, I fear that there's going to be a similar issue here. For example, in a low-light condition (i.e. in the morning when the supercap starts to charge from a depleted state) there won't be enough charge present to power the load. The supercap voltage will reach the enable threshold and turn the buck-boost IC on, which will then drain the cap and shut itself off, resulting in an oscillation.

Is there any way this DC/DC buck-boost IC Uvlo hysteresis can be widened somehow?
Like to allow the supercap to charge a bit (fully?) before the buck-boost IC is enabled, but still allow it to be drained far below this value?

[tom66]

If the converter has a PG output, you could use that to bias the EN pin if the EN pin is designed for analog sensing.  I am not aware of any part with both an analog-sensing EN pin and a PG output, but something like the LT3580 has a sensitive SHDN pin which trips around 1.2V and the specification is guaranteed in the datasheet, so with an external output good comparator you could implement this function.

Personally my suggestion would be to avoid messing with the hysteresis range of a converter like that and just implement the UVLO function yourself using something like a window comparator.  A device like TLV6710 is cheap enough to solve this problem without worrying too much unless you are going into very high volumes.

[mtwieg]

Yeah one thing I like about many LT controllers is that their enable pin has a well-defined threshold voltage (usually 1.21V), allowing designers to freely define their own UVLO threshold. Sometimes there's also a switched current source on the enable pin, which allows one to also freely define the hysteresis as well.

Unfortunately, I've never seen this sort of function on a TI controller. You would have to interface their enable pin to a separate UVLO circuit (which they will gladly sell as a separate chip).

[MasterT]

Make initially threshold high , than lower as converter starts up .

[Faranight]

This supercap bank can go up to 12.0V on my current design. I have an older lower-capacity circuit with a 5.0V supercap bank, and it used the MIC2779 voltage supervisor IC to adjust the operating voltage range.
Unfortunately, MIC2779 works only on supply voltages up to 5V. If only there was a 12V+ version of this IC.

A device like TLV6710 is cheap enough to solve this problem without worrying too much unless you are going into very high volumes.

The TLV6710 looks interesting, but how does one turn a window comparator into an adjustable Schmitt trigger?
Would it not make more sense to use a low-power analog comparator instead?
The good news is that EN pin of my PMIC seems to support voltages past 12V, so technically Vin can be used as a pullup source.

[tom66]

The TLV6710 looks interesting, but how does one turn a window comparator into an adjustable Schmitt trigger?
Would it not make more sense to use a low-power analog comparator instead?
The good news is that EN pin of my PMIC seems to support voltages past 12V, so technically Vin can be used as a pullup source.

You can set the hysteresis by connecting a resistor from EN to the resistor network. As the output turns on you can use this to bias the divider network in one way or another through a high value resistor.  One issue will be that the output of EN will go high when the hysteresis resistor needs to pull low, so you may need to use an inverter circuit (NPN transistor pulling EN low) and invert the logic of the window comparator.  I'd recommend modelling it in LTspice to see what values work for you.

[ArdWar]

Just tie the bottom divider into PG output. Assuming PG behaves exactly as it should be during startup that is...


EDIT:
Scratch that, this method is not applicable to this P/N. Turns out UVLO is tied to VIN. I thought it was tied to EN pin like several other TI converter.