I’m working on a project which utilizes a power-source which is liable to vary. I have a micro-controller sitting at the output of an LDO (Low Dropout Regulator).

I’m concerned about the LDO outputting a voltage below its fixed output if a voltage below specified input range is applied to it. Low-voltage could potentially damage my micro controller and disrupt the logic of the circuit.

Ideally, I’m looking for a device capable of operating under a wide voltage input range -- one that can compare an analog voltage value to some number I programmed into its “memory”. If the voltage input is greater than or equal to that number, it should produce an output which would allow the rest of my circuit to turn on. Basically this device would work as an external under-voltage lockout device, ensuring that my system operates within specified voltages.

  • \$\begingroup\$ Could you please clarify why you doubt the output of a LDO? If you doubt the working of the LDO, why would you trust the working of the electronics that will be checking the LDO's voltage? \$\endgroup\$
    – Huisman
    Commented Jun 15, 2019 at 19:16
  • 2
    \$\begingroup\$ Check internet for "supervisor IC". But I think they are as trustworthy as LDO's (provided both are operated within the limits of their intended use). \$\endgroup\$
    – Huisman
    Commented Jun 15, 2019 at 19:18
  • \$\begingroup\$ @Huisman It’s a doubt of the LDO. LDO’s have a low-dropout rating, there is a minimum input voltage required for the fixed output, and I have witnessed LDO’s drive a lower than rated output voltage if the minimum input voltage is not met. \$\endgroup\$
    – Tiana Pyre
    Commented Jun 15, 2019 at 19:20
  • 2
    \$\begingroup\$ Many microcontrollers have built in brown out detection. Is it too late to just choose a uC that has all the features you need? \$\endgroup\$
    – The Photon
    Commented Jun 15, 2019 at 19:35
  • 2
    \$\begingroup\$ @TianaPyre Why not arrange a voltage regulator that is capable of both buck and boost to get the most out of your input source. I am frankly surprised by your lack of information about your "power-source which is liable to to vary." I don't know why you are being so oblique and "hidden" about disclosing your source of power. But you should consider telling us more about it, I think. (This might be an X-Y question/problem.) \$\endgroup\$
    – jonk
    Commented Jun 15, 2019 at 21:08

3 Answers 3


The device you're looking for is called a "voltage supervisor" or sometimes "reset controller". It monitors the supply voltage and asserts the reset input on the microcontroller until the voltage is within a particular acceptable range. Some also add a short delay after the voltage becomes acceptable before they release the reset signal

Many microcontrollers contain equivalent functionality built-in, where it's usually known as "brownout protection". Check whether yours supports this.

As yet another option, some voltage regulators include a "power good" output which may accomplish the same thing. This is more common on switching regulators but some LDOs have it too.

  • \$\begingroup\$ This is not an AC brownout design -1 Its a variable current source into an LDO feeding SMPS phone battery charger resulting in variable dynamic oscillations of output undervoltage. (UV) \$\endgroup\$ Commented Jun 16, 2019 at 4:13
  • \$\begingroup\$ Your solution will just make it oscillate from no load rising voltage to overload under low light conditions. \$\endgroup\$ Commented Jun 16, 2019 at 4:23

I would recommend you research UVLO a bit more (Under Voltage Lock Out). As that is the thing you're describing. Are you sure your current chips don't already have some feature like this built in?

If not, it's fairly easily done with a comparator and a couple of resistors.

But there are a couple of challenges. Firstly when you power on, there could be a fairly significant initial voltage drop. You're measuring the input without the load, so it will look higher for a start and many things have an initial spike when they power up. You can solve this with a higher trigger point and a bigger hysteresis window.

Once you have figured out those numbers, you need to consider how long it will run for. Constantly oscillating every few 100ms isn't going to do you any good. This usually means you need some sort of energy storage with enough capacity for your task. That could take the form of capacitors or batteries.

The components, as I said are very simple. But working out suitable values for your needs can be tricky, especially with solar.


The problem

PV has enough power to charge a smartphone, but every day when power reduces two (2) phones failed ( in succession?) with damage due to PV charger.


When PV source dims and voltage drops, smart phone's charger cuts out. THen PV voltage rises to an adequate threshold and a smartphone charger turns on again with a surge demand power and PV drops in voltage, and the cycle repeats while the phone surge power increases losses in SMPS battery charger that appears to be the likely cause of failure with constant oscillation on and off and just below cutoff underload but above cutoff with no load. We don't know the exact component that failed but it is dead.

The fix

Boost the storage capacitance with a battery as intermediate low impedance large mAh voltage sufficient to drive the smartphone charger port. This will reduce the fast damaging cycle rate to perhaps completely charge the phone until the next day or at the very least reduce the cycle frequency to many minutes or have the capacity to fully charge at least 1 smartphone.

The question

Irrelevant to the problem ( aka XY problem)

  • My previous answer ( for advanced readers)

An external voltage with some impedance is just a load current based on the difference voltage. As long as it does not exceed the LDO’s internal OCP and/or OTP it comes down to a load regulation error based on the device specs. \$Zout = \Delta V/ I\$ The lowest impedance that a PV can supply is Voc/Isc =Zmpt

THe open circuit voltage Voc and short circuit current Isc and Z max power transfer is the impedance you want to match with your load ΔV/ESR=Imax for maximum efficiency. Then Zmpt rises rapidly as the sun sets as Vmpt drops slowly. ( see PV MPT curves for a better understanding)

The same applies for a negative applied external voltage I= ΔV/Rs.

If the external voltage is greater the Zout becomes open circuit.

A “DC OK” logic level, can be made with a reference voltage and comparator within your spec’s operating range and some desirable hysteresis.

You may be able to predict and prevent these load effects if you can define the loads and power sources. If the source drops below LDO headroom cutoff, this can also be monitored with a R ratio and another comparator.

If you want to optimize your battery charger and PV power, you will want the output voltage between 70% to 82% of Voc (no load) and disconnect load when storage cap drops below Vcharge acceptable tolerances. This will create an undesired Relaxation Oscillator effect as the voltage rises with no load and drops with insufficient low impedance to start the phone charger SMPS which tend to have a surge start current and have higher conduction losses at low input voltage or as you expected drop the PV voltage too low and cook the phone charger from it’s lack of protection for UV lockout.

There is also a design for a Window comparator using dual comparators for UV/OK/OV that may be considered. where OK >UV threshold and < OV threshold.

The only solution for this frequent problem is to have a local regulated battery to store PV power between phone charge uses. Then previously stored energy can be transferred to the preferably smaller mWh battery in the phone and still have a DC OK signal enable the charger connection. Matching the voltages as close as possible to the load at 82%Voc during max PV solar input is one design goal to minimize series drop power loss.

There must be commercial solutions with a PV battery bank that already does this. Maybe, search , find buy or make one.

Otherwise give all pertinent specs for all inputs and outputs to start a better design with links to datasheets in question.

  • \$\begingroup\$ This is the same as your previous answer. What shall weto do? (1) I delete both. Or, you delete both. (2) You roll back the first one, and delete this one. \$\endgroup\$ Commented Jun 18, 2019 at 22:57
  • \$\begingroup\$ @NickAlexeev again? I better go back to Desktop version only on the iPAd. I’ll take care of it TY. \$\endgroup\$ Commented Jun 18, 2019 at 23:06
  • \$\begingroup\$ @NickAlexeev do you agree with my solution (Battery pack to store energy) to prevent power oscillation burnout on smart phones? If so why not show your support or make a comment relative to the subject. TY again for the headsup but your 3 options suggest discontent. \$\endgroup\$ Commented Jun 18, 2019 at 23:11
  • \$\begingroup\$ @Tony, me agreeing or disagreeing with the answer has nothing to do with this. \$\endgroup\$ Commented Jun 19, 2019 at 0:14
  • \$\begingroup\$ Yes @NickAlexeev but I am requesting for your opinion. but if you would rather be unfriendly, just ignore it \$\endgroup\$ Commented Jun 19, 2019 at 0:21

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.