What is the minimum input voltage that I can use with an an LM7812 to obtain a clean, regulated output voltage?
\$\begingroup\$
\$\endgroup\$
2
3 Answers
\$\begingroup\$
\$\endgroup\$
5
Here's the drop-out voltage spec from the Fairchild LM7812 datasheet (provided by Mouser):
This means at least 14 V input is required to be confident of getting a well-regulated 12 V output.
If you're using a supplier other than Fairchild, you should check the datasheet from that vendor, but it's not likely that 12.53 or even 13 V will be acceptable.
If you need the input voltage to be less than 1 V above the output voltage for a linear regulator, you are looking for a low drop-out (LDO) regulator, which the 78xx series is not.
answered Dec 24, 2023 at 6:01
-
2\$\begingroup\$ There's no reason to presume that 1A is the load being driven, so the exact dropout condition cannot be inferred from that tabulation. Typically, at a few mA, 12.7V input might be enough. \$\endgroup\$– Whit3rdCommented Dec 24, 2023 at 9:02
-
5\$\begingroup\$ @Whit3rd Certainly not. Even at 0mA it typically requires about 1.2V headroom at 25°C (so no guarantees and performance for ripple rejection and temperature will be greatly degraded). A cautious designer should use more like the 14.5V figure in the output voltage spec above, so 2.5V. Or the 3V (15V in) where the ripple rejection is specified if "clean" output voltage is required. LM78xx/LM78Mxx regulators are great in many ways, but they are not LDOs. \$\endgroup\$ Commented Dec 24, 2023 at 16:12
-
3\$\begingroup\$ @Whit3rd, this 2 V spec is all we get from the datasheet. If you or OP has time to do additional experiments or characterization they may find a lower voltage that is good enough for them. If they want to design it and get it right without a bunch of extra effort I recommend to stick to what the datasheet says. Since they want to work with only 0.53 V overhead, that almost surely means finding a different, LDO, part and not mucking around with the 7812. \$\endgroup\$ Commented Dec 24, 2023 at 17:52
-
3\$\begingroup\$ That's leaving aside the uncertainty because this is not actually a spec but only a "typical" performance value. And that if you were to do extra characterization to find a lower overhead operating point, you'd also have to do incoming inspection on future lots to be sure they continue to perform like the ones used in the prototype. \$\endgroup\$ Commented Dec 24, 2023 at 17:56
-
\$\begingroup\$ See my answer which expands on what you say at lower currents. \$\endgroup\$– Russell McMahon ♦Commented Dec 25, 2023 at 10:05
\$\begingroup\$
\$\endgroup\$
If you require a "clean output", you care about "Line regulation", "Load regulation" and "ripple rejection". The datasheet specifies ripple rejection at an input voltage of 15V, so there is no guarantee that ripple rejection will be as good as specified at lower input voltages. Line regulation is specified at 500mA load current, and you see two performance values: The one including 14.5V input voltage is worse than the one starting at 16V. This might indicate worse line regulation at "low" input voltages.
To be safe, with that datasheet you should input at least 14.5V, preferably 15V. To me, connecting the 7812 input to a 15V supply seems acceptable, even if there is some fraction of a volt drop on the supply line from the 15V supply to the 7812. Note that the "load regulation" specification is only valid at the "default" 19V input voltage, so you need to qualify load regulation yourself at other voltages if that parameter is critical in your application.
The typical 2V dropout voltage specification means that at 1A load, the device drops around 2V. It might be 1.9V on some chips and 2.1V on other ones. As there is no maximum value specified, it might even be 2.4V on some rare chips, so this is no parameter you can rely on, but just a hint specifying a coarse estimate of the device performance. Using a "typical" performance parameter without any margin is not a valid design method if reliable operation is required under all conditions. In this case, transistors start to perform worse at high temperatures, so if the case temperature stays well below the maximum, the 7812 might be able to operate at a lower dropout voltage (which is confirmed by the graph quoted by Russel McMahon, which shows the maximum output current at a junction tempeature of 25°C), so for a one-off test circuit that stays well below 1A current and provides sufficient cooling the 7812, you might get away with even less than 2V dropout voltage, but if the voltage is not as clean as you would like, it's considered your fault, and not the 7812's fault.
\$\begingroup\$
\$\endgroup\$
Adding to 'The Photon's answer.
This is a crop from figure 3 on page 17 in the datasheet.
I've expanded the portion showing the input differential range of 0-5 Volt.
This is a typical curve. err on the high side with respect to voltage.
At zero current the "dropout" is just over 1 V.
At 0.5 dropout is 1.2-1.5V
At 1A about 1.9V.
This curve is at 25C and with the output having dropped 100 mV from the nominal regulated value due to loading.
answered Dec 25, 2023 at 10:03
a clean regulated output voltagein the question is vague. How much Line Regulation and Load Regulation does the application required? \$\endgroup\$