I'm trying to use a L7912 to output -12 V DC. The datasheet says to use a 2.2 uF tantalum capacitor on the input for stability. Further it says if an aluminum electrolytic cap is used, to have it at least ten times the size. The closest I could find was a 47 uF electrolytic cap. Without the cap the regulator works fine. If I put the cap in, (C2, + at the ground side, because it's negative voltage) the regulator seems to let the unregulated voltage through (higher, ripply).

What am I doing wrong?


simulate this circuit – Schematic created using CircuitLab


  • \$\begingroup\$ What have you assumed to be the pin number connections for the 7912? \$\endgroup\$ – Andy aka Nov 18 '18 at 16:27
  • \$\begingroup\$ @Andyaka I almost tripped over it, but unlike with the 78xx, first is ground, then input then output. Is that what you are asking? \$\endgroup\$ – rauschi Nov 18 '18 at 17:48

We'd need to see your schematic but I suspect that you may be relying on the 47 uF to do all your smoothing on a full-wave rectified supply. You need a large capacitance to do the primary smoothing - typically 1000 - 4700 uF. The local capacitor at the regulator is to avoid any high-frequency instability. The output capacitor is not optional either.

  • \$\begingroup\$ I have a 6.8 mF filter cap in front of it and a bridge rectifier in front of that, rectifying a 12 V RMS AC. That's pretty much it in terms of a schematic. \$\endgroup\$ – rauschi Nov 18 '18 at 14:48
  • \$\begingroup\$ And what's the voltage on C1? You're going to need 15 to 16 V. The top of BR1 should be connected to GND. (I assume it is.) \$\endgroup\$ – Transistor Nov 18 '18 at 15:15
  • \$\begingroup\$ The top of BR1 is connected to a very similar setup. I also forgot to mention, that the voltage source is a center tapped transformer and that I added a diode from input to output of the voltage regulator. I'll update the schematic. \$\endgroup\$ – rauschi Nov 18 '18 at 15:54
  • \$\begingroup\$ I updated the schematic. The voltage at C1 is 12 V RMS, so should be enough. As mentioned, it works fine without C2. \$\endgroup\$ – rauschi Nov 18 '18 at 15:58

I believe there is an error in the datasheet (you should provide a link to the datasheet you are relying upon). Edit: confirmed upon link added to question

The output capacitor is required for stability and should be at least 25uF if electrolytic.

Compare the TI datasheet for a similar product:

enter image description here

Unlike their 78xx positive regulator cousins, but like most LDO regulators, the 79xx series will by more than happy to oscillate if not provided with an output capacitor of appropriate ESR and capacitance.

Note: That might still not be your problem, if you have mixed the connections up, for example, they are not the same as the 78xx series, as Andy's comment suggests.

  • \$\begingroup\$ I added a link to the datasheet. It says the output cap is for improving transient response. \$\endgroup\$ – rauschi Nov 18 '18 at 16:09
  • \$\begingroup\$ The output cap is required for stability. Datasheet is in error. The input capacitor is required if the regulator is some distance from the filter capacitor. \$\endgroup\$ – Spehro Pefhany Nov 18 '18 at 16:57
  • \$\begingroup\$ It's different manufacturers. Are you sure they are not simply a little different? \$\endgroup\$ – rauschi Nov 18 '18 at 17:51
  • \$\begingroup\$ @rauschi I am 100% sure. \$\endgroup\$ – Spehro Pefhany Nov 18 '18 at 17:52

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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