Unexpected V-out From LM7812

Discussion in 'The Projects Forum' started by PGB1, Feb 2, 2013.

  1. PGB1

    Thread Starter Member

    Jan 15, 2013
    Good Morning!
    I have a question or two about unexpected voltage output from a circuit based around LM7812, if you don't mind a bit of Amateur Hour:

    For background, the project is for a part of my boiler's zone valve controller. The part of the controller I'm trying to repair operates a panel light, getting source power from the nominal 24vac transformer which controls the system.

    The existing control board for the light is LM7812 based. The bridge rectifier on it melted. (DB1f), opening the main control fuse- shutting down the heating system. (Only in January!)

    I re-built the panel light's board with new:
    Rectifier Harris #DB1F
    Regulator Fairchild #LM7812
    Both Capacitors (See attached drawing)
    Fuse 1/2 amp fast acting 250v (Wise, new addition to the circuit)

    I re-used the panel lamp which is marked as a 12 volt LED (green- remote panel mounted).

    The Main Question:
    While bench testing, I had 10.6 volts DC out of the LM7812 with no load.
    I tried 2 different LM7812, plus the original, and got the same results. These were bench tests with no other equipment on the supply transformer. Repeated with three meters (Same results +or- 1/10th volt)

    Some Specifics:
    Input Voltage = 26.1 vac (Held steady with all tests shown below)
    Voltage At Rectifier Input = 26.1 vac (Held Steady)
    Voltage At Rectifier Output = 22.9 vac (more drop than I expected)
    Held within + or - 1/10th volt for all tests shown below

    Voltage At LM7812 Output No Load = 10.60v
    Voltage At LM7812 Output LED Only 12ma Load = 10.12v
    Voltage At LM7812 Output Test Resistor As Load:
    1k Ω = 9.96 volts
    100 Ω = 9.82 volts

    Temperature of LM7812 in TO220 Case
    Measured heat sink at back of component body
    Start = 14.5C (Room Ambient)
    1 Hour With LED Load Only = 17.5C (Room 15.0 C)
    3 Hours With LED Load Only = 17.5C (Room 15.0 C)

    Schematic- Quick sketch of the project as-built & from a Fairchild datasheet (a copy is attached).

    So, why do all three LM7812 give me 10.6 volts output?
    Could it be 10.6dc plus the ac ripple of 0.9 equals 11.5 equals the regulator circuit is OK?

    I'm also wondering if a smoothing capacitor across the rectifier output + & - would help the dc voltage become closer to 12? Perhaps replacing the existing 0.33 μf with maybe a 1000 μf?

    I don't want to try a smoothing capacitor until I ask here, for fear of destroying the LM7812 or the rectifier.

    I'm stumped, even after looking at the Fairchild datasheet for operating parameters? What did I build (or measure) wrong?

    If it helps with diagnosis, the circuit board looks like a paper project board & the traces connecting the components are made of double layers of copper tape. The longest distance from one component to the next is ca 10mm

    Sneak-In Question: Why is the voltage drop across the rectifier 3.2 volts?
    The datasheet shows 1.1 volts max drop per element at 1 amp.
    Therefore, if I am correct: 2 diodes working at a time = 2.2 volt max drop at 1 amp.

    Thanks Very Much for your thoughts!
    PS: Not that it will help me solve the questions above, but wouldn't a voltage divider be a more appropriate circuit for this panel lamp?


    Mar 1, 2012
    If that's how the rectifier is physically connected, I don't think it's arranged the right way.

    Someone correct me if I'm wrong.

    You're 1.4V short, and a diode has a .7V drop, so having the rectifier connected wrong might be causing it??

    Just guessing here.
  3. BillB3857

    Senior Member

    Feb 28, 2009
    The 0.33uF capacitor is not providing much filtering. You say you are reading 26.1VAC at the rectifier output. Is that a typo or are you really measuring AC. What are you using as a reference point for your voltmeter. (should be the center lead of the 7812) Remember that without filtering, the output of your bridge will be half sine waves which go from 0 to peak and back to 0 120 times a second. Likewise, the output of the 7812 will be going from 0 to 12v 120 times per second. The zero time will subtract from the ideal 12v output.

    If all this board does is operate a light, does the light turn on? If so, AND it only drives an indicator, I wouldn't worry about it. If you just want a good solid 12V output from the regulator, put a meaningful filter cap in parallel with the 0.33 that is already there. Anything you have above 47uF would probably work well, depending upon how many indicators may be on at once.
  4. PGB1

    Thread Starter Member

    Jan 15, 2013
    Thank You, BillB3857 & GRNDPNDR, for taking the time to reply. I appreciate the education.

    Thanks, too, for your explanation of how the zero time effects the voltage. That's really cool. (And will be very useful to know in the future!)

    Yep- It's a typo on the rectifier output.
    I am reading 26.1 vac input & 22.9 vdc output. That's measuring directly at the rectifier's pins. I used 26vAC for the input voltage, because 26 matches the transformer operating the boiler & controller. (RMS meters)

    The reference point for the DC measurements was with the positive probe clipped to the board at the copper trace where the LM's 'Output' pin is soldered & the negative clipped to the 'Ground" pin's copper tape (center leg in this situation.)

    BillB's suggested capacitor worked great.
    I tried a 47μf capacitor across the output.
    I now have 11.8 volts DC with the LED as the load. That's right within the specifications from the datasheet. Then, I added some arbitrarily chosen resistors & every measurement was between 11.8 and 12.05. Can't ask for better!

    I am now guessing the datasheet drawing rightfully expects the input dc power to be already filtered & have little or no ripple. This leads me to ask what the purpose for the small capacitor they show is?
    Perhaps to cleanup any induced ripple picked up by the conductors between the dc supply's filter and the LM7812 input? (Just an uneducated guess)

    Is the large voltage drop across the rectifier just a characteristic, or did I arrange it wrong as GRNDPNDR mentioned?

    Since this is not yet permanently built, am I better off using a simple voltage divider instead of the LM7812?

    Thanks Again!
  5. bertus


    Apr 5, 2008

    There should be a buffer capacitor at the INPUT of the regulator.
    The value will be dependend on the current needed.
    Values between 1000 μF and 4700 μF are quite common.

  6. BillB3857

    Senior Member

    Feb 28, 2009
    The purpose of the small value caps on the input and output of the 7812 are to reduce the possibility of high frequency oscillations caused by the regulator getting into a fight with itself. Think of the cruise control on your car. If, when it is trying to accelerate, it applied too much throttle to fast, then when the car goes too fast because of that, it reduces the throttle too much and it slows down too much.. You can see what will happen with your speed. Same thing in electronics. The caps tend to keep things a little smoother.
  7. PGB1

    Thread Starter Member

    Jan 15, 2013
    Thank You both for sharing this knowledge.
    Learning is fun for me. Well, trying to learn is sometimes more accurate in my case.

    Is the buffer capacitor at the input of the regulator that you mentioned Bertus, installed parallel to the small 0.33μf from the datasheet drawing? What is its job in this particular circuit? Additional smoothing? (Sorry to pester with so many questions. I'm trying to learn all I can while I can before my brain's 'Learning Window' slams shut for another 20 years.)

    Thank you for your explanation about oscillations, BillB3857.
    That's a neat analogy about cruise control. It described the possible oscillation problem very nicely.
    Maybe my wife & I need a capacitor across ourselves & the house thermostats. 75, 65, 75, 65 ... (Except the stat in the Girl Cave. I'm too smart to mess with that one.)

    Thanks Again for helping guys!
  8. sheldons

    Active Member

    Oct 26, 2011
    Just to show you how the circuit is put together....
  9. #12


    Nov 30, 2010
    For a DC power supply, the first filter cap should be right at the output of the recticier(s), and it is a big one. There are other cases where the circuit is intentionally run on full ripple, but that's not what you're doing today.
  10. Audioguru

    New Member

    Dec 20, 2007
    The huge capacitor filters and smooths the 100Hz or 120Hz from the rectifier.
    Its high value makes it good at low frequencies but its large size causes it to have inductance so it is useless at high frequencies.
    The 0.33uF ceramic capacitor is small and is excellent at high frequencies so it prevents the regulator from oscillating at high frequencies.