I have a simple audio circuit that needs to be amplified.

The circuit is a CMOS based circuit, running on 9V coming from a 12V battery pack stabilized with a 7809 IC. I have also an amplifier circuit based on the Texas Instruments LM4950. The circuits are on separate PCBs.

I'd like to power the amplifier with the same voltage source as the audio circuit (the 12V battery pack.)

Now my problem is: when the circuits are powered with two different sources (two 12V battery packs) all is fine and the amplifier circuit sounds perfectly. When I connect it to the same battery pack (directly to the 12V) of the audio generator circuit it sends out only loud, interrupted noise.

I tried to opto-isolate the amplifier voltage source to eliminate the distortion and to filter the voltage with capacitors but it does not work. I also measured current to check if the current absorption of the amplifier is too high for both circuits but I measured only about 100 mA.

Maybe is a silly thing but I really can't find out to resolve it, hope that someone can help me.

CMOS Sound circuit CMOS Sound circuit

Amplifier circuit (straight from datasheet) Amplifier circuit (straight from datasheet)

Connections diagram Connections diagram

  • 3
    \$\begingroup\$ How about a schematic diagram of both circuits and a diagram of how they are connected to each other and the battery? \$\endgroup\$
    – JRE
    Sep 11, 2020 at 19:14
  • \$\begingroup\$ Try powering the amp and input source using different wires on ground. If anything changes re the noises produced then you are on a roll. \$\endgroup\$
    – Andy aka
    Sep 11, 2020 at 20:01
  • 2
    \$\begingroup\$ We're considerably in the dark without more info about your circuits, but if I had to guess it might be "motorboating". This is a kind of feedback -- like what you hear when you put a microphone too close to its amplified speaker -- but instead of sound waves, it is caused by ripples in the power supply voltage. The voltage fluctuates as the power amplifier draws current, and the initial CMOS-based circuit is sensitive to those changes, and around it goes. You either need a better regulated supply, or more decoupling between the stages. \$\endgroup\$ Sep 12, 2020 at 3:33
  • \$\begingroup\$ Thank you all, you are right, I added the schematics and the connections diagram @MarkLeavitt "motorboarding" describes the sounds I get quite good. \$\endgroup\$
    – simone
    Sep 12, 2020 at 15:04
  • \$\begingroup\$ @MarkLeavitt gives a very likely cause: Replace or add to C2 a big electrolytic capacitor, like 1000uF. \$\endgroup\$
    – glen_geek
    Sep 12, 2020 at 15:34

3 Answers 3


The L7809 is operating very close to the edge: it has a 2V dropout voltage meaning it needs +11V input to remain stable. You're giving it an unregulated +12V through a diode so very little margin - any spikes on +12V (due to amp current spikes) may freak out the L7809.

You can increase your margin by:

-making cap at L7809 input larger

-running Q1 collector from 12V (so current draw on 9V is CMOS only - negligible)

-using an L7808 or even L7805

But maybe best to just remove L7809 entirely, run CMOS from 12V too. CMOS is happy at that voltage and will be more tolerant to VDD fluctuations without a spastic regulator in the way.

  • \$\begingroup\$ remove L7809 entirely, run CMOS from 12V. Exactly. \$\endgroup\$
    – Janka
    Sep 12, 2020 at 19:19
  • \$\begingroup\$ Thank you @td127. I tried to remove entirely the 7809 as you suggested but unfortunately, nothing changed. The amplifier continues to work well with a different power and to produce noise if powered with the same battery of the CMOS circuit. I also tried to substitute C2 on the amplifier with a bigger capacitor (1000 uF) in parallel with the 1uF as other users suggested, but also in this configuration (with the 7809 removed) nothing changed. However, thank you for your explanation of the 7809 on 12v that is really useful. \$\endgroup\$
    – simone
    Sep 13, 2020 at 13:57
  • \$\begingroup\$ Interesting. There's really no good reason this shouldn't work so I think your initial thought that it's some "silly thing" is most likely correct, i.e. something isn't hooked up the way you think it is. What are your debugging tools? I assume you have a multimeter, but do you have access to an oscilloscope? \$\endgroup\$
    – td127
    Sep 13, 2020 at 18:34
  • \$\begingroup\$ Hi, @td127 I have an oscilloscope but I use it just only to "see" audio waveform. Honestly, I do not know how to use it to debug the circuit. \$\endgroup\$
    – simone
    Sep 15, 2020 at 13:06

Thank you for adding the schematic, it helps a lot, and supports the prevailing diagnosis that your sound generator is not getting clean, regulated 9V power when it and the amplifier are both running off the same 12V battery. As the amplifier drives the speaker, the battery voltage will drop (from the DC draw) and ripple (from the varying audio waveform).

Here are options to try:

  1. Use a separate battery for the sound generator. Unsolder the 7809 and short the IN to the OUT. Run + and - leads from each battery to its respective circuit, and run a twisted pair (or shielded cable) for signal and ground between the sound generator output and the amplifier input.

  2. If you don't want to use a separate battery, you could replace the 7809 with a low-dropout regulator like the LM2940. Check the pinout.

  3. Use a lower-voltage, and low-dropout regulator.

  • \$\begingroup\$ Hi, Mark I resolved using a transformer to "eliminate" the ripple isolating the voltage coming from the sound generator (I don't know how much technical accurate this is). Thank you for your explanation that was really useful to understand the problem. \$\endgroup\$
    – simone
    Sep 15, 2020 at 13:02

Here the solution I found: I inserted a 1:1 audio transformer connected between the output of the CMOS sound generator circuit and the input of the amplifier to isolate the input. With this configuration, the amplifier has no problems sharing the power source with the CMOS circuit and all the circuits work well. I used a Bourns LM-NP-1001-B1L but I think that any 1:1 audio transformer such as this should work fine as well.

Hope that could be useful for other people with the same "motorboarding" problem!


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.