couple line signal to analog microphone circuit

Discussion in 'General Electronics Chat' started by Dennisdb, Jul 14, 2015.

  1. Dennisdb

    Thread Starter New Member

    Jul 14, 2015
    10
    0
    I wish to couple two units that function as a light organ:
    The first is a microcontroller that plays an MP3 file. The output is a line signal meant for a headphone or something similar.
    The second unit is a discrete circuit using a simple microphone that picks up sound and steers the light intensity of a lightbulb accordingly.

    While the µ-controller runs on 5V DC the other circuit runs directly on 230V AC (without any galvanic protection)

    My problem is that the microphone also picks up ambient noise from people talking.
    First question: Can I just remove the electret condenser microphone and put the line signal to it?
    Second question: what are possible solutions to solve this?

    Thanks in advance!
     
  2. MikeML

    AAC Fanatic!

    Oct 2, 2009
    5,450
    1,066
    To make it safe (isolation from the line-operated organ), you should use a two-winding audio isolation transformer (like would be used in an old telephone modem). You have signal level to burn, so matching impedances is not necessary... Even an old transistor radio audio output transformer (500Ω ct to 8Ω) would work.

    The MP3 player puts out ~1v of audio. You only need a few mv at the organ end, so in addition to the transformer for isolation, some sort of attenuator (pot and resistors) will be required to drop the signal amplitude to go to the mic input...

    Like this:

    mic1.gif
     
    Last edited: Jul 14, 2015
  3. AnalogKid

    Distinguished Member

    Aug 1, 2013
    4,542
    1,251
    Agree with the circuit, BUT - neither of the transformers mentioned, and especially a transistor radio output transformer, are rated for 110 VAC isolation. An audio output transformer from a tube radio would have sufficient insulation (but no UL safety certification). A safer way would be to run the light organ through an isolation transformer, but even that will not prevent lethal voltages at the audio input point if the right combination of parts fail.

    ak
     
  4. MikeML

    AAC Fanatic!

    Oct 2, 2009
    5,450
    1,066
    Good point about the transistor radio transformers. The telephone transformers (from modems) I played with had a hi-pot test at ~6kV...
     
  5. Dennisdb

    Thread Starter New Member

    Jul 14, 2015
    10
    0
    Thank you

    Do you think I should combine the safety transformer for the mains as well as an audio transformer?
    How should I choose the transformer? Which parameters do I need to take into account? I found this block: http://www.mouser.com/ds/2/281/kmp_da100-185018.pdf
    Would it suffice, or should I just try to find an old modem? :)

    I'm honestly not too worried about personal safety at this moment, it's rather the safety of my microcontrollerboard that concerns me. ;)

    (I have some knowledge from my robotics & automation education, but am fairly new to the practical side of electronics. So forgive me if I ask the wrong questions.)
     
  6. AnalogKid

    Distinguished Member

    Aug 1, 2013
    4,542
    1,251
    I would go with the two-transformer approach. Besides the safety issues, it prevents ground noise problems between the two circuits.

    Sticking with the schematic in post #2, the audio transformer can step down the audio so the pot has a better adjustment range, and improve the driving impedance into the light organ for reduced noise. In round numbers, earphone outputs make volts and microphones make millivolts. The transformer can get you some of the attenuation you need, and the pot does the rest and adds adjustability.

    The transformer you linked to is for high speed digital data, not low voltage analog signals. A modem transformer would work, but is not optimal. The turns ratio usually is 1:1, where you want something more like 10:1; and they are made for a relatively narrow frequency band that does not have much bass. This is for blinking lights, not listening, but lights are slow responders, and the bass is what does most of the controlling that you can see. As mentioned above, a transistor amplifier output transformer will do nicely, something with a 1K or 2K input impedance and a 4 or 8 ohm (or 8 ohm center-tapped) output. This will get you 20 dB or more attenuation. Follow that with a pot of 1K to 10K.

    Mouser always has been better than Digi-Key for these kinds of parts. Their site is down; I'll look later.

    Can you provide a schematic of the light organ?

    ak
     
  7. Dennisdb

    Thread Starter New Member

    Jul 14, 2015
    10
    0
    I soldered the two circuits together with an audio transformer in between. Unfortunately the light organ did flicker a lot, and did not seem to relate to the played sound. After adding an RC filter between transformer and lightorgan it was way better and allthough I do not know exactly what I did did, the result is good enough for the application.

    The schematic is a light organ kit from Velleman (attachment) and the sound comes from an MP3 players' line output.

    I have a basic knowledge of electronics, and know what each component is capable of doing. However, I am wondering how the components in this circuit work together: I'm curious about the workings of the transistor-cascade do and how the RC's are dimensioned/designed for the low/mid/high filtering. Can you explain some more about these?
    (I understand the power conditioning circuit with diode, zener diode, L and C)

    Mod Edit:

    I have deleted your schematic because transformerless power supplies such as yours are a violation of our TOS. They are dangerous and can not be made safe. Use a wall wart or a transformer power supply instead.

    Wendy
     
    Last edited by a moderator: Jul 15, 2015
  8. Veracohr

    Well-Known Member

    Jan 3, 2011
    552
    76
    The transistors form a 2-stage amplifier.

    The low-pass filter cutoff is determined by 1/(2*π*R6*C4).

    The mid band isn't filtered.

    The high-pass filter cutoff is determined by 1/(2*π*C5*RV4).
     
  9. Dennisdb

    Thread Starter New Member

    Jul 14, 2015
    10
    0
    Thank you veracohr. :)
    I aso found some clear articles here, that explained a lot to me.

    You say the mid band isn't filtered, does that mean the full signal range is available on the mid range, or does the filtering on the other bands affect the mid band?
     
Loading...