TDA2003 amplifier. Tuning out the oscillation

Discussion in 'The Projects Forum' started by Fenris, Jul 26, 2008.

  1. Fenris

    Thread Starter Active Member

    Oct 21, 2007
    hi all

    I have recently completed a project which uses a TDA2003 amplifier. It works quite well. However It tends to go into oscillation if i go much over 50% on the volume control. I have the data sheet for the TDA and if i have followed it correctly then R15 in my circuit should be 20*R17 which is 94R. 100R is the nearest value and i have changed R17 from the 120R to 100R. It has allowed me to open the volume more but i get oscillation further up the scale. The oscillation is also, now, at a higher frequency it seems. The capacitor C18 should have been an 8n2 but these proved difficult to get hold of so i was advised a 10nF would suffice. Its my understanding that the capacitor effects the frequency range of the output. correct me if im wrong please. I understand the equation for Cx as seen on the data sheet but do not know what 'B' stands for in the lower line. anyway 82R is the next nearest down resistor i could try. Am i doing this correctly? The amplifier layout and components as seen in the attached picture is based on the velleman kit MK4001 7W amplifier. The PCB layout of the amp follows the original PCB that came with the kit. help and guidance would be appreciated, thank you.
    I have marked the resistors in question with the notations from the data sheet for clarification. the only thing i have notice is that C18 and R15 although strung between pin 2 and 4 of the TDA in the velleman schematic and the official datasheet for the TDA their positions are reversed is this critical?

    the signal to the amp is from a LM358 which passes through a 4066/7555 circuit which adds a chopping effect to the input from an electret mic.


    Last edited: Jul 26, 2008
  2. SgtWookie


    Jul 17, 2007
    C18 and R15 can be swapped or as-is. C18 simply blocks DC current. It doesn't matter on which side of the resistor the current is blocked.

    100R should be fine for R15.

    I have an idea that you are getting feedback from the speaker output to the input of the electret microphone. Try moving the electret microphone further away from the speaker.
  3. Audioguru


    Dec 20, 2007
    Is the oscillation an audio frequency or is it a very high ultrasonic frequency?
    If the oscillation is at an audio frequency then it is normal acoustical feedback because the mic can hear the speaker.

    Separate the mic from the speaker with them in separate rooms and a shut door.
  4. Fenris

    Thread Starter Active Member

    Oct 21, 2007
    I will do the separate rooms first. i should have thought of that. i thought it was a particularly high frequency my self. but first things first. thankyou chaps.

    Did the room/door test. I was able to open the volume wider. but then the sounds starts chirping and continues slightly after speech has stopped. when speaking with the volume lowered that chirp occurs on top of speech but stops when speaking stops. perhaps i should put that resistor back up to 120R as per the original?
    I am also going to change where the amp circuit picks up its -ve supply. currently its tagged to the -ve at the end of the main feed which feeds the 7555/4066 part of the circuit which generates the 'chop'. It may not be required but i will add a wire link instead that spans 8mm from the actual -ve feed pad to the -ve of the amplifier circuit. I think how i have it is slightly daisy chained which isnt good really. At least if i change it all areas of the circuit will have there own direct supplys.

    Further experimentation has revealed the source of the chirping. When the 2 filament bulbs are removed the chirping stops. So am i looking at back emf? heres a picture of the part of the circuit. L4 L5 are the links that connect the audio output to the sound to light unit. The small IC is an opto isolator. Q4, 5 are the mosfets that drive the bulbs. Is the solution as simple as a Diode in the positive and negative power feed to the sound to light unit thereby preventing back emf or is a capacitor in the power lines another solution?

    I have re-jigged the PCB pattern C7, C20. R26 ,R28 have been disconnected from the ground line they were on and now have their own. The power lines have also been completely severed so the sound to light is completely isolated from the rest of the circuit except at the opto isolator end. Now either a pair of wire links needs inserting to get power to the S2L or Diodes if required. It depends if the shared ground line was the cause of it or if it really needs diodes. Either way it means that all subsections within the unit take their supply directly from source rather than share here and there along the main runs. So do i diode or not? and what type would suffice? cheers again.
    Picture (left) shows the original S2L picture (right) shows the changes.

    Last edited: Jul 27, 2008