Electret microphone to ADC input

[caringin]

microphone adc

Hi,

I have a problem here. I'd like to use electret microphone as the input for ADC 0809, but the digital output is random. When I use DC input straight to input of ADC, the response is fine, I mean that if I give 5 V DC (not from the microphone), the digital output is 11111111 (assuming that the Vref+ = 5V and Vref- = 0V). I attach the schematic on the attachment. Please comment about the schematic, thank you.

 

[Borber]

mic adc

It is all wrong. Electret must be connected to Vcc trough a resistor (you have probably destroyed it). What is the purpose of graetz rectifier in front of ADC? It plays no role.

Explain what you want to design.

 

[paulux]

adc microphone

First, the microphone is better not connected to supply directly, it should be biased by a resistor.
Second, the function of 2 bjts (I guess) is your pre-amplifier, what is the input impedance and the gain of it? There may be problem in your pre-amplifier. Did you confirm the bias and structure of your pre-amplifier?

 

[caringin]

microphone + adc

Hi,

I have revised the schematic, I misplaced the VCC on the picture, but the actual is in the right place, thus I believe the ECM won't be damaged. I'd like to design the ECM as the input. For example: when I scream to the ECM, the digital output of ADC is 1111 1111. When I speak normally, the output 1111 0000. When I just whisper, the output 0000 1111. When I don't speak, the output 0000 0000. I know the reality won't be that nice, but I am sure you get my point.

The 2 BJT's are for pre amp section. I have tried using LM741, LM390, LM358 but all won't work. So, I just use BJT. There is AC signal from the output of BJT. That is what makes me sure that the preamp section works. Rectifier diode is for rectifying the current. Audio output is AC signal. I assume that ADC input must DC signal, so I use rectifier diode.

Thank you, I need more advices.

 

[Borber]

electret microphone adc

Few comments:
DC voltage on ADC input is probably more than 7V!
Bandwidth of preamplifier is only 20Hz because of 2.2u capacitor connected between collector and base of the last transistor.
Graetz rectifier is of no use. Actually only one of the diodes has rectification function, others are shortcircuited or constantly reverse biased. Upper right diode and capacitor may act as peak detector if input impedance og ADC is high enough (no data).

 

[caringin]

microphone adc interface

DC voltage on ADC input is probably more than 7V!

Do you mean ADC input from the preamp? No, the voltage after preamp is 1.5 V when there's no sound and nearly 4 V when I scream near the ECM. I measure from pin 26 ADC which is input analog DC and the measurement is precisely same as the voltage after preamp, but the ADC won't response what it should. The digital output is random. I have also eliminated the rectifier and use only 1 diode and it's still the same.

 

[Borber]

microphone with adc

Regarding DC level: it is obvious that your transistors are different than my models used in simulation.
It is better to use constant amplitude input signal to test ADC. Voice produces random signal and ADC output is each test diffrent.
Are you shure you want to measure peak value of voice?

 

[caringin]

lm741 schematic adc

Voice produces random signal and ADC output is each test diffrent.

I mean that when it's quiet (no sound), the ADC still went random.

Are you shure you want to measure peak value of voice?

No, I want to make sound controller, but only intensity of the sound, not the frequency as the sensor.

 

[Gorgon]

lm741cn electret preamp

Voice produces random signal and ADC output is each test diffrent.

I mean that when it's quiet (no sound), the ADC still went random.

Are you shure you want to measure peak value of voice?

No, I want to make sound controller, but only intensity of the sound, not the frequency as the sensor.

You say you measure random levels at no sound. From your drawing I assume that your grounds (12 & 5V) is not connected? If not, you will have a floating nonreferenced input. The capacitor on the input of the ADC is left floating too, you need some form of bleeder to discharge it, this is normally done with a resistor. The resistor value depends on the responsetime/ decaytime of your measurement and sample rate.

If your GNDs are connected, you only need a single diode to separate the capacitor from the amp. Due to the DC setpoint of your amp, you will always see a voltage over the capacitor, even with no sound in.

TOK

 

[caringin]

adc for microphones

Hi,

I have connected the ground as you told me so, but it was still the same. I have also used ferrite bead to separate the digital and analog GND and nothing went better. I use microcontroller to process the output of ADC. Do you have any experience having sound as the input for ADC? Please let me know about your experience.