filtering after digital to analog convert

[foxbrain]

hi
what is the best low pass filter can we use after adc dac conversation ?
thanks

 

[crutschow]

How many bits and what is the highest signal frequency?

How much do you need to filter the signal?

 

[foxbrain]

it's an 8 bit and i'm using dac800 i don't anything else about the frequency but i'm using it for human voice ....
my application is microphone->amplifier->adc->dac->amplifier->speaker but the sound is too awful and unclear so i need a good filter after the dac for it ....

 

[crutschow]

I doubt that a filter will solve your problem, but try a simple RC low-pass filter to start with a corner frequency of about 10kHz to see if it helps.

What do you mean you don't know anything about the frequency? Who designed the circuit?

 

[betwixt]

I agree with Crutschow. You could get 'aliasing' problems if the ADC/DAC speeds are too slow, and that can be reduced by filtering the input to some degree but 'awful and unclear' suggests something else is wrong. A filter after the DAC might remove 'sharpness' from the sound but it cannot add clarity. Can you post a schematic, or at least try temporarily removing the ADC and DAC and linking the amplifiers directly together to eliminiate the digital part as being the problem.

Brian.

 

[foxbrain]

What do you mean you don't know anything about the frequency? Who designed the circuit?

as i told you i'm using my voice frequency in this circuit.
for the schematic i'm using most of the circuit in the datasheet:
as i told you before my diagram is as this microphone->amplifier->adc->dac->amplifier->speaker
for the amplifier i've attached the diagram in the file circuit.doc
the adc i used the one in page 11 of the datasheet attached.
the dac i used is in the first page in the datasheet.

 

[betwixt]

I think most of us are familiar with the ICs. It's how you have connected them together we want to know.

I assume you have the ADC wired as in Figure 17 of the data sheet. How are you biasing the analog inputs?

Brian.

 

[foxbrain]

exactly fig 17 ....what do u mean with biasing ? i connected pin 7 to ground and pin 6 is the output of the amplifier. i connected leds to the output of the adc i found that the leds changes depending on my voice

 

[crutschow]

When you build a circuit without understanding it, it's not surprising it doesn't work properly.;-)

If you look at the data sheet for the ADC0804LCN you will see that the input signal goes from ground to V+, thus you must bias a normal audio signal, which goes above and below ground, at 1/2 V+ to keep within this signal range.

For maximum dynamic range, the peak-to-peak audio signal for the loudest signal should go from 0V to V+.

The data sheet also shows that the free run conversion rate is 8888 conversions/sec. The Nyquist theorem states that the maximum audio signal you can convert is 1/2 that or about 4.4kHz (which is low, comparable to phone bandwidths). Any audio frequencies above that will alias into the passband causing noise and distortion. So you need a high order 4kHz low-pass active-filter (probably 4-pole or greater) at the ADC input to prevent this aliasing.

Also 8-bits is a low-number of bits for sending audio (16-bits is typical) and will give a low S/N ratio as well as some possible audible distortion.

 

[foxbrain]

i don't think so because i went to the lab and i had put a function generator as a sine wave to the input and changed the frequency from 500hz to 4khz and more without any changes...........

 

[crutschow]

i don't think so because i went to the lab and i had put a function generator as a sine wave to the input and changed the frequency from 500hz to 4khz and more without any changes...........

Then I guess you don't need our help. ;-)

 

[betwixt]

Agreed - anyone who can defy the laws of physics either doesn't need our help or is beyond help!

Brian.