LM1973 acting strange

[Wynand]

Hi. I've been prototyping a 6 channel pre-amp for my PC.

LM1973 volume control IC controlled by an ATMega32 from Atmel.
I use a pot on one of the ADC pins for volume.

Q1. Yesterday everything worked fine, no pops. Today when I got to work and switched it on and changed the volume I hear pops. I take it this is when the new values is set in the LM chip.

The only thing mentioned in the datasheet for pops are JFET opamps on the output... I've got TL072's on it as buffers.

What could've happened?

Q2. Since the input impedence is "nominal 40K Ohm" , do I need input buffers, or is it just better to have them there.

P.S. All the opamps are just buffers, ie output to -in shorted.

[Wynand]

So... I've put the second channel on aswell.

I've noticed that the pops are way louder when there is no music
( ), so I'm thinking of being like a software designer...

It's not a bug, more like a feature... when there's no music then the pops will tell you how loud the setting is, so your warned!!

[paulb]

Maybe noise from the micro is getting into the audio path? Can you look at it with a scope?

[Wynand]

I can, but the only time the pop noises apear is when you turn the pot.

I've got a 20x4LCD on the mcu and the volume bar will only increase/decrease every 12 bits/value.
The volume on the other hand will change every 3 values and that is what I hear on the speakers.

This makes me conclude that it is the volume pot itself and not the MCU.

I'll give it a look anyhoo.

Thank for the reply

[Wynand]

I've had a look on the scope, but I don't know what I'm looking at.

With the scope dc couples the beam moves up/down on turning the pot.

With it set on ac the beam moves up/down and then settles on 0 again.

I was expecting to see a spike of some sort.

I dunno.

[Wynand]

You'd think with the fact that they say, click and pop less attenuation, and that the only time they mention how to reduce popping is with a jfet opamp, it wouldn't be that difficult.

[I_Forgot]

Try disconnecting the source from the input, or even shorting the input, then adjust volume again. If there is DC coming from the source you may hear pops. A blocking/coupling cap can fix that sort of problem.

I used the LM1973 a few years ago for a similar project. It worked great, no pops, but the board layout was a b**ch. I couldn't figure out what the designers were thinking of when they set up the pinout of that chip. The PGA2310 has a much more layout friendly pinout and equally great performance.

I_F

[Wynand]

So I've tried shorting the input, but to no avail...still popping.

[Wynand]

OK!!!

Success!

My circuit is a buffer followed with LM1973 and then an output buffer.

The problem was that even though I had a input cap before the input buffer, I had no resistor going to ground....

Everybody scream DC offset!!!

So, a newbie learns another lesson...

Thanks to those who replied.
After thinking some, I've decided to forgo the input opamps alltogether since the LM chip has a constant input impedance of 40k.

Later...

[jackinnj]

Make sure that the control lines have a 10K resistor to ground -- not for the popping but to remove "uncertainty" -- I have always found that in audio and data acquisition circuits the clock, data and chip-select lines like a logical zero to be a logical zero and not some 'tween state.

one of the folks from Parallax pointed this out in the 1990's -- same advice appeared in Nuts n Volts -- it costs next to nothing.