Help!!!! Volume Control Using TDA8424

Hi to all the gurus here,
I'm new to this forum. I hv been using TDA8424 Volume chip in my project. I manage to control the level using a pic controller.
However the problem i'm facing are... lots of noise. When the music are paused. Turn the device to a maximum level. The noise level is unacceptable. Beside when i sendind the data to the chip, while it is ramping the vol up or down, I can here some "ramping noise". I dunno how it called. It should be ramp smoothly without any additional noise.
I built this unit using 4 pcs of regulator 7812/7912 for the balance-unbalance input and unbalance-balance output. The ground of this circuit is analog ground and is supply from 1 transformer 909Vac. The other similar transformer 09Vac will supply 5V and 12V from 7805/7812 regulator respectively. The +5V/+12V will supply to the PIC and TDA8424. So here will be a digital ground. I had split the digital and analog ground before they join in certain spot of the board. Is there any problem with the power supply section? I'm quite happy with this chip. The only problem is i wans to get rid of those noise.
Even when i sending the data to increse the bass or treble.... the output will somehow mix with "some unwanted digital sound".
Please help....
Many thanks

Hi alvin,

Welcome to the forum. Just trying to help, but please post a schematic. It will make it easier to see what's going on with your circuit.


Torben

here is the drawing.
Let me know if i had did something wrong somewhere.
I manage to control the level.
But too many noise to handle.....

What is the nature of the noise? Can you describe it?

AG and DG should probably be hooked together at one point only, usually near the chip.

I dunno how to describe. The noise is quite unbearable. Okay..... when the volume level is at low level. The noise is barely noticeable (i dun think it is a hum). but when the signal ramps up.. the noise level escalate. Though in a smaller proportion to the signal level. But it is quite noticeable. Besides, when the ramping is in progress....the "ramping sound" is heard (quite noticeable too) it seems like a digital sound...every level it reach u will hear a noise. When the ramping stop, "ramping sound" will stop too.

mneary said:

AG and DG should probably be hooked together at one point only, usually near the chip.

Click to expand...

--------------------------------------------------------------------------------

AG and DG should probably be hooked together at one point only, usually near the chip.

I did tied both AG and DG together but to no avail. The noise still there.

My last resort will change to other Chip for a new trial. I may try out LM4832. Almost give up on this chip. I had two design on my PCB.
One is all share the same ground plane. Which is the shortest path back to the ground point.
Second is split both DG and AG and finally fixed at a point where both are jointed together

Try adding 200 ohm resistors between uC and TDA8424 on clk and data lines. Could use 1k resistors if you send data slow.

I know of an amplifier that has the analog audio parts on one PCB and the micro on a second board. The data lines have an opto-isolator on the analog board. There is no direct connection between boards that way.

There are several application notes on digital pots that talk about a “zipper” sound when you change the volume control. Some of the digital pots only change the settings on zero crossing of the audio signal to reduce this noise. I did not look up your data sheet to see if the TDA8424 does that.

Ohh....in my drawing i miss out the 10k pull up resistor for the I2C bus. You mean the 200 ohm place in a series connection? or Bridging the clk and data line?
The TDA8424 did not have any zero crossing detector. Would this be the reason behind for these "zipper" sound?

Adding a resistor to the clock and a resistor to the data line reduces the current flow. I know the inputs are high impedance so there should not be current flow. I am talking about current (AC) on the edges of data and clock.

Take a look at PGA2500. I know of one high end mixer that uses them on the mic/line inputs as cores adjustment. You set it to get the level close and use a different control for most level control after that. There are several parts like the PGA2500. Look at the application notes, there may be help there.

Not shown in your schema, but you are using bypass caps on your Vcc pins of your IC's are you not?
Other thing you might try is running your TDA part on 12Va, since is has analog paths.

ronsimpson said:

Adding a resistor to the clock and a resistor to the data line reduces the current flow. I know the inputs are high impedance so there should not be current flow. I am talking about current (AC) on the edges of data and clock.

Take a look at PGA2500. I know of one high end mixer that uses them on the mic/line inputs as cores adjustment. You set it to get the level close and use a different control for most level control after that. There are several parts like the PGA2500. Look at the application notes, there may be help there.

Click to expand...

Will try to add the resistor. Hope to get rid of the zipper sound first. Before troubleshoot on the other noise

Mikebits said:

Not shown in your schema, but you are using bypass caps on your Vcc pins of your IC's are you not?
Other thing you might try is running your TDA part on 12Va, since is has analog paths.

Click to expand...

I did had a bypass cap of 0.1uf on all the supply pins and locate as close to the chip as possible.
My initial design was run TDA on the 12Va. Which i feel that might not balance the load on the +V and -V. During those time the problem remain. Hence I add another 7812 to produce 12Vb. However all the problem still persist.

Hmm, Sounds like our doing the right stuff. Are you using point to point wiring, or pwb?
If point to point, the wiring is critical. Make sure all your analog signals are routed away from the digits. You have a good solid ground plane I assume. Do you have anyway to make sure your power signals are clean? I love my scope

The sound you hear may be the nature of the beast. Several digital audio volume control chips claim ‘ours don’t make a zipper sound’. Your ear should be able to hear a series of 1db steps.

Mikebits said:

Hmm, Sounds like our doing the right stuff. Are you using point to point wiring, or pwb?
If point to point, the wiring is critical. Make sure all your analog signals are routed away from the digits. You have a good solid ground plane I assume. Do you have anyway to make sure your power signals are clean? I love my scope

Click to expand...

The problem is i dun hv any scope to check if i had a clean power supply. I check the multimeter the 7912 had produce a voltage of -12.17 Whereas the posivite i got 12V. This two voltage is fall onto the balance-unbalance and vice versa. This could be the reason of noise introduce into the circuit. But before i intro these cct, i use to connect the signal directly to the TDA, the zipper sound is definitely there.
I think i do had a good solid ground plane. I read lots of article and this finally comes to my mind that a return signal needs to be as short as possible. Hence a good ground plane was incorporated onto my design.
Wat is pwb?

ronsimpson said:

The sound you hear may be the nature of the beast. Several digital audio volume control chips claim ‘ours don’t make a zipper sound’. Your ear should be able to hear a series of 1db steps.

Click to expand...

Yes this is exactly wat i hear.

ronsimpson said:

Adding a resistor to the clock and a resistor to the data line reduces the current flow. I know the inputs are high impedance so there should not be current flow. I am talking about current (AC) on the edges of data and clock.

Take a look at PGA2500. I know of one high end mixer that uses them on the mic/line inputs as cores adjustment. You set it to get the level close and use a different control for most level control after that. There are several parts like the PGA2500. Look at the application notes, there may be help there.

Click to expand...

Hi Ron
Thanks for your info. Your idea help indeed. The zipper had been very much lower compare the previous level. I use 220 in the first trial. The result get better. But it is still quite noticeable. Then for the second trial I used 3.3k and this porduce a way much better result eventhough it is not totally diminished but it improved a lots.

Besides, to all of you had been follow up this topic. I would like to thank you for your idea and explaination that guide me thru. Today I had manage to solve most of the noise problem. Finally I found the culprit that produce the noise. It is all due to the notebook that I'm using. I used HP pavilion media centre dv1000 notebook. These notebook produce too many noise that had affect my testing an had created so many problem to me. I'm using a USB - Serial convertor and the connection for the RS232 injected great noise into my board. Therefore i keep on suspecting my design problem.
After a few trial, I remove the 12Vb and tied it back to 12Va. The result is still good. To all of you many thanks for your effort.
Now, I'm still aim to totally remove the unwanted noise. There are still some though hardly hear from 2 feet away. Had to put my ears quite close to the speaker to listen to the noise. This is still need improve. Will learn more abt the grounding technique. I still belive it is the grounding issue.
Maybe will try other better chips available for my next prototype.
Thanks again guys.
Best Regards
Alvin

Besides, I think i need to incorporate RS232 isolator into my new design. Anyone with a recomendation?

Problem with TDA8424

I know its been a lot of time since the last post, but let me tell you that I am trying to implement a project with the same chip for university class, and I am not being able to establish a communication with TDA8424. In my case I am using a HC908JK1 microcontroller from Freescale instead of a PIC. However, it would be very useful for me If you could send me the ASM code you used to program the PIC, to know which are the step you performed to make the Volume control works.

Thank you,

Alan