Questions about TDA2003 and ground decoupling

Thread Starter

bloguetronica

Joined Apr 27, 2007
1,541
I'm projecting a stereo amplifier, but in the final stage of my project, I have questions on how I should hook up the grounds with the least humming.
My plan is to hookup the ground from the power supply first to the main filter capacitor, then to the input stage of the TDA2003 (the ground of the volume potentiometer and the ground of the feedback network), then to the amplifier ground and load ground, being those last two grounds decoupled from the former throught a thin wire.

Is this correct?
I tested this on a breadboard using only one amplifier and notice a very small hum.

The idea is to keep the input stage clear of ground pertubations. I'm using a buss configuration.

Here is the datasheet containig the circuit in which I'm based on: http://www.st.com/stonline/products/literature/ds/1449.pdf
 

lightingman

Joined Apr 19, 2007
374
Hi...Nasty thing "ground loops".... Thake all your high current grounds (power and output) to a central point (the main smoothing caps are the best bet).... Do not use this point for the ground on the input or volume control, but try using the ground on the amp PCB itself.....Any problems, let me know.....Daniel.
 

Thread Starter

bloguetronica

Joined Apr 27, 2007
1,541
Hi...Nasty thing "ground loops".... Thake all your high current grounds (power and output) to a central point (the main smoothing caps are the best bet).... Do not use this point for the ground on the input or volume control, but try using the ground on the amp PCB itself.....Any problems, let me know.....Daniel.
But what is the best order for a ground buss? I can't afford to have a metallic case in this project. It has to be made all in PCB. Also, I see that the PCB in the datasheet has the power and output ground (and also the amplifier filter cap as well), connected to the ground of the supply through a thin copper path. I think this is to avoid interference with the input stage, connected with ground close to the supply.

My plan is to use two capacitors, one that should be close to the amplifier (to avoid ground loops), and the other one close to the supply and input stage (the main filter capacitor). The important is to avoid motorboating due to voltage dips caused by an excessive current being drawn by the amplifier and output through ground. So I will use a thin wire between supply (and input stage) ground and output (and amplifier) ground, to limit a little the current going there. Is my assumption correct?

Sorry for not mentioning the cap that I will use close to the amp before.
 

lightingman

Joined Apr 19, 2007
374
Hi again,If you are using the PCB lay-out that is on the data sheet, then you will notice that they have done what I was talking about...I.E. seperate grounds for power connections and input....You need to use what is known as a star point for all the grounds (exept the input ground).....All grounds from the supply and output should be made with a reasonable sized wire 24/02 or something like that....to the star point....A non-screened case will not make any difference to ground loops, but only to input hum pick-ups....A good high value smoothing capacitor is worth using, and as you said, a couple of caps (100uF and 100nF) accross the supply as close to the IC as possible....IC amps are allways prone to motorboating problems, as the ground connection is 9 times out of 10 down 1 lead-out.....Good luck....Speak soon....Daniel.
 

Thread Starter

bloguetronica

Joined Apr 27, 2007
1,541
The problem is in the datasheet they don't specify where the supply ground is to be connected, if near the input or near the output. Also, I don't know if what I've said makes sense. So should I use the supply ground near the input stage (as I was planning, so the input will be attached to a more "solid" ground) or near the amplifier/output stage (here I thing the voltage dips caused by the amplifier and output will propagate to the input stage, which is farther from the supply ground)?

I noticed that there is a third hole in the PCB figure. Hence my doubt. Limiting the current going from the output/amplifier stage is an issue, since this project is to be supplied by an adaptor, which has a long cable, so I need to create a resistance a bit larger than the one present in the cable to ground the output stage without distorting the input stage. I know that the oposite is normally used (ground the higher currents, then the lower currents), but in the TDA2006 datasheet they use the supply ground near the input ground, and the output ground is farther from the supply ground.
 

lightingman

Joined Apr 19, 2007
374
The power supply ground needs to be connected to the thick ground track into the IC, not the input side grond.....I will send a PDF later this evening.....Daniel.
 

Thread Starter

bloguetronica

Joined Apr 27, 2007
1,541
Thanks. That's all I need to know. I was mistaken about the PCB represented in the TDA2006 datasheet. It is just like you are saying now. So I will use the output and amplifier grounds connected to the supply ground and the input stage ground connected to these grounds using a thin wire to allow decoupling.
If you send me the *.pdf I would appreciate, but there is no real need for you to bother.

Thanks by all the help, and thanks in advance, case you send the *.pdf.
 

Thread Starter

bloguetronica

Joined Apr 27, 2007
1,541
Thanks! That helped a lot. But since I'm using a PCB with all, I will use short and thick tracks to connect the output and amplifier stages, and then decouple the input stage from all the rest using a thin wire. I don't know if solid core is the best for this. I think that is the best way I can use the idea (project requires that I use a PCS to support all the components).
 
Top