Line out, Line in and Headphone connections

[Panu]

Hi,

There are no analog paths from inputs to outputs inside the VS10xx chips, but the ADC's and especially DACs are surprisingly good quality.

-Panu

[tvali007]

Hi Panu, Pasi,

Thanks for inputs. I finished my PCB few days ago, and tested the "Digital loop" trough VS1053.
I set (SM_RESET)|(SM_SDINEW)|(SM_LineIn) |(SM_ADPCM)|(SM_STREAM); The loop starts to work but with a lot of noise on background. The music is loud and clear, no clipping (except the noise) (like an AM Radio)
The output path works ok, I played some MP3 and there is no noise.
I mention that the ground is connected on both sides (VLSI board and Source board.)
The analogic schematic for Line in is show in attachament :

If I touch the 1nF capacitors the noise is attenuated
Did you encountered a similar situation so far? Do you think it has any connection that I didn't connected all AGND pins to AGND? I checked them with a multimeter and seems that they are common.

Thanks in advance for any feedback.

Regards,
Valentin.

[Panu]

Hi!

Sounds like a bad contact or something, but...:

Make sure that you don't have auto gain set and set digital gain to 1.

I don't see 1 nF capacitors, do you mean the 10 nF capacitors?

What is the amplitude of your sound signal in the analog input pins?

When it works, you should see about -85dB to -90dB noise floor when input is Line level (1 volt RMS).

Can you compare to the line input connection of the VS8053 Developer Board?
http://www.vlsi.fi/en/support/evaluatio ... board.html

-Panu

[tvali007]

Hi Panu,

The connections were ok, I followed your instructions, seems that AGC block was creating the noise problem.
It has no gain limit selected and I think that always amplify the smallest noise to maximum value it could.
I've set the gain to 1 and now is ok.
Regarding capacitor question, I wanted to say 1uF cap .

Now it works well, I inject from a PC for the moment, I have around 2V peak to peak.

Thanks for support,

-Valentin.

[ags]

Is there a recommended single circuit that will support both earphones and active devices (e.g. home FM transmitter) being connected to the VS1011E analog output signals? If the FM transmitter ground is isolated from the VS0111E ground (preventing a ground loop as discussed in the AppNote "VS10xxan_output" in section 3.2) will the simpler earphone circuit be sufficient (as provided in the Sparkfun breakout boards and included in the example connection schematic in the VS1011E Datasheet).

Thanks.

[Panu]

Theoretically yes, but can you in reality be sure that the transmitter is really isolated from the source as they might be connected together in a way that you cannot foresee (via a power suppy, via other cables in the system, via other apparatus, via physical contact,...)

There's not a single great way to do it. The simple way is to add coupling capacitors (up to 100 uF depending on how much bass you want to the headphone) to LEFT, RIGHT. This is safe, but you'll get a rather strong "pop" sound when you switch on the VS. Eliminating the "pop" ultimately leads to elaborate declicking schemes. You can use analog switches to cut off the VS-to-capacitor path for the time it takes to bias the capacitor via a small resistor over the analog switch. Or if your design uses negative analog voltages, you can use an op-amp as a differential buffer to re-center the signal aroung 0V. In that case you would use GBUF signal as the other differential input.

-Panu

[ags]

Panu:

I meant to thank you for the answer. I can be pretty assured of no ground loops in this current prototype because the driven equipment has its own supply which isolates it (through the transformer) from the supply to the VS1011e (also isolated through a transformer). I may try a more elaborate design in a later iteration.

[jcheriton]

Hi,

I know that this topic is getting a bit old now but the title fills my needs...
I am designing an Ogg Vorbis recorder/encoder around the VS8053 and will not be using the headphone output. Is it OK to leave these 3 pins (LEFT, RIGHT, GBUF) floating if they're not in use?
If not, what is the best way of dealing with these outputs?

John

[Panu]

Hi and thanks for your post. It's a very relevant question that comes up every now and then.

Is it OK to leave these 3 pins (LEFT, RIGHT, GBUF) floating if they're not in use?

No. If they are left floating, the analog driver may start to oscillate in some conditions, which will affect the power consumption and the performance of the IC.

If not, what is the best way of dealing with these outputs?

You need to have the RC components inside the blue circle in the original post. You can also check the required RC values from the VS1053 datasheet.

-Panu

[jcheriton]

Thank you Panu for your reply. The RC components are already on the prototype PCB but I was interested if they could be removed to give me a little more space and offer up a few sacrificial parts for the inevitable cost-cutting exercise that will precede the design of the production units.
I now know that they're necessary so will leave them in peace.
Hopefully this post will help others.
Could this information be outlined in the data sheet at some stage? ..or maybe it is but I've simply not seen it.
John