TAS5706: 15dB attenuation at 20kHz

Hi David,

In your test run, were any of the BiQuads programmed? 15dB attenuation at 20KHz is not expected, unless a bi-quad filter response is programmed for it.

For a quick check, please try the following...

Power up EVM, & connect  the GUI and then assert RESET on controller board to restore register defaults. Next, use I2C memory tool in the GUI to run the following initialization script & check the freq-response.

3652.TAS5706_Init_Script.cfg

Also, please specify the test details (Vcc, Load, Power, Input source, measuring equipment etc.)

-Ravi

Hi Ravi,

1-I found where the problem comes from: it is due to analog input of MC57xxPSIA board. It can be solved with short on input resistor. No attenuation is 3.7dB.

What should we suppose to have (there is no frequency response measurement in datasheet)?

2-I have another question: The system I want to test is as follow: 2.1 system: 2x12W in BTL for satellite speaker (TAS5706), and 1x40W (PBTL) for subwoofer (with slave amplifier TAS5602).

I have tried and this configuration and it works well if TAS5602 is in 2xBTL. I would like to use TAS5602 in PBTL on 3.2 Ohms load. Before testing it, I want to confirm with you: Do I just have to connect outputs in parallel? Is there a specific I2C configuration of TAS5706 to allow PBTL on slave amplifier? Do I have to change interchannel delays, or other config?

Thank you

David

Hi, David,

I'd like to chime in on #1... What resistor did you short? Is it R44 and R45 shown on page 19 of this document? I calculate the -3 dB point to be 160k, so I don't see this LPF causing a problem... Unless our vendor screwed up some component values.

For #2, I will let Ravi answer that.

-d2

Hi Don,

I can't open the document in your post, but you can have a look at schematics on TAS5706 EVM user guide document (p27 of  application-notes.digchip.com/001/1-2307.pdf). I have added jumpers JP2 and JP3 to short R39 and R40.

4.99k resistor with 0.01uF capacitor is lowpass filter of cutoff freq = 3.2kHz.

If I short the 5k resistor, attenuation @20kHz is reduced by 10dB (it was -13dB, and now it is -3dB).

I think that the 3dB attenuation that remains is due to PWM filter (I made my measurements on 4 Ohms load resistor). My speaker is 6 Ohms load, so it should improve a little.

Do you confirm my comments?

Thank you.

David

Hi Ravi,

Can you confirm the 2nd question about PBTL of TAS5602 amplifier?

Thanks a lot

David

Hi David,

On the analog-input response, you are right. The cut-off frequency is determined by the resistors R39,R40.

On the TAS5602, are you using the MC5601 controller board to generate the PWM's?

-Ravi

Don,

no I use tas5706 pwm output to drive tas5602 power stage.

i manage to make this work with tas5602 in 2x BTL.

i just wanted to confirm that i just need to connect outputs in parallel to make it work in PBTL. Or is there something to do in pwm driver configuration (tas5706) for PBTL?

thank you.

david

Hi Ravi,

no I use tas5706 pwm output to drive tas5602 power stage.

i manage to make this work with tas5602 in 2x BTL.

i just wanted to confirm that i just need to connect outputs in parallel to make it work in PBTL. Or is there something to do in pwm driver configuration (tas5706) for PBTL?

thank you.

david

Hi David,

To get the TAS5602 EVM working with the TAS5706EVM, did you blue-wire the connections from one board to the other? 

For PBTL mode, the outputs will need to be tied together post LC-filter. The register settings will depend on which audio channel  (left or right) you want to use to drive the mono-PBTL mode.

I am guessing that you are specifically looking for closed-loop amplifiers in your application (along with 192K Fs rate support with I2S). Some of our newer open-loop devices, allow a pre-filter PBTL, which can help reduce the BOM content. If you can share the system requirements, I  would be happy to look into the optimal device fit for the application. 

-Ravi

Hi Ravi,

Yes, we connected TAS5706 to TAS5602. As I said we managed to make TAS5602 work as salve of TAS5706. We tested sucessfully TAS5602 in 2xBTL output (controlled by TAS5706 PWM output).

We tried to connect in parallel the two BTL outpout  (of TAS5602EVM) and it works well. But I would like to confirm that my I2C script is correct and that there is not a specific configuration to change from BTL to PBTL.

Our system is 2.1: the two stereo channels are provided by TAS5706, and the mono channel for subwoofer is provided by TAS5602. So we mixed Left and right  for mono channel PBTL.

Can I send you by email my I2C script and details about our project?

thank you.

Hi, David,

Ravi is on vacation right now, but he'll be back on Monday.

Send him a friend request, and then you can send him files, etc, privately thru that interface. But, let's try to keep generic comments here for the benefit of the community. But, I understand that you don't want to publicly disclose too much about your project.

-d2

I encountered this issue too. In document SLOU299, 12/2010, page 19 schematic, when I installed JP2 and JP3, the analog input frequency response was quite flat (down about 0.25dB @20KHz). However if I removed those jumpers to enable the 6dB input attenuation, there was severe high frequency rolloff of nearly 6dB at 20KHz.

Eventually I found the design error that caused this. The attenuator formed by R39 and R41 (left) and R40 and R42 (right) has no frequency compensation for the .01uF load presented by C42 (left) and C41 (right).

If one places a 0.01uF capacitor across R39 and R40, then flat response is restored for the 6dB attenuation setting. Note that a low impedance source must be used for flat response in either case because of the heavy capacitive load.

HAL