ADAU1761 codec input-output attenuation

Hello naimark,

The DAC de-emphasis was enabled in your project. That rolls off the HF quite a bit. It is on the Digital Setup tab in the DAC section.

The level issue may be partially with your setup. Here is what I get with the Audio Precision we use in the lab. With this setup I see an 8dB loss.

When I send a 0dBV signal into the differential input I find when I look at the level with a meter in the DSP, the level coming into the DSP is about 1dB low. This is partially due to the loss that I have because the input impedance is fairly low on the PGA and the AP has a 40-ohm output impedance. Then there is the gain error of the ADC which is always on the side of not clipping with what is supposed to be a full scale input. So this 1dB is usually what I see.

Then from the DSP to the DAC there is a level difference due to the design of the DAC. 0dBFS is only 0.92Vrms so you lose about 0.75dB just due to the full scale DAC output. So we are roughly at a 2dB loss which leaves 6dB. I recall the designers setup the analog output mixer section to drop 6dB so that if you turned on both outputs from the DAC it would not clip the analog output stages. There may have been other reasons they did this but it is setup for a loss.

Now you are seeing -15dB. Place a meter on the signal in your DSP schematic and see what the level is in the DSP. See if it is too low coming in or the loss is on the output side. The meters in SigmaStudio are setup to read -3dB when you have a full scale peak to peak signal applied.

Then you can setup an oscillator and send that out of the DAC. The Oscillators are at exactly 0dBFS. Then see what you get on the output. It should be around -7dB unless there is an issue driving the load you are going into. I do not expect that but it can be an issue.

Dave T

Hello Dave,

Thanks for the fast response.

Per your comments:

I also using AP audio analyzer and familiar with generator output impedance problems.

That why I’m measuring levels directly on the DSP pins.

Per your recommendation, I placed level meter between input and output. With incoming level 0dbu the RMS meter showing -14db level, what is telling me that something is in ADC setup happening.

About DAC – thanks for pointing de-emphasis setup. It is quite aggressive. When I disabled it, I got the flat response.

Internal sine tone gave me 1.5db internal VU meter reading and the same analog balanced output level.

All of that telling me that I missed some kind ADC setting.

Thanks,

Mark

Hello Mark,

There are a few things you said that I am wondering about.

Which meter are you using in the DSP that gives you 1.5dB reading with the internal oscillator? Then you said it gives you the same output. Same as what? 1.5dB? -14dB?

Which Audio Precision are you using? We have the 2700 series here and some APx series. Let me know which you have and then I can setup a project in SigmaStudio and a test file for the AP and send you both. This way we can be comparing the same things. There are so many settings on the AP that can change your readings that we can spend weeks here going over it using the forum.

Dave T

Hello Dave,

The meter in Sigma Studio, I’m using, is Single Level Detector - reading level is =1.5db. Also, I add in series the Single Level detector with Numerical Display. Reading = -3db.

I am using external RMS meter for the level measurement. I have AP525 and usually confirming analyzer input readings with external meter or using the second channel to measure levels directly on a board.

The output level is 1.5db the same as internal sine source VU meter reading.

Thank you,

Mark

And one more thing:

When I applied 1KHz 0dbu on the Left balanced input, internal VU meters readings are:

Single Level Detector - reading level is = -14db

Single Level detector with Numerical Display Reading = -19db

Thanks,

Mark

Hello Mark,

Thanks for your info and your patience. You gave me a bit of a challenge today and ran into something I had not seen before. Also, usually when I am using this part to test things I use the single ended inputs and headphone or stereo single ended outputs. So I had to pull out the balanced cables and verify the test setup. I went between the two different series of APs to cross correlate. Finally after all this, I was seeing exactly what you were seeing. Then I studied the problem for a bit, scratched my head some then I turned on the power to the ALC block and poof! Problem gone! What is strange is that once I turn it on it is good even if I turn it off. I wonder if there is a digital bug that once it is turned on it cannot be turned off. I had to power down the part to test it again.

So go to the DSP page and turn on the power. Even though you are not using the ALC since it works by changing the gain of the PGA it must need to be powered on to use the PGA.

Now I get about -1.9dB on the output. I get -0.7dBV out of the DAC when using the internal DSP oscillator and so we are losing another 1.2dB-ish on the input side which is attributable to gain error and things like that.

The Meters. The Single Level Detector is one of our old routines and is not very accurate. The one with the numerical display is the one I usually use. It still is not meant to be a laboratory measurement instrument but it gives you a decent idea of what is going on in the DSP.

I will attach the test project I used, I think I did not save it with the ALC power turned on, But you can see my test project and we can use a common project if you have more questions.

J084_P_test-DT-Edits.zip

Thanks,

Dave T

Thank you Dave,

The bugs are happened.

Mark