RE: TAS2563: Multi-Stage Volume Ramp During Playback After Device Turn-On

Hi Joel,

I'm splitting this thread as it was no longer allowing me to reply, perhaps due to origin time expiring.

Notice that excursion is dependent on how much voltage is applied to the speaker. You're not usually pushing the same amount of voltage over all frequencies unless you play white noise or something similar.
The limitation in terms of voltage from the amplifier is given by 8.92V as mentioned before, if speaker is unable to reach its maximum level at the maximum voltage from the amplifier, then protection would not be needed from theory. So it basically depends on how sensitive is the speaker at its resonance frequency.

Best regards,
-Ivan Salazar
Applications Engineer

Quote from thread:

Hey Ivan,

In theory, should there be no output if I set Xmax to 0mm?

Reply:

That would be the expectation, but since that is not a real use case the algorithm will most likely behave different. Consider this is math running on the background and 0's can cause different kind of results depending on where these are applied.

Best regards,
-Ivan Salazar
Applications Engineer

Hey Ivan,

Our speaker is rated at 2.83 Vrms (4 V pk-pk for 1 kHz sine wave) @ 1 W. Does this mean that the speaker protection will never be engaged and that the limiter is kind of taking the role of acting as the speaker protection? Currently have the limiter set to 5V.

Is the limiter before or after the speaker protection?

Hi Joel,

I think the speaker rating is close to the maximum output from the amplifier, the protection may trigger only at the resonance frequency depending on the sensitivity.

I assume you're using the limiter from Device Control panel. This is applied after the protection algorithm.
You may notice there is also anti-clipping protection as part of the protection algorithm, you may change the settings of the limiter to double check if it's the anti-clipping or limiter acting.

Best regards,
-Ivan Salazar
Applications Engineer

Hello Ivan,

Is the signal after the equalizer normalized if boosting before it goes into the compressor? For example, if I send a 1kHz tone and there is a boost of +3 dB at that frequency and the amplitude is 1.0 max going in, would the level be normalized back to 1.0 max after the gain is applied?

Hi Joel,

The limiter will be applied after all preprocessing, and before that is the speaker protection. Any gain applied before each of these will not be applied finally at the speaker.

Best regards,
-Ivan Salazar
Applications Engineer

Is is possible to monitor the output of the EVM before the speaker (e.g. seeing a real-time read-out of DSP)

Hi Joel,

That would be the Echo Reference data (also called Audio Out). This has all processing applied to the input signal.

Best regards,
-Ivan Salazar
Applications Engineer

Hey Ivan,

Where does the boost converter reside in the signal chain? Before/after limiter?

Hi Joel,

Boost is looking at the signal envelope after the limiter, it is at a similar stage as the Echo Reference data.

Best regards,
-Ivan Salazar
Applications Engineer

Hello Ivan,

I am still having trouble completely removing the ramp...as it stands the most effect in reducing the ramp was setting all the thermal and excursion speeds to 20 ms...below this value i saw no change in ramp rate and above this value the ramp rate increased considerably. I will email you my current working ppc3 project.

Hi Joel,

I've got the details over email. I'll be working on that and will provide some feedback over next week.

Best regards,
-Ivan Salazar
Applications Engineer

Hi Joel,

It's been a while since there was any activity on this thread. Checking in to see if you are still waiting for support on this issue?

Brian

This issue has not been resolved. The conversation moved offline to an email thread with David McKinney where were were discussing sending speaker samples for TI to evaluate. I have not heard back from him yet.

Hi Joel,

I'm following up with David on the shipping address and requirements to complete the characterization and tuning.
You should get some feedback soon.

Best regards,
-Ivan Salazar
Applications Engineer

Hi Joel,

I completed the characterization on one speaker. I'm requesting Field team to send you a new PPC3 file with this data, can you please test on your side?

I'll continue with audio tuning, but want to make sure you're not getting the same ramping behavior as you did before. On my side I verified excursion stays within the limits with a few audio tracks, and estimated temperature is staying below 70C-delta, I'm checking on fine tuning to take it close to 80C-delta limit.

Best regards,
-Ivan Salazar
Applications Engineer

Hello Ivan,

I did get the ppc3 file from McKinney. Can you characterize all 4 speaker samples I sent? From my experience, the thermal parameter values varied greatly. I just looked at the parameters for the first sample and they seem to be relative in line with what I was getting. 

Hi Joel,

I'll be doing that for comparison, will let you know the findings.

Best regards,
-Ivan Salazar
Applications Engineer

By the way, have you test this on the speaker? Are you having the long ramp at the start like before?

Best regards,
-Ivan Salazar
Applications Engineer

Hey Ivan,

In have not built the firmware to run on device yet. I can do this today and get back to you on results.

Oren and I were attempting to verify the thermal and excursion prediction by using a Keyence LK-H052 laser. Do you have a laser and are able to also perform the actual vs. estimated excursion?

This is what we were seeing for actual vs. estimated excursion. I will send you an thermal camera image and the estimated temp in PPC3 verification as I do not want to share the picture here. The stimulus in this picture is at resonance (915 Hz) from characterization and 1 FS.

I am unsure why the actual excursion is biased high-side. Will try to determine.

Once we introduce the EQ, the estimated excursion is very high and shoots past the limit lines. Is this normal?

Hi Joel,

I received the pictures through email. I'll send you all the tuning files as well, all of them seem to be close to each other, there is only one that has a bigger difference on the estimated vs measured excursion vs frequency, but not affecting the parameters that much.

I have a laser on my side, I'm setting up a test report and will share with you by end of this week.
Your actual excursion measurement may be offset due to speaker being already heated up. After speaker is playing for some time there may be an offset on the position of the diaphragm.
I'm not seeing a similar effect on the estimated excursion when changing the EQ. Are you still using the EVM+LB2 during this test? Or is there something changing on the hardware?

Best regards,
-Ivan Salazar
Applications Engineer

Yes we are using the EVM+LB2 during this test

Hi Joel,

I'm sending you PPC3 files for all speakers now. I'm also adding one PPC3 file with some changes on EQ and DRC, can you please double check estimated excursion using this file as well?
I'm not able to see a similar effect like the above capture with higher excursion than expected, I think there may be a configuration problem and you're getting a different set of data instead of estimated excursion.

Best regards,
-Ivan Salazar
Applications Engineer

Hey Ivan,

Oren and I verified that the excursion and thermal protection appear to be functioning on the dev board. However, I still hear the ramp when building it on the device.

Hey Ivan,

I believe I may be overlooking some settings on the laser when trying to verify actual versus estimated excursion. I will be looking this today and make sure the settings are correct. Oren and I were seeing this on Friday. This is with the EQ I created...not the one you provided. 

Hi Joel,

So if I understand correctly, you observe the ramp on the actual device, but not on the EVM. Is this correct?
Any chance you can connect the EVM to the speaker mounted on the device already? Not sure if you did this already.

I'll double check the settings on the laser on my side, these are actually used as part of the speaker char if you select the laser measurement option for Bl parameter. I'll share the details later today.

Best regards,
-Ivan Salazar
Applications Engineer

Hey Ivan,

Oren can make a setup like that, we just haven't done that yet. By the way, what laser are you using for the characterization process? We are using a Keyence LK-H052 with a LK-G5001P controller

Joel,

Scale factor has to be configured to 10 V/mm on the laser controller.
I'm using these on my setup:

• LK-H050
• LK-HD500

Best regards,
-Ivan Salazar
Applications Engineer

Hey Ivan,

Would you able to share the settings you used for the laser since it seems to basically be the same hardware, but just a model older.

Do you have the LK-Navigator2 SW? You can control all laser functions instead of using the display buttons.

The software looks like this:

Hi Joel,

I'm not aware of a compatible GUi for our hardware, but I'll take a look at the settings and share with you anyways.

Best regards,
-Ivan Salazar
Applications Engineer

Hey Ivan,

This is the software: www.keyence.com/.../

Joel,

All HEAD settings are default on my LK-HD500, the "option" settings seems to be out of default on A, E and F parameters, I'll double check on that.
So far I'm still not able to replicate the estimated excursion data going high after changing EQ values.

Can you please confirm the PPC3 version you're using? You can share a capture of the "About" dialog.
Did you faced any trouble during installation? I wonder if there could be some problem with backend processing of the GUI messing up with the register configuration.

Best regards,
-Ivan Salazar
Applications Engineer

Hi Ivan,

I found that the Keyence laser has better S/N on its analog output when its scaling is set to 100 V/mm, as shown below:

After doing that the Actual vs. Estimated Excursion measurements in the Verification pane in the same order of magnitude.  The scaling setting was left at 10 V/mm.  These screen captures are for two different loudspeakers:

The laser's analog output has a step characteristic that Keyence says is expected due to its 2 mV resolution.  The sine wave shape is still reasonable at lower frequencies but gets very distorted as frequency increases.  How does this affect the learning board and verification?

Thanks, O.R.

Oren,

That seems to be receiving only the envelope of the positive movement of the speaker.
I'll plan to check each setting on the laser controller and share with you by next week. Not sure if all will match with yours due to different version, but should give an idea.
You may also share a few pictures of your setup with the EVM, LB2 and laser just to verify it looks OK.

Best regards,
-Ivan Salazar
Applications Engineer

Hey Ivan, 

Forgot to tell you that we do have the most recent version of PPC3 (3.2.1). I'll take a few pictures and post them. Oren has been working with a Keyence application engineer and we've been assured that we are operating it correctly.

Hi Ivan, what is the expected analog voltage range on the Learning Board 2 "LASER" BNC connector?  Can you please share a LB2 schematic?  Thanks, O.R.

Hi Oren,

I'll send LB2 files over email through or FAE, will also take some captures of the BNC signal for reference.

Best regards,
-Ivan Salazar
Applications Engineer

Hi Ivan,

Is the schematic in this thread the correct one?  https://e2e.ti.com/support/audio-group/audio/f/audio-forum/893084/pp-salb2-evm-schematics

My understanding of the ADS1271 ADC and its associated signal conditioning circuitry (p. 3) is that the input range at the BNC is +/- 2.25 V.  (U8A OPA1652 is set for a gain of ~1.1 and U6 THS4521 is set for a gain of ~2.)  Please advise if this is correct.

Thanks, O.R.

Hi Oren,

That schematic is correct.

I'll double check about the expected range of the signal, considering 10V/mm and 0.35mm displacement, the voltage would be +/-3.5V. Need to verify what level could cause saturation on the signal path.

Best regards,
-Ivan Salazar
Applications Engineer

Hey Ivan,

When will you be able to provide the reports on our samples?

Hi Joel,

I'll wrap up the test captures and share by this Thursday. Will show the difference before and after adding some gain, add excursion and temperature data to show it stays below the limits. Any difference from that behavior to the end application must come either from the connection to the speaker, or from the received data from the host.

Best regards,
-Ivan Salazar
Applications Engineer

Hi Ivan, have you been able to look into the Learning Board 2 and the Keyence laser settings?  The schematic shows that the +5.0 V rail is provided externally so there could be some leeway in positive-going headroom if the supply voltage is increased.  However, the -5.0 V rail is fixed from an on-board regulator.  Thanks, O.R.

Oren,

I'll share the laser signal captures here in the forum shortly.

Best regards,
-Ivan Salazar
Applications Engineer

Oren,

Attached are some captures from the laser output, I manually changed the position of the laser head up and down to where the maximum values were, the same limits are hared form Validation panel and the scope, I mean when the output signal from laser reaches its absolute maximum level, the validation panel shows a constant error delta.

Best regards,
-Ivan Salazar
Applications Engineer

LaserCaptures.pdf

Hi Ivan,

Thanks for taking the laser output captures.  I was able to get similar behavior with the following Keyence LK Navigator 2 settings, as shown in the screenshots.  Please confirm that they are consistent with the settings on your end.

Regards, O.R.

Hi Oren,

On the Display settings, I have 1/-1 on display value and 10/-10 on output value, they're multiples of what you have but not sure if it could make a difference.

(I'll send the speaker repot later today through our FAE)

Best regards,
-Ivan Salazar
Applications Engineer

Hi Ivan,

I'm assuming that you're referring to these:

I found that it doesn't make a difference as long as the ratio works out to 10 V/mm to match what PPC3 is set up for.

Thanks, O.R.

Hi Oren,

Yes, that's the parameter I refer to, thanks for confirmation it works as long as it is multiple of 10V/mm.
I send the PPC3 files along with the verification captures of each speaker through our FAE. Let me know if you have any questions regarding that information.

Best regards,
-Ivan Salazar
Applications Engineer

Hi Ivan,

We received the files and presentation, thank you.

- Why are you adding 9 dB of gain?
- What audio are you using?

Regards, O.R.