TAS2563: Speaker open and short detection in tuning mode

Hello Rekha,

Ivan is still tied up with some high priority tasks. He will look into this soon. We need to close out all the open issues that you have very soon.

Hi Rekha,

Try with w 98 1C 00 49 24 92 at least. After thorough testing I noticed that if I use any value lower than this, it detects 4Ohm as short, so LDG_RES_UT must be set to at least 8Ohm.

Best regards,
-Ivan Salazar
Applications Engineer

Hi Ivan,

As suggested, the LDG_RES_UT was changet to 00 49 24 92. The test was performed on both the custom board and EVK. 

1. Speaker short is not detected right in both the EVK as well as the Custom board.

Kindly suggest if the short case can be resolved as well. All the tests are done in Tuning mode from where we enter LDG mode.

BR,

Rekha 

Hi Rekha,

I'll double check this and get back to you by end of the week or early next week.

Best regards,
-Ivan Salazar
Applications Engineer

Hi Rekha,

I used a 4Ohm and the script below, I can see the proper detection of Speaker Short as 0x30.
What are you using to short across the speaker? Perhaps the method is not properly shorting causing the device to detect as open load?
Also, if you read 0x00 it would mean the detection sequence is not completed yet, bit 5 should be 1 if the diagnostic is completed.

w 98 00 00
w 98 7f 00

#Software reset

w 98 00 00
w 98 01 01

#Clear all flags and set required interrupt mask

w 98 00 00
w 98 30 1d
w 98 1b c7
w 98 1a fd

#Page 5 Upper and Lower Threshold

w 98 00 05
#ORIGINAL
#w 98 1c 00 2b e2 be
#w 98 20 00 1d 41 4d
#NEW TEST 083024
w 98 1C 00 49 24 92
w 98 20 00 1d 41 4d

#Load diagnostic routine

w 98 00 00
w 98 3d 0e
w 98 02 43
d 100
r 98 25 01

Best regards,
-Ivan Salazar
Applications Engineer

Hi Ivan,

We are using wires (jumper cable wires) for inducing the short across the speaker terminals.

Let us know if there is any other suggestion for inducing short and detecting the speaker short case correctly.

BR,

Rekha

Hi Ivan, 

1. As per your previous response, disconnecting one side of the load and shorting it directly to other terminal creates the open load situation. Am I missing something here?

2. Can we integrate the speaker load and short detection as part of the reg.bin or dsp.bin? If yes, how can we do so?

BR,

Rekha

Hi Rekha,

1. No, that is to create short load. Open load would be just disconnecting one terminal.
2. Load diagnostic routine is only working in ROM mode, so I don't think this would be part of the dsp.bin. Perhaps you can create a reg.bin with this script, but it should not be the same reg.bin as you use for normal playback as the configuration is different.

Best regards,
-Ivan Salazar
Applications Engineer

Hi Ivan,

1. As per your suggestion, this is how short looks like is it? Kindly check the image attached. I believe short is introduced between the speaker P and N terminals and not one single terminal. Am I missing something here?

2. We will look into incorporating as a new reg.bin. How can we add two reg.bin (one with the load diagnostic check and other the existing initialisation registers)? Kindly let us know the steps.

BR,

Rekha

Hi Rekha,

The connection is like this diagram:

Not sure about how to use 2 reg.bin files. As I understand it, you have a power up and power down sequence in the same reg.bin, perhaps you can add a new sequence. You may start a new question so we can assign to the SW expert team, they may have better comments on how to implement a new reg.bin sequence.

Best regards,
-Ivan Salazar
Applications Engineer

Hi Ivan,

I have tried as per your connection diagram to detect short but failed to do so. It would be great if we could get a snap of your setup for the short case with the EVK and speaker connections.

Regarding the 2 reg.bin integration we have created a seperate thread. 

Link: https://e2e.ti.com/support/audio-group/audio/f/audio-forum/1410576/tas2563-integrating-2-reg-bins-as-part-of-driver

BR,

Rekha

Hi Rekha,

I'll setup the test again and come back with the picture within a couple of days.

Best regards,
-Ivan Salazar
Applications Engineer

+ Shailesh, Savya

shailesh KutheSavyasanchi Shukla

BR,

Rekha

Hi Ivan,

Is this blue highlighted box below an inductor? What is the purpose of this here and the value used?

Instead of this RL combination used, can you use the speaker we had shipped to test with the EVK and let us know the results.

BR,

Rekha

Hi Rekha,

That's 33uH inductor, purpose is to emulate the resistive/inductive response of a speaker.

As far as I understand, the non-working situation on your side is the short-circuit, not the normal one. I'm I missing something?
We have several speakers shipped by your team; you may specify which one.

Best regards,
-Ivan Salazar
Applications Engineer

Hi Ivan,

Yes the non-working situation is the short-circuit case. The speaker used is CS28-02W57-07-1X (4-ohm circular one) for the test.

BR,

Rekha

Hi Rekha,

Got it. The speaker/load used for the short circuit test doesn't matter as the outputs are shorted to each other, but if needed I'll repeat the test with the actual speaker early next week.

Best regards,
-Ivan Salazar
Applications Engineer

Hi Ivan,

Did you get a chance to repeat the test with the actual 4-ohm speaker? 

BR,

Rekha

Hi Ivan,

I am still unable to detect the short circuit on the EVK. 

We are using the below version of the PPC3 tool. Kindly let us know if this version is fine or needs to be upgraded?

BR,

Rekha

Hi Rekha,

We're using the same version; however the version should not matter as we're only using the script mentioned in by previous post from Aug 30th:

Ivan Salazar said:

Also, if you read 0x00 it would mean the detection sequence is not completed yet, bit 5 should be 1 if the diagnostic is completed.

Fullscreen

Make sure you use only that script for testing, no other configuration.

Also make sure to short across the amplifier terminals, not speaker terminals. If you short speaker terminals and connect to a single terminal of the amplifier, the amplifier is actually seeing an open circuit. The amplifier only knows what happens to its own terminals.

Best regards,
-Ivan Salazar
Applications Engineer

Hi Ivan,

1. In case of short detection, we observe that before LDG is getting completed, Over current event is getting triggered. This inturn is shutting down the amplifier. We read the values in Page 0 reg 25 =0x00, reg 24=0x02 (indicates OCE interrupt), and reg 2=0x02 (shutdown). How can this be handled?

2. We have a sense resistor of 1K in the path as shown below. Does sense resistor have any dependency in load detection? 

BR,

Rekha

Hi Rekha,

You can try removing those resistors on sense pins just for testing.
However, I don't expect these resistors to cause over current. If you don't run load diagnostics and just use your normal configuration for audio playback, is it also detecting overcurrent?

Best regards,
-Ivan Salazar
Applications Engineer

Hi Ivan,

1. I have tried removing the sense resistors but the result is same as above. We observed overcurrent error.

2. In case of normal configuration, there is no overcurrent detected. The below is the dump: 0x24 reads 0c.

root@tegra-ubuntu:/home/ubuntu# i2cdump -f -y 7 0x4e
No size specified (using byte-data access)
        0  1   2   3   4    5  6   7   8   9   a  b  c    d  e  f    0123456789abcdef
00: 00 00 00 20 ce 22 09 02 30 10 f1 02 00 04 05 06    ... ?"??0???.???
10: 07 14 13 76 01 2e 60 0e 0c 00 fc a6 df ff ff 08    ???v?.`??.???..?
20: 00 11 00 00 0c 00 11 80 00 dc 4f 00 84 d3 00 a9    .?..?.??.?O.??.?
30: 99 40 81 34 46 b4 00 00 0a 0c be 58 7c 08 10 00    ?@?4F?..???X|??.
40: 4c 41 d8 c0 10 21 00 4c ec 52 d0 00 00 00 00 00    LA???!.L?R?.....
50: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00    ................
60: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00    ................
70: 00 00 00 f0 00 0f 00 00 4f 00 00 00 00 10 86 00    ...?.?..O....??.
80: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00    ................
90: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00    ................
a0: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00    ................
b0: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00    ................
c0: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00    ................
d0: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00    ................
e0: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00    ................
f0: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00    ................

 Let us know what can be done further.

BR,

Rekha

Rekha,

Do you also get OC for this test on EVM?
Previously you mentioned that you could run the diagnostic routine, but only short-circuit case was problematic. What changed?

Best regards,
-Ivan Salazar
Applications Engineer

 Hi Ivan,

1. Yes,  the Over current error (Reg 0x24) is seen in case of the EVK as well with short circuit case sometimes. This is only for short circuit.

2. I am able to run load diagnostic properly for normal and open case properly. The overcurrent error is only in case of short case and not in case of normal or open detection.  In case of short detection, we observe that before LDG is getting completed, Over current event is getting triggered in the custom board always.

Let us know how to handle this?

BR,

Rekha

Hi Rekha,

Are you playing anything during the load diagnostic routine?
You must not play anything during this time, the diagnostic routine itself plays a pilot tone.

Best regards,
-Ivan Salazar
Applications Engineer

Hi Ivan,

No, we are not playing anything during the load diagnostic routine. The script shared is only run and the registers from Page 0 are being read where we notice the OCE being triggered.

BR,

Rekha

Hi Rekha,

Just FYI... Ivan is out of office for next 2 and half weeks. He will get involved again when he gets back.

Regards,

Nayeem

Hi Ivan/ Nayeem,

Is there an update regarding the short speaker case detection?

We are considering speaker short to be overcurrent error and implemented as part of the load diagnostic check. If 0x24 register in Page 0 indicates Bit[1]=1 then we are indicating as short. Is this implementation fine. We are interested in the speaker terminal short detection.

BR,

Rekha

Hi Rekha,

Speaker short conditions would cause overcurrent event anyways, so I agree you can use them interchangeably, this is OK.
(Still not sure why you get overcurrent event before speaker short flag during diagnostics only, this is something we couldn't replicate on our side).

Best regards,
-Ivan Salazar
Applications Engineer