This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

TPA3136D2EVM: mono load EVM audio demo

Part Number: TPA3136D2EVM
Other Parts Discussed in Thread: ALLIGATOR, TPA3136D2

Hi,

I am  working on the audio demo using the TPA 3136D2 EVM. The load is a

PBTL mono load. 

I am using a 4 Ohm 3W resistor for this, resistance measures 4 Ohm on a multimeter. 

This resistor is connected across the OUTL+ and OUTR+ points on the EVM.

 

I have a 1 kHz, 480 mV p-p signal at the RINP TP of the EVM. I can see this on the scope. 

 

I have the voltage between PLIMIT and GND set to 1.25 V.

 

I am powering the EVM at 12V DC.

 

I cannot see anything on the scope when I connect a probe across the resistor ends.

I am unable to read any voltage across the resistor leads as well with a multimeter,

AC or DC.

 

The DC voltage between OUTL+ and GND measures 5.9V DC. Same voltage between

OUTR+ and GND.

 

Can you offer any pointers on this? I have attached the EVM configuration used

for a mono PBTL load. Does this need changing?

 

Thank you,

Best Regards,

 Priya

  • Priya,

    Please check using a scope if you see a switching signal on the output pins. The voltage you are seeing is probably an average voltage of the 50% duty cycle PWM which is in phase between the two outputs resulting in 0V differential across the speaker.

    Have you checked the Fault and SD pins?

    I remember your previous post where the EVM had been configured for pre-filter PBTL. Is this a new EVM or the same one from before?

    Regards,

    -Adam
  • Hi Pirya,
    Did you pull PBTL pin to High by opening the connector J13?
  • Adam/Shawn,
    SD and FAULT pins are measuring 10V. They are pulled to 12V through a 100K resistor.

    I have a replacement EVM with the configuration changes done for me that I am currently using.

    PBTL jumper has been removed.

    I tried probing the OUTL signal with the resistive load connected. I don't see anything but 0V on the scope.
    Same with the OUTR point.

    It makes no difference if I remove the load and keep the load open.

    The resistor does heat up when connected.

    I need to get this demo done. Looking forward to some more pointers.
    Thank you,
    Priya
  • Priya,

    Please send us a scope image of the output. One probe on each side of the speaker.

    Have you tried using a speaker in place of the resistor to check if there is audio?

    If the resistor is getting hot, there is some current flow through it.

    Regards,

    -Adam
  • The first picture is the input to the EVM, 1kHz, 490 mV signal. Two waves are seen because one is RINP, 

    the other is the function generator output.

    The loudspeaker when connected gives a loud tone, no clicking/ticking heard like from the past.

    The scope picture however is noisy, only the left part of the speaker shows a switching waveform,

    right part is 0 V (maybe this is how the mono output looks like?)

    I was hoping to snap a perfect EVM output waveform picture.

    Thanks,

    Priya

  • Priya,

    Do you have your scope grounds connected to the EVM? They should be connected.

    If a tone is heard then the device is operating correctly.

    Regards,

    -Adam
  • I have all grounds connected to the sleeve of the function generator probe (scope probes, EVM GND).The minute I hook up the loudspeaker, there is a lot of noise seen on the waveforms. I will now run a speech file and see what happens to the amplified sound. Hopefully, there won't be ticking, mangled noises.

    Thanks,
    Priya
  • Hi Priya,
    This issue is not expected. Could you please help to take a photo of your measurement setup and the EVM config? Where is the input signal from? Could you try with differential input mode and single-ended input mode(close or open J10)?
    Best regards,
    Shawn Zheng
  • For single ended input, is J10 open?
    That may very well be the problem.
    Thanks,
    Priya
  • Attached are 3 pictures of the test bench setup:

    1. Wavetek function generator provides 1 kHz sine wave input, 490 mV p-p.

    2. EVM shown with input connected to RINP. Note that the PBTL jumper is the only jumper in

    the EVM that is open.

    3. An RCA cable adapter is attached to the function generator probe. One end of this RCA cable is

    used for the input connections, the sleeve of the other end of this cable is used for the GND connections.

    Please review and offer your feedback ASAP. If the configuration needs changing, we need to know right away.

    Thank you,

    Priya

  • Priya,

    Your connections can be improved.

    I suggest connecting the function generator signal and ground through one cable using connector J9. You can then observe the signal using a scope connected to TP7/TP8.

    Alligator clips must Not be used for connections to the EVM, they are noisy and unreliable. Test points are not best for input signals, the RCA jack should be used when possible.

    The RCA cable is very long compared to the board traces, this long cable is likely picking up a lot of noise, please connect all grounds to J2 and remove the RCA cable.

    Regards,

    -Adam
  • Is it likely another EVM needs to be purchased? Are there simple measurements that can be done to verify the EVM is still working? There are a lot of accessories needed to use this EVM, that are only now being bought one at a time. If the EVM is damaged, it is not a very robust product, it seems to damage easily.
    Thank you,
    Priya
  • Can you suggest a part for the BNC to RCA probe?
    Thank you,
    Priya
  • Priya,

    I doubt your board is damaged.

    Using a normal BNC Male-to-male cable, use this adapter to convert to RCA:

    https://www.digikey.com/product-detail/en/cinch-connectivity-solutions-aim-cambridge/27-8110/367-1012-ND/385337

    Alternatively you can look up a BNC-male to RCA-male cable.

    Regards,

    -Adam

  • Adam,
    Thank you for the adapter part information.

    Today, I removed all alligator clips from the EVM and plugged in the RCA cable
    jack into J9. This removed noise from the input waveform.

    I connected the 4 Ohm load resistor across J3 and J8. The power supply voltage drops
    from 12V to 4V when I do this. The power supply reads a 0.4A current draw. Why does this happen,
    what does this mean? The resistor heats up.

    There is a noisy switching waveform that follows the period of the 1 kHz sine input. This again
    is better than before.

    Thank you,
    Best regards,
    Priya
  • I measured the current drawn by the 12 V powered EVM with no input or load connected.

    The meter reads 40 mA. I am told this should read 400 mA? I am attaching the data sheet

    of the power supply used. Let me know if this needs changing and how.

    Looking at the datasheet for the power supply, the current regulation for a load

    10-90% of rated voltage is <= 100 mA.

    PowerSupply.pdf

  • Hi Priya,
    Thank you for the update. I think ~40mA current on power supply during idle state on the EVM is expected. When applying the input signal on the device, how much is the input amplitude? How much is the output power? Looks like the power supply has enough current outputing ability, did you limit the output current on the instrument?
    Best regards,
    Shawn Zheng
  • Shawn,
    As posted in the scope picture on this thread, the input is at 490 mV p-p. I had to
    set the function generator at 10 dB attenuation to reduce the input to this voltage
    level.

    When I connect the 4 Ohm, 3W load resistor, I can see the DC supply voltage drop
    from 12V to 4V. I have set PLIMIT voltage to 1.25V DC.

    What is wrong with the given set of information? Please explain why the input
    power supply voltage is dropping.

    Thank you,
    Priya
  • Hi Priya,
    I don't think there is anything wrong in your EVM setting. Can you confirm there is no current limitation on your power supply instrument? How much is the output current from the power supply in this case? As I explained above, the output voltage from power supply may drop if the output current exceeds the limitation setting. Or can you try with another power supply?
    Best regards,
    Shawn Zheng
  • Shawn,

    Thank you for your reply. I changed the current limiting for the power supply to 1A. 

    Now when I hook the load, the power supply voltage does not fall down to 4V.

    However, I can smell the resistor burning. The scope waveform gets very noisy

    once I turn the power supply on. I am looking for a different power supply to try.

    Why is the load smoking?

    Thanks,

    Best Regards,

    Priya

  • I changed the load hook up to this connection. I don't see any switching waveform at OUTR+/OUTR-.

    I tried two 8 Ohm resistors in parallel and a single 4 Ohm resistor.I am using alligator clips to connect the

    resistive load. Let me know your feedback.

    Thanks,

    Priya

  • After re-doing the LC filter connections and soldering the filter and load together to eliminate
    all alligator clips, I seem to be getting an acceptable waveform across the right end
    of the load hooked up to the EVM.

    I will next test with a loudspeaker load. I will get the needed cable tomorrow headphone jack
    to RCA adapter. I will mark this thread resolved following this testing. Until then, I will keep
    it open for any questions.

    I hope there will be an app note about how to use this analog EVM. It helps to dumb things
    down for users like me .......

    Thank you,
    Best Regards,
    Priya
  • Hi Priya,
    I'm very glad that you make it work! I will close this post for now. If you have any more question in the following test, you could reply in this post and we will help you again. Thank you!
    Best regards,
    Shawn Zheng
  • Shawn,
    The success so far is only a preliminary step. The sound quality with a loudspeaker load is raspy and does not match the quality of the existing products in production. Existing product has 88 dB of sound energy within 3 ft distance along with the clarity and uses the same loudspeaker used on this testbench.

    I tried the filter from the post before. The loudspeaker sound fared better without this filter.

    The p-p voltage of the load was re-calculated for a 4 Ohm 3 W load V rms = sqrt(12), V p-p = 9.78 V. PLIMIT was increased from 1.25 V to 1.7 V. I sent the 1 kHz (0.49 V p-p) tone from the function generator. The loudspeaker sounded more raspy than before.

    I have ordered a new function generator to make sure the sine wave input is not generating noise that is getting amplified.

    Proper cables have been used on the testbench.

    I will post again when I try the new function generator with and without the LC filter. In the meantime, please post any insights. The time crunch is getting shorter.

    How will the sound energy reach 88 dB within a 3 ft distance? All calculations have been fixed at 26 dB gain for this class D part.

    Thank you,
    Best Regards,
    Priya
  • Hi Priya,
    Thank you for the info. I don't think there is any problem in your calculation. The gain is fixed for TPA3136D2. Did you measure the input and output signal by the scope? Make sure the input signal amplitude is indeed 490mVp-p and the output is 9.8Vp-p. Please use short cables between speaker and the EVM output connector - too long cables introduces higher serial resistance on the output. Make sure there is no current limit on the power supply instrument.
    Best regards,
    Shawn Zheng
  • Here is the picture with the new function generator. The blue signal is 490 mV p-p.

    The yellow waveform is 5V p probed across one leg of a resistive load. Two

    8 Ohm, 3W resistors were used in parallel with the LC filter from the previous posts.

    Current limit of power supply set to 1A. PLIMIT voltage 1.7 V. I think the 1.25 V at PLIMIT

    was cleaner.

    This is too noisy for acceptable performance. What can be done?

    Thank you,

    Priya

  • This is a scenario that can be easily set up at TI's end. Can you post an output waveform for a 4Ohm 3W load from this evaluation module set up in mono configuration?

    Thanks,
    Priya
  • I went back to PLIMIT 1.25V. The tone sent through a loudspeaker load with the new function generator is a little bit better, but not as smooth as expected.

    For a wave audio file, I measured the AC rms voltage and accordingly set the gain so the max AC reading was 0.17V. The sound quality is not acceptable. Maybe slightly less distorted but not as close as the existing product.

    I don't know what else I can do at my end. Looking forward to ideas.

    Thanks,
    Priya
  • Hi Priya,
    From the waveform which you shown above, the input and output amplitude looks normal for 26dB gain. So you are using an external LC filter on the output. Is it configured to 10uH+0.68uF? How about the connection and GNDing between the external filter and EVM board? 3W output power is not quite high, I think you can use the AMP on the EVM with ferrite bead configuration in PBTL mode. Set the EVM board to PBTL mode(without external LC filter), and listen to the output sound quality from the speakers. Do you have an audio measurement instrument, like AP? If yes, can you try to measure the output waveform with AP instrument? How much is the THD on the output? There are two speakers in parallel on the output, can you try to use one speaker at first to make the setup cleaner? TPA3136D2 is widely used in the market currently, and we never heard that it has performance issue. I believe there is something problem in your measurement setup.
    Best regards,
    Shawn Zheng
  • You mentioned that "The sound quality is not acceptable", could you please help to explain that? eg. The noise floor is high? Or single-tone is heard? If you can perform FFT on the output signal, you can find the root cause.
  • Besides, you are measuring the single-ended output signal. Could you please measure the differential ouput signal? The common common noise is actually cannelled in differential mode and will not shown on the speaker.

  • I am unable to see the scope waveform on the right leg of the resistor load any more.
    In the process of having a new LC filter built on a breadboard.

    In the meantime, I want equipment suggestions for all of the following. I will ask to have them rented:

    1. A pure sine wave generator. The one I have generates a noisy 490 mV p-p signal. I need a clean input tone.
    2. A good quality oscilloscope that comes with good probes. I may be able to find a better quality scope at work other than the one at my desk. I did check calibration on the probes and adjusted the probes.
    3. An audio analyzer to quantify distortion.

    Thank you,
    Best Regards,
    Priya