TPA3255: output PWM overlapping in BTL

Matthias Loy

Part Number: TPA3255

Hello at all,

I'm developing a small Amp for self-powered speakers.

My whole layout and circuit is running perfectly, but the TPA3255 does not. The amp is planned to run at 48V in BTL-mode at 600kHz but the idle-power is at 12 watts without any load!

The circuit and parts of the amp is exactly the same as in the EVM except the output-filter, there is a Coilcraft UA8014 for each BTL-channel (yes the coils are small, but the speaker impedance is 16ohms!).

I checked all parts, voltages and pcb twice without a result. In the picture you can see the overlapping PWM of outputs A&B ( C&D do exactliy the same). There is no DC-offset or signal on the inputs because while measuring all inputs have capacitors connected to ground (ground to cap and cap to input-pin). If I run the board at 36V the rising/falling edges of the PVM is crossing exactly at 36V/2 and idle-power drops to 8 watts without load.

There is no change in the PWM by changing the frequency to 450kHz.

I also changed the amp itself...

I hope you can help me?

over 3 years ago

0 Robert Clifton56 over 3 years ago

TI__Mastermind 20830 points

Hi Matthias,

12W power dissipation in idle is certainly too high for this device. There's certainly something going on here.

Would you be able to share the schematic? Just to double check to see if there's anything off with it.

Best Regards,

Robert Clifton

0 Matthias Loy over 3 years ago in reply to Robert Clifton56

Prodigy 25 points

Hi Robert,

here is the scematic:

Best Regards,

Matthias

0 Robert Clifton56 over 3 years ago in reply to Matthias Loy

TI__Mastermind 20830 points

Do you have DC blocking capacitors on the input? I don't see them in the schematic.

Best Regards,

Robert Clifton

0 Matthias Loy over 3 years ago in reply to Robert Clifton56

Prodigy 25 points

Yes there are 47uF DC blocking caps, they are in my analog-input sheet. they connect directly to the inputs of the amp.

Greetings

Matthias

0 Robert Clifton56 over 2 years ago in reply to Matthias Loy

TI__Mastermind 20830 points

Hi Mattias,

We actually recommend using a 10uF DC blocking capacitor. That might be causing some of the issues you are seeing. Can you change them out and try with the 10uF.

Best Regards,

Robert Clifton

0 Matthias Loy over 2 years ago in reply to Robert Clifton56

Prodigy 25 points

Hi Robert,

I tried 10uF but there is still the same fault.

0 Robert Clifton56 over 2 years ago in reply to Matthias Loy

TI__Mastermind 20830 points

Hi Matthias,

I've got another idea. Have you tried running the device without a load attached on the output?

Also can you reupload the schematic? I'm having a hard time seeing the values.

Thank you,

Robert Clifton

0 Matthias Loy over 2 years ago in reply to Robert Clifton56

Prodigy 25 points

Hi Robert,

the device is running without any load and consumes 12W!

Here is the scematic, I hope you can read is better now:

I have also made new pictures from the scope:

I hope you can help me!

Best Regards,

Matthias

0 Robert Clifton56 over 2 years ago in reply to Matthias Loy

TI__Mastermind 20830 points

Hi Matthias,

I did some testing on my end and found that the output switching like what you have seen doesn't appear to be abnormal at 48V.

How are you measuring the power losses coming from the TPA3255 specifically? Most of the time it would be difficult to separate the losses of one component from the rest of the system.

Best Regards,

Robert Clifton

0 Matthias Loy over 2 years ago in reply to Robert Clifton56

Prodigy 25 points

Hi Robert,

thank you for your answer!

In the datasheet on page 8 it says Pidle = 2,5W at 51V PVDD, which corresponds to ca. 50mA.

I measure the powerlosses by supplying the TPA3255 with a bench-supply just applying 12V and 48V. First running the TPA in reset and then in on-state without any load but all necessary components including the output filter. On 12V there are about 100mA of current which corresponds nearly to the datasheet. But on 48V are 250mA of current, causing hot coils, and my Pidle is about five times higher than in the datasheet.

I tried 10uH coils instead of 7uH and now my Pidle dropped to about 8,6W (180mA). But the overlapping PWM is still the same.

Best Regards,

Matthias

0 Robert Clifton56 over 2 years ago in reply to Matthias Loy

TI__Mastermind 20830 points

Hi Matthias,

This makes sense that you are seeing this change by changing the inductors. This could suggest that you're idle power ripple is higher than it should be. So if possible, the next step would be to measure the output current ripple to see whether it's correlating with the expected value.

Your information confirms that the additional power losses are exclusively on the output stage.

Best Regards,

Robert Clifton

0 Matthias Loy over 2 years ago in reply to Robert Clifton56

Prodigy 25 points

Hi Robert,

yesterday I tried to measure the current ripple and it showed an other strange reaction. By disconnecting all of the output filters the TPA draws much higher current and within a few seconds the current limit was reached.

To exclude my PCB, I did some rework on my design and reordered my prototype. I hope I can reply to you in a few days.

Best Regards,

Matthias

0 Robert Clifton56 over 2 years ago in reply to Matthias Loy

TI__Mastermind 20830 points

Hi Matthias,

Thanks for the update. Let me know when you receive it and have further updates.

Thank you,

Robert Clifton

0 Matthias Loy over 2 years ago in reply to Robert Clifton56

Prodigy 25 points

Hi Robert,

now the new PCB is here and I built a new prototype. All is still the same! It runns 2-3 minutes and my outputfilter is coocking! 100°C and more...with or without load.

What is wrong?

Best Regards,

Matthias

0 Robert Clifton56 over 2 years ago in reply to Matthias Loy

TI__Mastermind 20830 points

Hi Matthias,

I've got a few questions for you:

Did you ever test the speaker you are driving with the EVM?
Would you be willing to share the layout with me?
Did you ever test the device when it's in idle mode (no switching) and see if it's performing within the datasheet specs?

Best Regards,

Robert Clifton

0 Matthias Loy over 2 years ago in reply to Robert Clifton56

Prodigy 25 points

Hi Robert,

I tested my speaker with the EVM and my own board.

Yes I can share it with you. Where can I send it to you?

Yes i tested everything.

Best Regards,

Matthias

0 Robert Clifton56 over 2 years ago in reply to Matthias Loy

TI__Mastermind 20830 points

Hi Matthias,

Expect an email from me shortly!

Best Regards,

Robert Clifton

0 Robert Clifton56 over 2 years ago in reply to Robert Clifton56

TI__Mastermind 20830 points

Hi Matthias,

Maybe not as quickly. It seems that we are doing some maintenance on E2E. I'll send you a friend request to give you the information there.

Best Regards,

Robert Clifton

0 Dylan Yao over 2 years ago in reply to Robert Clifton56

TI__Genius 9140 points

Hello, as it is moved to email discussion, i will close this thread.

Audio

Audio forum

TPA3255: output PWM overlapping in BTL