Author Topic: TDA7498E output ringing  (Read 4694 times)

0 Members and 1 Guest are viewing this topic.

Offline MarcusSTopic starter

  • Regular Contributor
  • *
  • Posts: 55
  • Country: au
TDA7498E output ringing
« on: November 06, 2016, 01:22:02 am »
Does anyone on the forum have class d amplifier design experience.  I am getting ringing on the output sine wave of my class d amplifier.  This issue affects both channels at output voltage levels greater than 50V peak to peak.  The amplifier does not clip until 65V peak to peak (run from a 36V supply).

I suspect the issue is due to a silly mistake in my pcb layout.  My output inductors for the low pass filter are WURTH ELEKTRONIK  7443642200  Surface Mount Power Inductor, WE-HCF Series, 22 µH, 30 A, 18 A, Shielded, 2400 µohm.  Low pass filter capacitors are PET film Kemet 600V.

 
 

Offline PChi

  • Frequent Contributor
  • **
  • Posts: 264
  • Country: gb
Re: TDA7498E output ringing
« Reply #1 on: November 06, 2016, 10:40:52 am »
I have a few quections.
Is the ringing a problem?
Is the ringing with no load or with a load?
Is the ringing the same on both channels? The schematic appears to show a difference, SPKR-L+ is connected to R33 and C44 but SPKR-R+ is only connected to L1. Is that correct?
Is the Oscilloscope sampling rate high enough to show the true waveform?
 

Offline tatus1969

  • Super Contributor
  • ***
  • Posts: 1273
  • Country: de
  • Resistance is futile - We Are The Watt.
    • keenlab
Re: TDA7498E output ringing
« Reply #2 on: November 06, 2016, 11:35:31 am »
Your design is using all through hole components in the output stage. That means the filtering bandwidth of the LC lowpass may be limited. Your output inductors will have their parasitic capacitances (typically a few 10 pF) that transfer the switching transients to the output. The large capacitors will not be able to suppress these high frequencies because of their high ESL.

The consequence would be that you get a significant amount of RF switching residues at your filtered output.  Apart from high EMI emissions this doesn't harm the speakers. If the sampling rate of your scope is too low, then you get aliasing effects which may result in the picture that you see. Try using peak detection acquisition mode if the scope has that.

There is also a possible explanation why the effects are larger at higher output amplitude. I assume that you had a load connected during your test. If that is true, this means that the current flowing through the amplifier's switching transistors is highest at this time. High current through a switching MOSFET typically adds ringing, which in turn increases its emitted EMI spectrum. You should be able to see this by connecting your probe directly to the switch node, and zooming in.
We Are The Watt - Resistance Is Futile!
 

Offline MarcusSTopic starter

  • Regular Contributor
  • *
  • Posts: 55
  • Country: au
Re: TDA7498E output ringing
« Reply #3 on: November 06, 2016, 11:37:12 am »
Hi PChi

Yes the ringing is a problem as it can damage the output fets in the tda7498e and is audible through loudspeakers.  The issue is with load.  I have not tried it with no load.  The ringing occurs on both channels but appears to start at a slightly higher voltage if there is only one channel driven.  My oscilloscope is 100 Mhz it does show the true waveform and the waveform is clean up until about 50V peak to peak.  Once the volume or input source pushes the output above about 50v peak to peak then the waveform starts to show the spikes on the peaks. Spk right wiring on the schematic is the same as speaker left channel. Although I think you are tricked by the lack of connection blobs on the schematic.  I just did a net check and it appears to be the same as the left channel.  Which makes sense because you can see on the pcb layout that the left and right channel layout is identical.
 

Offline MarcusSTopic starter

  • Regular Contributor
  • *
  • Posts: 55
  • Country: au
Re: TDA7498E output ringing
« Reply #4 on: November 06, 2016, 11:41:45 am »
Your design is using all through hole components in the output stage. That means the filtering bandwidth of the LC lowpass may be limited. Your output inductors will have their parasitic capacitances (typically a few 10 pF) that transfer the switching transients to the output. The large capacitors will not be able to suppress these high frequencies because of their high ESL.

The consequence would be that you get a significant amount of RF switching residues at your filtered output.  Apart from high EMI emissions this doesn't harm the speakers. If the sampling rate of your scope is too low, then you get aliasing effects which may result in the picture that you see. Try using peak detection acquisition mode if the scope has that.

There is also a possible explanation why the effects are larger at higher output amplitude. I assume that you had a load connected during your test. If that is true, this means that the current flowing through the amplifier's switching transistors is highest at this time. High current through a switching MOSFET typically adds ringing, which in turn increases its emitted EMI spectrum. You should be able to see this by connecting your probe directly to the switch node, and zooming in.

The inductors are SMD the capacitors are through hole because they are film capacitors.  The noise may have some RF but it is actually audible through the speakers.  I have seen the issue on multiple scopes including a borrowed 1 Ghz scope and the waveform looks the same as my 100 Mhz scope.
 

Offline tatus1969

  • Super Contributor
  • ***
  • Posts: 1273
  • Country: de
  • Resistance is futile - We Are The Watt.
    • keenlab
Re: TDA7498E output ringing
« Reply #5 on: November 06, 2016, 12:00:35 pm »
Your design is using all through hole components in the output stage. That means the filtering bandwidth of the LC lowpass may be limited. Your output inductors will have their parasitic capacitances (typically a few 10 pF) that transfer the switching transients to the output. The large capacitors will not be able to suppress these high frequencies because of their high ESL.

The consequence would be that you get a significant amount of RF switching residues at your filtered output.  Apart from high EMI emissions this doesn't harm the speakers. If the sampling rate of your scope is too low, then you get aliasing effects which may result in the picture that you see. Try using peak detection acquisition mode if the scope has that.

There is also a possible explanation why the effects are larger at higher output amplitude. I assume that you had a load connected during your test. If that is true, this means that the current flowing through the amplifier's switching transistors is highest at this time. High current through a switching MOSFET typically adds ringing, which in turn increases its emitted EMI spectrum. You should be able to see this by connecting your probe directly to the switch node, and zooming in.

The inductors are SMD the capacitors are through hole because they are film capacitors.  The noise may have some RF but it is actually audible through the speakers.  I have seen the issue on multiple scopes including a borrowed 1 Ghz scope and the waveform looks the same as my 100 Mhz scope.
I would suggest 1) watching the switching nodes under load 2) adding RC snubbers to each of the four for a test (maybe 470pF ~ 1nF in series with 2.2Ohm) 3) adding SMD ceramic caps for a test, maybe 10nF / NP0 / 100V.

What is strange to me is that you say that the waveform is not aliased, and it is also audible. That contradicts my theory. My first thought when seeing your layout was, that it does not follow the general recommendation to make all high speed switching stuff as small as possible. You have significantly large current loops that emit dI/dt (H field), and you have large copper planes that emit dV/dT (E field). This could cause undesirable feedback effects from that emission, coupled back to the input, possibly leading to beginning instability of the amplifier. Have you checked for anomalies at the input side at pins like the inputs, ROSC, maybe pin SVR? In one of my last companies we were developing a sub GHz RF transmitter that would become unstable by feedback from the output stage into the external PLL components.
We Are The Watt - Resistance Is Futile!
 

Offline PChi

  • Frequent Contributor
  • **
  • Posts: 264
  • Country: gb
Re: TDA7498E output ringing
« Reply #6 on: November 06, 2016, 12:17:54 pm »
Thanks for the answers, they help.
I guess that the audio signal frequency shown on the oscilloscope is 1 kHz.
I think that the ringing will be inaudible but that the Pulse Width Modulation misbehaviour near clipping will be audible.
Have you tried increasing the supply decoupling right by the TDA7498E?

If ST Microelectronics have an evlauation board does the TDA7498 always behave in this way?
 

Offline MarcusSTopic starter

  • Regular Contributor
  • *
  • Posts: 55
  • Country: au
Re: TDA7498E output ringing
« Reply #7 on: November 06, 2016, 12:46:03 pm »
They do have an evaluation board and it does not behave this way.  The noise is audible.  this issue happens regardless of output frequency.  I have tried to copy my layout as close as possible to the evaluation board.

Here is the link to the application note

www.st.com/resource/en/application_note/dm00042610.pdf

Here is the link to the evaluation board gerber files

http://www.st.com/resource/en/board_manufacturing_specification/steval-cca044v1_gerber.zip

Here is the link to the evaluation board technical information

http://www.st.com/en/evaluation-tools/steval-cca044v1.html.html

 

Offline PChi

  • Frequent Contributor
  • **
  • Posts: 264
  • Country: gb
Re: TDA7498E output ringing
« Reply #8 on: November 06, 2016, 01:05:46 pm »
I have run out of ideas apart from a detailed comparison between the PCB layouts, probing the power supply and looking at the Pulse Width Modulated signals where the distortion is audible.
I guess that the leads to the speaker are the same length and gauge when comparing the ST Evaluation Board and the PCB.
 

Offline MarcusSTopic starter

  • Regular Contributor
  • *
  • Posts: 55
  • Country: au
Re: TDA7498E output ringing
« Reply #9 on: November 06, 2016, 01:09:36 pm »
I feel like I am missing something key from their application note and gerber files, but I cant seem to find it.  I'm hoping someone can see my layout compared to the evaluation board and say what it is I have done wrong.
 

Offline tatus1969

  • Super Contributor
  • ***
  • Posts: 1273
  • Country: de
  • Resistance is futile - We Are The Watt.
    • keenlab
Re: TDA7498E output ringing
« Reply #10 on: November 06, 2016, 05:10:53 pm »
I feel like I am missing something key from their application note and gerber files, but I cant seem to find it.  I'm hoping someone can see my layout compared to the evaluation board and say what it is I have done wrong.
The app notes go in the same direction as my suggestions: their output network is much more compact, and they use SMD filter caps for RF suppression. One more idea that may help finding the problem: solder a few inches of tinned wire to different points of the input network, and find out which one is sensitive to EMI.

I gave a number of hints and suggestions, that gives you several things to play with. Would be interesting to see the switch node signals under load, make sure to use low inductance probing for that. Although 100MHz scope bandwidth may not be sufficient. You need to make sure the switching transients are as "clean" as possible, with minimum ringing and overshoot.
We Are The Watt - Resistance Is Futile!
 

Offline MarcusSTopic starter

  • Regular Contributor
  • *
  • Posts: 55
  • Country: au
Re: TDA7498E output ringing
« Reply #11 on: November 07, 2016, 12:46:49 pm »
Turns out there is some ringing on the evaluation board as well above 52v peak to peak. 
 

Offline tatus1969

  • Super Contributor
  • ***
  • Posts: 1273
  • Country: de
  • Resistance is futile - We Are The Watt.
    • keenlab
Re: TDA7498E output ringing
« Reply #12 on: November 07, 2016, 05:37:13 pm »
Turns out there is some ringing on the evaluation board as well above 52v peak to peak.
I tend to be cautious with eval board designs. I worked for a big semiconductor company time ago. Don't expect them all being designed by the most knowledgeable seniors...
We Are The Watt - Resistance Is Futile!
 

Offline FrancoisC

  • Newbie
  • Posts: 1
Re: TDA7498E output ringing
« Reply #13 on: December 19, 2017, 08:24:42 am »
Hello Marcus!

Have you been able to fix this issue?
If yes, possible to share your new schematics?

How did you generate the waveform?

Thank you!

Regards,

Francois
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf