I have been testing a commercially made OEM Mobile audio amplifier that uses a TPA3116 IC. this is supposed to be a 50 watt per channel amplifier made by Ntouch. The TPA3116 according to the spec sheet should easily be able to output 50 watts per channel into a 4 ohm load...But 2 samples of this product both output about 16 watts before some really nasty negative spikes start to appear. this will NOT produce 50 watt per channel at any distortion level.
I suspect that the tiny switching power supply inside of this product is to fault. they are not using a conventional SMPS type supply. it appears to be some type of switching voltage multiplier type supply. it is so small and a 4 layer board so very hard to fully understand what is going on...but...It appears they are stepping up the 12-15vdc input to something around 21-27vdc.
Has anyone see this issue where spikes are produced only on the negative going peaks? Is this a symptom of a sagging power supply or is this cause by some other issue?? granted i know it's hard to tell on this specific product, but this maybe a common issue other people have seen using these IC's??
anyone seen this issue before??
Top trace is the output of the amplifier. Bottom trace is the input signal going into the amplifier.
I suspect that the tiny switching power supply inside of this product is to fault. they are not using a conventional SMPS type supply. it appears to be some type of switching voltage multiplier type supply. it is so small and a 4 layer board so very hard to fully understand what is going on...but...It appears they are stepping up the 12-15vdc input to something around 21-27vdc.
Has anyone see this issue where spikes are produced only on the negative going peaks? Is this a symptom of a sagging power supply or is this cause by some other issue?? granted i know it's hard to tell on this specific product, but this maybe a common issue other people have seen using these IC's??
anyone seen this issue before??
Top trace is the output of the amplifier. Bottom trace is the input signal going into the amplifier.
Attachments
There are 2 things you should consider:
The TPA3118 should always be measured symmetrically, i.e. the difference between both bridge legs - single ended they measure quite badly.
Having said this, it is not obvious that these ugly oscillations (I have seen similar) really contribute to audible distortion.
You should do a real THD analysis before blaming the amp.
The TPA3118 should always be measured symmetrically, i.e. the difference between both bridge legs - single ended they measure quite badly.
Having said this, it is not obvious that these ugly oscillations (I have seen similar) really contribute to audible distortion.
You should do a real THD analysis before blaming the amp.
bb
If you measure fully differential, this "negative" spike would turn into a "positive" spike on reverting the inputs.
Thus this attribution is irritating.
Following this, you have some asymmetric distortion in your system that does not cancel out measuring across the btl output.
Next thing to consider is checking the signal at the inputs of TPA3118.
btw - do you have some full working samples as well?
In that case a pin by pin comparison might help.
If you measure fully differential, this "negative" spike would turn into a "positive" spike on reverting the inputs.
Thus this attribution is irritating.
Following this, you have some asymmetric distortion in your system that does not cancel out measuring across the btl output.
Next thing to consider is checking the signal at the inputs of TPA3118.
btw - do you have some full working samples as well?
In that case a pin by pin comparison might help.
Last edited:
ok, so can you provide some full working samples for comparison?
do things turn out better by removing the dummy load?
and find out the single ended source of distortion, i.e. pos/neg output, or both, pos or neg halfwave...
Last edited:
There is something odd, as it happens only on one halfwave.
Over the years I did a bunch of TPA3118 designs and never experienced this special problem.
Could you measure each half-bridge output separately?
Ok I did some further testing today. and I'm not sure why it's happening, but I figured out that if I swap the INPUT to the amplifier, that the spikes are now on the top! but the scope shows a clean signal. I am beginning to think my Audio Precision ATS-1 is showing some signs of age. If I place a .015uf cap from the positive phase to ground. the problem goes away! I can then test the amp to near full power. My power supply craps out before I can get to 50 watts but I get a nice 40 out of one channel! My big power supply bit the dust sometime ago and I just haven't gotten around to fixing it or replacing it.
It might be interesting that I experienced a similar strange and unexpected noise source that degraded the overall measurements. After some investigation it turned out that my sound card produces a lot of out-band-noise beyond 20kHz. An RC-lowpass 560Ohm/3.3nF between soundcard line-out and amp input vastly improved measurements by lowering the noise curtain by about 20dB!
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.