PAM8403 Clipping

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Hello everyone,

It is rare for me to go on forums for help, as I am usually able to figure things out with all my little projects that I find myself doing, but this time I am at my wits end :eek: lol

My setup is using i2s dac on a raspberry pi, which is hooked up to a PAM8403 amp running on its own power supply (I have also tried the PAM8406, but nothing has changed). The amp is then output to some small Tectonic 10W full-range bookshelf speakers (Tectonic TEBM28C10-4/A 1-1/8" BMR Full-Range Speaker 4 Ohm)

The volume is quite low, from what I understand that is because of the low sensitivity of these speakers. So, I am trying to "max out" the volume from the PAM8403/6 as much as possible. There are a few things that I don't understand about why I am getting clipping at high volumes.

First is: I thought that clipping only occurs when the amp is trying to output more power than it is supplied. But I am running the amp off of a 5V 3A power supply, so shouldn't that be plenty to supply the 3W or 5W of the PAM amplifiers?

Second is: The clipping is really only occurring during "complex" sounds. For instance, I can output a sine-wave with "speaker-test" at maximum digital gain and maximum amp pot gain, and I won't get any clipping. In addition, during normal music, I can have the gain maxed out and experience no clipping for most of the song. However, the more "complex" the songs is (with range of frequencies) the more clipping I get and the more I have to turn down the gain. That is the part I don't understand about all of this.

For some songs such classical songs, with the full orchestra playing at once, I have to lower the gain all the way down to about 70% to avoid clipping. Or during hip-hop songs, I will only get clipping when there is bass. So to avoid having to constantly adjust the volume, I just have to leave my volume set to 70%. The problem is that this is significantly quieter than 90-100% volume level, and I am certainly not satisfied with the volume output at this level. My laptop builtin speakers are just as loud, if not louder than this (granted less quality of course). I have heard of people running these PAM amps off of batteries and computer usb ports, so why is this happening for me using a dedicated 3A power supply?

Does anyone have any tips on what I can do? Or an explanation as to why I am having this experience? I was looking to have a nice easy budget project, but at this point I am in about double the cost than anticipated, and 30+ working hours which is much longer than I wanted to.
 
First is: I thought that clipping only occurs when the amp is trying to output more power than it is supplied.
For these bridged class D clipping occurs when the instantaneous output voltage should be greater than the supply voltage, which is impossible.

For 5V that means any signal peak > 5V is clipped. Much music has a high crest factor - the peaks are a lot more than the average power would suggest.

Typical low sensitivity speakers would normally be paired with an amp with supply rails of 35 to 70V or similar, not 5V.
 
The speaker has a sensitivity of 80dB/W/m. For a lower power project this could be limiting.
Yes, I guess that is the root cause, otherwise I wouldn't have to "push the limits" of the amp. The part I am still not really understanding is why I can max out the gain (digital and amp) of sine waves without getting any clipping, but I get a lot of clipping when I play actual music. If anyone cares to explain, I am always happy to learn, but I suppose it doesn't matter anyway.

Either way, I have put so much time into this, I am just going to invest in a proper amplifier TPA3116 and be done with it.
 
Typical low sensitivity speakers would normally be paired with an amp with supply rails of 35 to 70V or similar, not 5V.
Thanks, I just saw this after posting my previous reply. I ended up purchasing a tpa3116 and a 24v 5A power supply. Hopefully this will be sufficient?
For these bridged class D clipping occurs when the instantaneous output voltage should be greater than the supply voltage, which is impossible.

For 5V that means any signal peak > 5V is clipped. Much music has a high crest factor - the peaks are a lot more than the average power would suggest.
Thanks, I will try to understand what this all exactly means haha. When you say signal peak >5v, you are saying signal coming out of DAC or signal coming out of amp?
 
Mark has just explained this , although I am no expert on class D if you go above the actual voltage levels of the power supply both in the positive and negative in your output signal then clipping will occur . We are talking VOLTAGE levels here not power levels plus and minus 5 Volts =10 Volts peak/peak
 
Mark has just explained this , although I am no expert on class D if you go above the actual voltage levels of the power supply both in the positive and negative in your output signal then clipping will occur . We are talking VOLTAGE levels here not power levels plus and minus 5 Volts =10 Volts peak/peak


This makes a little more sense to me, I think it also explains why I don't get clipping on pure sine-wave signals. I still feel its odd that people run these amps with batteries and have no complaints. I guess maybe they just have more realistic expectations than I did, and probably better choices on driver sensitivity.
 
diyAudio Moderator
Joined 2008
Paid Member
If you had a more efficient speaker, say 95dB, then that 3W would be acting like 100W. Also, small baffles can add to the challenge of higher efficiency.

Anyway, your decision to increase power is probably helpful. There definitely are speakers that do well on low power, but they are less typical.
 
why I can max out the gain (digital and amp) of sine waves without getting any clipping, but I get a lot of clipping when I play actual music. If

A sine wave has the same amplitude above or under the X axis and You adjust volume just before the clipping onset.

Music can be very dynamic, so on loud passages You get clipping. This happens for example with Vinyl. Modern music is highly compressed doesn't have the same dynamics & tend to sound louder because of this.
(Google loudness war)

A simple experiment You can do:

Have a seat in front of a vintage system with analog VU's. Pioneer Sansui Akai, You choose it. Make sure there aren't DSP's inbuilt, just pure analog transistor stages.

Observe the VU movement with a Vinyl record or cassette. VU needle moves left & right.
Now, do the same with the same CD album. the VU needle moves always in the half right region.
The online streamed version is worse ..
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.