Class D amplifier audible hiss

Hello,

I've been playing with little PAM8403 amplifier and I have problem with a noise. I have to say I didn't analyze the spectrum in any software but by hearing it I would say it is something close to white noise. It doesn't change at all with turning the volume knob up or down or shorting the inputs. So I started looking on the internet for possible causes.

I found out that it could be the chip itself (PAM8403 doesn't have the best SNR value) but then I found some tests of different cheap boards with the same chip (not the PAM8403) and there were big differences between them in term of audible noise, similiar to what I'm hearing. And since then I wonder:

What causes this audible level of noise and can I get rid of it? It's a broad question, I know, sorry for that. I have two ideas on my mind - It's problem with filtering the power supply rails - or - The chip is causing this and those cheap chinese boards, whose tests I've seen, were using just bad clones of original chip.

I bet it's much more complicated than just this. So thank you very much for any useful informations or materials. I really appreciate it!
 
PAM8406 is also ultra low budget but it is a slightly better performer.

I was not aware of chinese clones but one better uses original ICs.

Do you happen to use a cheap SMPS as power supply!?
 
Last edited:
A decent linear PSU for such an ultra light power amplifier is built in a breeze with usually better results than an SMPS with whatever filter. What is not created does not need to be filtered. It is surprising (no it is not) that I can predict beforehand in many cases that an SMPS is used when the complaint is noise. Should at least tell something about those wonders of technology 🙂 We are talking 5V 2A max. here and many would probably use 5V 1A 😀

BTW you state that the PAM8403/6 are not OK with SNR but they ar not so bad in that aspect really.
 
Last edited:
Thanks for your reply @jean-paul.

The main reasons I'm using SMPS is size and weight. I wanted to build it as small as possible. Can't imagine how could I fit any transformer there. Maybe SMPS + linear regulator would do the thing but I didn't have any suitable SMPS for this purpose.

With reference to SNR, I thought 80 dB is not much good since I've seen chips with very similiar specs and SNR well over 90 dB (PAM8008 for instance).

I'll try more ways to reduce that noise and see which will come up best 🙂