TPA3116D2 Amp

But could I be certain that the hiss would disappear completely with a 5k pot? And what would happen to the sound experience if I went for a 5k solution pot? It's a lot of work for me to dismantle the active speakers so I don't wanna go for a 10k pot only to have to do it once again for a 5k.. The speakers have decent sensitivity and are used on my desk next to the computer screen, so ideally there should be no noise at all. This is no high end setup by any means so would a 5k work?
 
Being a bank holiday got a couple hours to start some modding. Replaced the DC caps with some 470uf 35V, and as I pulled off the heatsink a few nights ago to test temperature of IC, I yanked out the 1nf ceramic smd DC cap bypass and dropped in a 1nf 100V (not smd).

Have to say, just by listening the vocal range of frequencies jump forward 12inches. And not in a bad way. More spacial. Even standing off to the side of drivers, virtually 90degrees off axis the sound has exploded.

Wanted to add some 100nf replacements at same time but those I got were too massive to use. Gonna add PP 100nf next week. Probably some WIMA.
 
Hi,

I have not been active on here for a few years, but now fancy a little tinker. I ordered one of the black/blue boards today and I am starting to sort through all of my components to find out what I do and don't need to order. I have a couple of 3.3uf Ampohm polypropylene input capacitors and wondered if these will be suitable for this board?

http://www.farnell.com/datasheets/327360.pdf

Many thanks

Richard
 
As your listening position is quite close to the speakers, I recommend reducing internal gain of TPA3116 to its minimum as the best option to reduce hiss.

this makes some sense to me, with no educated reason why it would, other than the harder it works the more it might lose fidelity... Though loudness is still wattage, regardless of changes to gain right?


anyway, back to the questions I was trying to get a hold on a few posts back, reducing gain increases input impedance. Wouldn't that increase the hiss, if it really is the low current input, by doing that?

I guess I'm having trouble believing that's the source of my hiss though. I should ground all my audio inputs and take a listen....
 
Disabled Account
Joined 2010
anyway, back to the questions I was trying to get a hold on a few posts back, reducing gain increases input impedance. Wouldn't that increase the hiss, if it really is the low current input, by doing that?
.

No, the rise of internal input impedance does not contribute to additional hiss.
It is a minor mod - just check it out!
;)

Just for that reason I set my all TPA-amps to lowest gain available.
 
Last edited:
Changed my gain down to 26dB and it definitely helped a lot. Less hiss, still the same thing with the volume knob but less of it and when the volume is up, it's not audible.

I also read the gain structure guy, informative. Thanks

On the impedance matching, I'm not sure if I get that. Surely you can't "match" impedance, since output are single digit or less ohms and inputs are in the kohm? Or is this back to the volpot resistance vs input impedance thing? I didnt quite get that either.
 
which cap is this?

If you mean which 1nf cap I am referring to, its the 1nf smd cap under the heatsink which is part of the DC power cap circuit going into pins 17/18/19. Cap C23 on the tpa3116 data sheet schematic.

I know everyone seems to think modern smd, minute technology is a god send, but personally I don't.
I replaced it with a larger leaded Panasonic 1nf ceramic with shortest legs I could manage to safely solder, the hiss halved, and the music became more spacial. I did replace at same time as changing the DC power caps to 470uf 35V, so can't be sure of the caps improvement on its own.
 

Attachments

  • Fotografija-0067.jpg
    Fotografija-0067.jpg
    648.1 KB · Views: 474
Last edited:
Hello,
If you have another power supply of the proper DC voltage and amperage try that.
With the volume pot turned all the way down at first of course.

Other, smarter people will no doubt be able to tell you how to measure this board to see if it's functioning properly. Hope they chime in.
Good luck!
Mark