I've only listened to 3116 (which has identical silicon) and in stock form the TDA8932 beats it on bass, but loses out on HF cleanliness. TDA2030A loses out across the board to both these classD chips but does have cleaner HF than stock TDA8932. The trick with the 8932 is to clean up the analog supply (pin 8) with better filtering, it then stomps ahead of the other two in everything.
Following up I do think this amp has lower noise and better low level detail than my wiener monos, but not as full in low frequencies. Seems to have a little better separation of instruments and depth. However, the monos have more bass output. Otherwise the 3251 has the same sonic attributes. The 3251 has more power though and since I may pursue more power hungry speaker builds in the future, I will probably keep it as my main Amp!
There are various ways of addressing this. The simplest of all is to increase the value of Cvdda which is 100nF in the application schematic.
The next step up is to design an LC filter - meaning Rvdda is replaced with an inductor.
Beyond this its possible to use an IC regulator like LM317, preferably preceded by an LC or RC filter to eliminate the highest frequencies.
I have a hunch that the optimum solution will involve a discrete regulator and LC filter, but that's something to try when I build my next TDA8932 prototype amp.
@Scott - no, I haven't focussed on getting better inductors. When I've tried lower loss Ls I've struggled to notice any improvement. Contrast that with upgrades to the supplies which are instantly noticable.
Attachments
So basically changing the 5 100uf caps for 220uf correct ?
Scott
QUOTE=abraxalito;5077495]There are various ways of addressing this. The simplest of all is to increase the value of Cvdda which is 100nF in the application schematic.
The next step up is to design an LC filter - meaning Rvdda is replaced with an inductor.
Beyond this its possible to use an IC regulator like LM317, preferably preceded by an LC or RC filter to eliminate the highest frequencies.
I have a hunch that the optimum solution will involve a discrete regulator and LC filter, but that's something to try when I build my next TDA8932 prototype amp.
@Scott - no, I haven't focussed on getting better inductors. When I've tried lower loss Ls I've struggled to notice any improvement. Contrast that with upgrades to the supplies which are instantly noticable.[/QUOTE]
Scott
QUOTE=abraxalito;5077495]There are various ways of addressing this. The simplest of all is to increase the value of Cvdda which is 100nF in the application schematic.
The next step up is to design an LC filter - meaning Rvdda is replaced with an inductor.
Beyond this its possible to use an IC regulator like LM317, preferably preceded by an LC or RC filter to eliminate the highest frequencies.
I have a hunch that the optimum solution will involve a discrete regulator and LC filter, but that's something to try when I build my next TDA8932 prototype amp.
@Scott - no, I haven't focussed on getting better inductors. When I've tried lower loss Ls I've struggled to notice any improvement. Contrast that with upgrades to the supplies which are instantly noticable.[/QUOTE]
No, I've experimented with increasing the main supply capacitance (which means the caps connected to pins 20 & 29 on the IC) and while it makes an improvement for sure, the low hanging fruit is on pin 8 (the signal circuitry supply, not the power stage supply).
Best move the discussion to one of the TDA8932 threads - for example this one : http://www.diyaudio.com/forums/class-d/277130-fasten-seat-belts-tda8932-pessimistic-review-9.html
The photo shows 103 marking on the resistor which is 10K ohms. It's probably connected across one of the inputs to ground. Easy enough to check with a multimeter. If so it would be advisable to fit one otherwise the input impedance will differ between channels. You could use a 1/8W through hole resistor across the input terminals if SMD is an issue.
10k I believe is normally there, you can see on other 3 ampboards? Only if you want to use mute pads, you need a connection/resistor there.
edit: It connects gvdd, 6 or 7Vdc, to one of the mute pads, when mutepads are connected together that Vdc gets to mutepin and tpachip will mute. Doesn't affect inputsignal.
edit: It connects gvdd, 6 or 7Vdc, to one of the mute pads, when mutepads are connected together that Vdc gets to mutepin and tpachip will mute. Doesn't affect inputsignal.
Last edited: