Those "optional" components degrade the sound, so it's no go for those components. R5 stays where it is, it more one the output path than signal. AndrewT I don't understand what are you trying to tell me with the last post. I've come up with really nice sounding amp with no DC offset problems (one channel 30mV the other one 42mV), maybe some different component values still can be used, but I definately won't be using the ones that are most commonly used or the ones that are recommended by National. AndrewT what kind of experiances do you have with LM3886?
Regards Aleš
p.s. When I said I'm still learning about this I meant learning about amplifiers and their functions, not about building one.
At the very least include the decoupling caps on the power supply. The chip should be decoupled by an electrolytic cap (10uF and up - like the C4, C6 you have) where the power enters the board and a 100 nF~1uF polypropylene close to the chip. Some people add a good ceramic cap (100pF~1nF) in parallel with the polypro as well.
These decoupling caps are used to lower the supply impedance and are used to prevent the amplifier from becoming unstable (i.e. oscillate). They are not optional components.
If you use AC coupling on the input (like you have in your latest schematic), the sound quality is likely to be dominated by the input cap. Use a good polypropylene cap here.
If you DC couple the input (as in your original schematic and in National's data sheet), you need to AC couple R5 in your schematic.
A third option is to use a DC servo.
~Tom
Member
Joined 2009
Paid Member
Looks interesting.
The lower impedance feedback loop generally reduces distortion (in simulations of discrete Lin topology amplifiers) but it does require a large feedback shunt capacitor. If this forces you to use an electrolytic then I'd recommend making it a larger value - look at D. Self's books or latest issue of elektor magazine - measurements of distortion in electrolytics show a distinct improvement if you oversize the cap and even better if you use a non-polar electrolytic cap (if you already have regular electrolytics you can place two of them back to back in series to fake a non-polar cap).
I don't view a cap on the input as too detrimental, it can be of small value allowing you the option of a very high quality cap at reasonable price (probably a better cap than is in the output of your source!).
The lower impedance feedback loop generally reduces distortion (in simulations of discrete Lin topology amplifiers) but it does require a large feedback shunt capacitor. If this forces you to use an electrolytic then I'd recommend making it a larger value - look at D. Self's books or latest issue of elektor magazine - measurements of distortion in electrolytics show a distinct improvement if you oversize the cap and even better if you use a non-polar electrolytic cap (if you already have regular electrolytics you can place two of them back to back in series to fake a non-polar cap).
I don't view a cap on the input as too detrimental, it can be of small value allowing you the option of a very high quality cap at reasonable price (probably a better cap than is in the output of your source!).
With risk of going a little off-topic - I've seen where D.Self presents examples of cap distortion, however I haven't seen anything on polar vs non-polar caps in this situation, can you point me to where I can read up on that?
Also, speaking of Self's books - the constant-current source on the output is exactly what I was reading recently - "Crossover Displacement" and it's a nice technique, if a little strange. In fact, the PIMETA headphone amp I've got does exactly the same thing and it works pretty sweet.
In fact - here:
Cambridge Audio-XD
and
Some more from Cambridge Audio
Point being, this is not biasing in class AB, at least not in the way you'd usually put a large bias on a normal output stage. And it should work better, up to a certain power level.
Last edited:
It has been a while since I last posted in this thread. Meanwhile I did made an amplifier from page 2 and works great, sound is also great
. I have ~60mV DC offset on both channels. The input cap is tantalum capacitor 10uF/16V and I think it is the most neutral of all types available. Also I did took some advices from you and modified PCB.
Best Regards,
Ales
. I have ~60mV DC offset on both channels. The input cap is tantalum capacitor 10uF/16V and I think it is the most neutral of all types available. Also I did took some advices from you and modified PCB. Best Regards,
Ales
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.