"Yuanjing" Gainclone 3886 - eBay amazing value ?

[AndrewT]

I agree on Carl's point 2)
there is no need for the buffer.
I'll go further there is no need for the input gain stage for normal Line Level Sources.

Make the gain stage bypassable, or apply it to a dedicated input that is reserved for low level source equipment.

The 22uF on the NFB leg is far too small.
The four 2u2F could be a little on the low side for deep and extended bass reproduction, but should be OK with normal restricted bandwidth speakers.

[carlmart]

About the preamp stages, if you use a volume pot it might be better to implement either a line stage OR a buffer stage, to isolate the pot from the 3886. But just one stage.

Don't get me wrong: I have used my amps, including a 3886 Penasa version, with no buffer between pot and amp, but if the stage is there as on this case, why not try it? But just one stage.

Even if I agree with you on the 22uF cap being a bit small, it is what National puts on their datasheet.

Also agree on the 2.2uF caps being small, but that I think was improved with 3.3uF our friend is using, isn't it?

What I do not agree on the power supply design is about the high current stage feeding the 3886 also feeding the regulator.

If the schematic is right, which I think it might not be, some things could be modified:

1) Simple mod. Add an extra cap, 220uF or what could fit there, between 22k and 470 resistors. Another cap, between 100uF and 220uF (also what could fit on copper side), after the 470 resistor and before entering the regulator. That should filter any fluctuations that might happen on larger current pumps.

2) Complicated mod. This is what I would have done. Add another diode bridge and capacitors to feed the regulators. The CRC filtering could be kept. To do this you would need a small separate board.

Carlos

[carlmart]

Sorry, another comment on the supply.

That 22K resistor is too large. But OTOS you need an RC filter there to isolate the regulator from fluctuations due to the main amp current demands.

An RCRC should follow the main 10000uF capacitors, but space limitations should say what to put in.

[Haz]

Thanks for your reply's and suggestions they are much appreciated. I don't really want mod the board right now its really nice as it is, even though they maybe some flaws in the design.

What do you think of these boards? It seems to be exactly the same as the LM3886 example circuit on the data sheet.

eBay - The UK's Online Marketplace

For the power supply I was think of this.

eBay - The UK's Online Marketplace

[carlmart]

The boards seem fine. They are designed for discrete and better diodes on the supply.

Whether the circuit is the same as on the datasheet, it's quite likely, but not for sure. It's the one most people are using, recommended by the manufacturer, and this is a commercial product.

It's up to us, DIYers, to try to improve on things and see what happens.

[Haz]

I need some help with these LM3886 boards I was talking about in my last post, I know its going off topic a little as the thread is for the Yuanjing Gainclone.

I'm unsure which value of resistor to use for R1 (see the pic) On the data sheet the seller emailed it was expressed as 0.15. After seeing ready made boards on ebay they were using 0.1R 5w resistors.

Somebody on YouTube told me its 10R resistor with 10 turns of coil wire in parallel.

So I'm a little confused, I will also post the schematic here too but I'm 95% sure its the wrong one.



http://i317.photobucket.com/albums/m...oardPCBDIY.jpg

[troystg]

Originally Posted by Haz

.......

Somebody on YouTube told me its 10R resistor with 10 turns of coil wire in parallel.

Just wrap 10 loops around the resister and solder the two ends to each end of the resistor.

[AndrewT]

Is that resistor in series with the route from outpi#ut pin to output terminal.
Or
is R1 in parallel with the output load with one end connected to ground?

[Haz]

Hi Andrew the resistor is in series with the route from output pin to output terminal.

[AndrewT]

Then it is a resistive buffer between amp output (after feedback) and the feed to the speaker cables. This is there to help prevent the reactive load affecting the amplifier feedback and thus changing stability margins with changes in frequency and reactance.

The resistor value is usually quite low, typically 0r1 to 0r2.