Oh no, its possesed.

[karma]

ok got it thanks

[Jean]

made a new thread about listening impressions.

[aHobbit]

Originally posted by carlosfm
If you use NI topology without the caps from RI to ground (ex: Ci in LM3886 datasheet) you risk having big DC offset.
On one channel you may have 50mv, on the other you may have 100mv.
The chips have a huge production drift.
The problem is, those caps (Ci) usually affect the sound in a negative way.

That's why I like the Inverting topology, very easy to control dc-offset, without caps. Inverting + input buffer is the way to go!
And regulated... another step forward!
Sorry, I must be a weird guy , but this sounds so superior that I can't believe people are still insisting on a copy of the Gaincard...

I recently made the lm3886-based amp, but started with the datasheet values 20k/1k (Rf/Ri) in NIGC config. Read also from LM3886 datasheet the following:

"These instabilities can be eliminated through multiple bypassing utilizing a large tantalum or electrolytic capacitor (10 µF or larger) which is used to absorb low frequency variations and a small ceramic capacitor (0.1 µF) to prevent any high frequency feedback through the power supply lines."

I instead put 0.22uF at the reg's output, and since it is too close to amp, I just thougt that will take of the 0.1uF being asked for in the LM3886 rail. I chose to experiment with Tantalum, being small and with less value, per LM338, could equal that of electrolytic cap with higher value, and yet tantalum is so small and has lower impedance (see quoted datasheet below):

"In general, the best type of capacitors to use are solid tantalum.
Solid tantalum capacitors have low impedance even at
high frequencies. Depending upon capacitor construction, it
takes about 25 µF in aluminum electrolytic to equal 1 µF
solid tantalum at high frequencies."

As a result, got a very small and light board that is not too big an issue for the LM3886 pins to carry. the only big part there is just the 33uF cap for the Rm (mute).

Have made my first listening, using a pre-amp from the pre-out of my Yamaha RX-V630. My DC offset is just 30mV. Tried to find issues on sound, and all I can say is, the amp is dead quiet on my wharfedale diamond 8.3, and sound many times better than my Yamaha 630 amp itself. I did not use any caps in series with the Ri.

I will have to finish my second amp based on Joe Ras design, like the one Nuuk implemented in his VBIGC, use the same small tantalum caps. My rail is +/- 29Vdc. Will check if DC offset is way better as reported by Nuuk in his website, about 1mV (if I remember it right).

Afterwards, I will produce 6 amps to feed on to the pre-outs of my Yamaha amp.

I appreciate the discussion you made as you experiment on your amp especially the PSU portion.

[carlosfm]

[aHobbit]

I was surprised to learn that the tantalum I used on my LM3886 rail is in fact 1uF tantalum only (not 10uF as recommended by the datasheet). I remember, if carlosfm made ue of 33uF in the rail, and a 1uF tantalum is basically equivalent to 25uF in performance, then this is no issue.

However, the sound I have presently is too good, even the bass is so much better than my typical commercial amps nowadays. I want to treat the amp negatively so I will not be biased, and be able to identify issues but cant find any anomaly in the full frequency spectrum.

Funny though, at first power up, there are some sparks in my LM3886 board, then followed by thin smoke. even after putting it off, some sparks continue (probably because of capacitor charges before the reg). Looking at a magnifying lens (my eyes is no good), I saw no burnt component - so I though perhaps some small issues got burnd in the process - or was it the break-in you guys are refering to???

And some audible hiss to. this continue for 2nd and 3rd switching ON. After those spark ended, the amp just sound as usual. After a week of some occasional fireworks display on my LM3886 board, the amp became quiet, no hiss, no hum, no fireworks, and stabilize at 30mV.

Just my unusual experience with the amp!

[Greg Erskine]

Hello Jean,

I don't know if it makes a difference but I noticed when you measured the second channel you had the dummy load plugged in.