My "audiophile" LM3886 approach

Ultimate BOM 2.0 RC

After the research for the My_Ref Fremen Edition Release Candidate BOM I've also updated the TP My_Ref RevC BOM accordingly.

You can find it attached as PDF to this post or on my RevC Google folder.

Pay attention: it's a Release Candidate BOM so subject to some fine tuning.
 

Attachments

  • My_Ref Ultimate Rev C BOM 2.0.pdf
    9.5 KB · Views: 420
i put in 560ohm caddock to220 types in place of carbon film 1k for r1 & r4 spots and it's a clear improvement. there's more air, and the notes have become so much smoother that it actually blooms now! just like a tube amp!! i don't know if it's the value change or just the brand change but i am sure of one thing...

people would be more surprised that this is not a tube amp, rather than that it sounds like tubes. ;)

the 560 value was from the forum member protos, written on this thread long ago, on how he changed the value to 600ohm and noticed a similar improvement to mine. 560 was the closest value i could find that caddock offers. i haven't put any heatsinks on and it's only lukewarm to the touch.
 
Last edited:
Hey man, I feel your pain. I just had a similar experience comparing an Alps pot to the eBay chinese SMD unit. They both have strength and weaknesses that are so dependent on what's being played through the system. Our friend Andrew has a point when he suggests there is a lot of tail chasing with all the parts swaps/adjustments, but if there was a perfect circuit there would only be one commercial and/or DIY amp and only one forum to support it.

It is said the key to a good life is partially based on selecting the best set of compromises - and it appears to also be true in the search for "Audio Nirvana".

I'm suspecting the best amp configuration is the one that's on front of us when we tire of the search.;)
 
well said, bob.

i think the main problem with my amp now is that it just sounds so marvelous whatever i do to it. this makes it difficult to subjectively weight in the small changes i make to the sound, because the initial impression is always 'marvelous!' ;)

560ohm did make it less musical, but it was still bloody musical enough for me not to notice it quick it enough. also doesn't help that my hearing's been suffering from the noisy environment for the past month... (i was dumb enough to rent an apt right next to a four way. doh! ) i'm moving out next week, though. the blackgate nx's will be getting a full workout then. ;)
 
Last edited:
... and it had definitely made the sound less musical. i only figured this out after an extended listen. :( koa 1k ohms are back in and caddock 1k will be ordered :p
wrong move, I guess... ;)

In this case I bet that what made it sound "less musical" is not the (not so much) different value of the resistors, but rather their different construction technology. Carbon resistors usually tends to give a "warmer", "richer", "more musical" touch to the sound, while on the contrary metal film ones tends to give a "dryer", "cleaner" and "colder" one.

For this reason, to get a more "balanced" sound Thorsten suggests to use carbon resistors on solid state gears and metal film on tubes ones.

Thus in this circuit I'd suggest trying to use carbon compound resistors such as Kiwame or similar just about everywhere it's possible.

In general, use the highest wattage resistors that you can fit on the PCB: that will reduce transient thermal effects (distortions). This is particularly important on the NFB network(s), where these (and other) subtle effects gets hugely amplified by the loop gain.
 
Last edited:
hey
unixman

im sorry. 1k was carbon film resistors. (im starting to lose track of them after too many swaps) thick film caddock definitely had more refinement. notes bloomed more
noticeably also, and there was less grain. this i think i can blame on material composition,
because those are the general traits of thick film resistors. no doubt 1k caddock would be the best choice for me.