My "audiophile" LM3886 approach

Jump to Latest
#2,881
In fact this is new PCB design and not tested yet!
However as I know, no Rev.C PCB is available at the moment. I've asked from Russ many times ago for a new collective PCB purchase (Rev.C) and have no response until now.
So I decided to make my own PCB for Rev.C with some mods which are;
- The basic PSU design is replaced with a LT1083 based regulated PSU. Based on CarlosFMs snubberised regulated (LM338) PSU design.
- On the board, simple zener diodes (feeding LM318) replaced with 7815/7915 regulators.
I expect to have a better amplifier but as mentioned I havent got a chance to test it yet.
🙄
 
#2,882
Hi,

As far as I've seen from some of the posts in this thread, the reason Mauro used zener diodes and no snubbers etc. is because all of these things had been tried and tested during the development of the amp(s).

That is to say that Mauro has already calculated the best possible configurations to achieve RevA / B / C and found that adding things like snubbers didn't seem to add any real benefit to the circuit (mainly because this amp topology isn't the same as a GC, so the same rules don't apply.)

I'm not trying to put anyone's work down here, but it's just a personal preference to want to build a RevC amp as this seems to have received the majority of the excellent reviews so far. So, please don't get me wrong, I'm all for changing things and seeing what sounds better, but I'd like to start with a "known" base layout first, then modify from there.

I'm tempted to have some RevC boards made up using Russ' design. I'll look into it again later (at work atm - trying to sneak a post in!) 😉

OzOnE.
 
#2,883
OzOnE_2k3 said:
Hi,

As far as I've seen from some of the posts in this thread, the reason Mauro used zener diodes and no snubbers etc. is because all of these things had been tried and tested during the development of the amp(s).

That is to say that Mauro has already calculated the best possible configurations
Click to expand...

...so why Mauro himself would have replaced the zeners with active regulators in his "my_Evolution"? 😉

The "my_ref" is indeed a quite careful and optimized design, but among its original design goals there were also simplicity and reduced costs...

Of course this is not to say that it's possible (let alone easy) to improve such a design by means of "random", clueless mods. Nevertheless, something like improving the filtering and regulation of both chip PSU lines are IMHO not bad ideas, and worth trying.

Of course, if you want to know whether such mods are for the better or the worst you should have the original as a reference... but that's another story. 🙂
 
#2,885
Dxvideo said:
So that means youre proping my idea is it?
Click to expand...

sure, IMHO your idea is worth trying (otherwise I would have told you to forget about it before... 😀 ).

Although should have I done it myself, perhaps I would choose different schemes for the regulators, at least for the "control" stage (LM318), which is the "core" of the system and the most important and sensible part of it.

There I would rather use a good, wide-band CCS (e.g. something like Gary Pimm's ones) followed by some good, low-noise, shunt type regulator (with a good, relatively small by-pass cap in parallel to make sure that the PSU impedance stays low up to RF...).

Unfortunately such approach can not be applied to the 3886s too. But at least the scheme you've chosen is (supposedly) proven and well sounding with the same chip...
 
#2,887
Honestly, the zeners worked great in the RevC, and if you can acheive cleaner output from 78/7912 regs, then great. I think spending too much time messing with them, however, is going to be a lot of work for little (if any) audible improvement. I think the 7812/7912 will be just fine.

If you don't like it after you build it, then think about changing it, but I am willing to bet you won't ever feel the need.
 
#2,889
Dxvideo said:
In fact,
If I take consider; my LT1083 has his own PSRR (not lower than 60dB) and also LM318 has (>65dB).
Click to expand...

the key question is: at what frequency?

if you look at the PSRR vs frequency plots, for most op-amps you'll find out that it goes down very quickly. And that's even more so for typical chip regulators.

About noise: yes, Zeners indeed are noisy. But be careful, many integrated regulators are noisy too, some even more than simple zeners!

In the my_ref design the LM318 is what "controls" the overall NFB loop, that is what eventually control the output... thus that part of the circuit (and its PSU) deserves the maximum care.

I don't like 78xx/79xx for this application, they are noisy and have a ridiculous bandwidth... chances are that the simple zeners as for the original design would be better!

I guess not by chance on the my_evo Mauro have rather used a simple discrete regulator here. 😉

Believe me, a CCS + shunt regulator (even the very same zeners...) is the way to go here. And perhaps go up to +-15V instead of just 12.

That provides the maximum isolation from any power supply noise and fluctuation (the CCS high impedance is sort of an "open circuit" there...) while providing low impedance and stable voltage to the chip.

That way, effectively isolating the 318 from the 3886 (and the PSU modulations that it creates) thus preventing any undesired chip interaction and "indirect feedback" of sort you may actually get some real improvement.

Of course for this to be effective at least the CCS (but hopefully also the shunt reg.) must be able to work correctly up to (well over) the overall (closed-loop) amplifier bandwidth.

(well, as a quick hack test, if you have a regular my_ref you may try powering the LM318 off some batteries, of course with good cap by-pass too... I would be curious to hear your feedback :cannotbe: ).
 
#2,895
#2,897
Ok..
I have returned to the original again. Just zeners!
So this is the last layout;
An externally hosted image should be here but it was not working when we last tested it.

But still I have some questions;
- If the EMI from output components is that much important then its a better idea to use noninductive type thin metal resistor for current sensing (0R47) Then I will use a TO220 %1 0R47 for this position. Any objection?
- Mauro uses 2x24v center connected secondary trafo. However mine is 2x30v and independent (because I use a regulated PSU as you know) ... So if you look closer to this;
An externally hosted image should be here but it was not working when we last tested it.

The power supply for the speaker protection circuit is based on a center connected trafo.. To adaptate it for my trafo, should I use a bridge rectifier and a bigger value of series resistor (like 560R?)
Like that?;
An externally hosted image should be here but it was not working when we last tested it.

- What kind of relay should I use for this circuit?

Thanks in advance for the replies..
I'll start very very soon and will inform you about the result 😀
 
You must log in or register to reply here.
Top Bottom