X-Calibre : Hot Rodding Mauro Penasa's LM3886 design

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
X-Calibre: Hot Rodding Mauro Penasa's LM3886 design

Mauro Penasa designed a very "different" LM3886 amp
http://www.diyaudio.com/forums/showthread.php?s=&threadid=54571
This is a really great sounding amplifier and tweaking it is allot of fun

There has been allot of speculation as to "tweaking" this design and unfortunately it has been shot down several times.

Here is a tweakers board, you can try allot of different configurations on it even build the original design on it, if you want, Minus the bridge rectifier

I will be doing a bit more documentation on it and how to build it,

My design uses
2 x 550VA transformers
8 x 10 000 Sprague 40v caps
16-x MUR 860
And all other caps are Rubycons

Regards
Rudi
 
Front view
 

Attachments

  • front.jpg
    front.jpg
    60.4 KB · Views: 4,703
Hi Dr.H

I do not want to give Mauro the idea that I am being disrespectful to his intellectual property. But he will have to give us something new because this one (for me at least) is much better. There is no way that I am going back to the old design - this one beat it hands down. in one word CLEAN

Sound staging is wide, while keeping a pinpoint accurate focussing
It has a depth that i did not pick on the original.
The entire sound spectrum cleaned up, bass is tight well defined. mids are sweet and very clean and tops are just spectacular,

The basic specs are the same as the original. i have just changed board features

* Possibility for larger input caps. The big boxy ones normally sound better
* Bipolar feedback cap configuration
* Parallel main caps. - You can get the advantages of more expensive caps by doubling up two caps
* snubber - if you want that. I tried it and it worked well, gave some life back to the amp
* Series or shunt regulation for the LM318 or even both if you want.
* Discrete diodes for the bridge rectifier – MUR 860

for the rest. It is exactly the same as the ref C, the main reason for me making this board available is that I thought that you can really make this amp sing. And that is where Mauro had the vision and I just want to give it the drive, as you can see in essence it still Mauro layout and 90% of his design. I just added the BLING
 
In the spirit of DIY, thanks to Mauro

I've listened to Rudi's X-amp. It is very very good. We are all here to DIY and in that sense I think that Rudi has taken a great design and gotten even more performance out of it.

The sound of X:

Most noticeable was the quietness/blackness of the background, with no noise/hash present.

Extended, detailed and airy highs-lovely to listen to, no hint of grain. This is a real strong point in the amps performance, giving a real sense of musical timing, especially in acoustic jazz feauring cymbal work.

Mids were wonderfully detailed and present in the performance. In one instance, for the first time, I heard voice vibrato on a well-known piece by a female vocalist.

Bass was fundamental, i.e. not rolled off, but deep and powerful. The body of an acoustic/upright bass takes on more flesh/realness than previously.

Staging and presence were reference class.

Shortcomings? The amp needs a good pre to show its best. We started the listening session with a passive ALPS, but the pot tended to some warmth/bloom in the lower mids/upper bass. Replacing with a passive NOBLE cured the problem.

Bottom line is that I was intending to build REVC to replace my 3875-based design. I will now be building X-Calibre, the jump in performance is compelling!
 
Yeah been talking to Rudi on the phone, actualy had to nag him to post the amp, seeing how there are so many people who are quick to dismiss anything.

I realy like what he did to the amp, although it is certainly no longer a low cost one. Will be one of those multiple month projects for me.

I was in two minds about adding that much capacitance, but Rudi was so convinced and compelling after actualy hearing his amp, there was no containing him.

I look forward to ordering my boards

First we gave you the heart transplant operation, now the x-calibre... proudly souf efrican
 
rudi said:
* Possibility for larger input caps. The big boxy ones normally sound better
* Bipolar feedback cap configuration
* Parallel main caps. - You can get the advantages of more expensive caps by doubling up two caps
* snubber - if you want that. I tried it and it worked well, gave some life back to the amp
* Series or shunt regulation for the LM318 or even both if you want.
* Discrete diodes for the bridge rectifier – MUR 860

Hi Rudy,

You are obviously in the right direction, and you start to solve some of the issues that made me have absolutely no interest about this amp, it would not perform to the levels I'm used to with my LM3886.
This is not any critics to Mauro, which I respect, but I see this amp since the beginning as an exercise.
There are lots of things to improve there (and still are), but I stepped out of the discussion because:

1. I find the design unappealing. You can do better with less parts.
2. There was some 'animosity' on the other thread against whatever I (and others) said about the design, even if it was a suggestion to help improving it.

Yes, you have better results by snubberizing those big caps.
Yes, you have better results with discrete diodes, and the MUR860s are very good. This was reported years ago by Peter Daniel.
What are you guys doing, sleeping?
We are discussing this for years, working hard and sharing it here and you guys go all the way back and make the same mistakes all over again and then 're-discover' the ways to improve the amp, while this is old story for us that have extensive experience with these chips.

PS: Ryan, listen to the amp with the AD815 pre. ;)
 
Well, if the alternative is just making the same mistakes and not moving on, then I think this is a step in the right direction...

In Mauro's defence he has always stated that some of the design principles were compromises, ones he was ok with, and also to some extent I'm sure some of these compromises included not haveing to spend as much on components.

There may yet be more elegant soutions as you alude to, but in the mean time us hobby types who didn't have the priviledge to have an electronics education, are haveing fun, learning good fundementals (sometimes the hard way), and most importantly building some decent amps.

Yes, you have better results by snubberizing those big caps.
Yes, you have better results with discrete diodes, and the MUR860s are very good. This was reported years ago by Peter Daniel.
What are you guys doing, sleeping?

Well these weren't in the Mauro amp, and now they are... I would hardly call that sleeping!

Have a good evening.
 
Hey Carlos, great to see you on here:cool:
Thanks to Rudi for the new board layout and for opening a tweakers thread for this amp

After all the hype etc for this amp and seeing as though I built both Carlos and Pedja's chipamps why not try this one too, I already have the biggest part of the bits knocking around so I decided to etch the boards and give it a try

I first stook to the RevC parts list etc but used Rudi's layout because of the dual rectifiers which I thought would be better in the long run.
The amp is actually not bad but I was not happy with the bass to be honest
I've always felt Carlos amp had by far the deepest tightest bass out of any of the chip based amps and Pedja's had the best top and sheer clarity, this is MY opinion and in MY system

I now have swapped the standard diodes to MBR10100 Schottkies, removed all the small 100nf decoupling caps off the top board and cut the legs much shorter and soldered them directly to the LM318/LM3886 pins undeneath, the zener for the LM318 has been removed and replaced with seperate diy boards with LM317/337 TL431 set to output +/- 15v, added 2x 470uf back to back low ESR caps in the fb
I prefer the amp already to the original, I will also of course try the snubber too next;)

Thanks again to Mauro for sharing another amp to play with and to Rudi for the layout and fb tweak and to Carlos,Pedja etc for all the advice and nice diy amps to build in the past:D

This will probably be my last chipamp project after a few more mods to this one, I'm going to try the Hpex classD modules next
 
Nordic said:
Well these weren't in the Mauro amp, and now they are... I would hardly call that sleeping!

I hope you haven't taken my post personally, it was not meant to.
But it was funny to see everyone defend this as an untouchable design, and the best chipamp ever (!!!), and defending the (finally admitted) compromises as good, including those that have only one experience with the LM3886 chip: this amp.

I repeat, I respect Mauro (despite his attitude against me since the beginning, I don't know why) and I think this amp was initially created as a pure exercise, he didn't expect it would go on to a commercial offer and so many people building it.
 
Hey T, couldn't help but notice about the scotkies, (forgive spelling) When I spoke to Rudi yesterday, he mentioned specificaly that he used them, but that they sounded horrible to him, if memory serves correctly, he did like the MUR860 or 680s...

Have you tried those?

Quite frankly the amp is not that bad, in that we are mainly just changeing the power supply circuitry, with the amp basicaly still in basic config.
 
boards or kits.

I have had several request for boards or kits.

first of all i will not be able to do any during the december timeframe. sorry guys. it is summer in south africa and evertyhing is closing down during Dec.

I will have to do a investigation to the shipping cost to several countries. the south african postal system is a ripoff and i need to secure delivery if I were to do kits.

the goodnews is that i do have 400 sprague caps available that do add to the performance of the amp

please mail me if you are interrested and i will file all request and act on it during Jan 2006 (which is not that far away)
 
Nordic said:
When I spoke to Rudi yesterday, he mentioned specificaly that he used them, but that they sounded horrible to him, if memory serves correctly, he did like the MUR860 or 680s...

actually i said that the MBR 10100 gave a very good sound and were very clean but the MUR 860 gives you the bass. most speakers would need this . T. has transmission line speaker and he can sacrifice the bass a bit in order to get more clarity, and the MBR10100 did give me that, it is very nice but robbed the bass a bit
 
Nordic said:
Hey T, couldn't help but notice about the scotkies, (forgive spelling) When I spoke to Rudi yesterday, he mentioned specificaly that he used them, but that they sounded horrible to him, if memory serves correctly, he did like the MUR860 or 680s...

Have you tried those?

Hi Nordic,

I've tried several including the MUR860 and MSR860, remember this was in my system and my opinion, I personally find the MUR860 to have too much of a sonic signature compared to the schottkies, some people report the schottkies to have no bass and be a bit lifeless but I find them to make the sound less coloured, the Mur860's do seem to add more bass and mids, its down to you if you want this or not:)
 
t. said:
I've tried several including the MUR860 and MSR860, remember this was in my system and my opinion, I personally find the MUR860 to have too much of a sonic signature compared to the schottkies, some people report the schottkies to have no bass and be a bit lifeless but I find them to make the sound less coloured, the Mur860's do seem to add more bass and mids, its down to you if you want this or not:)

And what is your reference when you compare those two types of diodes (by saying one adds, the other does not)?
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.