Is LM3886 a good power IC?

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Re: Re: Re: Chip amplifiers

janneman said:



This is conflicting. A high cap supply will appear very close to a regulated supply. A low cap supply will heavily modulate output on peak signals, which both a high cap and a reg supply will do (much) less. Are you sure you haven't switched your statement around?

Jan Didden

Here is my understanding of more detail of the standard discussion:

The LMxxxx chips sound better in the mids and highs when they see lower capacitance (1500uF or so) rather than the 10,000uF or so in a standard power supply. As noted in the other posts this could be for inductance reasons etc.

The initial approach that started the DIY interest was the original Gaincard implementation that used smallish power supply capacitors. This was the original DIY approach that gained much DIY interest. Since the LMXXXX can handle a pretty noise PS, this works without significant hum etc.

The downside is that in demanding situations (difficult speakers, low end bass etc.) the reserve in the caps is to small (arguably) to effectively handle transients etc. So whether or not the classic gainclone 1,500uF approach sounds good has a dependency on the load you are driving.

The regulated approach has gotten a lot of interest recently because one can use large (2x4700uF) caps before the regulator, but small ones at the chip PS pins (33uF-100uF). The idea is that the chip sees low capacitance (or associated low inductance) but can get the current to handle transients and difficult loads.

So the move to try regulated has less to do with PS ripple and more to do with having a reservoir of power that appears to the chip as low (capacitance/inductance).
 
AX tech editor
Joined 2002
Paid Member
Re: Re: Re: Re: Chip amplifiers

demogorgon said:



Not with GC's. as everyone who have played with theese chips know, they sound like **** with high large reservoares of capacitance, or rather the inductans that the capacitans brings with it, to more or less extent.

I'm being rather bastant (as in dead certain that my knowledge is aqurate) here now, but this is regarded as a common fact concerning GC's is it not?

regards
marius

Common fact? How so? Might as well call it common delusion, no? I think you can find the same number of people that agree with you as those that don't. Carlos makes strong statements about this, and people start to believe it. But really it is nothing more than one persons personal opinion. I am not bashing Carlos, just pointing out how these mechanisms work.

However, the fact that low cap values do measureably and audibly modulate the output signal is 101 Electronics, not one man's opinion.

Now of course it is possible that you are right, but the probablility is slim in view of the conflict with repeatable and established demonstrations. What I most miss is the critical attitude of people scratching their head how such an apparent conflict could be possible. (Well, one exception. Dimitri DOES try to find an explanation).

Instead all follow unquestioningly the first statement. Even saying, yeah, but GC's are different. Hogwash! How would a GC which is a simple, outdated design amp of which there are 100's of similar types around, suddenly turn the laws of physics upside down just because they are called that way?

There are other threads on this forum where people make strong statements that an amp with a quasi-symmetrical output stage like so many GC's is horrible to listen to!


Jan Didden
 
Carlosfm,

I think one notion is that it always comes across better when some one says:

"In my experience" or "In my extensive testing I have found"

rather than:

"I tested alot of configurations and therefore approach 'x' will never work because it did not work for me".

Not only is it better because people internalize your argument better, it is also more civil, leading to better dicussion.

Someone may hit on a great combination that is counter to your current implementations and priciples. You may think the chances are low but they are likely not 0%.
 
one person personal opinion

OK, TDA7294 power supply was optimized up to Shottky bridge, then CLC filter, 10000uF, 50mH, 10000uF.
But later, after 3xLM317/337 (in parallel) voltage stab addition all this iron can be thrown away, as now the filter has no influence, no more.
 
As someone who has built more than his share of rotten sounding amps......

I can tell you that what you learn in Electronics 101 does not always apply to making an amp sound good. Especially wrt to cap values in the PSU.

Of course, there are plausible explanations for that.

Oh......wait...some of the hobbyists turned chip amp gurus will remind you that us engineers don't know how to "think outside of the box", and that only they can. Because their minds weren't corrupted by industry.

Yep...........being a chip amp guru.........priceless. (Real end of commercial.)

Jocko
 
janneman said:
Carlos makes strong statements about this, and people start to believe it. But really it is nothing more than one persons personal opinion.

Oppinion?
Yes, why not?
Instead of throwing theories and ashes for the fire why don't you try it?
Believe who you want. I'm certain of what I say, you are not.
You guys seam to make amps "by the book", everything is so "datasheet"-like and so obvious, isn't it?
That is P-A's problem, he did a one-shot design, tried it, it works, let's sell.
No research, no refinement, just happy, just spot-on on first try!
Thats the Microsoft approach, one week after releasing people say it's no good.:clown:
While some of the best amps take months or even years of refinement and research to be made, this guy just throws a chip into a fancy green PCB and expects to make a revolution...

Jan, you should really listen to these chips with a high capacitance PSU.
You could do what you wanted, like I did, it would only sound good if you snubb it.
Otherwise, it's all generic theories.

moving_electron said:
Carlosfm,
I think one notion is that it always comes across better when some one says:

"In my experience" or "In my extensive testing I have found"

rather than:

"I tested alot of configurations and therefore approach 'x' will never work because it did not work for me".

Not only is it better because people internalize your argument better, it is also more civil, leading to better dicussion.

Someone may hit on a great combination that is counter to your current implementations and priciples. You may think the chances are low but they are likely not 0%.

Great, be it as you want.
Let the others do the miracle, I don't believe in miracles.
Also, as you never tried what I am saying, you don't have any idea of who's right.
Do you?
So why discuss?

This thread was opened in MY NAME with words I never said, so the spanking has just begun.
Give me the opportunity to open a thread in P-A's name and I will do that.
Fair play?
Where?:confused:

roibm's thread was a review and a sharing of his experience with this amp.
I didn't know he was going to open a thread here, I was surprized.
He also opened that thread on the Gainclone forum to share his impressions on the sound of the amp.
He was banned from diyAudio.com!
 
carlosfm said:


Great, be it as you want.
Let the others do the miracle, I don't believe in miracles.
Also, as you never tried what I am saying, you don't have any idea of who's right.
Do you?
So why discuss?


I was summarizing the past discussions so that they would not have to be rehashed. I am actually working on an amp per your recommended values. I am expecting that it will sound very very good. I am kind of puttering along a bit each day so it will be a while till I hear it.

So I have great faith in your reseach and results. I am at the same time open to the concept that there will be other implementations down the road that run counter to the priciples in your recommendations that will sound very good as well. Perhaps even better. So I am avoiding the dogma of thinking any one approach or assumptions will necessarily hold for ever. The history of design and science runs counter to that.
 
Re: Re: Re: Re: Re: Chip amplifiers

janneman said:


Common fact? How so? Might as well call it common delusion, no? I think you can find the same number of people that agree with you as those that don't. Carlos makes strong statements about this, and people start to believe it. But really it is nothing more than one persons personal opinion. I am not bashing Carlos, just pointing out how these mechanisms work.

However, the fact that low cap values do measureably and audibly modulate the output signal is 101 Electronics, not one man's opinion.

Now of course it is possible that you are right, but the probablility is slim in view of the conflict with repeatable and established demonstrations. What I most miss is the critical attitude of people scratching their head how such an apparent conflict could be possible. (Well, one exception. Dimitri DOES try to find an explanation).

Instead all follow unquestioningly the first statement. Even saying, yeah, but GC's are different. Hogwash! How would a GC which is a simple, outdated design amp of which there are 100's of similar types around, suddenly turn the laws of physics upside down just because they are called that way?

There are other threads on this forum where people make strong statements that an amp with a quasi-symmetrical output stage like so many GC's is horrible to listen to!


Jan Didden


Attack with full force it is then?

I know what my hearing tells me. im not the guy who usally buy into hypes or what not, and i can tell you for certain that my GC sounds notoriously worse with the 6800uf per rail i'm using now, as opposed to the 2200uf+470uf per rail i initially tried. lacking something in the mids, as many have said before me.

this observation, and the fact that many other share this result makes me belive that we (as in those of us that shares these resoults from testing ourselves) may be onto something here, and the only plausible explonation i'v heard so far is capacitator inductans. now, i do not have the knowledge required to give another explonation, i did not even know that caps had inductance until recently! so that should be your reson for "lacking critical thinking right there.. chipamp builders usally isn't electro engeneers.

now, i will be willing to test other topologies, part quality's and anything else brought to my attension, that is why i'm here. I'm even willing to state that i may be wrong in my current oppinions, as new information and input always has a way of changing things.

so there you have it :bigeyes: .
I was wrong to be "dead certain" when saying GC soundquality decreases when capacitans is added. but i may be right as well.

PROVE ME WRONG!:smash:
 
Re: Re: Re: Re: Re: Re: Chip amplifiers

demogorgon said:
I know what my hearing tells me. im not the guy who usally buy into hypes or what not, and i can tell you for certain that my GC sounds notoriously worse with the 6800uf per rail i'm using now, as opposed to the 2200uf+470uf per rail i initially tried. lacking something in the mids, as many have said before me.
both the "low capacitance > high capacitance" and vice versa arguments are fundamentally flawed... you can't just say that until you've fully analyzed the application and its layout.

I'm thinking 2200uF + 470uF, with the 470 crammed up close to the LMxxxx, will be better than a distant 6800.

High capacitance is a good thing to have, but keep it back by the rectifier. Keep small value caps next to your load to provide adequate decoupling. I think anyone should know this.

I also think people on here should take some time to try and decouple a RF amplifier. Believe me, you learn things about the practical world *real* fast...
 
Re: Re: Re: Re: Re: Re: Re: Chip amplifiers

gmarsh said:

both the "low capacitance > high capacitance" and vice versa arguments are fundamentally flawed... you can't just say that until you've fully analyzed the application and its layout.

I'm thinking 2200uF + 470uF, with the 470 crammed up close to the LMxxxx, will be better than a distant 6800.

High capacitance is a good thing to have, but keep it back by the rectifier. Keep small value caps next to your load to provide adequate decoupling. I think anyone should know this.

Yes, i should have nemtioned:

the 2200+470uf where in the exact same spot the 6800's are now.
8cm away from the rectifiers and 15cm away from the chips. they have 1000uf on board as well.
 
Re: Re: Re: Re: Re: Re: Re: Chip amplifiers

gmarsh said:
High capacitance is a good thing to have, but keep it back by the rectifier. Keep small value caps next to your load to provide adequate decoupling. I think anyone should know this.

And what have I been telling here for a long time?:bawling:
100uf on the chips, bypassed with 100nf.
Big caps on the rectifier diodes/bridges.
Even this way, you need a snubber on the big caps.
1R+100nf poly.
Otherwise you will run from the amp as the plague, it will not sound good.
You will call it... a chip amp.:clown:
Murky midband and treble.
You will not know how these chips can sound.
Read my lips: High-End.
For those who tried a chip-amp with LM chips and didn't like, here's an easy and fast opportunity to give it another try.

Guys, it is much easier and faster to try the big caps and snubbers than reading tens of pages of threads like this discussing the sex of the virgin.:rolleyes:
 
Re: Re: Re: Re: Re: Re: Re: Re: Chip amplifiers

demogorgon said:


Yes, i should have nemtioned:

the 2200+470uf where in the exact same spot the 6800's are now.
8cm away from the rectifiers and 15cm away from the chips. they have 1000uf on board as well.
just a thought - having a lot of bulk capacitance after the rectifiers will cause the conduction angle of the rectifiers to go down.

Then, fourier kicks you in the face - the conduction current transients will cause more higher frequency garbage, *and* I expect that the 6800's will do a worse job decoupling those than the 2200+470uF pair... not to mention the PSRR of the LMxxxx chip gets worse.

Maybe this is why 1500uF seems to work very well, despite common theory... and also explains carlos' snubber suggestion.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.