diferences between the lm3875 and the lm3886?

Hi all,

I'm just trying to get my head into all this chip amp stuff, and I’m wondering what the differences between the lm3875 and the lm3886 BrianGT kits are, audibly and physically.

I did a search and it seems that the 3875 is the slightly better SQ chip. I plan on building a couple of stereo BrianGT amps to run front and rear speakers.

Any recommendations??

Any input greatly appreciated,

Cheers.
 
Internally, there are big differences. The output current limiting of the 3886 kicks in at around 11.5 A, but the current limiting of the 3875 kicks in at 6 A. The 3886 has a higher rated instantaneous power output. Obviously, the 3886 is a lot more robust, and a lot better suited to low impedance loads (anything below 8 ohm). The rated noise and distortion figures are very similar. If your speakers are 'easy' to drive and are rated at 8 ohms or higher nominal impedance, then try the 3875. If you have notoriously difficult speakers, or they are rated at less than 8 ohms, try the 3886. Of course, you could build both types and compare for yourself, but you could do a lot worse than to choose the more appropriate chip based on the load it will be driving.
 
I can relate my personal experience here, I have built two virtually identical stereo amps using both chips. Both had identical power supplies which consisted of one 500VA/25V dual secondary trafo with a subberized PS from Brian at chipamp.com.

My experience opinion is that the 3886 sounds better all around, but when you have very efficient speakers (8+ ohms) the 3775 is quite close. But even for efficient speakers I liked the LM3886.

But by far my favorite AMPs are my dual mono paralelled LM4780s. I had these driving a pair of very hard to drive DIY 3-4 ohm speakers and they really sing.

Best bang for the buck in my opinion is the 4780 in parallel, but if you are doing P2P it would be a lot harder than the 3886 or 3775. If I were doing P2P (which I have) I would choose the 3886 over the 3775 simply because it sounds great, and you will have more options open to you down the line.

Just my 2C

Have fun! :)
 
If everything is done well and the system is suitable to show differences, LM3875 is still the best sounding chip.

LM4780 in stereo is not that far behind, but it lacks the last bit of refinement and precision. Same goes for LM3886. The opinion is based mostly on response I'm getting from people using the kits, as well my own personal observations.

LM4780 in parallel mode has more driving power, but shows even less of that refinement, which in most systems isn't that obvious anyway.
 
Some info on snubberized supply can be found here: http://www.diyaudio.com/forums/showthread.php?s=&threadid=43423&highlight=

A comparison between snubberized and not snubberized supply was posted here: http://www.diyaudio.com/forums/showthread.php?s=&threadid=56077

I would suggest trafo at least 2 x 22V and 300VA. Best way to buy it locally to save on shipping.

The snubbers are easy to apply, the first impression is usually positive. Later, you might come to conclusion that they alter the sound a bit, but again, it greatly depends on personal preference and equipment used.
 
My objective is to emulate with modern techniques ( as the OPamp ) the sound of the old technology.

Hmmm. I don't really like the sound of 'old technology' if this in any way means vintage tube sound. Or maybe i just like using tubes to emulate the sound of 'new technology' :)

Sound prefences apart, the chip amps don't sound bad on the Signums, they just sound short of breath, dynamics and bass. The 3886 is better in this respect but not by much. Increasing the filter capacitance also achieves little although with high capacitance the snubbers Carlos suggested help a lot. It is really strange that a low power tube amp sounds so much more powerful.
 
Hi, analog_sa!

As a matter of fact, if I speak up now I am just showing off foreign clothes... Because the ideas that follow are originally not mine, but my friend's. Who, hopefully, will talk up himself here, one time.

So, it is just a fresh experience for me, but seems to fit so much your case, that I simply can't resist.
If you want to transform your GC into that tube-like sound, than please, try the schematics described in here; figure 2.

If you look closer, you will notice, that it differs from a classic NIGC by ONE resistor! So, you just leave your original clone as it is [22k/680ohm], lift the return sense point of the negative feedback [the bottom end of the 680 ohm] and put it on top of the .22 ohm resistor.
This will, as described, give You at around 7 ohm output impedance.

[And it works like this, I have measured.]

Now, I would have not believed myself, if my friend does not push me.. And still, I would not dare to suggest it in general.. but, given the fact, that you actually like the sound of your speakers with your tube amp, please, give it a try..

How did you say? Breath? Dynamics? Bass?
By the way, we tried it with the original, 1000 uF config. Stick to it, with more caps it got worse.

Also Mauro mentioned this kind of configuration several times already now. [His Stasis buffer has an output impedance of 3 -4 ohm]

Ciao, George
 
Hi George

An interesting suggestion. Of course 7ohm of output impedance seems a bit extreme even for a tube lover - my 2A3 amp has less than an ohm and several other NFB tube amps i've tried are even lower. What makes things more complicated is that the SF Signum are generally known to sound good with either tube amps or with monstrously current capable SS amps. Any ideas why?
 
general distribution THD, NFB and damping factor ( in all the audio range ). This 3 mixed elements together is the sound of an amplifier of any technology.
When faces the problem with of the OPamp, the first thing that is looked for is of unload the great NFB on a parameter doesn't criticize for to depart audio, and the way the more easy is bring it on the tide. If a output stadium is used to transimpedance , the big part of the NFB works on the dumping in specular way to the load, and is demonstrable that the interfacing is improved with it. On the other hand widens it to mono -triode uses the same begin (transimpedance). the only problem is that remains still a great Gain margin that creates only instability, but this is an other history.
I mine proposed "stasis" ( if uses the resistances after 0.1 % ) introduces in going out the theoretical impedances ( bring in the chart ).
Have not intention to teach you as is you notice this parameter, you go at school even you!
Appreciate your involvement in the diatribe on the audio quality, but have tired me to behaviour the part of the unprovided thing.
On the one hand feel me say to non-being a "audiophile", from the other there are un discreet cognizant of technical that examine my encircled without have found suitable theories.
I don't expose him because have not a good follow-up of the English and the structure of the forum not it allows.
Goings to do the examinations to Pass in her forum, if from it are able! without grudge ( this is my end post ) goodbye

Mauro
 
analog_sa said:
Sound prefences apart, the chip amps don't sound bad on the Signums, they just sound short of breath, dynamics and bass. The 3886 is better in this respect but not by much. Increasing the filter capacitance also achieves little although with high capacitance the snubbers Carlos suggested help a lot. It is really strange that a low power tube amp sounds so much more powerful.

analog_sa said:
What makes things more complicated is that the SF Signum are generally known to sound good with either tube amps or with monstrously current capable SS amps. Any ideas why?

On the LM3886 do you use a zobel on the output?
 
Peter Daniel said:
If everything is done well and the system is suitable to show differences, LM3875 is still the best sounding chip.

LM4780 in stereo is not that far behind, but it lacks the last bit of refinement and precision. Same goes for LM3886. The opinion is based mostly on response I'm getting from people using the kits, as well my own personal observations.

LM4780 in parallel mode has more driving power, but shows even less of that refinement, which in most systems isn't that obvious anyway.

The LM4780 is just a pair of LM3886's in one package.

The internal schematic for the LM3875 is virtually identical to the LM3886 with the exception of the mute function.

The LM4780 in parallel mode has the ability to drive lower impedance loads. Into a nominal 8 ohm load, the bridged and paralleled LM4780 configurations are both capable of 120 watts (with a whoppingly large heat sink.)