LM3876 Gainclone, Bass shy

[pinkmouse]

goldyrathore,

Have you measured your speakers?

It seems most likely to me that they have a lumpy impedance plot, and this is sucking all of the power out of the bottom end. This effect is much more significant than components or construction, and should be ruled out first.

[Morse]

Hi goldyrathore;

I'm with Pink Mouse on this.

Check out the spec sheets on both chips - the LM3876 is pretty much dedicated to 8r loads (pg 9, "output power vs load resistance"), while the TDA2009 actually does better with 4r loads than 8r loads (pg 5, "output power vs supply voltage" - check out the 4r vs 8r curves!!). Thus if you're using 4r speakers, or if there's a big dip in the impedence somewhere in the midbass, it's going to suck the life right out of the system.

Good luck on it and all the best!
Morse

PS Yes, breakin CAN make a big improvement in the sound of some components IME, but the differences between those two chips performance into low z loads is eye-opening!

[ashok]

Hi Goldyrathore,
Can we just call you Goldy ?
Break in of components was something that I was very skeptical about a few years ago. I have a CD67SE that was 'awful' and a Denon amp that sounded flat when I got them . As a result I hardly used them since I have other alternatives. Then the 'alternatives' went out of service due to upgrades and broken parts. So these units came back into service. Both went through opamp changes - obviously - but did not improve as much as desired. However they had to be used till the others came back on line. That was about 9 weeks or so. The transformation has been quite dramatic. I am sure more tweaks will improve this further , but it DID make a difference. So sit back and relax. Pass judgement after at least two weeks of use.

About the heatsink. Bass transients occur fast and if the heat sink is not adequate it will raise the junction temperature dramatically causing all kinds of problems. However when the transient is gone the temp falls. The duration is so short that it will not raise the temp of the heat sink significantly and you might hardly see any long term effect on the heatsink.
So use as big a heatsink as you can accomodate - within reason of course.
In addition , if you are using a 1000uF power supply cap as some people do , you probably will hardly get any bass due to power supply drop. Especially if the speaker impedance is low. Use at least 4700uf or preferably 10,000uF. I can hear lots of protests , but it is up to you to determine what sounds best in your system.

I am still working on my LM3886 gainclone with 4700uF supply caps per channel ( 4700+4700).
Cheers.
Ashok

Don't forget that the load on your preamp is 10K ohms. So check the output cap on your preamp. The one on the power amp is big enough but the one on the preamp output may be not big enough- if there is one over there.

[carlmart]

Originally posted by ashok

In addition , if you are using a 1000uF power supply cap as some people do , you probably will hardly get any bass due to power supply drop. Especially if the speaker impedance is low. Use at least 4700uf or preferably 10,000uF. I can hear lots of protests , but it is up to you to determine what sounds best in your system.

I am still working on my LM3886 gainclone with 4700uF supply caps per channel ( 4700+4700).

As I have been away from this Forum for some time, what I am going to say may have already been discussed.

The question on the capacitance being small if you want more bass can be solved by bi-amplifying. These chip amps are simple enough to use two of them with separate pots at the input to balance the levels of both sections.

Such a setup would work better if you are using a subwoofer, that is a 2-way for highs + mid/mid-low and a sub for the low.

Then you can use just 1000uF for the 2-way amp and 10,000uF or more for the sub amp.

A simple 2nd order filter at the input might do the cross (besides the passive xover inside the 2-way) or you may try an active one. My preference is for the former, to keep things simple.

Then you would have the best of both worlds.



Carlos

[DeonC]

Hi there

May I add one more suggestion. In your first post you wrote:

1> I used a 15-0-15 5 Amp transformer for the supply

If I am correct that translates to a 150VA trannie. Have you considered upgrading the trannie to a 220VA or better? I wuld go as far as suggesting you go as big as you can get. It is known that a 700VA trannie for a PRE-amp makes a HUGE difference over a 300VA, so my suggestion is not totally mad.

I would also suggest upping the total capacitance of the PSU. Adding some larger caps could just be beneficial to the low-end.

One last suggestion is this- if you're going to stay at low voltages, try experimenting with battery power. Find a shop that sells 2nd hand car batteries, and use that as the basis for your PSU. I would still suggest bridging the battery with a handful of small caps (1000uF down to 4.7uF) as this will just improve the 'speed'.

I hope this helps.

Enjoy,
Deon (who is far out there )

[Nuuk]

Have you considered upgrading the trannie to a 220VA or better? I wuld go as far as suggesting you go as big as you can get. It is known that a 700VA trannie for a PRE-amp makes a HUGE difference over a 300VA, so my suggestion is not totally mad.

The amp that I built based on the same circuit that Goldy uses (but with the 3875 chip) produced almost too much bass. It had a single 120VA transformer!

I am sure that 700VA for a preamp makes a difference over a 300VA - but an improvement?

As others here, I suspect that the problem may lie with the speaker matching.

[DeonC]

Hi Nuuk

I am sure that 700VA for a preamp makes a difference over a 300VA - but an improvement?

Well, here's what Neil McBride had to say:

To cut to the kill, the big coil does add a bigger, more expansive sound with noticeably bigger and deeper bass.

The part that got my attention was:

... noticeably bigger and deeper bass.

As a confirmed bass-freak, that equals HUGE in my books

Enjoy,
Deon

[tg3]

You have tried the speakers with another amp, right?

[goldyrathore]

Yes, I do have 4 Ohm speakers and that is screwing things up.
I will try another gainclone with LM3886 and keep you guys posted in the same thread.

Ashok, Do you have PCB track layouts for 3886 gainclone with you that u can share.

Thanks in advance,
Goldy