LM3886GC, Does it prefer 4ohms or 8ohms?

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
8 Ohm definately.

I recently completed 4 channels and made the mistake of getting the plastic parts for convenience. Also, I'm running at 35-0-35 when you should run at 28-0-28 if you use 4 ohm. I'm running it on my Jamo's which are 4ohm and these chips are getting HOT! I have a very big heatsink on them, but the losses on the plastic package is too high, so it cannot transfer the heat quick enough. They sound quite good though, for incase you didn't know ;)
 
I guess I don't understand the power supply voltage issue. I heard a few times now that 4 ohm loads work better with lower voltages. Is this because the load is drawing more power and a lower voltage power supply keeps things under control? Even if you have a high voltage power supply, couldn't one compensate by reducing the gain...playing with the feedback resistor? Excuse my ignorance...I'm very new at this :D
 
Again, I'm speaking from the experience I had and from what I read.

Apparently you keep the voltage lower into 4 ohms to prevent heat spots building up on the chip. There is not a hell of a lot of dissipation area on the chip, so there is not a lot of contact surface area to dissipate heat. You can have as big a heatsink as you want, the thermal footprint of the chip is still small!

I think my heatsinks are big enough ;)
Picture185.jpg


By using 28-0-28V supplies, you limit the dissipation in the chip if you want to run 4 ohms. When I first read it I didn't take it too seriously, but now that I've run it for a while I realised they were quite serious about it! :hot: Anyway, mine won't be used on 4 ohms permanently, they will run on my studio monitors once I'm done building all my speakers.

The point I was trying to make (like an idiot, I must say!) is that if you were to run 4 ohms, you need to have your PSU at 28-0-28V AND you shouldn't use the plastic package. Use the metal tab version instead.

I hope I haven't confused the issue any more
 
Think of the load at the end of your amp as being powered....
remember all the power doesn't go to the load, but it DOES go to the chip, so if you hit the chip's maximum level, you will likely overheat it it and burn it out.

The minimum rateing for lm3886 is about 7 AMP (you can look it up in datasheet) - this means with variations between chips, they should not blow up till they reach at least 7 ampere - in fact it shows a possible maximum up to 11A depending on the chip in your hand.

now take your load (speaker/driver) and divide it into the supply you got hooked up.
40V (almost max for lm3875 level) would on lm3875 run into 8 ohms. 40/8 = 5amps at 40V

for the lm3886 you can see that 4ohms x 7ampere = 28V. (as it is the minimum figure - remember some could make it to 11A) so it could in theory run onslightly higher rails....

But it should make the whole thing transparent to you for future use.
 
henkel said:
Hi!

Planning to build my LM3886 soon, also planing to build a pair of speakers for it. Does the LM 3886 prefer 4ohms or 8ohms drivers?

Thanks.

Just from my experience. LM 3886 can work both but it's recommend to reduce the supply volatge not more than +/- 35V (ideally +/- 28V) to make most use of 4 Ohm while avoid activating the protection circuit which heavily degrade the sound.

I also have 3886TF with +/-35V on 4Ohm speaker. The chip run really hot, and my big heat sink get warm pretty quickly.


Note, LM3886 can deliver ~40W@ 8Ohm and ~ 70W @ 4 Ohm on paper. However, the difference is hardly observed in pratice -- at least in one of my system.
 
AX tech editor
Joined 2002
Paid Member
CarlosT said:
Again going back to a related question...if you had such a high voltage power supply that would drive the LM3875 into the self-destruct territory, couldn't the gain be simply adjusted down into a safer range by tweaking the resistor values of the feedback loop?


You cannot always decrease the gain, there is a minimum below which the amp becomes unstable.

But even if you did, the max output doesn't change - it would just take a bit more on the volume control!

Jan Didden
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.