Max power from LM3886TF

[Arx]

What is the maximum power you could expect to get from a LM3886TF assuming extremely good cooling?

Judging by the datasheet with a 40V supply, you should be able to get a little over 80W into an 8 ohm load.

Has anyone tried this?

What's the highest continuous wattage anyone has ever reliably attained and measured?

Also, has anyone ever burned one out trying for high power?


Any answers from anyone who's actually tried it would be best.
No suggestions to use the T please... I know it's better thermally. I want to know what can be done with the TF (I have lots)

Anyone tried sanding them?

-Nick

[BWRX]

Originally posted by Arx
What is the maximum power you could expect to get from a LM3886TF assuming extremely good cooling?

(I have lots)

Then you shouldn't have any problem sacrificing one in order to answer your question

[peranders]

Originally posted by Arx
Any answers from anyone who's actually tried it would be best.
No suggestions to use the T please... I know it's better thermally. I want to know what can be done with the TF (I have lots)

It isn't important really if the max power is 75 , 80, or even 90 W. The interesting figure is which power region you are. The difference between 75 and 80 W is 0.28 dB! You won't hear the difference in sound pressure at least. If you are thinking that 50-75 W is too little youmust consider 100-200 W instead.

If you are interesting in high continous power youmust use the T model.

[Nordic]

Sadly with these chips, when you reach a certain power threshold, the THD figures goes to hell... (look at datasheet) As with most chips I think the result is best at about half of the max figure...

[AndrewT]

Hi,
the worst case condition into a non reactive load is not maximum power. In fact testing to maximum into an 8r0 load should never break a properly designed amplifier.
Half power is worse for chip stressing.
Reactive loads are much worse for chip stressing.

If you want 60W into a real speaker then you should be aiming for 50 to 60 degree phase angle. This is worse than halving the resistive load value.
What I am suggesting here is that your 60W into 8ohm amp MUST be capable of 100W into 4r0 as a minimum spec for driving an 8ohm load.
If you have 4 to 8ohm speakers, that resistive test load becomes 2r0.

Now try that through a TF package (using a cold sink and for less than 1second).

[Arx]

Originally posted by BWRX

Then you shouldn't have any problem sacrificing one in order to answer your question

I don't have any problem sacrificing one. I just figured if someone else already had it'd save me the trouble.

Originally posted by peranders
If you are thinking that 50-75 W is too little youmust consider 100-200 W instead.
If you are interesting in high continous power youmust use the T model.

That's true, assuming I'm talking about squeaking out only 5 W. I don't even really know what's expected from those chips though, so if I'm talking the difference between the 50W they seem to be using as a reference for 8 ohms in the datasheet, and 80W that's a significant difference

Originally posted by Nordic
Sadly with these chips, when you reach a certain power threshold, the THD figures goes to hell... (look at datasheet) As with most chips I think the result is best at about half of the max figure...

Whaa??? Are you reading the same datasheet as I am? I don't see a strictly THD graph, but it looks to me like the THD+N continues to drop right up to the point of clipping.

Originally posted by AndrewT
Hi,
Half power is worse for chip stressing.

Half power may be worse if your testing is very simplistic, but testing with a real load and a variety of signals I don't think that will be the case.

-Nick

[AndrewT]

Hi,
half power might be very common (or even predominant) for an amp driven to clipping on low dynamic range material. Then the heating could be beyond any chipamp to heatsink interface.

[Arx]

Originally posted by AndrewT
Hi,
half power might be very common (or even predominant) for an amp driven to clipping on low dynamic range material. Then the heating could be beyond any chipamp to heatsink interface.

A good reason to avoid clipping.