Different between LM3886T & LM3886TF

[Malquis]

I've a four channel amp (2XLM3886T + 2x3886TF)with 38V per rail 76V total and I'cant notice any difference in sound and almost none in temperature the only thing I noticed is that building with metal case is more difficul because of isolation, and I think in the end the plastic pakage is better cause there is no pad nor heat transfer substance and the transfer of heat betwen the heatsink and chip is better I didn't use a pad on the metal encapsulated ones and I must be carefoul cause that heat sink has -38V !!!
I isolated the hole heatsink instead.
cheers.!!
Malco

[woolly]

this is my 3rd post and i just realised i never said hello yet. HELLO everyone. Loads of good advice here. Thanks for all!

Im half way building 4 x monoblock gainclones using 3875T and 3886T with bare metal on chips. Im deciding on heatsink design, and wondering..
1 What is the metal back connected to? Im guessing one of the power inputs. thats the reason it needs to be isolated. If it was ground there would be no problem right.

2 Is there more heat dissipated with or without the paste/grease? If it is better metal directly against metal, can i use 4 separate and isolated heatsinks? That would actually be easier to fit in my housing

I have some original ideas for some parts of the design. I will post when i check all works!!

[AudioFreak]

Originally posted by woolly

2 Is there more heat dissipated with or without the paste/grease? If it is better metal directly against metal, can i use 4 separate and isolated heatsinks? That would actually be easier to fit in my housing

You need to use some kind of thermally efficient interface such as thermal paste otherwise it won't be metal on metal, you actually will have lots of little air gaps which are very poor thermal conductors.

[AndrewT]

Originally posted by DcibeL
You have to consider the entire thermal path. When using the chip with plastic case, you do not need an insulator between chip and heatsink, and one less layer of thermal paste. I think in the long run the plastic case is definitely worth it, and thermally will not be much worse if at all than the metal tab version.

Other than the package type, the two chips are identical.

Originally posted by burnedfingers
........The uninsulated one doesn't run much hotter when used on the SAME heat sink. The difference is about 2 degrees ...................I believe that a lot of problems are caused when trying to use heat sinks that are way to small. I have used +-40 volt supplies with a 4 ohm loads without any problems....

The TF is much inferior to the T when both are connected direct to an adequate sink with appropriate thermal compound thickness to eliminate all air gaps in the interface.

The TF is still inferior to the T when the T is electrically isolated using 0.5C/W thermal conductor and the TF is compounded direct to the sink.

I am surprised that there was only 2Cdegree difference between the two types.

I strongly support Burnedfingers call for using adequate heatsinking.

The table in the National datasheet is the minimum when driving that power into a resistor load. I think there is considerable confusion in the way National present temperature information.
They repeatedly talk about temperature protecion at 150degC, but I think they are referring to chip temperature (=Tc). The ratings they use are driving Tj to 250degC if I have correctly interpreted the data. At this temperature the Spike protection will repeatedly interrupt the steady state signals into resistor loads.
Into a real speaker, which is reactive, the Spike protection will trigger much earlier. The only way I can see to prevent complete breakup of the audio signal when high levels of output signal are demanded, is to lower the Tc AND Tj well below the test values used by National.
To get there I suggest we all use very efficient speakers to reduce signal demand or approximately double the minimum requirement for the heatsink.

There it is, I got on my hobby horse again, and repeated the same message.
KEEP THE CHIPAMPS COOL.

AN1192 compares max power for the T & TF packages. Table 2 Tf=75%T power dissipation.