lm3886 design

[pacificblue]

The trouble is, most manufacturers seem to think that publishing the thermal data of CPU heatsinks is a breach of etiquette, so you hardly ever know, what you get. Some DIYers seem to think the same about calculating heatsink sizes and that Trial-&-Error is the only way to go.

What little thermal data is available indicates that CPU heatsinks are indeed in a reasonable range to fit a chipamp. Most show 2-4 K/W without a fan and 0,5-1,5 K/W with fan. But again, you hardly ever know, what you get, and therefore you absolutely should provide the fan.

[ttan98]

Originally posted by pacificblue
The trouble is, most manufacturers seem to think that publishing the thermal data of CPU heatsinks is a breach of etiquette, so you hardly ever know, what you get. Some DIYers seem to think the same about calculating heatsink sizes and that Trial-&-Error is the only way to go.

What little thermal data is available indicates that CPU heatsinks are indeed in a reasonable range to fit a chipamp. Most show 2-4 K/W without a fan and 0,5-1,5 K/W with fan. But again, you hardly ever know, what you get, and therefore you absolutely should provide the fan.

If one wants to DESIGN PROPERLY, then it always best to design on the CONSERVATIVE side esp. when is no data on the heatsink. Hence a fan is necessary in all occasions.

I for one has a different policy, I don't like to over design, if the heatsink is not even warm when amplifier is idle and I am using easy load and don't play very loud, should turn on the fan permanently? Ans is NO.

[AndrewT]

Originally posted by ttan98
If one wants to DESIGN PROPERLY, then it always best to design on the CONSERVATIVE side

No.
If one wants to design then one must define the specification that must be met.
Any design that meets the specification is a good design.
If the specification is lacking then that does not make for a bad design, it indicates the lack of foresight on the part of the team defining the specification.

There is no need to be conservative nor to over design nor risk unreliability. It all should be defined.

[ttan98]

Originally posted by AndrewT
No.
If one wants to design then one must define the specification that must be met.
Any design that meets the specification is a good design.
If the specification is lacking then that does not make for a bad design, it indicates the lack of foresight on the part of the team defining the specification.

There is no need to be conservative nor to over design nor risk unreliability. It all should be defined.

Ideally yes, it does not occur in every circumstances, this is the real world.

[gabanyayaya]

Originally posted by ttan98


Ideally yes, it does not occur in every circumstances, this is the real world.

This is true. Parameters meant as a base guide on every things we do but the implementation is up to the builder to decide as long as safety precaution taken for consideration even not 100% as stated to the manual....remember this is a hobby where we spends time and effort on something that we passion about and this is how diyers get creative...according to their pockets...

IMHO....

[AndrewT]

Originally posted by gabanyayaya
get creative...according to their pockets...

and that should be part of your specification.