I need to cool down four LM3886 (+-35V, 4/8Ohm load each) but most probably only two of them will be dissipating maximum heat at any given time so the occasional excess of heat could be treated with a thermo-regulated fan.
So the choice is seen in the picture. One big aluminum heatsink (actually got two of them but the other one is not nearly as nice as this one) or four of those little pure copper CPU heatsinks. The big one could be cut etc if there is a need to. I've also scanned a standard PC case fan to better illustrate the size.
Btw, haven't yet tested the aluminum heatsink but those cute little sinks can easily dissipate about 20-30W and with a small fan running at low rpm one such sink easily copes with LM3886 at full power (actually tested with a clipped to square signal with voltage on 4 ohm load dropping to 23V - that's at least 130W!). It barely gets warm...
Aesthetics is also a factor as this amp will most probably be placed in a transparent box.
Any thoughts?
So the choice is seen in the picture. One big aluminum heatsink (actually got two of them but the other one is not nearly as nice as this one) or four of those little pure copper CPU heatsinks. The big one could be cut etc if there is a need to. I've also scanned a standard PC case fan to better illustrate the size.
Btw, haven't yet tested the aluminum heatsink but those cute little sinks can easily dissipate about 20-30W and with a small fan running at low rpm one such sink easily copes with LM3886 at full power (actually tested with a clipped to square signal with voltage on 4 ohm load dropping to 23V - that's at least 130W!). It barely gets warm...
Aesthetics is also a factor as this amp will most probably be placed in a transparent box.
Any thoughts?
Attachments
Thermo regulated fan switching is OK, but by the time it triggers, the chip/heatsink have already become hot. There was an article in Electronic Design about 6 months ago on controlling fan speed based on the current drain in the power stage of the device to be cooled.
The article used to be on line, and probably still is, at EDonline. Look for artcile ID no. 5695. Here is the schematic...
Now if you really want to get nutz about it, you can measure the thermal time constant of your chip/heatsink combo, run the audio signal into a delay line that delays the signal to the amp by that much time, with the undelayed signal controlling the fan. The fan will kick on and start blowing BEFORE the loud signal that heats up the chip arrives at the chip!
On the other hand, it seems like an awful lot of trouble to go through to replace a larger, cheap, passive heatsink...
I_F
The article used to be on line, and probably still is, at EDonline. Look for artcile ID no. 5695. Here is the schematic...
Now if you really want to get nutz about it, you can measure the thermal time constant of your chip/heatsink combo, run the audio signal into a delay line that delays the signal to the amp by that much time, with the undelayed signal controlling the fan. The fan will kick on and start blowing BEFORE the loud signal that heats up the chip arrives at the chip!
On the other hand, it seems like an awful lot of trouble to go through to replace a larger, cheap, passive heatsink...
I_F
Attachments
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.