split heatsinks?

[george a]

Hi All

I have a query about heatsinking.
I'm using a square across section tunnel heat sink with a fan at one end .
The tunnel is made of two identical halves which lock together to form the square cross section.

With fan cooling the heatsink is rated at 0.30degC/W as a pair (tunnel configuration- joined) or at 0.35degC/W individually, ie: could possibly be used as 2 heatsinks at 0.35degC/W

With two amp channels using the joined heatsink I feel I would be better of splitting the tunnel and getting approx half the temp rise.
Am I missing something here??
How far apart should I keep them to still enable sharing of common fan

Thanks in advance for any suggestions
Cheers

[GRollins]

I'm not clear on why you want to throw away part of your heat dissipation capability. Are you trying to get away from using a fan?
The possibilities are thus:
1) Heatsinks together with fan
2) Heatsinks together without fan
3) Heatsinks apart with fan(s)
4) Heatsinks apart without fan(s)

1--Heatsinks together with fan will work very efficiently, but with the arguable downside of fan noise
2--This one I would not recommend, as the air flow will be somewhat restricted. The very thing that makes option 1 work is against you here. With the heatsinks together, the air flow from the fan is forced to stay within the confines of the tunnel where it will do the most good. Without the fan, you're dependent on convection and will need all the air flow that you can get.
3--This will work, but somewhat less efficiently than a tunnel. You still have fan noise.
4--Less efficient at heat transfer, but convection is silent.
I would vote for either option 1 or 4, depending on fan noise issues.

Grey

[george a]

Hi Grey

Thanks for your response
I'll clarify the issue
The heatsink is for an Aleph 5 pumping out X watts of heat if both heatsinks are mated together they appear to act with a capacity of 0.3degC/w with fan and a temperture rise of some nominal amount.

The manufacturers literature to my eyes implies that each half of the heat sink has a capacity of 0.35degC/W on its own-not much worse than when mated together- possibly due in inefficient thermal coupling along the join when mated.

if this is the case one could reduce the tempertare rise by directing 1 channel to each independent "half" - thereby heat load is X watts/2 into each 0.35deg heatsink versus X watts into 0.3deg heatsink ie: approx half the temperature rise........

heatsink details
http://www1.jaycar.com.au/productVie...Max=&SUBCATID=

Have i missed something or am I just proving I'm an idiot (again)
The A5 will be in a cup'd (vented via fan also) and I'm trying to keep the temp down to stop my 70deg C thermal trips from cutting in.
BTW plastic face temp of irf240 is 60degC

Cheers and thanks

[wuffwaff]

Hi,

the difference between the two options is Rth 0.6 / ch or Rth 0.35 /channel. For a 25° temperature rise this means 25/0,6 = 41 watts or 25/0,35=71watts

Both are a bit low for an Aleph5 (34V, 2.1A gives 143watt/channel). This means a temp rise of 86° or 50°

You need a sink with around 0,1°K/W for an Aleph 5 (stereo)!

William

[bowdown]

hey George, if it helps i use the exact same heat sinks from jaycar for my test bench, in one of my tests when ambient temp was 30ish deg C and the heat sinks are in a tunnel config without a fan, within about half an hour of a sine wave of 1khz @ 200 wrms and @ 8 ohms the heat sinks get up to about 65 degrees C and rising but as soon as i put a normal 12v fan on it,it drops from 65 to 45 degrees C in about 10 min and after an hour the heat sinks stay at a constant temp of about 40 degrees even after 2 hrs the temp is stil around 40 degrees. hope this helps...

Cheers

Bowdown

[george a]

Hi all andthanks for responses.
How do i confirm bias current on the Aleph

I have it throttledback at moment to give 24deg temp rise (that is plastic transistor face not heatsink) above ambient so less rise than that on heat sink.

Bowdown did the 200w load go direct into the sink??


How do i confirm actual level of idle current on the A5- i have about 1.7V across the 1 ohm reistors on the output mosfets ? (6 per channel total)

Cheers and thanks

[wuffwaff]

1.7V over 1 ohm is 1.7A (V=IxR)..........

Is the 1.7 the added value of the three output fets?

Normally you measure the voltage over the three source resistors, add the values and devide by the resistor value. This is the total current.
The dissipation is VxI is two times the supply voltage x total bias.

William

[george a]

actually make that 0.17v not 1.7 volt across each resistor

so prob 0.17A X12
= apprx 2A


sounds good though!!!!

[wuffwaff]

eh,

sounds not so good! you must only measure the resistors of the output or the current source. Not all of them cause then you will measure the same current twice.
Do you have 6 or 12 fets per channel?
If 6 then you´l have 0,17x3/1= 0,51A bias wich is enough for 4watts into 8 Ohms or 2 watts into 4 Ohms
if 12 then you still only have 1A bias wich means 16 watts 8 ohms.

William

[george a]

Wuffwaff

are you sure??