pls help w. figuring out heatsink (fins) size for a LED aluminum slab heatsink

[liesandlies]

Howdy y'all

in general, how much are cfm requirements lowered, for each 1x increase in cooling surface area due to adding finned heatsinks?
(example.. if a 100cfm fan can cool an aluminum slab.. and heatsink fins are added that double the slab's surface area.. what would the cfm new req's be)
((intenting to use 3m thermal adhesive tape of .6W-m/k or 1W-m/k, though hopefully some general guideline can be fairly simple and basic ))


Situation in question is wishing to quietly cool.. a 48 led panel, using 90w (actual), attached to a 10"x9" aluminum slab.

The factory fan is too noisy. Installing spare fans found that a fan of 70cfm, is the minimum that will cool a 60watt portion of the LEDs.
That 70cfm fan is also a bit louder than preferred, so using a higher cfm fan to adequately cool the full 90watts is a negative. (and the light case will only fit a 120mm fan.. and keeping all the parts inside the case would be preferable if it's possible)

It seems that attaching finned heatsinks is one idea.
And.... there are dozens of sizes of heatsinks, so I'm wanting to have some general idea of how much heatsink surface area would be needed, so that a ~60cfm fan can cool the full 90watt array.
Can it be done?

 

[Doer]

In general, rules of thumb, i.e., (more formally named as,) heuristics, are used.

And SpuraSPL has refined these for us and I extrapolated water from that.

75cm^2 - completely passive, no fan.
25cm^2 - active fans, constantly on
5cm^2 - flowing water

So, you use a measurement called the perimeter per inch. And you calculate how many inches in length (converted to cm) with give the area you need to dissipate the watts, (more than 1/2 the total watts, light is the rest of the work)

 

[Doer]

Now that you see this, CFM is not what we need. I has to be in Linear Feet per Minute. I did a bunch of calculation on this in my build thread early on. (a few weeks ago. )

https://www.rollitup.org/t/cxa3590-build-plan.863105/

You need another factor of how fast the air goes past your fan per the fan diameter per minute, based on it's CFM.

 

[liesandlies]

Okay, so.. a lot of equations there, .. I'd try to calculate it out, but I can't find the exact equations to use. And I was thinking a general idea might be enough. I'm reminded of one webpage I saw that had a crazy amount of equations and nomenclature, it was like a advanced physics chalkboard :b

... Like for a 120mm 90cfm fan, blowing downwards onto a 90w 10"x9" aluminum slab (90square inches)... cools it fine.
... And if you add heatsinks and increase the surface area to a total of 180 square inches... then the equivalent 120mm cfm would be lowered to __ cfm
... If you triple the surface area to 270 square inches... the cfm would drop to ___

Planning to use double-sided thermal tape/adhesive... to allow for removal and future changes. With the syringe adhesive, once placed, it's permanent, so no ability to change it if it's not adequate.

cfm is not lfm, and this leaves out factors of c/w and such... but just some general ideas.


water is attractive, but a bit more complicated and likely cost more than air (as well as requiring water).
side-note.. One idea that's been done with computers is to use a large water reservoir, many gallons, and cycle the water through that.. (reservoir water, pumped to water blocks, and going back to the reservoir). Instead of using a fan to cool the water (dissipate heat from it), the large reservoir acts the same as a huge heat-sink, without needing active cooling. During lights off time the heat can dissipate a bit, or ice cubes can be added, or warm water removed and cold water added.

 

[CatSpeaker]

I went with the slab of aluminum, 23 3/4" long, 9" wide, and .52 thick.
covered it with Dell heatsinks, cooled by half dozen 60mm fans.
Glued on with Arctic Silver Alumina Epoxy.
Massive overkill, running 4 CXA3070s.
I suck at math so I went with the overbuild the fuck out of it mentality.
At work now, will post pictures Sunday night, unless GOT interferes.

 

[Doer]

Okay, so.. a lot of equations there, .. I'd try to calculate it out, but I can't find the exact equations to use. And I was thinking a general idea might be enough. I'm reminded of one webpage I saw that had a crazy amount of equations and nomenclature, it was like a advanced physics chalkboard :b

... Like for a 120mm 90cfm fan, blowing downwards onto a 90w 10"x9" aluminum slab (90square inches)... cools it fine.
... And if you add heatsinks and increase the surface area to a total of 180 square inches... then the equivalent 120mm cfm would be lowered to __ cfm
... If you triple the surface area to 270 square inches... the cfm would drop to ___

Planning to use double-sided thermal tape/adhesive... to allow for removal and future changes. With the syringe adhesive, once placed, it's permanent, so no ability to change it if it's not adequate.

cfm is not lfm, and this leaves out factors of c/w and such... but just some general ideas.


water is attractive, but a bit more complicated and likely cost more than air (as well as requiring water).
side-note.. One idea that's been done with computers is to use a large water reservoir, many gallons, and cycle the water through that.. (reservoir water, pumped to water blocks, and going back to the reservoir). Instead of using a fan to cool the water (dissipate heat from it), the large reservoir acts the same as a huge heat-sink, without needing active cooling. During lights off time the heat can dissipate a bit, or ice cubes can be added, or warm water removed and cold water added.

Yeah, there is active cooling of water thru a radiator and a water ballast tank, like I use. Of course you can add all kinds of hybrids to this, both active and passive.

I had100 gals in the shade and it was good for sinking a 1000w hps.

I wish I could tell you about your fan calculation. I started my thinking there, I really did. It is not exactly the right question, somehow. Air is massive and compressible. The force equations are Cubes and Cube roots according to speed, temperature and density. If you are firing straight down into a heat sink CFM turns into LFM in about 1 inch...bam, it hit the bottom and fans out. So, it doesn't scale like 2x is 2x better.

You add the heat to the grow and increase the ambient and lower the sink's capability, that reaches an equilibrium that is difficult to predict. There are so many other factors.

It is a main reason why I have always chosen to transport heat out. But, in some places, you need that heat.

The way I ended up looking at it, is build cost per watt of PAR. All is traded off as you can see in my build thread, and in the end I convinced myself of a few things.
- efficiency is the reason I am doing this. Light is light.
- COBs and drivers cost the same, use 1 driver for several COBs, not 1:1.
- GO BIG on the cooling. It is the least of your worries, unless too small. Then it's failure.

 

[liesandlies]

I went with the slab of aluminum, 23 3/4" long, 9" wide, and .52 thick.
covered it with Dell heatsinks, cooled by half dozen 60mm fans.

lol, yes, this is the opposite direction for me.
I'm wanting to get the cooling to be silent from 5' away.

 

[liesandlies]

Yea, my original intent was something simple and small and quiet. I was looking at led strips, but the cost of the strips, power supply, and parts.. was about the same as a 90w 6 band manufactured light using 3w chips, and a lot of proven grows and reliability. The fan decibels is the one sticking point. It came with a fan of about 33dba.

cooling led strips can be done passive. but cooling 90watts in a 10"x9" area takes more active cooling.
I could either:
add some small amount of heatsinks, inside the case, and use a quiter fan.. but without being sure of airflow design
or
take the slab out of the case and use a lot of heatsinks, and use a 250mm fan to blow down on the entire area

I'm not able to get any new parts for a while.. so it's all just for imagination anyway