Mau5Capades: builds & grow journal

ttystikk

ttystikk

Well-Known Member
Growmau5 said:
Thanks Robin.
-36" L 5.886 HSusa profile: 9lbs
-111mm Pin heatsink : 1.25lbs entire array with driver 8.772lbs

I am ok with a margin of error for this video. However, if guys that know whats up like you Robin, Pos, GG, SDS, Purplebuzz, Alesh ,Church, SUp tell me my methods are flawed. Ill pull down the video.
Click to expand...
I think it's certainly convincing in terms of reasonable standards. Research projects attempt to create datasets and draw all but unassailable results, I don't think that standard is necessary here, nor will more rigorous methods yield substantially different results.

Pins better than fins, with fans or without! Cheaper and lighter, bonus!

And yet another well produced video, making sure this stuff is accessible to a wide audience is great stuff! That's something densely designed and written research papers don't do well at all.
 
tags420

tags420

Well-Known Member
Nice. for everything that it is for...basic DIY setups, it is all you need and is plenty good for making a general decision.

Now I'll nit pick and add more positive info about what could be going on....only because you asked. Like I said before...this is plenty good enough for it's intended purpose.


-We have always considered 36" of 5.88 HSUSA to be an active setup, with the failsafe capabilities to run passive if needed. as well as using the extra profile for spread of the build. And did/does exactly that. As far as the chip safeness/performance...many production units wish they could get those temps so nothing to worry about for those who do go the HSUSA route...they also work fine and how they should.

-That was often based on surface area, efficiency, and experience(of a small group). When looking at them on a more mathematical way, thermal resistance, you can see the differences take place, and see whole different strokes for different folk could level the field to be how they were designed...omni-directional heatsinks(pins) or orientation based(extrusions).
  • Pretty easy to get for your mechatronics sinks, thermal resistance is 1.11 right? or 1.06 or somewhere in there. And does not vary too much with orientation from their stated resistance.
  • 5.88's resistance is 1.75/3". That is based on a spot source, so we can break the 36" into 4 peices of 9". Length effect the thermal resistance, and the correction factors(http://www.aavid.com/thermal-tools/length-correct) can be applied. So (1.75 x .6)=1.05...but that is in the vertical position where gravity is helping the airflow and maximizing the sinks design...and where they come up with the 3" resistance in the first place. But that is not how most of us use them. We run horizontal, so let's get that resistance....
Screen Shot 2016-05-02 at 1.22.26 PM.jpg
Screen Shot 2016-05-02 at 1.31.37 PM.jpg
http://www.lrc.rpi.edu/programs/solidstate/pdf/Jayasinghe-SPIE6337.pdf

When in a horizontal position, you can expect only ~70% of the original vertical orientation thermal resistance performance.
images.jpg
http://www.diracdelta.co.uk/science/source/h/e/heat sink/source.html#.Vye4TRUrI_U

With 1.05÷.7=1.5c/w now. ...matching almost perfectly with your test...~41c rise over ambient.
 
Last edited:
BobCajun

BobCajun

Well-Known Member
Apparently the pin heat sinks perform especially well when tilted, as with a spotlight. When vertical, they're similar to other circular heat sinks. At least that's what MechaTronix says in their pin type section. But look at these primo ones on their site for high bays.
 
Airwalker16

Airwalker16

Well-Known Member
I'm wishing my setup was passive pins now! But I'm running 1750ma. & wAnna go to 2400 next run. So I need to find big extruded that are black anodized.
 
Airwalker16

Airwalker16

Well-Known Member
BobCajun said:
Apparently the pin heat sinks perform especially well when tilted, as with a spotlight. When vertical, they're similar to other circular heat sinks. At least that's what MechaTronix says in their pin type section. But look at these primo ones on their site for high bays.
Click to expand...
Damn those are loooong. Takes away some vertical space in a tent. But those could easily run 2400ma. Says 100-120watts ! That's almost a CXB3590's max!
 
Airwalker16

Airwalker16

Well-Known Member
BobCajun said:
Apparently the pin heat sinks perform especially well when tilted, as with a spotlight. When vertical, they're similar to other circular heat sinks. At least that's what MechaTronix says in their pin type section. But look at these primo ones on their site for high bays.
Click to expand...
And those are radial sinks. Not pin.
 
Nu-Be

Nu-Be

Well-Known Member
Growmau5 said:
Thanks Robin.
-36" L 5.886 HSusa profile: 9lbs
-111mm Pin heatsink : 1.25lbs entire array with driver 8.772lbs

I am ok with a margin of error for this video. However, if guys that know whats up like you Robin, Pos, GG, SDS, Purplebuzz, Alesh ,Church, SUp tell me my methods are flawed. Ill pull down the video.
Click to expand...
This is just food for thought, not a criticism, per se.

I don't think you should pull down the video, but I also don't think you necessarily controlled for all variables.

Physics agrees with you about the PIN heatsinks, and it's why CPU heatsink manufacturers have been making incremental variations on two primary themes for awhile: the first is creating a more conductive/convective array of material - adding pins or pipes; the second is getting the mating surface completely flat. (Material type is a possible third, but that's immaterial here.)

I mention this because years of CPU heatsink testing and review have shown that any irregularities in the mating surface of heat source and heatsink will have dramatic effects on the cooling performance and capacity of a given sink. I think it's a pretty reasonable assumption that a machined pin heatsink surface is far more likely to have good and consistent contact with the heat source (chip) than an extruded surface, no matter how much that extruded surface is polished. This is the case regardless of the TIM used. The extruded surface would have to get milled and polished to be equivalent.

This is mostly an academic argument - it's the sort of rigor required when you submit to a peer-reviewed journal. Once you can control for this in your documentation, you'll be closer to publishable. I've seen this issue cause a much greater disparity in CPU heatsink performance than the one you described here, but I'm not going to claim it's the cause in your case. Pin heatsinks are simply better, overall, and thermodynamics proves that.

So, how's about a solution, eh? Your machine shop & fabricator buddies/teachers should be able to help. :)
 
Growmau5

Growmau5

Well-Known Member
@tags420 thanks for the post and the links to do some further research on the topic. I agree, the 5.886" was the go to for a lightly active (small fan on 5v) with the passive safety net. Seems like these days everyone is obsessed with "being passive" for one reason or another, while expecting to get Supra's efficiency numbers at high Tc. And I agree, nothing to worry about with this tried and true set up.

@Nu-Be I have seen high-end PC guys go to great lengths to get a matting surface that was flat within 0.001" I can machine a surface close to one or two thousands, but man, its a lot of time and work. You bring up a great point, long extruded heatsinks aren't perfectly flat unless surface machined, its just not in the nature of the extrusion process with aluminum.

hsusa5886.jpg

pin sinks perform best when angled or tilted. but what if the pins were hollow? would this be enough of a convection current or draft to improve its performance, i guess we shall find out.

hollow pin.jpg
 
Airwalker16

Airwalker16

Well-Known Member
Growmau5 said:
@tags420 thanks for the post and the links to do some further research on the topic. I agree, the 5.886" was the go to for a lightly active (small fan on 5v) with the passive safety net. Seems like these days everyone is obsessed with "being passive" for one reason or another, while expecting to get Supra's efficiency numbers at high Tc. And I agree, nothing to worry about with this tried and true set up.

@Nu-Be I have seen high-end PC guys go to great lengths to get a matting surface that was flat within 0.001" I can machine a surface close to one or two thousands, but man, its a lot of time and work. You bring up a great point, long extruded heatsinks aren't perfectly flat unless surface machined, its just not in the nature of the extrusion process with aluminum.

View attachment 3671791

pin sinks perform best when angled or tilted. but what if the pins were hollow? would this be enough of a convection current or draft to improve its performance, i guess we shall find out.

View attachment 3671799
Click to expand...
Omg you're FABRICATING your OWN Hollow Pinned sink?!?!
 
You must log in or register to reply here.
Top