DIY Passive cooling with PIN Heatsinks SST120 and SST140

robincnn

Well-Known Member
This 1 hour build took me almost 2 hours.

There is nothing Ideal about those Ideal holders.:spew:

Right side you can see Ideal holder for 3590 COB and the Reflector adapter. You can see the + on the holder.

On the left Top. The ideal holder for CLU048 has the +... ok
Left bottom you can see the reflector adapter covers the + on holder and has its own + marking. When i mounted the reflector adapter i did not notice this and + landed on - side.
Had to remove everything and re wire the whole thing all over again. :finger::wall:

upload_2016-8-21_20-14-52.png
 

robincnn

Well-Known Member
GrafTech Graphite thermal pads, get them right the first time. No rework.
I tried to take it off and it is not easy to remove. Tried iso propyl alcohol and still cannot clean the back of cob.

upload_2016-8-21_20-33-58.png

Also these do not cover the cob on corners
upload_2016-8-21_20-35-54.png
 

Airwalker16

Well-Known Member
This 1 hour build took me almost 2 hours.

There is nothing Ideal about those Ideal holders.:spew:

Right side you can see Ideal holder for 3590 COB and the Reflector adapter. You can see the + on the holder.

On the left Top. The ideal holder for CLU048 has the +... ok
Left bottom you can see the reflector adapter covers the + on holder and has its own + marking. When i mounted the reflector adapter i did not notice this and + landed on - side.
Had to remove everything and re wire the whole thing all over again. :finger::wall:

View attachment 3763132
That's a 3070 holder and adapter right?
 

Abiqua

Well-Known Member
Hey @robincnn there's a new citi cob calc on the cob calc thread. According to the data the 90cri clu058-1825s are around 10% less efficient than the 80cri. What are your thought? Because I have never hoped my eyes were deceiving me so much since I checked my lotto ticket last night.
High cri are photon filters, this has been known for a while....you might have higher output in certain nm's over 626, but your overall photon output is decreased and hence efficient as a direct result....

The whole CRI thing makes no sense, we throw away and deem useless other anthropogenic measurements [lumens], CRI is probably even more useless imho...but i digress...should be fun nonetheless.....I like to fuck around with spectrum too, but so far after hundreds of combo's,still haven't seen anything "concrete" :peace:
 

Airwalker16

Well-Known Member
They are similar
Citizen CLU-048 with Ideal Holder 50-2204CT, reflector adapter 50-2100AN
Cree CXB 3070 with Ideal Holder 50-2234C, reflector adapter 50-2100AN
Oh wow, so ideal makes an entirely different chip lok holder for the citizen line huh? I could tell the reflector adapter was the same though.
 

JorgeGonzales

Well-Known Member
High cri are photon filters, this has been known for a while....you might have higher output in certain nm's over 626, but your overall photon output is decreased and hence efficient as a direct result....

The whole CRI thing makes no sense, we throw away and deem useless other anthropogenic measurements [lumens], CRI is probably even more useless imho...but i digress...should be fun nonetheless.....I like to fuck around with spectrum too, but so far after hundreds of combo's,still haven't seen anything "concrete" :peace:
CRI pushes the main peak up to about 630nm, adds a bit more above 700nm, and reduces the relative amount of green by quite a bit. Basically a bit closer to solar levels of relative blue, green, and red, so not entirely just for people. All decent things for plants, theoretically, not sure how much weed cares.

I am also not sure where 10% less efficient comes from, since I haven't looked at the new software, but that doesn't sound right. 3500K in 80 and 90 CRI should be about equal in number of photons emitted, believe it or not. You can ask master @guod how that works, but presumably energy is lost, but the number of photons remains the same.

Since plants really care about photons, not energy, it seems to work out OK. @nevergoodenuf did a head to head, 90CRI was basically equal for yield and lab r sults, with the only real difference being less popcorn.

I think @robincnn is the only one so far to test 80 vs 90 CRI under a PAR meter, and from memory it was a close contest which ended up being within 5% after adjusting for unmeasured photons above 700nm. So within binning margins, basically. But software, even truncating 400-700nm should have them even closer.

Anyway, food for thought. I don't think green light is evil or anything, but it's unnatural to have a higher amoung relative to red, and if we can push the red up a bit into the sweet spot without losing photons, why not?
 

coreywebster

Well-Known Member
High cri are photon filters, this has been known for a while....you might have higher output in certain nm's over 626, but your overall photon output is decreased and hence efficient as a direct result....

The whole CRI thing makes no sense, we throw away and deem useless other anthropogenic measurements [lumens], CRI is probably even more useless imho...but i digress...should be fun nonetheless.....I like to fuck around with spectrum too, but so far after hundreds of combo's,still haven't seen anything "concrete" :peace:
Thanks Abiqua, I guess I missed that lesson, its been a steep learning curve for me. I have been trying to read so much and take so much in. I only just had my head around the cxbs and then the CLU058s were in the lime light. There are many occasions I have felt very stupid and I'm sure there will be more moments to come. Appreciate your comments.

@JorgeGonzales I think there are some errors or I am using the new calc poorly, I spoke to Malocan last night to ask his thoughts and he pointed me to some of your charts. I had somehow not come across them. So they look a lot closer than the calc suggested. Which helped me sleep last night. Thanks for putting all that information together Jorge. Your a top man.

@robincnn very sorry for hijacking your thread, I had a panic moment last night and a big feeling of dread washed over me. The calculator I was referring to is the new Citi calc on here in the thread DIY cob calculator. Anyway after reading the comments above and speaking to Malocan I feel more reassured.
My order from you is due today so I'm looking forward to receiving. Thanks again for the top customer service. I appreciate you making the purchase less painful from a European perspective.

Thanks again folks :clap:
C.W
 

pop22

Well-Known Member
"Meanwell LPC100-2100 = $35*2= $50" New Math or discount..........lol!


Khatod 90 Silicone lens $16*2 =$32
Bridgelux Vero 29 $22*2=$44
Meanwell LPC100-2100 = $35*2= $50
Junction box and glands $10
AC Power wire $1.5
140mmx80mm pin Heatsinks $19*2= $38
Aluminum L channel $10
Total = $185
with Some wagos and screws total ~ $200


View attachment 3758334 View attachment 3758335 View attachment 3758336 View attachment 3758337

Heatsink center is 55C and heatsink edge near the wire is 50C. Would have been nice if there are a little more clearance between the heatsink and driver wires.
View attachment 3758338 View attachment 3758339 View attachment 3758341
Grounded the frame
View attachment 3758342
 

pop22

Well-Known Member
so one for each chip? I actually like that idea. I realize the plus and minuses of it, I may go that way with one of my lights
 

Airwalker16

Well-Known Member
so one for each chip? I actually like that idea. I realize the plus and minuses of it, I may go that way with one of my lights
Hlg-320H-C2100b runs 4 cxb's at 75watts ec. And it's $80-$100. Save yourself the clutter of 4 drivers..
 

Airwalker16

Well-Known Member
GrafTech Graphite thermal pads, get them right the first time. No rework.
I tried to take it off and it is not easy to remove. Tried iso propyl alcohol and still cannot clean the back of cob.

View attachment 3763140

Also these do not cover the cob on corners
View attachment 3763141
You'd think using a graphite pad square that covers the ENTIRE back would be better. Because once you screw it down, it'll clear that hole and give you thermal interface on all points of contact. Even those tiny little edges around the hole. Just need a square that's the citi cobs full length mm X full length mm.
 

Abiqua

Well-Known Member
CRI pushes the main peak up to about 630nm, adds a bit more above 700nm, and reduces the relative amount of green by quite a bit. Basically a bit closer to solar levels of relative blue, green, and red, so not entirely just for people. All decent things for plants, theoretically, not sure how much weed cares.

I am also not sure where 10% less efficient comes from, since I haven't looked at the new software, but that doesn't sound right. 3500K in 80 and 90 CRI should be about equal in number of photons emitted, believe it or not. You can ask master @guod how that works, but presumably energy is lost, but the number of photons remains the same.

Since plants really care about photons, not energy, it seems to work out OK. @nevergoodenuf did a head to head, 90CRI was basically equal for yield and lab r sults, with the only real difference being less popcorn.

I think @robincnn is the only one so far to test 80 vs 90 CRI under a PAR meter, and from memory it was a close contest which ended up being within 5% after adjusting for unmeasured photons above 700nm. So within binning margins, basically. But software, even truncating 400-700nm should have them even closer.

Anyway, food for thought. I don't think green light is evil or anything, but it's unnatural to have a higher amoung relative to red, and if we can push the red up a bit into the sweet spot without losing photons, why not?

All I am arguing is that CRI is derived for human use, technically. The caveat that the increased red spectrums of the high CRI, probably are most hoped to enhance more of the Emerson Effect, but truly that is a merely a by product of selecting for a rendering closer for the human eye and relies on the fact that these wavelengths are in fact acting as described in the Emmerson, but that still isn't known, unfortunately.
Without looking, has there ever been a paper done specifically on Emerson and kanna? Damn, might have to hope over to SciHub and dig......

But I agree with the spectrum anecdotes wholeheartedly, I just wish that we could work towards a true CRI ....Cannabis Rendering Index :) :peace:
 

pop22

Well-Known Member
Although convenient, I've yet to see a thermal pad actually worth using. And the data I've seen shows that their performance is always less than thermal paste.

I spent years scraping them damn bubblegum pads off of cpus.........

Too bad they don't make a spray on coating, would be just the thing with a template. Most people put the paste on way, too, thick. It is meant to fill voids in the contact surfaces, not make up surfaces that aren't flat by more than a few thousandths. I've spent hundreds of hours working with TIM when my company was building custom, over-clocked gaming computers. We found that a layer thin enough to see thru was the best. We applied it to both the heatsink and the cpu. We then scraped it completely off with a single edge razor blade. Next, we reapplied another coating to the cpu that was very thin, so thin it looked like a smudge. testing showed improvement of .5-1.0 C. We did this with the video cards also, some of those cards were hitting 350 watts!
When building these machines, every degree you can shave off is important!

At that time, Arctic Silver was the TIM of choice.






GrafTech Graphite thermal pads, get them right the first time. No rework.
I tried to take it off and it is not easy to remove. Tried iso propyl alcohol and still cannot clean the back of cob.

View attachment 3763140

Also these do not cover the cob on corners
View attachment 3763141
 
Top