Heatsinks for DIY LED lamps

[qwerkus]

I didn't know there was a newsletter?

It used to be stock updates only. But recently' it appears he added prices and comments making it look more like a newsletter.

 

[littlejacob]

Bonjour
10 $ between 3070 & 3590...so 40 $ per 200w panel...you have it back at the first crop...in electricity saved and in yield...!
Have a great day ★

 

[Jerry Liebler]

Bonjour
10 $ between 3070 & 3590...so 40 $ per 200w panel...you have it back at the first crop...in electricity saved and in yield...!
Have a great day ★

First off: The CXB3070 is high bin, (BD) and the CXB3590 is high bin (CD). The lower bins of CREE are about as bad as the Veros
You would need to really be in a high $ cost power area!!! If we dim the CXB3590 (CD bin) panel to produce the exact same amount of light as the CXB3070 (BD bin) panel running all out (109.4 PAR watts) we will find it uses 5.8W less power! Lets say our power costs $0.25/KWhr (higher than most of the US), to recoup $40 requires saving 160,000watt hour or 160,000/5.8=27483 hours. If each crop takes 4 months, the lights will be on 12 hours x 30 days x4 months =1440 hours. More like 19 crops to recover!
As for greater yield, I doubt that too as there is less than 3% difference in output.

 

[SupraSPL]

If I'm going to "abandon" the Veros for CRE due to efficiency, the CXB3590 CD bin isn't the only choice, though it appears to be the most efficient. Perhaps The CXB 3070 BD bin is more price friendly & still offers great efficiency ( 54+%).

If you can source the CXB3590 CD bin for a fair price, there is nothing that can beat it in terms of efficiency OR price. It is a common misconception that the CXB3590 is expensive, it is actually the cheapest/PAR W at those levels of efficiency, which is why we use and recommend them.

 

[littlejacob]

Bonjour
I guess 3% difference is about 8% more buds!
CU

 

[EfficientWatt]

If you can source the CXB3590 CD bin for a fair price, there is nothing that can beat it in terms of efficiency OR price. It is a common misconception that the CXB3590 is expensive, it is actually the cheapest/PAR W at those levels of efficiency, which is why we use and recommend them.

I think I know an exception and is the setup i'm going for ... :

With a 150W driver (ELG-150 in my case, but applies to all 150W drivers):

3 CXB3590 CD @50W vs 4 CXB3070 BB @37.5W have very close to identical performance. (les than 1 lum/W difference ...)

However, with KB new prices, the CXB option is cheaper. The difference in price increases if you count in holder and lenses ...

(And you have more point sources .. always better right ?)

Can you see any reason to prefer the 3590 in this example @SupraSPL ?

 

[Rahz]

I figure this is a good place for TIM stuff. I typed in Vero 29 and the first link was to www.graftechled.com Oddly enough they don't have a Vero 29 product, but they do have Vero 18 and all the large CXB lineup covered.

What's there? TIM sheets made for COBs. Unfortunately the link to the datasheet isn't working. If someone is curious enough perhaps they will contact Graftechled and get them to fix the link.

 

[ttystikk]

I think I know an exception and is the setup i'm going for ... :

With a 150W driver (ELG-150 in my case, but applies to all 150W drivers):

3 CXB3590 CD @50W vs 4 CXB3070 BB @37.5W have very close to identical performance. (les than 1 lum/W difference ...)

However, with KB new prices, the CXB option is cheaper. The difference in price increases if you count in holder and lenses ...

(And you have more point sources .. always better right ?)

Can you see any reason to prefer the 3590 in this example @SupraSPL ?

Dude, quit- you're going to ruin my strategy here... I'm waiting until there's another, even sexier chip than the CXB3590- when suddenly it will be yesterday's news and finds itself sitting unwanted on the bargain shelf... and it will still be every last bit as good then- possibly better with higher bins- as it is right now.

So stop giving everyone funny ideas about how they don't ALWAYS need the LATEST & GREATEST in order to have great lighting at good prices. Okay?!

 

[littlejacob]

Bonjour
Graphtech made some thermopads!
CU

 

[SupraSPL]

True in that example the 3590 has only 2% efficiency advantage and that is probably within the margin of error of the estimate. Assuming Jerry can get CXB3070 3500K BB, that seems like a good design. The only thing that I could think of to improve it is that you could step up from 91% to 94% driver efficiency using the HLG-185 for the same driver cost/W. That would also allow you to fit 4 CXB3590s on one driver and reduce the lens/holder cost/PAR W.

 

[EfficientWatt]

True in that example the 3590 has only 2% efficiency advantage and that is probably within the margin of error of the estimate. Assuming Jerry can get CXB3070 3500K BB, that seems like a good design. The only thing that I could think of to improve it is that you could step up from 91% to 94% driver efficiency using the HLG-185 for the same driver cost/W. That would also allow you to fit 4 CXB3590s on one driver and reduce the lens/holder cost/PAR W.

Looking at the datasheet, it looks like between 80% and 100% load, the ELG is ~93% efficient, although a lot of description text talks about 91% or 92% ... confusing

I don't know if it's because of improvements over original release, or if there is a mistake somewhere ... . Maybe that's just the 500mA version too ?

 

[SupraSPL]

Interesting I never noticed that, I was just going by the first page of the data sheet but I have never got my hands on one to test. So looks like it could be 93% ELG vs almost 95% HLG, not as big of a difference, but what advantage does the ELG offer?

 

[EfficientWatt]

Interesting I never noticed that, I was just going by the first page of the data sheet but I have never got my hands on one to test. So looks like it could be 93% vs almost 95%, not as big of a difference, but what advantage does the ELG offer?

I have a quote for them at 23$ a peice, ~10$ shipping depending on quantity ...
so about 33$ before VAT => ~40$/pc all included and delivered to my doorstep.

That's by far the cheapest W/$ driver I have found with very high efficiency.

 

[SupraSPL]

Dang $23 is great, is it a legit source? Looks like retail is ~$42 ea. At $42 that is $0.28/W. The HLG-185H-C is about $58 retail or $0.31/W.

 

[EfficientWatt]

I made a mistake, its actually 40$ (39.93$ estimated) to my door step, so about 36.5€ ...

And I can't find HLG185 delivered in europe for under 60€ or 65-70$ ...

I guess if I could find them in a shop or have them delivered for 58$, I'd consider it .. until then ELG looks like a good choice to me

EDIT : They seem legit, they say they are original MW supplier and will honor the 5 year warranty from shipping date ..
Marked as a "2 year gold supplier" on alibaba ..

 

[Greengenes707]

I made a mistake, its actually 40$ (39.93$ estimated) to my door step, so about 36.5€ ...

And I can't find HLG185 delivered in europe for under 60€ or 65-70$ ...

I guess if I could find them in a shop or have them delivered for 58$, I'd consider it .. until then ELG looks like a good choice to me

EDIT : They seem legit, they say they are original MW supplier and will honor the 5 year warranty from shipping date ..
Marked as a "2 year gold supplier" on alibaba ..

Check with these guys...
http://lsdpower.com/en/pro_show.asp?id=90

 

[EfficientWatt]

Thanks @Greengenes707

I've seen this site (previously linked by you maybe ?), but gave up when not able to view any of the datasheets ..

They do have a wide choice of options and currents, interesting ...

... I sent an email

 

[Greengenes707]

Thanks @Greengenes707

I've seen this site (previously linked by you maybe ?), but gave up when not able to view any of the datasheets ..

They do have a wide choice of options and currents, interesting ...

... I sent an email

Probably me, I meet them at a conference and they were nice and had a good product in person. Someone on here grabbed a couple.
The options are great. Any and every combo. Some are on hand, while others they have to make.

 

[EfficientWatt]

(Maybe we should move this somewhere else, but I'm not sure which topîc ..)

I used PCT for my example, but maybe you could help with your personal data

Like said in my example ealier: I can see with PCT => if powered at 150 W total : 4x CXB3070 BB are === 3x CXB3590 CD

It looks like, if running the same combo harder, wih 200W, or 250W but same total wattage for eaxh combo : 4xCXB3070 are actually getting a little more efficient than 3xcxb3590.

PCT doesn't give CXB3590 data for the low side, but I suspect that CXB3590 will gain more when running softer than point in my example ....

Does your data confirm those efficiency curves 'crossing over' ?

Thanks !

True in that example the 3590 has only 2% efficiency advantage and that is probably within the margin of error of the estimate. Assuming Jerry can get CXB3070 3500K BB, that seems like a good design. The only thing that I could think of to improve it is that you could step up from 91% to 94% driver efficiency using the HLG-185 for the same driver cost/W. That would also allow you to fit 4 CXB3590s on one driver and reduce the lens/holder cost/PAR W.

 

[SupraSPL]

PCT does not agree perfectly with the specs in the data sheet and PCT is based on case temp despite that it says junction temp. That said, I do see a narrowing in the specs between the 3070 and 3590 as you increase current, but not a crossover. I would like to try and verify that behavior now that I have an Apogee and the correction factors I should be able to test this first hand.