...All Things Vero...

Would you consider buying a VERO after reading through some of the posts?


  • Total voters
    357

stardustsailor

Well-Known Member
And some info about suggested Clamping Loads For the Vero series :
(Flat head screw ,w/o washer )

Vero 10 : 2 x M2.5 x 23 N*cm torque = 92 Kg total ( = 202.8 lbs )
Vero 13
: 2 x M2.5 x 34 N*cm torque = 136 Kg total ( = 299.8 lbs )
Vero 18
: 2 x M2.5 x 34 N*cm torque = 136 Kg total ( = 299.8 lbs )
Vero 29
: 4 x M3 x 45 N*cm torque = ~300 Kg total ( = 661.4 lbs )

http://www.engineersedge.com/calculators/torque_calc.htm

*****With the use of flat washers , total clamping loads,
could be increased up to 5x,depending on the washers used.

(Wider washer = higher torque= higher clamping load. )

Of course ,this is allowed only for the Vero series ,of which the substrate is made from aluminium.
Regarding other types of COB arrays
with ceramic substrates -i.e. Cree CXA-CXB -there's no way ,
that such high clamping loads could be achieved / reached .

( My DIY V series : 1330+ kg !!! ( 2930+ lbs ) of total clamping load ,per Vero 29 )

The array will break in pieces ,as it is made of fragile ceramic material.
It gets clear ,that for high thermal performance of a high power COB ,the chosen substrate should be made from Aluminium or Copper ,as they offer high thermal conductivities-thus lower thermal resistance between junction & case- ,
while they are able to withstand excessive "clamping loads " from Array holders /screws ,
thus insuring better surface contact ,which results in higher heat conduction .
Ceramic substrate used for COB LED arrays , AlO =Aluminium oxide=Alumina ,is a bad choice for high power LED COBS.It is very fragile,not able to withstand high clamping loads ,as also it has poor thermal conductivity ,vs Aluminium or worse vs Copper.

- For comparison : Ideal CHIP-LOK 50-2234C array holder for CXA/B 3070 =
2x M3 x 0.5 N*m torque =166.66 kg total clamping load ( ~367.4 lbs ) . MAXIMUM -

But alumina has low heat expansion coefficient ,and thus a material with / of low mechanical stress ,due to heat cycling .

Still...
Higher thermal conductivity ,means better cooling .
Better colling means ,more efficient real-time operation( lower Vf droop ) .
Better cooling means higher light output maintenance,thus longer service life ( lower Tj ).
Better colling means higher efficiency at higher driving currents.


You'll be suprised ,when in real life surfaces contacting ,can be as low as 20% or even lower , of the "theoritical"surface area ,especially with low ( or w/o ) clamping load applied .
Like :

cxa heat sin k.jpg
In blue color is the bottom ceramic super smooth surface of a Cree CXA3070 array .In red is the untreated
(only degreased ) aluminium surface,magnified many-many times ...
The white gap in between is supposed to be decreased by sanding the heat sink surface and/or apply lots of clamping force/load and of course what's left then ,to be filled with TIM .
Cheers.
:peace:
 
Last edited:

DonPetro

Well-Known Member
ordered 10 vero10 5000k, two drivers (5 on each), two 36" x2" heatsinks and some 50mm fans and 9v supplies. Time to rock out some veg lights! woot.

Driving them at just under 300ma. Should get good efficiency that way.

building two fixtures one for each level of veg.
Im also planning a 5- Vero10 fixture but will likely upgrade to it to 10-Vero10s since i have a 600mmx140mm heatsink coming. Will run passive with the 5 and bump up to active cooling with 10.
 

SupraSPL

Well-Known Member
To clarify, there is a big difference between alumina ceramic and aluminum nitride ceramic, which has a thermal conductivity up to 200 (W/m K)
From the Cree thermal management PDF
thermal conductivity.png

This article goes into it in more detail about aluminum nitride ceramic COBs


This article quotes Cree's Vice President Applications Engineering Mark McClear:
Cree also exploits the lower thermal resistance ALN (aliminum nitride) substrates offer to produce packages for demanding uses. "This is particularly important in high temperature, high drive power, high current applications that also demand high reliability," McClear underlined. Streetlights again are a good example of such an application. "Bottom line, choose the right LED for the right application, and the substrate is part of that choice."

From the same article:
Ceramtec "The packaging started with AL2O3 (alumina ceramic), but is switching to ALN (aluminum nitride ceramic). We have upgraded our standard ALN to 200 (W/m K)"

So we should keep in mind that Cree engineers are no fools and they chose to continue using the ceramic base for the CXBs. Different architectures will have different pros and cons. For example, if using the ceramic base allows them to achieve higher efficiency from the start, that can override the small difference in temp droop. The tradeoff, it seems to affect current droop characteristics, as we have seen the CXA is the best choice for those who are looking for very high efficiency designs, the underdriven Vero simply cannot achieve the efficiency/cost levels of the underdriven CXA, no matter how much money you throw at it. And the other side of the coin, those who are looking to minimize up front cost and pack huge amounts of light into one COB, the Vero29 dominates.

That said, the difference in cooling performance between the CXA3070 and Vero29 is small in practice, even at 2.4A where the CXA is almost maxed. And that is not a fair fight, I will have to check the thermal performance of the CXA3590 to get a better idea what the aluminum nitride ceramic is capable of.

Another important point, the cooling performance does not even come into play for applications where the COBs are run relatively soft. For example, at 800mA (28W) I measure .75% temp droop on the CXA3070. So why would Cree want to sacrifice that efficiency for a slight thermal improvement that does not translate to much higher output? Also, I was shocked to discover, as long as the entire surface is sufficiently coated with thermal paste, there is no thermal benefit to sanding the heatsink, even at 2.55A.

As far as fragility, I have handled hundreds of CXAs and I have only ever had 1 crack, which was my fault for prying on it when trying to remove it. The funny thing is it still works and did not lose any output, I use it all the time :)

So they have different characteristics, that is a great thing. You just have to choose the one that most closely suits your design goals.
 
Last edited:

stardustsailor

Well-Known Member
To clarify, there is a big difference between alumina ceramic and aluminum nitride ceramic, which has a thermal conductivity up to 200 (W/m K)
From the Cree thermal management PDF
View attachment 3356133

This article goes into it in more detail about aluminum nitride ceramic COBs


This article quotes Cree's Vice President Applications Engineering Mark McClear:
Cree also exploits the lower thermal resistance ALN (aliminum nitride) substrates offer to produce packages for demanding uses. "This is particularly important in high temperature, high drive power, high current applications that also demand high reliability," McClear underlined. Streetlights again are a good example of such an application. "Bottom line, choose the right LED for the right application, and the substrate is part of that choice."

From the same article:
Ceramtec "The packaging started with AL2O3 (alumina ceramic), but is switching to ALN (aluminum nitride ceramic). We have upgraded our standard ALN to 200 (W/m K)"

So we should keep in mind that Cree engineers are no fools and they chose to continue using the ceramic base for the CXBs. Different architectures will have different pros and cons. For example, if using the ceramic base allows them to achieve higher efficiency from the start, that can override the small difference in temp droop. The tradeoff, it seems to affect current droop characteristics, as we have seen the CXA is the best choice for those who are looking for very high efficiency designs, the underdriven Vero simply cannot achieve the efficiency/cost levels of the underdriven CXA, no matter how much money you throw at it. And the other side of the coin, those who are looking to minimize up front cost and pack huge amounts of light into one COB, the Vero29 dominates.

That said, the difference in cooling performance between the CXA3070 and Vero29 is small in practice, even at 2.4A where the CXA is almost maxed. And that is not a fair fight, I will have to check the thermal performance of the CXA3590 to get a better idea what the aluminum nitride ceramic is capable of.

Another important point, the cooling performance does not even come into play for applications where the COBs are run relatively soft. For example, at 800mA (28W) I measure .75% temp droop on the CXA3070. So why would Cree want to sacrifice that efficiency for a slight thermal improvement that does not translate to much higher output? Also, I was shocked to discover, as long as the entire surface is sufficiently coated with thermal paste, there is no thermal benefit to sanding the heatsink, even at 2.55A.

As far as fragility, I have handled hundreds of CXAs and I have only ever had 1 crack, which was my fault for prying on it when trying to remove it. The funny thing is it still works and did not lose any output, I use it all the time :)

So they have different characteristics, that is a great thing. You just have to choose the one that most closely suits your design goals.
Thank ,Supra.
A deeper look into Cree's CXa/B series.
Aluminium Nitride ,eh ?

"
AlN was first synthesized in 1877, but it was not until the middle of the 1980s that its potential for application in microelectronics was realized due to its relative high thermal conductivity for an electrical insulating ceramic (70–210 W·m−1·K−1 for polycrystalline material, and as high as 285 W·m−1·K−1 for single crystals)
"

Good material choice then for the CXA/B arrays.
 

doctorflux

Well-Known Member
Good material choice then for the CXA/B arrays.
Cree has to use a ceramic substrate for those CXA arrays because the chips inside use a flip chip architecture (no wire bond) which requires eutectic die bonding to the substrate. This process requires higher temperatures which will damage an aluminum substrate.

Veros on the other hand are epoxied to the aluminum substrate and wire bonded together
 

AquariusPanta

Well-Known Member
To clarify, there is a big difference between alumina ceramic and aluminum nitride ceramic, which has a thermal conductivity up to 200 (W/m K)
From the Cree thermal management PDF
View attachment 3356133

This article goes into it in more detail about aluminum nitride ceramic COBs


This article quotes Cree's Vice President Applications Engineering Mark McClear:
Cree also exploits the lower thermal resistance ALN (aliminum nitride) substrates offer to produce packages for demanding uses. "This is particularly important in high temperature, high drive power, high current applications that also demand high reliability," McClear underlined. Streetlights again are a good example of such an application. "Bottom line, choose the right LED for the right application, and the substrate is part of that choice."

From the same article:
Ceramtec "The packaging started with AL2O3 (alumina ceramic), but is switching to ALN (aluminum nitride ceramic). We have upgraded our standard ALN to 200 (W/m K)"

So we should keep in mind that Cree engineers are no fools and they chose to continue using the ceramic base for the CXBs. Different architectures will have different pros and cons. For example, if using the ceramic base allows them to achieve higher efficiency from the start, that can override the small difference in temp droop. The tradeoff, it seems to affect current droop characteristics, as we have seen the CXA is the best choice for those who are looking for very high efficiency designs, the underdriven Vero simply cannot achieve the efficiency/cost levels of the underdriven CXA, no matter how much money you throw at it. And the other side of the coin, those who are looking to minimize up front cost and pack huge amounts of light into one COB, the Vero29 dominates.

That said, the difference in cooling performance between the CXA3070 and Vero29 is small in practice, even at 2.4A where the CXA is almost maxed. And that is not a fair fight, I will have to check the thermal performance of the CXA3590 to get a better idea what the aluminum nitride ceramic is capable of.

Another important point, the cooling performance does not even come into play for applications where the COBs are run relatively soft. For example, at 800mA (28W) I measure .75% temp droop on the CXA3070. So why would Cree want to sacrifice that efficiency for a slight thermal improvement that does not translate to much higher output? Also, I was shocked to discover, as long as the entire surface is sufficiently coated with thermal paste, there is no thermal benefit to sanding the heatsink, even at 2.55A.

As far as fragility, I have handled hundreds of CXAs and I have only ever had 1 crack, which was my fault for prying on it when trying to remove it. The funny thing is it still works and did not lose any output, I use it all the time :)

So they have different characteristics, that is a great thing. You just have to choose the one that most closely suits your design goals.
Hey now, don't be bringing in your Cree propaganda; this is strictly a Vero-based thread..... ;).

I've yet to have my cup of tea this morning but I'll attempt at making sense of the point you brought up.

So basically, Cree products avoided the thermal advantages that the Vero line boasts because they like their products operating under 2A, as they depend on selling their products based on the fact that they are the most efficient COB on the market today (when ran at low currents).

It's obvious that the Vero 29 wasn't designed to be ran below 1A, rather it's target was for +2A applications, where it could handle the extra heat without sacrificing efficiency and longevity.

If only Cree and BridgeLux came together to work on a special project that encompassed COBs for indoor horticulture enthusiasts...

:weed:
 

stardustsailor

Well-Known Member
Cree has to use a ceramic substrate for those CXA arrays because the chips inside use a flip chip architecture (no wire bond) which requires eutectic die bonding to the substrate. This process requires higher temperatures which will damage an aluminum substrate.

Veros on the other hand are epoxied to the aluminum substrate and wire bonded together
So it might be these kind of dies ,ok ?
http://www.cree.com/LED-Chips-and-Materials/Chips/Chips/Direct-Attach/DA1000-LED
or
http://www.cree.com/LED-Chips-and-Materials/Chips/Chips/Direct-Attach/DA1000-LED

The DA (Direct Attach ) die series from Cree...
Could be ...
But ...
Capture.JPG


Chips seem to be of "elongated rectangle "shape -like the DA series -,but still those wire bonds ,
should have not been there ,if the dies where "Flip Chips" ,ain't so ?

Another die family that fits to eutectic attachment ,but it needs bond wires (not a flip-chip) is the TR family ,like :
http://www.cree.com/LED-Chips-and-Materials/Chips/Chips/TR/TR5270M-LED


Cheers.
:peace:
 
Last edited:

stardustsailor

Well-Known Member
Not hating on the Vero by any means. My COB loyalty goes to the one that gets growers the most bud for their design and sometimes that will be CXA sometimes Vero. Seems like they will be neck and neck so it makes sense that each has their own niche.
I totally agree!
They are so different COBs regarding their semiconductor substrate technology as also regarding their packaging tech.
But ,at the same time ,they are like ...twins ...,in plenty of ways ...

Check this out :
fut1.JPG

fut 2.JPG

http://apps1.eere.energy.gov/buildings/publications/pdfs/ssl/ssl_mypp2014_web.pdf
 
Last edited:

AquariusPanta

Well-Known Member
Not hating on the Vero by any means. My COB loyalty goes to the one that gets growers the most bud for their design and sometimes that will be CXA sometimes Vero. Seems like they will be neck and neck so it makes sense that each has their own niche.
Any knowledge of anything is welcome here, and you kind sir are a beacon of wisdom.

Sailor, you already know you're like the yoda of the COB forces... :lol:

Keep up the healthy, enlightening debates, fellas!
 

stardustsailor

Well-Known Member
That's just awesome and i am so ripping that pic!
PHHHrRRRR!!
:spew:

Quick,sloppy ,full of mistakes job ..
3 out of 10 ..And I'm being kind.

:P..
No ,really ,that's crappy job ...
Very quickly done ,lots-lots of mistakes & left undone work ..
But anyway ,big Thanx !!!

(AP :Could I...err..yes .one more ..? I know .last one before ...but ...ehmmm ..mmm .ok ..forget it ... )
 

coolj

Well-Known Member
I was looking at Mr Flux's 'Step into my room' thread, its fabulous! but thats too many wires for me. i was thinking is i shunk that down

I was thinking 1.5x1.5 or 2x2 flat piece of aluminum , not sure the thickness or the best fan and chip option, drivers etc.

I can buld the thing just not sure about the roght parts and the and all this math is freaking greek...lol, but dam you guys are freaking scholars . at this stuff!

i hope im not trying to produce a unicorn, lol
 

AquariusPanta

Well-Known Member
PHHHrRRRR!!
:spew:

Quick,sloppy ,full of mistakes job ..
3 out of 10 ..And I'm being kind.

:P..
No ,really ,that's crappy job ...
Very quickly done ,lots-lots of mistakes & left undone work ..
But anyway ,big Thanx !!!

(AP :Could I...err..yes .one more ..? I know .last one before ...but ...ehmmm ..mmm .ok ..forget it ... )
I think Cococola36 would approve of your work... ;-).
 
Top