CXB3590 1500W

SupraSPL

SupraSPL

Well-Known Member
ttystikk said:
@SupraSPL or anyone who feels confident with the material; I ordered CXB3590 3500k CD bin 72V, to go four at a time with the appropriate 200W Meanwell driver. This combo has been spec'd out to be 56% efficient. Now I'm hearing about a 5 x CXB3590 combination running at 61% efficiency. Can you or anyone explain this five chip setup in a bit more detail? Would it be worth the switch from my current combination?
Click to expand...
Here are some possible combos, assuming you are using 72V class CXB3590s

(4) CXB3590s 3500K CD 72V
(1) HLG-185H-C700A
COBs will run up to 54W ea, ~56% efficient

(5) CXB3590 3500K CD 72V
(1) HLG-185H-C500A
COBs will run 37W ea, ~61% efficient

(6) CXB3590 3500K CD 72V
(1) HLG-185H-C500, dimmed to 400mA
COBs will run ~27W ea, 63% efficient

(6) CXB3590 3500K CD 72V
(1) HLG-120H-C350A
COBs will run 25.5W ea, 63.5% efficient

But if you are using 36V class it can be much simpler, as @Stephenj37826 pointed out:

(5) CXB3590 3500K CD 36V
(1) HLG-185H-C1050A
COBs wil run at 37W ea, 61% efficient

(8 ) CXB3590 3500K CD 36V
(1) HLG-185H-C700A
COBs will run at 25.5W ea, 63.5% efficient
 
Last edited:
NapalmD

NapalmD

Well-Known Member
organic-fanatic said:
What is the difference between the A & B version of the hlg drivers ?
I notice most have chosen the b, but my gut tells me to go with A.
thank you .
Click to expand...
The A drivers have an internal dimmer built into the driver that you can turn with a little screw driver. With the B drivers you wire up an external dimmer to the driver which the only difference as far as I know is you can dim lower.
 
ttystikk

ttystikk

Well-Known Member
SupraSPL said:
Here are some possible combos, assuming you are using 72V class CXB3590s

(4) CXB3590s 3500K CD 72V
(1) HLG-185H-C700A
COBs will run up to 54W ea, ~56% efficient

(5) CXB3590 3500K CD 72V
(1) HLG-185H-C500A
COBs will run 37W ea, ~61% efficient

(6) CXB3590 3500K CD 72V
(1) HLG-185H-C500, dimmed to 400mA
COBs will run ~27W ea, 63% efficient

(6) CXB3590 3500K CD 72V
(1) HLG-120H-C350A
COBs will run 25.5W ea, 63.5% efficient

But if you are using 36V class it can be much simpler, as @Stephenj37826 pointed out:

(5) CXB3590 3500K CD 36V
(1) HLG-185H-C1050A
COBs wil run at 37W ea, 61% efficient

(8 ) CXB3590 3500K CD 36V
(1) HLG-185H-C700A
COBs will run at 25.5W ea, 63.5% efficient
Click to expand...
What's interesting from an applications standpoint is that with each incremental step up in efficiency still comes a reduction in overall PAR watts emitted, thus requiring a higher density of CXB3590 chips to reach the desired intensity;
4 chips @54W = 216W x .56 = 121PAR W
Correct? Understood that I'm not accounting for lens losses, etc.
So,
5 chips @37W = 185W x .61 = 113 PAR W
And,
6 chips @27W = 162W x .63 = 102 PAR W

More money, more chips, less total intensity per kit, thus needing more density- which requires yet more expenditure to reach a given PPfd level across a given surface area.

What's the payback? Power savings... and heat reduction, leading to more power savings at larger scale. But, how much?

Assuming light intensity is held constant, we can just use the efficiency ratings. But just multiplying watts times efficiency doesn't get us there; as efficiency goes up, the number of emitters goes down, but as PAR watts per emitter falls, the number of emitters increases.

There's a conversion factor for this wattage reduction as efficiency increases over a given area, I'm just not sure how to figure it.

I dunno... it's late, maybe this is just chasing my tail...
 
Stephenj37826

Stephenj37826

Well-Known Member
ttystikk said:
What's interesting from an applications standpoint is that with each incremental step up in efficiency still comes a reduction in overall PAR watts emitted, thus requiring a higher density of CXB3590 chips to reach the desired intensity;
4 chips @54W = 216W x .56 = 121PAR W
Correct? Understood that I'm not accounting for lens losses, etc.
So,
5 chips @37W = 185W x .61 = 113 PAR W
And,
6 chips @27W = 162W x .63 = 102 PAR W

More money, more chips, less total intensity per kit, thus needing more density- which requires yet more expenditure to reach a given PPfd level across a given surface area.

What's the payback? Power savings... and heat reduction, leading to more power savings at larger scale. But, how much?

Assuming light intensity is held constant, we can just use the efficiency ratings. But just multiplying watts times efficiency doesn't get us there; as efficiency goes up, the number of emitters goes down, but as PAR watts per emitter falls, the number of emitters increases.

There's a conversion factor for this wattage reduction as efficiency increases over a given area, I'm just not sure how to figure it.

I dunno... it's late, maybe this is just chasing my tail...
Click to expand...

I think that running them at 1050 could be worth while. It really depends on the electrical cost where you are located. 700 seems to be alot more cost for not that much efficiency gain honestly. I personally like the approach of either 1400 or 2100 with dimming. More horse power when they are bulking up. I like the idea of slowly ramping up the light for the first 3 weeks of bloom then ramping back down the last 10-14 days. I believe this will lead to higher g/kwhr if you will. It seems with full spectrum lighting things tend to finish a little sooner. The possibility of 1 added harvest per year really changes the metric when it comes to roi in my book. I don't think you can go wrong with any of those choices above. Lots of food for thought though :)
 
ttystikk

ttystikk

Well-Known Member
Stephenj37826 said:
I think that running them at 1050 could be worth while. It really depends on the electrical cost where you are located. 700 seems to be alot more cost for not that much efficiency gain honestly. I personally like the approach of either 1400 or 2100 with dimming. More horse power when they are bulking up. I like the idea of slowly ramping up the light for the first 3 weeks of bloom then ramping back down the last 10-14 days. I believe this will lead to higher g/kwhr if you will. It seems with full spectrum lighting things tend to finish a little sooner. The possibility of 1 added harvest per year really changes the metric when it comes to roi in my book. I don't think you can go wrong with any of those choices above. Lots of food for thought though :)
Click to expand...
Eight week bloom cycles and I can turn the room around in a matter of hours. I'm READY! LOL

I was trying to quantify the watts per square foot of each combination when held at an equivalent PPfd value. From there I was going to look at the heat difference between them and extrapolate likely differences in cooling requirements for the rest of the environment.

I suspect you're right in that somewhere right around 50-60% efficiency there's another one of those 'sweet spots', beyond which little is gained even at a large incremental increase in expense. Good for racing, not for production, lol

I have another question, while we're talking efficiency; how much difference does operating temperature make on performance? How much better would a given CXB3590 chip perform at, say, 100F than at 130?
 
Stephenj37826

Stephenj37826

Well-Known Member
ttystikk said:
Eight week bloom cycles and I can turn the room around in a matter of hours. I'm READY! LOL

I was trying to quantify the watts per square foot of each combination when held at an equivalent PPfd value. From there I was going to look at the heat difference between them and extrapolate likely differences in cooling requirements for the rest of the environment.

I suspect you're right in that somewhere right around 50-60% efficiency there's another one of those 'sweet spots', beyond which little is gained even at a large incremental increase in expense. Good for racing, not for production, lol

I have another question, while we're talking efficiency; how much difference does operating temperature make on performance? How much better would a given CXB3590 chip perform at, say, 100F than at 130?
Click to expand...


Here is a really cool tool to help you answer those questions.


http://pct.cree.com/dt/index.html

As temps rise efficiencies go down. Also to take into consideration is vf goes down as well so you end up with even less light because of less overall wattage draw. Not huge on the vf thing but in a room of several hundred cobs it could make a difference.
 
nogod_

nogod_

Well-Known Member
Higher density of chips = more even spread across the canopy = maximum yield per sq/ft and per watt

Suffice it to say I believe wholeheartedly in the value of biting the bullet and driving more chips softer (25-35w)

You don't have to sacrifice ppfd at all, you are just ensuring your photons are coming from more points over your growspace (and getting more photons for your [edit: power bill] dollars).

Take your example and adjust....

4 @ 54w = 216w x .56 = 121parwatt
8 @ 27w = 216w x .63 = 136parwatt

Same theoretical driver, less heatsink, twice as much $$$ on chips.

Let's assume 1000w build, how long you think it will take you to recoup the cost of 20 extra cxb assuming at least 7% bonus per cycle?



ttystikk said:
What's interesting from an applications standpoint is that with each incremental step up in efficiency still comes a reduction in overall PAR watts emitted, thus requiring a higher density of CXB3590 chips to reach the desired intensity;
4 chips @54W = 216W x .56 = 121PAR W
Correct? Understood that I'm not accounting for lens losses, etc.
So,
5 chips @37W = 185W x .61 = 113 PAR W
And,
6 chips @27W = 162W x .63 = 102 PAR W

More money, more chips, less total intensity per kit, thus needing more density- which requires yet more expenditure to reach a given PPfd level across a given surface area.

What's the payback? Power savings... and heat reduction, leading to more power savings at larger scale. But, how much?

Assuming light intensity is held constant, we can just use the efficiency ratings. But just multiplying watts times efficiency doesn't get us there; as efficiency goes up, the number of emitters goes down, but as PAR watts per emitter falls, the number of emitters increases.

There's a conversion factor for this wattage reduction as efficiency increases over a given area, I'm just not sure how to figure it.

I dunno... it's late, maybe this is just chasing my tail...
Click to expand...
 
Last edited:
Stephenj37826

Stephenj37826

Well-Known Member
nogod_ said:
Higher density of chips = more even spread across the canopy = maximum yield per sq/ft and per watt

Suffice it to say I believe wholeheartedly in the value of biting the bullet and driving more chips softer (25-35w)

You don't have to sacrifice ppfd at all, you are just ensuring your photons are coming from more points over your growspace (and getting more photons for your [edit: power bill] dollars).

Take your example and adjust....

4 @ 54w = 216w x .56 = 121parwatt
8 @ 27w = 216w x .63 = 136parwatt

Same theoretical driver, less heatsink, twice as much $$$ on chips.

Let's assume 1000w build, how long you think it will take you to recoup the cost of 20 extra cxb assuming at least 7% bonus per cycle?
Click to expand...
Unfortunately when you run these cobs at lower currents the vf drops meaning you dont get the same wattage draw.3590 vs Vero.png
 
ttystikk

ttystikk

Well-Known Member
nogod_ said:
Higher density of chips = more even spread across the canopy = maximum yield per sq/ft and per watt

Suffice it to say I believe wholeheartedly in the value of biting the bullet and driving more chips softer (25-35w)

You don't have to sacrifice ppfd at all, you are just ensuring your photons are coming from more points over your growspace (and getting more photons for your [edit: power bill] dollars).

Take your example and adjust....

4 @ 54w = 216w x .56 = 121parwatt
8 @ 27w = 216w x .63 = 136parwatt

Same theoretical driver, less heatsink, twice as much $$$ on chips.

Let's assume 1000w build, how long you think it will take you to recoup the cost of 20 extra cxb assuming at least 7% bonus per cycle?
Click to expand...
In direct electrical consumption costs? Forever! In terms of a yield increase, that's dicier, because you won't get 7% more yield, at best you'd move one bar up the PPfd chart, a very small incremental increase. For twice the money in chips it just doesn't seem worth it, even when considering years for amortization.

This becomes a more immediate concern when dealing with the heat. Lower HVAC operating costs are where most of the power savings will actually come from.

Which leads me right back to a question I asked last night; how does one calculate performance changes due to chip temperature?

Just for grins, what would happen in terms of efficiency if we could run a chip super soft AND keep it really cool, like 70°f or even cooler?
 
ttystikk

ttystikk

Well-Known Member
PurpleBuz said:
law of diminishing returns come into play (how much $$ for the extra cooling and hardware cost versus the return)
Click to expand...
Indulge me...

How do we quantify the performance improvement vs running a COB at Cree's test spec of 77°f?

For laughs, twenty five degrees higher is 102°f, and there's a performance drop. Exactly what is that drop?

If we now reduce the operating temperature of the chip by 25°f, do we get to turn all of that above loss into a performance improvement vs test conditions? If not, how much do we get?
 
sanjuan

sanjuan

Well-Known Member
ttystikk said:
Indulge me...

How do we quantify the performance improvement vs running a COB at Cree's test spec of 77°f?

For laughs, twenty five degrees higher is 102°f, and there's a performance drop. Exactly what is that drop?

If we now reduce the operating temperature of the chip by 25°f, do we get to turn all of that above loss into a performance improvement vs test conditions? If not, how much do we get?
Click to expand...
Like this?
Cree 11C-25C-39C.jpg
 
Shredderthirty

Shredderthirty

Well-Known Member
nice nice nice, like all the talk over here and can't wait to build a 4x8 full of lights...

but right now i'm looking at vacuum chambers, sorry to be way off topic but i'm looking at a 5 gallon shatter vac container, or glass vac, or just the regular best value vac. Anybody used any of these or have a recommendation on which is the best or any better ones i should be looking at?
 
You must log in or register to reply here.
Top