Par mapping the world

[CobKits]

Even a sphere number can be deceiving since the Gavita has a beam angle of 139 degrees (or something) and it hangs a lot higher.

well sphere is *all* reflection, the detector is shielded from direct light by a baffle. that said lights of vastly different geometry do perform differently and as someone mentioned sphere diameter is a factor. sphere is semi-voodoo and our understanding is in part fundamentals and in part practical experience. i have a whole book on the subject, will upload

 

[wietefras]

well sphere is *all* reflection, the detector is shielded from direct light by a baffle. that said lights of vastly different geometry do perform differently and as someone mentioned sphere diameter is a factor. sphere is semi-voodoo and our understanding is in part fundamentals and in part practical experience. i have a whole book on the subject, will upload

Sphere is indeed *all* reflection, but in a grow it's far from all reflection. That's the difference where COBs gain a bit extra.

 

[CobKits]

well yeah but were trying to get a handle on gross output, no? all cobs radiate at similar angles and directivity is left up to the user (reflectors and lenses with theor various angles, # of cobs in an array, mounting height, current, etc)

i see both sides. i understand how OP wants the measurements to be replicable but also i think you would agree reflective walls is more of a "real world" grow environment

 

[OneHitDone]

well yeah but were trying to get a handle on gross output, no? all cobs radiate at similar angles and directivity is left up to the user (reflectors and lenses with theor various angles, # of cobs in an array, mounting height, current, etc)

i see both sides. i understand how OP wants the measurements to be replicable but also i think you would agree reflective walls is more of a "real world" grow environment

Not really when a lot of growers are moving into larger spaces where aisles are code for becoming legal and complaint?

 

[CobKits]

thats fine, those growers will need optics to keep their light out of the aisles, that will be site specific

im just talking about how to measure gross output in a way thats useful. as soon as you put any chip into a lens/reflector/array and try to measure it against any other light that isnt exactly the same setup there will be bias. i beleive you also mentioned the reflectors are "similar but different" - boom another variable

esp when comparing hid with cobs and such, as the radiation patterns differ so greatly

thats why its fundamentally incorrect to suggest "chip A is x% more efficient than chip B" with a test like this. its too specific. "a fixture of C number of chip A with reflector/optic R in X pattern is more efficient than a fixture of D number of chip B with reflector/optic Q in Y pattern"... which adds a host of variables. its as much a measure of the fixture layout itself than it is of the chip efficacy

bugbee's take on flat-plane integration vs spheres is worth a look, as is the goniophotmeter style of measurement

when i measure individual chips i always make sure to have no optics. i wont even use BJB holders as they have a mild reflective effect

 

[wietefras]

well yeah but were trying to get a handle on gross output, no? all cobs radiate at similar angles and directivity is left up to the user (reflectors and lenses with theor various angles, # of cobs in an array, mounting height, current, etc)

i see both sides. i understand how OP wants the measurements to be replicable but also i think you would agree reflective walls is more of a "real world" grow environment

Well this would be the default tent for that model HPS and people want to replace it with a COB fixture. So that would be a reasonably common situation.

Perhaps 600W in a 4'x4' is even more common. Still, lot's of Gavitra Pro 1000W in use too in a 5'x5' tent.

 

[pirg420]

Hey @CobKits do you have a thread documenting your 12 inch on axis test? I dont understand the details of your test and id love to have more info if you could link me to the thread? if you dont have a thread do you mind explaining the process you used in detail, just for my own curiosity, thanks dude.

 

[CobKits]

open space, par meter at 12" on axis from cob. various cobs on identical heatsinks tested in similar ramp of 0.1 up to 3A and plotted.

ive since moved to using a 20" integrating sphere for single chip measurements but the results are similar

 

[pirg420]

Up next is a brand new 1000 watt hortilux vs a used Ushio 1000 watt bulb (6 months 12/12)

Used same exact reflector for both tests, just changed the bulb.

Reflector is a used but cleaned cheapo batwing reflector.
Ballast is a digital DigiLume 1000 watt

I am assuming that when the Ushio is brand new, it puts out the same or similar par numbers to a brand new hortilux, but this hasnt been tested, this test is more to show the differences between used and brand new bulb.

I realize i should have used a used hortilux, but could not get my hands on one.

Used Ushio bulb at 24.75 inches:
actual wattage 1017 watts
4x4 ppfd: 428
5x5 ppfd: 364

Brand new Hortilux at 24.75 inches:
actual wattage 1010-1015 watts
4x4 ppfd: 515
5x5 ppfd: 432


Summary:

The brand new hortilux puts out 20.3% more par then Used Ushio on a 4x4
The brand new hortilux puts out 18.68% more par then Used Ushio on 5x5

I used hps for years, and hardly ever changed my bulbs, waited for them to burn out basicly. But this demonstrates that you should be changing your bulbs well before the 6 month mark i would say.

 

[pirg420]

open space, par meter at 12" on axis from cob. various cobs on identical heatsinks tested in similar ramp of 0.1 up to 3A and plotted.

ive since moved to using a 20" integrating sphere for single chip measurements but the results are similar

not totally sure what you mean by 'on axis'. How many readings did you take at each wattage level? 1? (im imagining the sensor is head on, 12 inches away from the chip, and a single reading is logged)

thanks for clarification

 

[Photon Flinger]

I think you got it. Measurement at 12" directly under light source.

 

[CobKits]

not totally sure what you mean by 'on axis'. How many readings did you take at each wattage level?

one. unless i rerun the whole chip. every curve is made up of 20 or so datapoints so if an individual measurement is off you can see it immediately visually as its off the curve. typically anything above 5 watts its pretty much a good measurement, in the free air measurements stray light is the enemy and has more of an effect when you are reading less than 100 ppfd.

sphere has improved this low-light measurement aspect (vs me taking measurements in a big room at night with a headlamp to see the meter and record data)

 

[pirg420]

This is 12 vero 29d's vs 12 cree cxb3590 db's in a reflective 10x5 grow tent


12 Vero 29 D's
actual wattage 616 watts
Total PPFD: 16237
10x5 PPFD: 324.75
PPFD Per Watt: .527



12 Cree cxb3590 DB's
actual wattage 616 watts
Total PPFD: 16528
10x5 PPFD: 330.56
PPFD Per Watt: .536


Summary:

The cree DB Fixture puts out about 1.8% more par then the Vero 29 D fixture.

Both fixtures are identical, with identical drivers. Reflectors are exactly the same. Both fixtures are square with 12 chips.

 

[CobKits]

thats right about where i would expect it.

i still scratch my head over why anyone would select a vero D when vero B is the same price and more efficient.

 

[pirg420]

Next up in the 10x5 grow tent is 16 cxb3590s vs Gavita 1000 vs HPS 1000 Batwing


Tested these at 24":

16 CXB3590 DB's
actual wattage 1021 watts
Total PPFD: 26722
10x5 PPFD: 534.44
PPFD Per Watt: .523

Gavita 1000
actual wattage 1060 watts
Total PPFD: 21664
10x5 PPFD: 433.28
PPFD Per Watt: .408

1000 HPS Batwing
actual wattage 1013 watts
Total PPFD: 15398
10x5 PPFD: 307.96
PPFD Per Watt: .304


Did a follow up test at 60" for Gavita and Cree DB

16 CXB3590 DB's
actual wattage 1021 watts
Total PPFD: 20520
10x5 PPFD: 410.4
PPFD Per Watt: .401

Gavita 1000
actual wattage 1060 watts
Total PPFD: 15484
10x5 PPFD: 309.68
PPFD Per Watt: .292


Summary:

24 inches:
The cree DB Fixture puts out about 23.3% more par then the Gavita 1000 Fixture
The cree DB Fixture puts out about 73.5% more par then the 1000 HPS Batwing
The Gavita 1000 Fixture puts out about 40.6% more par then the 1000 HPS Batwing

60 Inches:
The cree DB Fixture puts out about 32.5% more par then the Gavita 1000 Fixture

 

[pirg420]

thats right about where i would expect it.

i still scratch my head over why anyone would select a vero D when vero B is the same price and more efficient.

according to the datasheets the D is quite a bit more efficient then the B?

 

[pirg420]

I wanna do some par testing with active vs passive cooling, has there been a thread on this yet?

Im assuming the active cooling keeps the Tj/Tc much lower and is better for efficiency, but how much lower is the question

 

[CobKits]

according to the datasheets the D is quite a bit more efficient then the B?

maybe by current or something, but not at a given wattage

example (per b-lux excel tool) for 3000k 80 cri, 55 Tc

vero D at 1400 mA (50.0W) = 158 lm/W
vero B at 1400 mA (70.4W) = 158 lm/W

or alternatively
vero D at 1400 mA (50.0W) = 158 lm/W
vero B at 1016 mA (50.0W) = 164 lm/W

easy way to remember vero29 by voltage/efficiency

vero C = 68V = most efficient
vero B = 50V = intermediate
vero D = 36V = least efficient

the ranking of CBD should be easy for even the stoniest of brahs to remember

FWIW the efficiency isnt really a product of voltage, it just lays out that way as the higher voltage chips in the vero29 series have more dies so at a given wattage its less current per die

 

[pirg420]

I must be missing something, or the datasheets dont line up with the tool?

 

[Hybridway]

Up next is a brand new 1000 watt hortilux vs a used Ushio 1000 watt bulb (6 months 12/12)

Used same exact reflector for both tests, just changed the bulb.

Reflector is a used but cleaned cheapo batwing reflector.
Ballast is a digital DigiLume 1000 watt

I am assuming that when the Ushio is brand new, it puts out the same or similar par numbers to a brand new hortilux, but this hasnt been tested, this test is more to show the differences between used and brand new bulb.

I realize i should have used a used hortilux, but could not get my hands on one.

Used Ushio bulb at 24.75 inches:
actual wattage 1017 watts
4x4 ppfd: 428
5x5 ppfd: 364

Brand new Hortilux at 24.75 inches:
actual wattage 1010-1015 watts
4x4 ppfd: 515
5x5 ppfd: 432


Summary:

The brand new hortilux puts out 20.3% more par then Used Ushio on a 4x4
The brand new hortilux puts out 18.68% more par then Used Ushio on 5x5

I used hps for years, and hardly ever changed my bulbs, waited for them to burn out basicly. But this demonstrates that you should be changing your bulbs well before the 6 month mark i would say.View attachment 4015922

I used to go only 2 grows per bulb the first yr., then started going 3 (6-7mo).
It's nice to see an estimated light loss over that time. Deff need a used Hortri though.
Good to see mapping.