PICO's DIY Thread - Advise, Ideas and Technology - NO PANEL REPS!

[lax123]

what I was saying AS5 8,9; momentive: 0,84 how are These numbers Close ?

 

[smokey the cat]

Head desk moment - didn't notice the blasted decimal shifting. Clearly I'm not smoking enough weed, lol. You're completely right. 0.84 puts it on par with quite bad thermal paste.

For a CPU with it's high clamping strength I'd guess 3-4*C difference between good paste and bad paste. But as this will need to provide all the clamping effect as well as heat transfer I have no idea what I'll see.

Hmmmm. Not what I was thinking when I ordered it. That damn decimal place, lol I do have some Arctic Silver and Ceramique somewhere in the house, and screwing is definitely gonna get a firmer contact. How well do self-tapping screws work out in cheap Al heatsinks do you think?

Then again, it maybe worth experimenting. On the plus side: Vero are rated to run hella hot. It's a shame my missus ruined my thermal probe by sticking it in the dishwasher - this would be the perfect use for it.

 

[MrFlux]

The silicon glue will be just fine. It helps to use glue clamps to push out the excess of glue. Pushing really hard with just the fingers does not come close. After that it will be mostly the vacuum force that keeps it together. Since there is so little glue in between in doesn't matter much what the thermal conductivity of the glue is, almost anything that fills the air gap would do.

Compared to a CPU die the Vero board has a much larger surface (about 10X) while dissipating less heat so this makes it much easier to cool. And besides, even Intel uses just a thick slab of cheap goo inside their integrated heat spreaders.

 

[algrow12]

I have been trying to wrap my mind around this the more I read the more I get confused. I was thinking about using the vero13s for my small space but I can get the vero 10s for half the price. I could use more of them for more points of light. Im thinking 6 2700k and 2 5000k. My question is can someone tell me from this list which driver I need and how many of them or would it be cheaper to just use LDD-500h.

Driver link - http://www.meanwell.com/search/LPF-60D/LPF-60D-spec.pdf .

 

[smokey the cat]

Vero 10
Nominal Current 350mA, 26.7V, 9.3W

Vero 13
500mA, 32.5V, 16.2W



Gidday Algrow,

I had orginally looked at stringing some of the smaller Vero together, but if you want to use the meanwell drivers they are difficult to build an array with. You can't run them parallel, and the voltage of the emitter is so high it's difficult to find a driver that can handle two or more in series.


I ended up ordering a couple of V-18 and two lpf-60d-36.

 

[algrow12]

Smokey how did it work with your 100kohm trimpot or will I need to get something else to dim it with. Also what color did you go with.

 

[caretak3r]

Vero 10
Nominal Current 350mA, 26.7V, 9.3W

Vero 13
500mA, 32.5V, 16.2W



Gidday Algrow,

I had orginally looked at stringing some of the smaller Vero together, but if you want to use the meanwell drivers they are difficult to build an array with. You can't run them parallel, and the voltage of the emitter is so high it's difficult to find a driver that can handle two or more in series.


I ended up ordering a couple of V-18 and two lpf-60d-36.

borrowing from the CXA thread, I think this driver would do well (700ma/264V or 500ma/370V) (if you can wait on the long ship time since no one has them in stock yet) - hlg-185-c

 

[algrow12]

I was wondering if this pot would work with the lpf-60d-36 and if so how would I wire it up.

http://datasheet.octopart.com/P231-QC20BR100K-BI-Technologies-datasheet-5355129.pdf

 

[smokey the cat]

Here we go - someone holler if you see anything amiss, or chime in with jargon I'm missing or misusing.


Looks fine - get 100kohm for Mean Well.

Linear taper style will evenly increase resistance from 0 to 100% of rating as you turn it.
Logarithmic type will only slowly increase at first before increasing fastest at the end - your volume knob on a stereo is probably logarithmic.
Either will work.
(The taper style for that part are Linear, and two versions of logarithmic: "audio", and "reverse audio" - see page 2)

Looking at the data sheet from Meanwell the dimming outputs look nice and even - so I'm getting linear as I think it should give nice even dimming across the dial. 25% on the trimpot = 25% driver output. Audio or reverse audio would be useful if you're expecting to make fine adjustments at very low or very high driver output.

There'll be a bunch of shaft options - which would be important for someone wanting the right height knob for a nice case. Get whatever you like, what you do with your shaft is noone's business but your own, lmao

Wiring -
Trimpots have three connections - the outsides are at either end of the 100kOHM coil resistor, the middle is the "Wiper" which moves from one end of the resistance coil to the other. Connect your dimming line to one side of trimpot and the wiper - have a play with it with your multimeter once it shows up and you'll see how it works.

Wiper is always in the middle - if you were to you connect to the "wrong" side of the trimpot your dimming knob will be working in reverse, and you can lol at how you're turning it backwards to increase the power. No drama.


tl;dr - 100K ohm, hook up to side and middle.

 

[lax123]

hmmm, no expert but sure u want to use a pot? i think lot of power will get wasted as heat...and u Need a high W pot...
but maybe its different with constant current sources, too wasted to think about it id think you need comparators + poti or jut pwm

 

[smokey the cat]

iI thought the wholepoint of it was that circuit detected resistance, or PWM or 10V. Why would sensing resistance versus sensing a 10V signal waste power?

edit/
Checked data sheet - dim cable leads to a described "detection circuit" - it shouldn't be using any extra power there.

 

[lax123]

ahhh ok...its a part of ur cc Driver and meant for sensing...sry thought u had something different in mind

 

[SupraSPL]

do you think 5 veros (29s) would outperform a 600 watt HID?

When we design an LED lamp we have to be careful to maintain efficiency otherwise we defeat the purpose. As much as I am impressed by the Vero29, I would run them no higher than 1.4A (51watts - 34% efficient). Even better at 700mA (38% efficient)

Once you go past 1.4A the current and temp droop will penalize efficiency to the point you might as well just use more emitters. At 2.8A dissipation is 110 watts and efficiency is 27.8% (assuming Tj 85c which is probably generous).

So to answer your question, (5) vero29s at 2.8A would dissipate 550 watts at 27.8% efficiency so 153 radiant watts versus the 600 HPS at 40% efficiency 240 watts radiant. The HPS is putting out 57% more photons and creating 10% less heat. Granted the LED can direct the photons to the canopy better and the LED can keep the heat out of the canopy better, but considering the difference in cost the economy of the design would be questionable.

 

[NewGrower2011]

http://www.amazon.com/LENBO-White-W...dp_s_cp_5?ie=UTF8&refRID=1XYB0WM4JR0FPXGP3SFK


Are these LED floods you can find out there worth looking into at all? I've seen cheap ones stating both 110v and 12v - so I assume a driver is include if stating 110v...

 

[NewGrower2011]

I'm aiming for a 600/750 equivalency myself I think so I'm interested in these COB/matrix deals... I need a little spread for the footprint I'm targeting so I think I'll go with a larger # of LED matrixes and figure out all the series/parallel/# of drivers once I get a better plan on the LEDs themselves.. I do want to go remote ballast though...

 

[NewGrower2011]

Another question... What contributes the actual heat (aside from the LED itself of course) - the DC power supply/conversion or the LED driver(s)?

I'm guessing the power supply is the bigger of the two if not all of it... I was wondering if the LED flood-lights that have a driver already in them would still benefit from a 'remote ballast' for the power supply if I went with the DC versions... also people are using them like Malibu lights... Would one of those landscape lighting (DC-based) replacement units be cheaper than just a DC supply from one of the DIY LED sites? Lowes/etc?

 

[andyhot]

Hi everyone.
I am currently in the planning stage of a diy led light for a 1m x 1m x 2m tall tent.
my plan so far is for 24 vero 10 leds in 5 rows in a pattern like below.
x-x-x-x-x
-x-x-x-x-
x-x-X-x-x
-x-x-x-x-
x-x-x-x-x
(x= 1x v10, -= case, X= 2x v10)
i would be using 6 x 120v 350ma drivers and run 4 vero with each. By my calculations this should give approx 26000 lumens at 240w if I used all 3500k vero10.

I was hoping for a bit of advice what the general consensus is on what colour/k is considered best for a light to be used from seed to harvest. Would I be best with a mixture such as 12 2700k and 12 4000k or all the same such as 3500k.

Thanks in advance for any help and advice.
Andy.

 

[SupraSPL]

http://www.amazon.com/LENBO-White-W...dp_s_cp_5?ie=UTF8&refRID=1XYB0WM4JR0FPXGP3SFK


Are these LED floods you can find out there worth looking into at all? I've seen cheap ones stating both 110v and 12v - so I assume a driver is include if stating 110v...

6500K probably has too much blue to be ideal for anything except cloning and 100LM/W in a cool white means relatively poor efficiency, not that I trust their lumen output figures and who knows if that is at operating temps.

A 110v unit will likely have an AC constant voltage -> DC constant current conversion in one step (70-90% efficient). The 12v DC version probably has a DC constant voltage -> DC constant current (85-95% efficient). If you used your own 110V->12v DC constant voltage power supply (70-80%) you will still need another step. So your best bet is to try and get 90% efficiency in one step with a 110v constant current driver.

 

[PetFlora]

Though I agree that 5000/4500 is better, still, I have grown nice seedlings with 6500

6500K probably has too much blue to be ideal for anything except cloning and 100LM/W in a cool white means relatively poor efficiency, not that I trust their lumen output figures and who knows if that is at operating temps.

A 110v unit will likely have an AC constant voltage -> DC constant current conversion in one step (70-90% efficient). The 12v DC version probably has a DC constant voltage -> DC constant current (85-95% efficient). If you used your own 110V->12v DC constant voltage power supply (70-80%) you will still need another step. So your best bet is to try and get 90% efficiency in one step with a 110v constant current driver.

 

[PICOGRAV]

When we design an LED lamp we have to be careful to maintain efficiency otherwise we defeat the purpose. As much as I am impressed by the Vero29, I would run them no higher than 1.4A (51watts - 34% efficient). Even better at 700mA (38% efficient)

Once you go past 1.4A the current and temp droop will penalize efficiency to the point you might as well just use more emitters. At 2.8A dissipation is 110 watts and efficiency is 27.8% (assuming Tj 85c which is probably generous).

So to answer your question, (5) vero29s at 2.8A would dissipate 550 watts at 27.8% efficiency so 153 radiant watts versus the 600 HPS at 40% efficiency 240 watts radiant. The HPS is putting out 57% more photons and creating 10% less heat. Granted the LED can direct the photons to the canopy better and the LED can keep the heat out of the canopy better, but considering the difference in cost the economy of the design would be questionable.

This is an inaccurate statement, nothing wrong with running the vero 29 up to 1.5x the rated drive current.