713 watt LED build log

So this is the start of my log. I might end up posting it somewhere else for a bit of cash in which case I would provide links from this page, but until then I will post here. There won't be much to look at for the next month or so as parts arrive. I have already received the heat sinks and purchased some angle and c-channel aluminum and will be sawing them up today. Once the frame is complete and the drivers arrive I can start playing with positioning the hardware which will give me a better idea of how long the connectors need to be and how much wire I will need. I'll cover the technical aspects in the next update.

Basic stats: 23 units consisting of a 50 watt radial heat sink (ebay), Vero 18 emitter, Mean Well PLM-40-1050 LED driver. Total output in watts: 713. Total output in lumens: 86862. Efficiency: 122 lumens per watt - 34.85% efficiency.

After much thought I've decided to use a 2:1 ratio of 2700K/4000K emitters. I think the Veros at 2700 probably peak very close to 640nm so I'm taking a chance on them. After growing a couple 'mater plants with them I'll reassess the spectrum and consider making adjustments if I think I need to.

Here is a pic of the sinks in position and the aluminum frame material.

Good luck with your project..been waiting for someone to build a big light!

passively cooled 713w?

Yes, passively cooled, 30 watts on each 50 watt sink, about 20 watts of heat. I have a strong vent fan and good circulation pattern... fingers crossed.

I'm sure you've read this?; https://www.rollitup.org/led-other-lighting/702193-bridgelux-vero-29-led-test.html

Thanks for the link. Yes, more than once. Lots of useful information in that thread. Based on that I decided to remove the mylar and use white walls. I already get some burnt leaves on the outer edge using t5s. It's difficult to get the mylar perfectly flat without gluing the stuff down.

I think my emitters are placed close enough that I'll get an even light spread everywhere except the outer edges. If you look at the pic I posted earlier the 2700Ks will be in the vertical groups of 3 and the 4000s in the vertical groups of 2. Distance from like row to like row is 15 inches. Light intensity for these emitters at 12 inches and 7 inches out is about 50%. So, 14 inch spacing between like rows would be ideal. I managed 15 which is very close. Based on that I'm hoping the 4000Ks will manage stretch well enough. I'm confident I will at least be getting more blue than I currently am with my 3000K t5s.

The frame will be very minimal with no casing so I suspect those radial heat sinks will do a fine job and are the perfect profile for the air movement I'm dealing with. I will have a couple small fans for mixing air around which I've always done with my t5 panels anyway.

I haven't really put much thought into far red or blue supplements. I need to do some more research, but I'll probably not modify anything until I've picked a few tomatoes. It could be something I add in, but it wouldn't affect the current schematic. The only problem I might have dealing with a future modification is too much light. Being able to spread the source out, I would consider it a waste if I have to raise the lamp higher above the canopy.

I had a 60w draw multichip on a similar sized heatsink. The fan hit it indirectly and it'd get really hot. 30w might be okay but a bunch together might have a oven effect and be heavily dependent on air exhaust. Would need a fail safe like a thermostat in case of fan failure I'd think.
either way I'm sure you'll figure it out! Nothing like money spent and parts laying around for motivation...lol

That is a ton of power it will be awesome. To give fair warning though, I agree with positivity the heatsinks need more surface area to keep your Tj in a value point range (85C or less). For reference, I use 110cm2/dissipation watt for passive cooling to achieve a Tj of ~50C in the case of a lamp that is 40% efficient. Heatsink temp (35c) is just a bit above ambient (25C).

With temp droop, current droop and vF shift all taken into account, the 3000K Vero18 @ 1050 puts out 107 lm/W (32% efficient) and that is assuming a Tj of 50C. So higher junction temps create even more heat because of loss of efficiency from increased temp droop. At Tj 85C the Vero18 drops to 100 lm/W and 29.7% efficient. I have not worked up the numbers for the 2700K and 4000K but it would be very easy to do so let me know if you need them.

Found the sinks on eBay and calculated each to be about 750 sq cm. So for a 32% efficient lamp, assuming we need 125sq cm/dissipation watt, each heatsink is good for only 6 watts. PITA I know. Its hard being a passive heatsinker. I have been unable to beat Heatsink USA prices and their prices keep rising. I pay 38.50 shipped for a 50w heatsink so that works out to 77cents/watt versus $2.20/watt for the radial eBay heatsink.

SupraSPL said:

Found the sinks on eBay and calculated each to be about 750 sq cm. So for a 32% efficient lamp, assuming we need 125sq cm/dissipation watt, each heatsink is good for only 6 watts. PITA I know. Its hard being a passive heatsinker. I have been unable to beat Heatsink USA prices and their prices keep rising. I pay 38.50 shipped for a 50w heatsink so that works out to 77cents/watt versus $2.20/watt for the radial eBay heatsink.

Click to expand...

Thanks for looking it up and doing the math. I've went ahead and ordered an emitter for testing purposes. The lamp my design is based off of with the same sinks uses CXA 2520s @ .7A, so 25 watts and was described as 'barely warm', so I'm not at all sure what to expect now.

PetFlora said:

My 2 cents

Staggering the rows, just like in your picture, will max uniform canopy coverage

I think 2:1 @ 2700/4000 is too much red. Consider 2:3, or 1:2

Click to expand...

Still more blue than a lone HPS though right?

what would be the ideal spacing for a 4X vero18 setup?

You'll get even coverage with up to 14" spacing. Moving them closer will give you more intensity with a smaller overall footprint.

4 of them running at 1.05 amp evenly spaced in a 2x2' (emitters in a 10-12" square) area would be a good veg lamp. 4 x 4000K and a 5th 2700K in the middle turned on at flower would be a good setup from start to finish for a single bush. It would be 38 watts per square foot during flower which is pretty respectable.

Vero 29 @ 1.4A uses 50 watts, so 4 of those would provide 50 watts per square foot in the same area. All the grows I've done so far have used t5 tubes, about 35 watts per foot yielding .5 GPW in hydro. Someone with actual LED experience might have a better viewpoint.

How much blue will be somewhat strain dependent, so if you can run some blue on a separate o/o switch you have more flexibility

Rahz said:

Still more blue than a lone HPS though right?

Click to expand...

I have drivers to play with! After setting up a test unit and running it for several hours my test probe was registering a temp of 120F.

Thoughts?

Also a bummer... all of Bridgelux's distributors are out of the 2700K Vero 18 @ 80 CRI. Nobody is receiving new stock until the first of March.

Do you have a way of measuring vF at room temp vs vF at stable operating temp? That should help give you some idea of Tj. Based on your 50C measurement Id guess your Tj about 85C. Where did you measure it and what was ambient temp?

A friend of mine has the tool to measure vF, I'll get in touch with him about it. Ambient temp was 74 degrees F. I measured from top center of the heat sink with a piece of electrical tape holding the probe down. This was with only ambient air flow. I will get better air movement in the chamber, so I'll be testing in that environment soon.

85 at Tj is a bit higher than I'm comfortable with. If you're right I'll have to consider other options.

I haven't gotten any new temp readings, but in the chamber the lamps are staying only warm to the touch. The aluminum frame also helps dissipate some of the heat. I have two fans blowing across the rig and am happy with the temperature situation. The lamp works great. I'm on my 2nd grow with it and would be happier if I had waited for the 2700K emitters. I ended up using 3000K for the reds and there's a bit more blue going on that I would like. So if anyone builds a lamp with the Vero emitters I would recommend 2700K supplemented with 4000K.

That being said, veg is fantastic and growth explodes at 12/12. First grow came out to .8 GPW and that was with some rough clones. Of the three one had trouble starting and never grew to full size. I expect to pull 1 GPW for my current grow which is at week 3, already frosting up.

I got some bleaching at 4 inches last grow, plants are now happy with the lamp 6-7 inches above the canopy.

I'm still unsure of the effect of COB size in relation to lamp placement. Despite the hype of the bigger COBs I suspect the larger the COB the further away from the canopy they will need to be to prevent bleaching. For this reason a larger number of smaller cobs seems ideal. I didn't use Vero10s because the build would have been more complicated and more expensive. The Vero18s seem like a good compromise between lots of small emitters and fewer large ones.

In this post I'm including a pic of the frame -w- drivers and heatsinks. Sorry I didn't have everything wired up in the pic but you get the idea. I've also included the final circuit schematic. I went with the PLM series driver from Meanwell because it has PFC (power factor correction) which makes them less lossy resulting in better power bills and less heat output. Another deciding factor was the screw down terminals which meant I didn't need to solder a single thing.


Final cost for 720 watts of LED covering a 3x7 area (20 square feet): About $1400. I'm not sure what else to add, so feel free to ask questions.