DIY multi LED grow light

would say youre fine in case of the LH351H then, it may dont work out with other 3535s,
my 660 3535 would quit their service if the heat sink pad is connected to GND.
thermals will be good anyway due to the nearby alumimium base so you could consider to isolate the slug from gnd to make the board compatible with a wider range.
normally the sink gets some "ears" (square pad) left n right to deduct the heat to aluminium, it of course have its drawbacks as its not spreading the heat that good over the top copper layer.
if it will be to measure is another question.

here you can see one led example.

cobshopgrow said:

would say youre fine in case of the LH351H then, it may dont work out with other 3535s,
my 660 3535 would quit their service if the heat sink pad is connected to GND.
thermals will be good anyway due to the nearby alumimium base so you could consider to isolate the slug from gnd to make the board compatible with a wider range.
normally the sink gets some "ears" (square pad) left n right to deduct the heat to aluminium, it of course have its drawbacks as its not spreading the heat that good over the top copper layer.
if it will be to measure is another question.

here you can see one led example.

https://www.prolightopto.com/storage/ftp/DS-0517.pdf

Click to expand...

The prolight design is now quite dated and is a throwback to the old LUXEON LUX leds that always used an active slug, many of the crappy early chinese lights use that old P2 style design chip that also used an active slug. This is rare to find now.
I would think if you want the best solution for 660nm, Samsung is not a great choice

Cheers

its maybe dated and i am aware that isnt that common (anymore) but these design are sold till today and get led die updates.
leds who have their slug seperated internally do work also fine if you seperate them also on the mpcb,.
you maybe have a better guess how big the impact on the thermals will be, would think its small if the transfer pads are big enough, considered there is aluminium right underneath.

@bob0816 btw. how you plan to solder the 600mm mpcb?

cobshopgrow said:

normally the sink gets some "ears"

Click to expand...

This is my basic 3535 footprint. I used the one suggested by Osram for their Oslon Square/SSL series as model.

https://dammedia.osram.info/media/resource/hires/osram-dam-5537535/LUW CQAR (EQW)_EN.pdf e.g.
Copper is higher heat conductive then aluminium so I will expand the copper planes around the LEDs as big as possible.Just like normal copper fill but expanding from the heat pads and connecting the anode and cathode pads of consecutive diode footprints instead of vias (vias drawen "under" them for KiCad to know they are connected). I will post the new design today probably, maybe tomorrow.

welight said:

I would think if you want the best solution for 660nm, Samsung is not a great choice

Click to expand...

Did you plot the data yourself? I wanted to compare the diode specs but I didn't know how, did you read the values of the graphs and extrapolated? It does look more like exact data. Looks nice.
Unfortunately I already ordered. I thought getting them all from one brand and in the upper price range, felt like it was a guarantor for longevity (well with my first light board maybe not the smartest, but a bold move).

cobshopgrow said:

how you plan to solder the 600mm mpcb?

Click to expand...

It will all be reflow soldered in a normal oven. I will get through all reflow data and find the best overall temperaturse and times and then craft it with the help of a roast thermometer.
The price for the solder stencil is already included in my calculations.


Does anybody already conformal coated an LED build and knows which silicone coat is suted for LEDs? For now I only know to use silicone because of its thermal conduction and translucency, but not which brand is applicable without being a risk of oxidation through gasing out volatile organic compounds (VOCs, see in the attached PDF) will have to check with the manufactorers I guess.

bob0816 said:

Copper is higher heat conductive then aluminium so I will expand the copper planes around the LEDs as big as possible.Just like normal copper fill but expanding from the heat pads and connecting the anode and cathode pads of consecutive diode footprints instead of vias (vias drawen "under" them for KiCad to know they are connected). I will post the new design today probably, maybe tomorrow.
YES Copper POUR is your best solution, but also thicker traces, ie use 3 oz min


Did you plot the data yourself? I wanted to compare the diode specs but I didn't know how, did you read the values of the graphs and extrapolated? It does look more like exact data. Looks nice.
Unfortunately I already ordered. I thought getting them all from one brand and in the upper price range, felt like it was a guarantor for longevity (well with my first light board maybe not the smartest, but a bold move).

Does anybody already conformal coated an LED build and knows which silicone coat is suted for LEDs? For now I only know to use silicone because of its thermal conduction and translucency, but not which brand is applicable without being a risk of oxidation through gasing out volatile organic compounds (VOCs, see in the attached PDF) will have to check with the manufactorers I guess.

Click to expand...

Just like normal copper fill but expanding from the heat pads and connecting the anode and cathode pads of consecutive diode footprints instead of vias (vias drawen "under" them for KiCad to know they are connected). I will post the new design today probably, maybe tomorrow.
YES Copper POUR is your best solution, but also thicker traces, ie use 3 oz min

Did you plot the data yourself? I wanted to compare the diode specs but I didn't know how, did you read the values of the graphs and extrapolated? It does look more like exact data. Looks nice.
Unfortunately I already ordered. I thought getting them all from one brand and in the upper price range, felt like it was a guarantor for longevity (well with my first light board maybe not the smartest, but a bold move).
Yes data is generated from led data sheets

USE Cree Chemical compatibility guides, they have complete lists of materials that are suggested to use
cheers

copper is almost twice as good as a thermal conductor then aluminium, but that doesnt mean aluminiun is bad.
it depends a lot on the area which heat is transfered over, from a certain size there will be hardly any benefit going bigger or for a better material.
the most important part is to get initial heat close to the led away quickly, there some copper is important.

always good to check the thermal conductivity of a material to get some idea.
and imagine that even some with a lower number are considered average heat conductors, like water with its 0.60 K.

in the end a thermal cam tells you what going on, but its not hard to get in to a safe range if you dont drive them too hard, which isnt good anyway.

i hope your oven is good, regular household ovens lack a bit in their heat distribution qualities.
its hard to get a even temp, maybe you find something to improve the distribution?
i did mpcb repair reflows in a oven, was quite a pain and the boards wherent as long as yours.
the white paint will tell you, when its darken you know you where over the sweet spot .
i would consider at least one PCB as "practice" board, maybe get some of the cheapest 3535 you can get too, just to test the reflow process.

welight said:

Just like normal copper fill but expanding from the heat pads and connecting the anode and cathode pads of consecutive diode footprints instead of vias (vias drawen "under" them for KiCad to know they are connected). I will post the new design today probably, maybe tomorrow.
YES Copper POUR is your best solution, but also thicker traces, ie use 3 oz min

Did you plot the data yourself? I wanted to compare the diode specs but I didn't know how, did you read the values of the graphs and extrapolated? It does look more like exact data. Looks nice.
Unfortunately I already ordered. I thought getting them all from one brand and in the upper price range, felt like it was a guarantor for longevity (well with my first light board maybe not the smartest, but a bold move).
Yes data is generated from led data sheets

USE Cree Chemical compatibility guides, they have complete lists of materials that are suggested to use

https://www.cree.com/led-components/media/documents/XLamp_Chemical_Comp.pdf

cheers

Click to expand...

...me gustan de 4oz...

Usually, the copper thickness of a standard PCB is 1oz to 3oz.

Thick-copper PCBs or heavy-copper PCBs are the types of PCBs that the finished copper weight is more than 4oz (140μm).


y las pcb flexibles para interconectar paneles o leds...a Grow Light Australia creo le vendrian bien...para interconectar paneles...a los de HLG tambien... en diy que decirte...



...recuerdo a Kanna con la cinta amarilla y cinta de cobre haciendo sus conexiones asi... y a Supra usandola tambien para sujetar leds...las famosas cintas Kapton amarillas...ouuu yeahhh... forgeted ones?...

...las vueltas que da el mundo ehhh Kanna...juasss

os deje un regalo de navidad a los bros... me da que muchos no lo habeis visto...pues me ignorais...



...mucho lio y ajetreo ....juasss ustedes os lo perdeis...

...tienes razon Wilson...espiritu navideño...este tampoco lo borrare yo...

Saludos desde Tenerife

V. 1.2

welight said:

USE Cree Chemical compatibility guides, they have complete lists of materials that are suggested to use

https://www.cree.com/led-components/media/documents/XLamp_Chemical_Comp.pdf

cheers

Click to expand...

Thanks, that's very helpful!

I won't go over 1 oz copper or with copper core because both are way more expensive and I think copper core could be to soft a metal with the board length.
Will consider the temperature thoughts on the reflow soldering but I think with an convection oven and a thermometer I will be on the safe side, will keep both eyes open.
I'm sure as long as I stay around optimal current level there will be no need for active cooling but will check the temperatures for different currents to drive it efficiently.

even a convection oven isnt really good for reflowing something, i used one too.
a cheap trick to guess the melting point is to put a piece of solder somewhere in the oven you can see well, if it melts you know there are f.e. 220c or whaever the melting point of your solder is.
its doable in a oven but really needs some practice to get a nice result, especially for pcbs with a white solder stop.

Hm, maybe I'll try with bottom heat then. To have a test run is a good advice, thanks.
Do you have a picture of your light or a link?

you really need to try a bit would say.
bottom heat is probably good, convection and bottom heat sounds good to me.
its a bit ago i reflowed some kingbrite lmb301b pcbs.
i sacrified one board for the tests to see where the solderstop becomes beige.

as i didnt want to hassle anymore with a oven for reflowing mpcb i made me some simple fr4 based strips for my reds n uvas.
theyre spread well appart and have lots copper surface to spread the heat, beside that theyre driven so low anyway, there is really no need for aluminium, the reds f.e. never see more then 0.2W each 3535 led.

i did long going tests with a thermal cam, low 30s celcius, mid to high 30s on the led die max. as far i remember.
when you zoom a bit in you see the selfmade strips (red, blue) in between bridegelux eb3s, the custom ones fit in between the eb3.
below is what i am recently using since a year or so.

Nice rig, how many W is that?
Did you buy them assembled or how did you solder the FR-4?

Where you able to reflow solder without damaging the soldermask after the test run?

I'm also considering a hot plate but I don't have a infinitely variable plateau like hot plate available, only infinitely variable ceramic stove or gradually hot plates, seems both rather troublesome.

i could max out the rig to about 400w, but ill never ever will do that in that growspace.
soldered them with a smd hotair, fr4 is quite easy to solder compared to mpcbs.

never dealt with a hotplate, basically it sounds like a good idea.
i remember flat electric pans for crepes, quite big ones.
probably one thing to keep an eye on is the contact to the hotplate, plate should be quite even for good transfer (no airgaps).
maybe a combination of hotplate and hotair is a good trick too, i never tried!
if you really want i am sure you can make it work, mpcb need to be heated entirely and quite even to get good results but with a lil sense it can be done for sure.

I saw these two before I was starting blueprinting:

Reference pads (09:21); electric oven reflow soldering (10:51)

inductive cooktop reflow soldering (07:39)

I think I will go with a convection oven and figuere out it's actual temperatures according to set temperatures with a cooking thermometer first and then do two strips first and then the other three.
The MCPCBs will be ordered with stencil, just have to check the data sheets for thickness again to get contact=heat transmission as good as possible

Microcontrolled electric oven as reflow station, A nice toy/tool and food for thought

i did used a pizza oven myself more then a decade back,they work good.
if i remeber right i used some extra aluminium inside to spread the heat better.
there is lot instructions out regarding good pizza ovens, its just theyre too small for you?

with a stencil it will be very easy to apply the solder paste, there i wont have any sorrows,
its not a really dense populated board.

the guy with the atari shirt explaining very well and he nailed the hotplate/hotair mix method.
this video is a good find.

who knows you may find some good settings quickly for your oven, as long youre aware of what could happen, its half the way.

oven....hotplate/hotair mix method... i like the diferent techs ...

im old...my preference its hotair method but ...actually...maybe i would try or learn more ....new hardwares for me


...yeah for some things a hotplate similar to the prev video its a nice hardware for electronic diyers...

...ahora los que nos venden pcb tambien suelen vender los stencils para esos pcb...
...mira bien la fecha de caducidad de las pastas de soldar cuando las compres...y mantenla bien cerrada y en una bolsa al vacio en una nevera...

Si Wilson que mire posts del Bro SDS.... aunque actualmente yo como sonda sensor de calor elegiria un PT1000...

PD...bro SDS ...los PT1000 tambien necesitan amplificadores...MAX31865...me da que tienes algunos por ahi...para reflow..... ;--)







y volvemos a mute Wilson...

Saludos desde Tenerife

If your using MCPCB then I would always use a hot plate if you do not have professional production equipment, you can more easily observe the reflow process with a hot plate to gain some experience about the reflow process and also move leds while solder is molten if needed, if you use FR4 then convection oven may make more sense. in your pad design you may also wish to consider relief pad design to help evacuate excess solder.
I suggest this because we use a large conveyor reflow oven, but if we have any issue and need to rework, we have to use a hot plate to do that, a convection oven for rework is no use as parts/solder need to be molten for you to be able to fix, so better to use a hotplate for both your reflow and ability to repair
Cheers

very valid points.
i mean its doable to rework a mpcb in a convection oven.
you just need to be quick and be able to withstand the heat blasting in your eyes when opening the door, you have a 10secs to work with your tweezer.
its a pain, but i where able to repair a few this way, replacing 301b leds, wont do again.
if the budget allows and youre able to find some nice hotplate, i would also say go for it.
technically hotplates arent thaat complicated.
if you cant find something, maybe think of repurposing some 24v 3d printer heatings elements underneath some custom aluminium you can get locally, just a quick idea, as the 60cm are quite long.

edit; maybe such a generic 120mm pcb hotplate is all you need, they cost nothin also.

V 1.2.1

Added a reflow control pad on the bottom left and a 3.0x2.4 cm logo in the middle

cobshopgrow said:

i did used a pizza oven myself more then a decade back,they work good [...] its just theyre too small for you?

Click to expand...

Yes, 40 cm is too long for a small oven.

salmonetin said:

my preference its hotair method

Click to expand...

welight said:

better to use a hotplate for both your reflow and ability to repair

Click to expand...

cobshopgrow said:

if the budget allows and youre able to find some nice hotplate, i would also say go for it

Click to expand...

Unfortunately I don't have a hotair gun. I want to do it as easy and clean as possible. With the Metal Core PCB it won't heat up enough with just the hothair gun (because the aluminium will spread the heat which is what I want for the running heat of the LEDs) and put more heat stress on the part.
Probably I will try and find a hot plate with a good cost-benefit ratio so I can do repair if necessary and maybe a small hotair gun like the dude in post #55 if they are cheap, but for starters I'll try convection oven reflow soldering.

cobshopgrow said:

maybe think of repurposing some 24v 3d printer heatings elements

Click to expand...

I think they won't get hot enough, only found plates ending at 100 °C / 212 F which is too far from 200 °C / 392 F.

welight said:

in your pad design you may also wish to consider relief pad design to help evacuate excess solder.

Click to expand...

All pads are designed with excess pad space for excess solder like in post #45 or with noses to the short sides.