DIY LEDs - The write-up

Status
Not open for further replies.
dwzm,

I am afraid I disagree. If you had a power supply (that was well regulated for voltage) that was at the right voltage (say 3.7 for one amp to drive an LED) and 16 amps. You could hook up 16 LEDs safely. It one blows you are still save. Th voltage would stay at 3.7 volts and the current would drop to 15 amps. This is the way most power supplies work. So daytripper is that supply constant current or constant voltage?

I agree it is not the best voltage for LEDs. But it would work just a little dim.
 
It would be interesting to go back through all the builds that have been posted and determine typical costs per LED for the heatsink material people have used. I've always estimated $1/LED but clearly there's a pretty wide variation.

67¢ / LED for the 10x7x3 monsters I got on EBay. ~1300²" total surface area and each one weighs about 6 lbs. Didn't really think about how tall 3" was until I had them in hand... They are surplus and have various holes in the base plate, would you believe not one of the holes infringes on my grid?
 
http://www.alltronics.com/cgi-bin/item/22Z014S/25/28-mm-sq-x-12-mm-Heatsink-Pkg-of-10

39 cents each! Use one per LED. It mentions "adhesive backed" though, and I don't think I'd trust a random adhesive on a surplus heatsink.

These look like the ones that I saw on terahz's fixture, which prompted me to get some for my project (mouser). 80 cents apiece but not adhesive backed.

He used them because you can cable tie them to a rod, then angle each as necessary over the tank. A couple fans ensures good cooling, but I had them running in a test setup without any fans and they only got about 15 degrees above ambient when running the LEDs at 700ma.

I am making a square tube (aluminum 3 sides with acrylic on the bottom) just large enough to put 40mm fans at each end push/pull, with the LED/Heatsinks on a rod inside the channel so I can angle them depending on where over the tank each is mounted - that way I can avoid lighting the glass, give extra over a clam, etc. Cooling should be fine with ducted airflow over the heatsinks. I will be mounting a temp sensor in the channel to alert the controller if a fan quits, to prevent overheating.
 
Cooling should be fine with ducted airflow over the heatsinks. I will be mounting a temp sensor in the channel to alert the controller if a fan quits, to prevent overheating.

Wow no posts in a whole day. Anyway. Your idea sounds great. One thing to note however is. The temp sensor will only read air temp as you designed. If there is no airflow (fans not running). The heatsinks may heat up, and not transfer to the air around them.
 
Thanks for the recommendation. I went and picked up this exact model and type K thermocouple... it's a very small bead type probe that seems to work fairly well.

I found that with just the LEDs on the temperature of my heatsinks and star PCBs is staying around 100. When my T5s come on the heatsinks get quite a bit hotter (havent measured at the hottest yet), so I have overall canopy ventillation issues, the issue isn't with my heatsinks or heatsink fans. I have some fans on top of my canopy blowing OUT, but it looks like I need to add some fans blowing air in also to get a push/pull setup going.


One thing I noticed about the 82139 meter: In Diode test mode it doesn't push enough power to light up a Cree XR-E like my tiny radioshack meter does. Do you find the same thing??


You can get a Craftsman DMM that reads thermocouples AND does all these voltage & current measurement things I do to my LEDs.

Front panel says: 82139 ~$30 from Sears.

I used omega "fine" wire thermocouple wire in some of my testing.
As Kcress pointed out, the TC itself can throw off the measurement if you are not careful.

The BEST way to test these thing would be to use the smallest TC wire & run it under the Star, but you would need to cut reliefs in the bottom to accomodate the wire.

Stu
 
Which MCPCB format do you prefer???

I went with the Cutter 10mm sq for the XRE only because there was a cheap optic that fit it.

I then when with the Cutter 10mm circle for the XPG just because.. uh, well... just because. There was no cheap optic I saw for any of the XPG

Any thoughts? The only down-side I can see to the smaller 10mm vs the 20mm is perhaps a more difficult time soldering the leads on.
 
555PWM.jpg
]
ok so I made this circuit. My only problem is when I attach the dim
+ i get nothing. If I go before the resistor I get full on. No dimming. What am I doing wrong?
 
Last edited:
Grim, you got me thinking on "surplus" heatsinks. A few minutes of googling yields some interesting potentials:

http://www.alltronics.com/cgi-bin/item/22Z014S/25/28-mm-sq-x-12-mm-Heatsink-Pkg-of-10

39 cents each! Use one per LED. It mentions "adhesive backed" though, and I don't think I'd trust a random adhesive on a surplus heatsink.

Some of the "medium" heatsinks from the place you posted could work for a small number of LEDs each, too.

This place has a TON of weird shapes:

http://www.alliedelec.com/fans-motors-thermal-management/heatsinks/?gclid=CPOnwMu9tKACFSFy5QodmRXCTw

I just order a mess of those heat sinks so I hope they work out.
 
I did some temp testing after running my T5s and LEDs for several hours using the thermocouple.

With only 2 exhaust fans installed on my canopy, my heatsinks were getting to a warm/hot 119F, warmer than I'd like. So I just installed two more fans blowing into the canopy. After an hour or so of runtime, the max temp reading I get on my heatsinks is 90F. So canopy airflow can be more important than the actual airflow across your heatsinks.... blowing warm air across them won't do any good. I'd also like to mention that my canopy has tons of holes in it. Even the 4" holes I used for the newly added fans were already there, so you can't really rely on the air to just replenish itself with fans exhausting. I think all the warm air was collecting inside the top of my canopy and the fans on my LED heatsinks were essentially pulling from that hot air supply. With more fans in the canopy that heat doesn't build up anywhere specific and gets removed more efficiently.
 
... The temp sensor will only read air temp as you designed. If there is no airflow (fans not running). The heatsinks may heat up, and not transfer to the air around them.

Actually, the temp sensor will be on one of the heatsinks. This won't measure junction temp of the LED, but will be better than air temp.
 
With only 2 exhaust fans installed on my canopy, my heatsinks were getting to a warm/hot 119F, warmer than I'd like. So I just installed two more fans blowing into the canopy. After an hour or so of runtime, the max temp reading I get on my heatsinks is 90F.

Any photos or diagrams to show what the exact configuration is? I'm curious to see how it compares with what others are doing.
 
555PWM.jpg
]
ok so I made this circuit. My only problem is when I attach the dim
+ i get nothing. If I go before the resistor I get full on. No dimming. What am I doing wrong?

Do you have a multimeter that can detect duty cycle or frequency? Or at least voltage?

I'd start by reading the voltage coming off the LM317. Make sure it's 10v. Then I'd test the output (where the DIM connections are supposed to be). Make sure you're getting something reasonable.

FWIW that circuit is "backwards" compared to what I'm used to for a 555 so I'm having trouble deciding how to troubleshoot it.

Did you solder the IC in place or use a socket?
 
Whats the recommended gauge wire to use for a led set-up? As I look through all the threads I have seen various sizes used. I am going to start my build this week want to make sure what size wire to use between:
1. power supply and buckpucks
2. buckpucks and leds
3. led to led.
Thanks for any assistance.
 
Any photos or diagrams to show what the exact configuration is? I'm curious to see how it compares with what others are doing.

Sure, this picture was taken awhile ago when I installed my T5s:

start.jpg


When I installed the T5s I moved the two rear 120mm fans to the top 2 holes blowing air OUT. I mainly moved them because I read that blowing air on the center of a T5 bulb could actually reduce its output.

So I recently put two 120mm fans back where you see them in that photo, pushing air in. I used Noctua NF-S12 fans which are virtually silent. They also put out such a wide diffused airflow that it won't do much "spot cooling" of my T5 bulbs, just gently forces air into the canopy to mix with the hot air which the Icecap fans on top pull out.
 
Has anyone tried the 4 led board cutter has put together?

2 Cree XPG whites and 2 XP-E Royal blues on about a 5" board. What do you think? 31 and change per board.
 
Status
Not open for further replies.
Back
Top