1st build started today Quin SST-50

[moderator007]

The current determines wire size not voltage. In series you have a high voltage with what ever your estimated drive current will be. I am assuming your going to run at 5amps. 22 AWG would probably work fine but 20 AWG or below would be better. The bigger the wire the less resistance. If you can get some silicone high strand count wire that would be best. You can find it at RC hobby shops or online sold by the foot.

 

[Curdog92]

Made a little progress on my project...here's a few pics.

First thought was to mount LEDs on pads, then solder wiring to the underside, then encase the wiring in JB Weld. After further thought, I now plan to mill out a groove in the JB Weld to create a solid copper wire inlay so that the LED will mount just like it would as a SMD. If it works out as planned, it should be a very clean install for the LEDs.

 

[wquiles]

Great photos - looking forward to seeing more of the project

Will

 

[mikesantor]

This project is looking great. Thanks for sharing!

Sent from my DROID BIONIC using Tapatalk

 

[CKOD]

Nice work, the JB weld should work out great. Ive done comparative testing and had stars mounted to heatsinks with JB weld, and it handled reflowing just fine. Cant wait to see the copper inlaid and turned down.
As big and thermally massive as that heatsink is, I'd reflow the LEDs on a stovetop. Add a temperature sensor (the infrared laser ones wont be accurate with bare aluminum!), And if youre using solder paste, apply your paste, heat the whole thing to ~100-115 C, place your LEDs carefully, and turn the stove up and watch closely. Once the solder starts melting, turn off the stove since the temp will keep rising a bit as the heat on the bottom makes it way to the top. let it cool enough for the solder to solidify, then move it to another burner to cool fully (if you have an electric stove. If you have a gas one, I'd imagine its fine just letting it sit where it was)

Dont quench it in water, or cool it very quick, as the ramp down rate is just as important(if not more important) as the ramp up rate.

 

[Hoop]

Is that a cnc mill? If not, how did you machine the circle without a rotary device?

Press fits should generally be around a few tenths to a half thousandth to avoid galling and bent pins. (pin should be .0002 to .0005 larger than the mating hole)

 

[easilyled]

Exciting project. You have enviable skills.

I'm just wondering why you choose to use SST-50s instead of XM-Ls? XM-Ls are significantly more efficient than SST-50s and produce less heat. Even though you can drive SST-50s harder you get as much light output from driving an XM-L at 3.0A as you do from driving an SST-50 at 5.0A and at that drive the SST-50 consumes far more power and produces far more heat.

 

[Curdog92]

Thanks for the comments. I tested the stovetop method tonight with success with a few non LED SMDs--worked fine, although it was on a much smaller scale. I am still testing different methods of creating the copper traces--having trouble with just about everything I tried. I tried solid copper wire milled to 0.7mm, pressed copper wire, even milled copper block. But the best results have come from cutting a 2 mm cross section of 1 inch copper tubing, then cutting the circular piece, flattening it out, and then shaping it to fit from LED to LED. The wall thickness is 0.8mm so I essentially end up with a rectangular wire (2mm X 0.8mm) that is easily bent and shaped into place and will sit upright once I refill the groove with JB Weld.

I am using a non-CNC, 1988 model RF 31 mill I picked up for $500 a few years back. It has tons of run-out, and several hundredths of slop/backlash on every axis. It's very hard to get any precision out of it--tons of compensation required while working with it. At the same time I'm working on the flashlight, I am slowly restoring the mill.

I created a little jig with a pointer (that can be seen in the pictures clamped to the mill table) and rotate the head by hand. While I realize its not safe to mill something being rotated by hand, it is aluminum, I'm using a small bit, and I'm taking very small cuts on each pass.

I divided the head into 5 parts using a compass/protractor initially and was off by large amounts--several hundredths. I finally achieved reasonable accuracy using simple geometry to calculate the distance between the 5 points on the star, then used a digital caliper with the distance locked to scribe the 5 marks around the perimeter. The head is rotated on the jig using the pointer to get my 5 sides.

SST-50s and reflectors cost $50. I thought that was a great package deal.

This is more of a learning process than anything else for me.

 

[Hoop]

Good choice on the SST-50 imo. XML's have nasty green tint that people fail to mention. SST's are quality in comparison.

 

[Curdog92]

Little more progress... I gave up on creating a solid copper connection from one LED to the next in one fell swoop. Instead, I decided to create individual pads that can be placed more precisely. Once the pads secured in place with JB Weld, I'll solder a wire link from pad to pad, then encase all the copper in JB Weld. I milled some 10 ga copper wire into what I'll call "angle iron" type copper bar. The grooves cut into the heat sink are 2 mm deep, and 4 mm wide. The copper wire was milled to 2.5 x 2.5mm, then milled down to create the 0.7mm x 6.5mm LED +/- pads.

Way too much work going into this...but it's a fun learning process. I'm bound and determined to mount the LEDs as SMDs...would probably have finished by now had I went with stars

 

[burro breath]

Way too much work going into this...but it's a fun learning process. I'm bound and determined to mount the LEDs as SMDs...would probably have finished by now had I went with stars

if its worth doing it is worth overdoing. great work and cant wait for more

 

[wquiles]

Way too much work going into this...but it's a fun learning process. I'm bound and determined to mount the LEDs as SMDs...would probably have finished by now had I went with stars

I can't begin to tell you how much I have learned by being "stubborn" and determined to make something work - way too many times

Keep up the good work

Will

 

[Mettee]

Man that is some very cool work, cant wait to see more.

 

[gt40]

great use of jb weld...

 

[Curdog92]

Hopefully I'll have LEDs burning tomorrow. Got the +/- terminals completed.

Solder points created for jumper wire.

Hopefully later tonight I'll get the wires in, cover with JB Weld, cure overnight, then mount the LEDs and have 5 SST-50s burning tomorrow!

 

[mikesantor]

Man that is some very cool work, cant wait to see more.

I could not say it any better. I dont watch many build threads on cpf but this is good. Great job my friend.

 

[Curdog92]

Here's everything wired up. I removed the wires in the thru-hole and replaced with solder pads (forgot that I'll have this on a stove--would have melted.)

And the final product, with continuity checked, awaiting LEDs. Now I'll be watching the superbowl wishing I was mounting my LEDs

Thanks for all the positive comments.

 

[The_bad_Frag]

Wow that looks awesome.

 

[FRITZHID]

that is some beautiful workmanship and ingenuity, i can't wait to see the finished product. something like this is more a work of art than anything else, and should be appreciated as such. functional art is always a plus in my book. does a job but is nice to look at and appreciate the work that went into it. kudos.

 

[Curdog92]

Thanks for the kind words. I did a test heating run on the stove with a grill temperature guage inserted into a newly pressed copper core in the center (to be temporarily used for the purpose of monitoring temp.) I turned off the burner at 200 C but the temp kept rising to about 250 C, enough to boil solder out of the JB Weld. I'm thinking this would may have fried the LEDs. Looks like I will have to invest in a contact temperature probe for my multimeter.

With a nudge from a spoon, these solder beads rolled off leaving no visible signs of where they came from (not even a pinhole.) Had I not seen them appear first hand, I would have no idea where they came from.