DIY : Soldering Leds to heatsink ,without MCPCB.Doable ?

stardustsailor

stardustsailor

Well-Known Member
Another Alternative for Epoxy :
Adtech ..
http://www.adtechplastics.com/pc-61-39-ec-440-high-temp-base-resinhardener-for-n-61-n-6-ceramic-filled-casting-system.aspx
http://www.adtechplastics.com/pc-54-38-el-336-high-temp-laminating-system.aspx

And it's ceramic fillers :
http://www.adtechplastics.com/pc-177-43-n-61-ceramic-powder-mix.aspx
http://www.adtechplastics.com/pc-176-43-n-60-ceramic-powder-mix.aspx

...that improve filler compaction and 'specific heat'* properties of...
Click to expand...
***The specific heat is the amount of heat per unit mass required to raise the temperature by one degree Celsius. The relationship between heat and temperature change is usually expressed in the form shown below where c is the specific heat. The relationship does not apply if a phase change is encountered, because the heat added or removed during a phase change does not change the temperature.
 
stardustsailor

stardustsailor

Well-Known Member
So...Up till now ....

-Heatsink will not adhere with epoxy ,if it is anodised (black,of course ..)
- 80-110 um total thickness of epoxy+ceramic filler dielectric layer is desirable .
-Probably ,spraying pre-mixed A+B liquid epoxy,is the best way .
-Leds can not be soldered easy using reflow oven .Manual hot air reflow soldering is the only way,to utilise ..
(Hand -made like true Ferr..ok ...You got the point .. )
 
stardustsailor

stardustsailor

Well-Known Member
Now ..
That's a quite easy step .....

Since the first " Half " of dielectric layer is cured ..( 40-50 um )..
Then another fresh epoxy layer is applied and then a copper plate/sheet of 1/2 to 3 oz ,
( that is per square what ? feet ? inches ? meters ?...... http://org-www.doosan.com/dse/attach_files/pdf/Heat_Materials.pdf ......)
is attached ....

1 oz is 35 microns thick
1/2 oz is 17.5 um thick
2 oz is 70 um thick
3 oz is 105 um thick ...

Copper FOIL ....(NOT PLATE OR SHEET ...!!!)

Copper :
I.e.
Google "Copper Sheet " ...
http://basiccopper.com/
http://www.onlinemetals.com/merchant.cfm?id=93&step=2
http://www.whimsie.com/coppersheetwire.html


Over the copper plate ,a flat & rigid board is placed (i.e wooden one ..)..
( of bigger surface than copper plate )


Clamps are fitted and tightened,so to apply pressure between copper plate and heatsink ,while second layer of epoxy is curing ...

clamped.jpg
Clamps will be removed after epoxy has completely cured ...

Edges of copper plate will be trimmed and sanded to match the edges of heatsink (profiled)..

After this step ,the heatsink will have attached a copper surface on it ,
but they will be electrically insulated between them ....
Though ,heat will pass relatively easy from copper plate to aluminium heatsink .
Only ~0.1 mm of epoxy +ceramic filler ,stands between them ...

Not any particular trouble with this step ....

***Bubbles should not occur /minimised if epoxy cures at high temp . ( 80-100°C ) for a short while ..
With the clamps on,into the oven for 15' at ~80°C,towards middle of curing time (after 3-6 hours of epoxy parts mixed firstplace)
....
***Caution needed not to over-wrap/twist the copper sheet-while handling and working with - that would cause "wrinkles " to form ..
 
stardustsailor

stardustsailor

Well-Known Member
This step is kinda tricky .
Some skills in image processing / cad designing software are needed ....
Those of you having them ,can create a "library "of different circuits arrangements ...

I.e ...Dedicated to a fellow RIU DIYer ,who goes the Oslon Way ( I've turned yellow-amber , from jealousy !! ).
A really fast designed (rough -print !!! ) circuit ....

(With Leds To be Manually soldered,with hot air solder reflow gun .
Reflow oven can not be used here .)

What is visible in black color is going to be copper ...
the rest ,bare epoxy ..(de-copperized !!!.. )..

Of course the laser print has to be flipped horizontally ,so to transfer correctly to the copper plate ...

Here :for guod new.jpgfor guod new2.jpg
 
tenthirty

tenthirty

Well-Known Member
stardustsailor said:
Solution #2 :thin flexible circuit board material bonded to the HS
????

Any links there ?

Can it be bought unetched ?
(fully copper covered ..)

How to you bond that with heatsink ?
???



Edit :
Of course carbon conducts electricity ...
Drawing a line-between two contact points - with a simple pencil on a pcb ,actually is like placing a resistance ....
Carbon fiber was pointed ,just to show that is very easy for the epoxy to form very thin layers ( 30-60 microns thickness ),
even if applied -carefully- with a normal brush ....
Click to expand...
http://www.allflexinc.com/flexible-printed-circuit.shtml

You can have rolls of sticky back of what ever you want, multi layer etc......

Roll out a length, cut it off the roll, solder on your diodes, feed wires, peel the sticky back and stick to the HS.

Easy as pie!

You could still use the epoxy if you would like.
 
H

Hosebomber

Active Member
Sorry I was under the impression that you was using the West Polyester resin and not the actual epoxy. (mix ratios change greatly and very on a number of things like cure time and temp).

Adding a additional layer of copper under the thermal slug that is the thickness of the resin may take care of the height difference more effectively (cost, time, and thermally speaking) than some of the other options you have listed.

Likewise, it would be easier to vacuum bag the copper on to the epoxy rather than wood and clamps. Epoxy resins adhere to wood rather easily. With a single good vacuum system you could easily make a cleaner smoother surface. If you really wanted to get fancy with it you could make a small autoclave to vacuum bag and heat cure at the same time.
 
stardustsailor

stardustsailor

Well-Known Member
Hosebomber said:
Sorry I was under the impression that you was using the West Polyester resin and not the actual epoxy. (mix ratios change greatly and very on a number of things like cure time and temp).

Adding a additional layer of copper under the thermal slug that is the thickness of the resin may take care of the height difference more effectively (cost, time, and thermally speaking) than some of the other options you have listed.

Likewise, it would be easier to vacuum bag the copper on to the epoxy rather than wood and clamps. Epoxy resins adhere to wood rather easily. With a single good vacuum system you could easily make a cleaner smoother surface. If you really wanted to get fancy with it you could make a small autoclave to vacuum bag and heat cure at the same time.
Click to expand...
Now,that's a snack of tasty ideas !
Thanx ,brother !
 
stardustsailor

stardustsailor

Well-Known Member
When designing the circuit for printing :

-The printed circuit on paper has to be flipped horizontally ("mirror image " ) ,in order to be correctly transferred on the copper plate .
-Laser Print size has to match the actual size and measurements of led contacts /slugs ,heatsink size ,leds distances & placing ,ect .
So Image resizing is needed .( scale 1:1 )
 
Bumping Spheda

Bumping Spheda

Well-Known Member
GIMP is a nice program and it's free. You can make the image whatever size you want and then tell it to print per your specified dimensions in a plethora of units (pixels, inches, millimeters, etc).

There are also a few free programs that can convert JPG to DXF. Easy way to draw your diagram in MS Paint and get quick vector info out of it.
 
stardustsailor

stardustsailor

Well-Known Member
Now ....
Even a little girl can do that !

[video=youtube;lxRNQbEGwm4]http://www.youtube.com/watch?v=lxRNQbEGwm4[/video]


Toner is styrene acrylate co-polymer ..

1)Printed circuit can be transferred quite easy from paper to copper ,with a bit of ...ironing ...

2)Then (relatively safe) Iron (III) chloride is used (can be purchased from electronics shops, in liquid or water soluble solid form ),
to " dissolve " un-masked copper,away ....(A small brush can be used ,at this point ...)

3)With some Aceton ,toner "mask" is cleaned off(dissolved) ,and shinny copper is visible now and ready to be soldered ....
 
stardustsailor

stardustsailor

Well-Known Member
Bumping Spheda said:
GIMP is a nice program and it's free. You can make the image whatever size you want and then tell it to print per your specified dimensions in a plethora of units (pixels, inches, millimeters, etc).

There are also a few free programs that can convert JPG to DXF. Easy way to draw your diagram in MS Paint and get quick vector info out of it.
Click to expand...
Also there are some free electronic circuit designing software ....
Some google search needed....

I.e . http://www.circuitstoday.com/electronics-circuit-drawing-softwares

Here's a "25 Watt " design for Oslon leds ,made with Photoshop in less than 30 ' .....
25_oslon preprint.jpg
 
stardustsailor

stardustsailor

Well-Known Member
So...I really do think that it is not so difficult to design and transfer a circuit for leds ,"directly" to a heatsink ...
Without the use of any kind of MCPCB ....
With a max thickness of 80-110 microns for the dielectric layer
and ranging from 15 up to 100 microns for the copper layer ....
But still , it has to be put into test ...


Notes
Materials needed :
-Copper foil ( 1/2 ~ 3 oz @ 18 ~ 110 um )
-Quality Epoxy resin & hardener in low viscosity liquid form
-Ceramic filler for the epoxy resin
-Iron (III)chloride
-Heatsink
-Acetone
-Soap water
-Isopropyl Alcohol
- Rinsing Water (tap water will do fine .)

Tools needed :
-PC ,software & Laser Printer
-Oven
-Air compressor and paint gun/sprayer or air-brush.
-Extra fine sandpaper.
-Clamps & flat rigid boards or " air tight bag "system .
-Knife / scissors
-paper towels
-soft brush (for applying Iron chloride .)
- Hot Iron or ironing press .
-gloves/protective glass .
 
ganja 2

ganja 2

Active Member
stardustsailor said:
Now ..
That's a quite easy step .....

Since the first " Half " of dielectric layer is cured ..( 40-50 um )..
Then another fresh epoxy layer is applied and then a copper plate/sheet of 1/2 to 3 oz ,
( that is per square what ? feet ? inches ? meters ?...... http://org-www.doosan.com/dse/attach_files/pdf/Heat_Materials.pdf ......)
is attached ....

1 oz is 35 microns thick
1/2 oz is 17.5 um thick
2 oz is 70 um thick
3 oz is 105 um thick ...

Copper FOIL ....(NOT PLATE OR SHEET ...!!!)

Copper :
I.e.
Google "Copper Sheet " ...
http://basiccopper.com/
http://www.onlinemetals.com/merchant.cfm?id=93&step=2
http://www.whimsie.com/coppersheetwire.html


Over the copper plate ,a flat & rigid board is placed (i.e wooden one ..)..
( of bigger surface than copper plate )


Clamps are fitted and tightened,so to apply pressure between copper plate and heatsink ,while second layer of epoxy is curing ...

View attachment 2477985
Clamps will be removed after epoxy has completely cured ...

Edges of copper plate will be trimmed and sanded to match the edges of heatsink (profiled)..

After this step ,the heatsink will have attached a copper surface on it ,
but they will be electrically insulated between them ....
Though ,heat will pass relatively easy from copper plate to aluminium heatsink .
Only ~0.1 mm of epoxy +ceramic filler ,stands between them ...

Not any particular trouble with this step ....

***Bubbles should not occur /minimised if epoxy cures at high temp . ( 80-100°C ) for a short while ..
With the clamps on,into the oven for 15' at ~80°C,towards middle of curing time (after 3-6 hours of epoxy parts mixed firstplace)
....
***Caution needed not to over-wrap/twist the copper sheet-while handling and working with - that would cause "wrinkles " to form ..
Click to expand...




Sets
temperature
pressure
time
:lol:
 
stardustsailor

stardustsailor

Well-Known Member
Dielectric layer materials : Low viscocity-( high wetting )Epoxy + Ceramic Filler(s) .

Dielectric layer thickness : 80 -110 um (microns ) - 0.0031496 " to 0.0043307 "

Characteristics of ceramics that are used as filler for MCPCB
( http://hkpca2.intimexptx.com/ptxcms/website/hkpca2/gallery/d3030c1b-8f00-456e-879c-b0b5f1c08633.pdf ):

Alumina : Dielectric constant { Dk } of 9.7 -Thermal Conductivity of 33 W/m-K

Aluminum nitride :Dielectric constant { Dk } of 10 -Thermal Conductivity of 117 W/m-K


Beryllium oxide ( Beryllia - BeO) :Dielectric constant { Dk } of 6.7 -Thermal Conductivity of 251W/m-K


-Electrical insulator with a higher thermal conductivity than any other non-metal except diamond.
Actually exceeds that of some metals.Like all beryllium compounds, BeO is carcinogenic and may cause chronic beryllium disease.
Once fired into solid form, it is safe to handle as long as it is not subjected to any machining that generates dust-
Click to expand...


Boron Nitride :Dielectric constant { Dk } of 4.2 -Thermal Conductivity of 55 W/m-K

https://www.epotek.com/SSCDocs/whitepapers/Tech Paper 42.pdf

http://wings.buffalo.edu/academic/department/eng/mae/cmrl/Increasing%20the%20thermal%20conductivity%20of%20boron.pdf



More :

[url]https://smartech.gatech.edu/bitstream/handle/1853/11285/CPWongIEEE13.pdf?sequence=1




[/URL]
 
stardustsailor

stardustsailor

Well-Known Member
Ganja...
What do you think ? Are we going to try it ?

This is the ASL 25 ( as substitute to current ASL 24 ) ...
Astir ASL_25 print.jpg



An optional ASL 30 is being designed ...
But I'm afraid it's going to push the heatsink to it's limits ..
Most probably ...


Αυτό που με ρώτησες ...
http://www.hobbyarmos.gr/16-filla-metalou-sirmata/fillo-xalkou-30x40-0.15mm.html

300 χ 400 .. 2,25€ το πανελ για χαλκό από το βιβλιοπωλείο ...(0,15 χιλ )
Χοντρό στρώμα ...Κάτι σε 0.1 -0.7 θα ήταν ακόμη καλύτερο ..

Βρίσκεται εύκολα όμως ....

Άλλο 300χ 400 χ 0.1 χιλ

http://pacoartcenter.gr/product.php?productid=987

200 χ 300 χ 0.12 χιλ .

http://www.aplada.gr/20-by-30-0-12mm.html
 
You must log in or register to reply here.
Top