Killer PCBs

This instructable demonstrates the process for making printed circuit boards with features as small as 0.005suitable for LQFP or QFN ICs using negative dry film photoresist. This will enable you to handle just about any kind of integrated circuit available--even ball grid array! Pictured are boards with a TSSOP-14, QFN-40 packages using a .65mm pitch and zero insertion force flex sockets with .5mm pitch.

Step 1: Background

After experimenting with home PCB fabbing for a while, I've finally worked out a process that produces reasonably consistent results that actually look pretty good. I spent lots of time trying to use the toner transfer method with varying degrees of success (OK, varying degrees of failure might be more accurate). I also tried Philmore/Datak negative photo resist spray with consistently horrible results (the stuff eventually melted the spray nozzle that came with it and leaked all over the place). Not Green & not recommended. Now I could have purchased presensitized boards and saved a lot of trouble, but I find the material to be too costly for the volume of boards I'm producing. I eventually tried dry film photo resist and I won't be going back! I'm not going to go into the intricacies of schematic capture, or etching since those subjects are well covered by other instructables. No volatile compounds are used--only simple bases which can be rendered environmentally safe by filtering solids and neutralizing with HCl (see manufacturer instructions for proper disposal procedures). This process, when combined with a Peroxide/Cupric Chloride etching process forms an environmentally responsible, Green PCB development process.

If you haven't tried the toner transfer method, do so. Unless you are blessed with magical toner and/or paper, the dry film resist method will yield better results, but the process is a bit more involved. If you are satisfied with the toner-transfer results, by all means, stick with that method. Naturally the standard warnings apply: PCB etching and dry film processing involve caustic materials--be sure to use protective equipment and have an eye-wash station handy (or at least a bucket of water). Also note that dry film developing and stripping involve strong bases--keep them far away from your etching chemicals, or they may react violently.

Thus far, I've used three types of dry film resist, all of which performed well:
--MG Chemicals 416DFR Dry Film Resist About $20.00 for 12" by 5 feet at Frys, Altex and online. MG refused to quote larger quantities, and will not divulge the manufacturer of their film.

--Dupont Riston M115 available at Think & Tinker Excellent resist, much more economical than MG if you want larger quantities (12"x50ft for $96.75, 12"x100ft for $116.26). Outstanding outfit, very helpful, friendly people and lots of great info. Terrific site!

--Kolon Dry Film Resist Korean manufacturer sells for somewhat less than Think & Tinker's Riston, but with a minimum of 500ft cases.

What you will need
- Laser Printer
- Home/Office Laminator
- Laser Printer Transparencies
- Spray Adhesive
- Negative Dry Film Photo Resist
- Resist Developer (sodium carbonate)
- Resist Stripper (sodium hydroxide)
- Glass Sheets
- Clear Tape
- Yellow Bug Light
- Light-Safe Area

Optional
- Vacuum Bag or Vacuum Frame
- Collimated UV Exposure Source
- Rotary paper trimmer
- 21 step Stouffer Sensitivity Guide for Calibration

Comments

[wi10008]

I've just tried printing laser print on gloss photo paper and placing print directly against photo-resist. I left the paper glossy, without applying oils even. Then exposed as normal. This gives much better resolution than transparencies! The laser printer can get a lot more toner down at a hotter temperature on gloss paper. If the toner is directly against the board any scatter of the UV has little effect on the resolution.

[wi10008]

Another option for darkening transparencies is to use a black dry erasable marker. Colour in the entire sheet with the erasable marker and allow to dry. Wipe off again with soft tissue. The ink remains trapped between the dots of the laser print but wipes away cleanly every where else.

[wi10008]

Regarding the double-print method to get a darker mask I've found that a number of websites recommend a "toner darkerner" or "toner opacifier" spray. Some forums say that this is just a clear acrylic sprays such as "Krylon Acrylic Crystal Clear".

Apparently, the laser print is sprayed with the acrylic/solvent and then rolled with a smooth roller. The solvent softens the toner and the pressing spreads the dots of toner out slightly filling in the gaps. The solvent is then allowed to evaporate leaving a very dark print. I intend to try this soon.

Thanks fo the great write-up!

[sleemanj]

Applying the dry film is THE hardest part of the process.  I've tried a number of methods, and here is the best one for me so far, using a laminator.

1. Clean board very clean, as described already here.

2. Prepare a "carrier", I take a piece of thin flat stiff cardboard such as a breakfast cereal box,  I place the board in the middle of the cardboard, and at each of the board (the top and bottom as you would feed it to the laminator) I stick down a popsicle stick (or such like) onto the carrier with double sided tape to form "end stops", to keep the circuit board from moving as the laminator first bites it, imagine [stick][board][stick] - the sticks must be longer than the edge they are butted up against.

3. I take a piece of dry film, cut so that there is about 5mm extra all around (or at least top and bottom).  I place the film curl-side down on top of the board so that it covers the board and the 5mm at each end overlap the edge of the popsicle sticks, then I put a piece of sellotape over one end of the film to stick it to the popsicle stick to make a hinge across the top edge of the film.

4. I take another piece of cardboard (wider than the film), and I tape that to the other end of the film to make a "tail" from which you can hold the film.

5. Fold the film over backwards using the sellotape at the top as a hinge, now stick a piece of tape on a corner of the bottom, and separate the lower protection layer, removing it completely.

6. Holding the cardboard "tail" you attached, lift the film up vertically, and then feed your carrier into the laminator, lowering the film by the "tail" as the laminator joins it to the board.

That's it.  Sounds complicated, but it makes it MUCH easier to get a good clean bond.  Run it back through the laminator a couple of times if you think it's necessary (experiment with your laminator).  Don't have the laminator too hot or the film can be damaged, and likewise too cold might not bond well.

Once you've made the carrier (all of 2 minutes) you can of course re-use it for the same length of board.

Leave the bonded board in the dark for a good 10 or 15 minutes before exposing it, to allow it to cool and the bond to strengthen, the UV polymer is quite soft when warm (which is why it needs to be warmed, so that it conforms to the board's microscopic contours).

[sleemanj]

I'd like to add, that in my recent experiments with dry film, I didn't use Sodium Hydroxide to strip it, I just dropped the board in the Sodium Carbonate solution which I had used to "develop" it, and when I came back in about an hour or so, all the resist had simply floated off the board.

I expect that Sodium Hydroxide would work pretty much instantly, but if you can wait half an hour, there seems little point messing with that. Sodium Carbonate (Washing Soda) is perhaps more readily available than Sodium Hydroxide (although both are pretty easy to come by), the stuff I bought was sold as a "natural fabric softener" and is 100% Sodium Carbonate, just look on all the packets of laundry supplies in your supermarket, there's sure to be one hiding there somewhere.

For developing, about 5 minutes is the longest I would leave it in the solution, longer than that and it risks damaging the traces too easily. Don't use a scrubbing pad or anything abrasive, the traces can be damaged, just use a wad of toilet paper to firmly wipe the board after it's soaked a couple of minutes.

The hardest part of this process, is getting the film onto the board evenly, without any trapped bubbles or wrinkles or dust, and then getting it through the laminator without it slipping.

NB: I didn't particularly measure the amount of Sodium Carbonate crystals I dissolved, so your mileage may vary.

[robielyn]

Nice writeup!

I have a question though.  Since this is a "negative" process the black areas on the photomask are going to be copper free areas on the board.  Toner density problems occur in the central portions of large black areas.  However, these are the places where copper is going to be removed and I'm wondering what difference does a bit of "speckling" in the copper free spaces make since there is no chance it will cause shorts.  The clear areas of the negative will be the various footprints and traces.  The spaces between traces will be very narrow lines and the laser printer can deal with that with no difficulty.

It seems like I'm missing the point of why you are going to all the work of making two copies and gluing them together.  Can you help me out with this?

[purpulhaze]

Nice instructable. I got some good pointers on making my own pcbs. I have a quick question. How do you cut your pcbs?

I've not been able to get a really clean straight edge. I use a rotary tool with cut off wheel but have also tried using a hack saw, scoring and tin snips but can't get a clean edge. I've been thinking about purchasing an old heavy duty paper cutting, wet tile saw or 8" metal shear (link below).

What do you use or think?

http://www.northerntool.com/shop/tools/product_200309554_200309554?cm_mmc=Google-pla-_-Metal%20Fabrication-_-Benders-_-143363&ci_sku=143363&ci_gpa=pla&ci_kw={keyword}

[Culturedropout]

Hi. First off, your Instructable was very informative and useful; thanks for all the work that obviously went into producing it. I've been testing some of the same (or very similar, I believe) etch-resist film you have, with the idea of etching a brass plaque after exposing it through a pair of identical laser-printed transparencies (for better contrast). After two or three failed attempts, I found that throwing the whole thing in a zip-lock bag and zipping it shut on the nozzle of my shop-vac worked pretty nicely to keep everything tight together, and after seeing some sort of banding or striping in the exposure when I tried a fluorescent black-light tube about 10" above the target, I ended up just tossing it out the back door into the sunlight for about 10 minutes and that seemed to work pretty well. I may have to experiment more with exposure time, but it's starting to get pretty close.

The point where everything went to heck was when I tried using this technique to salt-water etch the developed brass plaque.  The resist film pretty much boiled off the surface of the brass after a few minutes, seeming to start around the exposed brass areas left after developing, and bubbling out from there.  So...  am I to conclude that this particular film is not suitable for salt-water electro-etching?  Is  there possibly a different kind that might be?  The article on electro-etching uses a CNC-cut vinyl decal as the etch resist, so I'm sure it's a lot thicker than the film I'm using.  Do you think that's the issue, or does the salt water break down this film?  If so, what's the least toxic etchant you might recommend that would work well with both brass and this film?  Any suggestions would be most appreciated.  Thanks!

-Bill

[ffsman]

alternate method used in INDIA:

goto: file menu in Eagle
select: Export
then select: image

A dialog box appears asking where to store the IMAGE of your board
tick: monochrome
(change resolution if you want or leave it to default value )
Select: browse and then save your image file

the image saved is positive image

locate the file
open it using paint
press: ctrl+i ->this will convert positive image to negative image
then save it

[c14nz]

* how much Na2OH3 (sodium carbonate) solution should be diluted in 500ml of water? NaOH (I know that Peak Out containing 30% NaOH) is not good for development or is too strong (I have seno4007 but I don't know if it's good for this type of film, I know it's good for spray Positiv20)

* can you tell me in what products can be found Na2OH3

thanks

[FrozenFire]

Na2OH3 doesn't exist - if you mean sodium carbonate, that's Na2CO3. Washing soda. Chances are your supermarket sells it.

[mrwolfe]

I find that a wax printer (e.g tektronix phaser) or a colour laser (e.g HP laserjet 1600) gives pretty good density. Ordinary black and white laser printers are OK, but you have to turn the density up as far as you can. Even then you still get pinholes.

[GM2009]

thanks
I did my first two runs - step that I thought is easiest (laminating) proved to be most difficult ... for now. I've used UV led box, exposure 3-4 minutes, drain cleaner 1:20 dissolved in water, everything was fine except laminating.

I had a lot of, as you say, blisters. Laminator is same as one on the picture (soverign). I didn't see your reply so I've tried setting 7 and 5 and both were bad.

I am laminating 0.05 brass sheet, not making pcb, so MG method (overlapping) cannot work.

Now I will try to run machine cold and that will probably help. The fact is that film has to adhere to copper perfectly. Any imperfection will ruing etching (dust etc.).

Working with copper clad is probably easier than thin brass sheet but as with everything new practice helps.

[incoherent]

 Oops, I accidentally replied to myself below so be sure to read that first! Forgot to mention one more step also:

Press all the excess material together to completely seal it.  Then trim with scissors to be sure no uncovered material is exposed as this will stick to the laminator rollers.

Finally run it through the laminator on the cold setting two or three times.

Good luck,
incoherent

[incoherent]

 Here's my advice based on countless rounds of laminating/etching:

First put on gloves and eliminate all traces of oil from the substrate by scrubbing with softscrub w/bleach, then dry quickly with an air compressor to avoid oxidation as much as possible.  Do not touch the substrate with bare hands from here out.  I switch from dish gloves to surgical gloves at this point.

Next, cut the photoresist to a sheet large enough to cover the front and fold over to the back (the long way--if not square) with 1/2" extra all around.  peel back just the first half inch of the covering (the dull side on the inside of the curl).  Lay the material over the board, making sure that it is sufficiently aligned to wrap to the back without veering off the board.  Use your thumbs to press the 1/2" exposed section to the board evenly starting at the middle and sliding towards the sides.  If the board is too cold, this won't stick well, but at about 80 degrees F, it will stick quite well.  Unless you're in an igloo, this should be no problem.

Once you've got the first 1/2" adhered, turn the board around so that the adhered portion is facing you and the still-covered film is facing away from you.  Pull the film back one more 1/2" between the board and the free film.  Carefully press the film to the board starting at the middle and working out to the edge being sure to avoid bubbles.  Repeat this process 1/2" at a time until the front of board is covered, then flip the board and do the same on the back side of the board.

Substrate temperature is pivotal--less than ~75 degrees F, and it doesn't stick well enough and bubbles creep back in after you press the film down.  If the temp is ~90 degrees F or more, it will stick too strongly without even pressing, and may stick prematurely and entrap bubbles.

Hope all goes well.  Be sure to post pics--is your UV LED box featured in one of the other instructables?

[incoherent]

Frankly, I've found that it works best cold--any heat at all, and I get a sort of blistering here and there.

[GM2009]

What is laminator setting (temperature)? Thanks.

[UltraMagnus]

that doesn't look like a PCB done in eagle, what program did you use?

[incoherent]

Yes, I did it in Eagle except for converting to negatives--that was done in Inkscape. BTW, which part looks non-Eagle-like?

[UltraMagnus]

I guess the curvyness of the tracks threw me off. I take it you didn't use the autorouter then?

[incoherent]

I can't stand the autorouter--maybe I'm just doing something wrong, but I never seem to get more than a few wires routed before it quits. I'd like to try to make an autorouter myself someday, but that's a big job...

[josheeg]

I used eagle autorouter and kicad with freeroute.
Both if the wires are small enough to make it and the grid the routing is done on is small enough and their is enough room it does it free route may take  a wile but my linux computer laptop was dated.

[pfred2]

I never had a problem with the autorouter in Eagle. I never make too challenging boards though I guess. I do remember one bug with filling lands in the program though. I wonder if it ever got fixed?

[perkinsb1024]

Stay far, far away from autorouter. It will never yield the results that a patient hand and clever eye will. I've never liked the curvy lines though, I use the 45 degree bend lines

[pfred2]

Personally I rather like Eagle's autorouter. Some circuits are affected by sharp cornered traces. I doubt circuits much care what you like or not.

[UltraMagnus]

don't know what problems you people have had with the autorouter, but I have found it works fine once you fiddle with the settings a little.

[pfred2]

The autorouter has worked fine for me. I have done some complex boards, single sided, with hardly any uncompleted traces using it. Anyone having troubles with Eagle's autorouter is not using the program correctly. There are tutorials available that teach how the program works. I have used them myself to learn aspects of the program that I did not understand. Though I have to say, learning how to manipulate the routing rules wasn't one of them!

[jeff-o]

Hmmm, I wonder if you could use galvanic (aka electrolysis) etching for this step...

[drleavsy]

I suggest to use 100-200 UV LED (12V power supply) in the form of array as it was described in http://www.instructables.com/id/How-to-make-a-printed-circuit-board-PCB-using-th/, then there is no need to take care about collimation and about fragile tubes and stuff. I believe the time of exposure can be really short about 1 min as it described in aforementioned source . 

[GM2009]

I have Samsung ML-2510 and printing on transparent foil was not so good (not dense and visible printing lines - toner was quite used). Also tried my office printer (prof. quality). It produced very uniform black (fresh toner) but not dense enough. Two foil would be just perfect. Very good output was produced with HP C5580 InkJet printer. Used 1200dpi resolution, grayscale, Black Ink Only, Ink Volume heaviest. Result is really good, sharp. Double overlapping foil is not needed. Printer is brand new though, I do not know if result will change over the time. Still didn't finish complete process. Etching is not used for PCB (even though I might later) but for Brass etching (model photoecthed parts). I do not know if other InkJets will produce good result to. Somebody else can add more tests. Selection of printers is kind of limited. For example, many Epson models will not print on transparency anymore (printer will not recognize media). Most of HP printers do print on foil. Most of Canon printers have limited B&W resolution of 600x600 dpi. Not good enough for fine work. As for etching I used Peroxide and Muriatic acid for test. When it was fresh it was fine but after one day it seams like it is not strong enough? Etching brass of 0.1-0.3mm thickness is much slower than cooper clad CB and it is bit more complicated since two sided etching is required and sometimes embossed as well.

[JoeyJunior]

Nice job. One question though, why did you use curved lines instead of 45 degree angles?

[incoherent]

Just for fun really. I've been toying with writing an autorouter using a spring & damper model like that used in Graphviz. Curvey lines can be applied to accommodate obstructions, frequency/EM sensitivity increases tension, etc.. That effort is a ways down the list right now, though...

[mikeg103002]

NECESSITY is the mother of invention... not adversity...

[incoherent]

Yeah, publishing at 3AM has its drawbacks! I realized that after I posted it, but I just haven't gotten around to rewriting it.

[john-louis]

incoherent,

Thanks for that, I have had a look and see what you mean about the channeling / frame construction. I have been thinking about the glass vs bag approach and think maybe a double sided frame with glass on one side and film / thick transparrent plastic on the other - kind of like yours but transparrent plastic - may give the best results as one side can stretch and "vacuum form" to the shape of the board.
I will measure up and start playing when I get back from uni in a week (last week of exams =D)
Regards,
Alex

[john-louis]

Hi, thanks for replying!
Alex is indeed my real name, john louis the middle bits =D
I have spent a good while thinking about this frame, I need to do some .5mm pitch surface mount stuff soon so was planning a complete UV box rebuild - the current one is a classic scanner enclosure (using lid for pressure) and philips solarium lamp job. I plan to use an atachie case with a raising lid (on collapsable telescopic style stilts) with the UV lights in the bottom of the case, along with vacuum pump and ballists and (possibly) a collimator - if I can figure out what is really needed for one.

A removable double sided glass vacuum frame for doing double sided work near the lid / stored in the lid - but slidable up and down the stilts.
I was thinking of getting two really simple picture frames and milling a trough (moat) around the outside of the glass and then putting a thin rubber seal around the outside of the trough, then threading a fitting into the trough from the outside, allowing the glass surfaces to almost touch when vacuumed together, hence working kind of like a flower press on the PCB. My only worries is that the glass may crack or that because neither of the surfaces have any "give" the requires pressure will not have the desired effect...
From your description I assume the U section channel is on the top bit (with the plastic sheeting) and the pipe tapped into the U channel?

Cheers,
Alex

[incoherent]

Alex,

Here are some pics of my vacuum frame. Incidentally, I scanned a pre-registered, double-sided transparency while working on the PCB instructable and found the scan made it appear to be very poorly registered (see 4th pic in the series). I think this was due to the very-wide-angle nature of the scanner optics as well as the the separation of the layers despite the lid being closed. The effect could have been caused by total internal reflection even if the layers were tightly pressed together, but I never tried it again with a vacuum bag/frame to diagnose the problem. The point is that the scanner lid may not press as effectively as you might think. Oh, yeah (ADD moment here), you might be able to make a kind of hybrid vacuum frame/bag by placing a transparent plastic film between the glass and the frame.

Regards,
incoherent

[incoherent]

I forgot to mention that you can try out the vacuum bag approach too. This may be even better than a vacuum frame in that the flexible bag presses the artwork to the board instead of non-conforming glass (especially good for doing solder mask where you've already got an uneven surface on the board). The vacuum bag presses so tightly (and conformingly), that if you hold it up to the light, you can see little bumps where little particles of dust are trapped between the board and the bag! Just get a large ziplock-style bag (the kind without the little plastic slider). Next get a spool of thread, non-silicone adhesive/caulk and some tubing (fishtank tubing seems to work well). Working from the inside, poke the tubing through a corner of the bag oposite the zipper and pull all but 2" through the hole. The bag should be big enough to lay flat on the exposure unit without the zipper or tubing under the glass. Apply some caulk inside the bag around the tubing in the corner, then use the thread to tightly bind the bag to eliminate leaks. Next place your PCB/artwork sandwich in the bag and zip it up, then vacuum out the air (leave the vacuum running throughout the exposure process to keep it tight). You can also use a vacuum bag such as this for doing fiberglass or carbon fiber composites to tightly squeeze the layers together and squeeze out the excess polymer--just be sure the vacuum won't suck it up!

[incoherent]

john-louis/Alex (if that's your real name:), I've been working on a video instructable for the vacuum frame, but, of course, that means a whole new process. I'll summarize though: Go to the local home improvement shop and ask about window screens. They should have extruded aluminum channels designed making window screens. They are about 3/16" x 1/2" (5mm x 12mm) with a u shaped channel along the inside. You'll also need the plastic corner pieces and the rubber strips used to mount the screen to the frame. You'll also need ~4mil plastic sheeting, foam tape (the kind used for weather stripping), and non-silicone adhesive/caulk. I think that the stuff I used is acrylic-based, but they don't say for sure--it smells like acrylic, though. Cut the channels about 1/4" smaller than your glass sheet on both sides and top/bottom. I drilled holes along the inner edge all around and plumbed brass tubing through the corner pieces to ensure that the vacuum would flow evenly with trapping pockets of air (although this step might be overkill). Run foam tape around the bottom (just outside of the holes (if you made them). Finally liberally coat the foam tape and all joints with the acrylic caulk (no silicone, because silicone doesn't stick to silicone--go figure...). You'll need to make multiple layers (and keep them kind of thin so that it will set up. Use a crayon to coat the glass sheet with wax and place the frame on the glass after applying the caulk--this will ensure a smooth, flat surface for your seal. Keep applying adhesive until you stop all the leaks. I hacked a fishtank air pump to to draw a vacuum on one side and pump on the other. just figure out where the air intake is and try to hack it to feed from your vacuum tubing. I used part of an old pill bottle to make a sort of plenum which I sealed over the air intake on the bottom of the pump. I'll try to get the instructable together just as soon as I figure out how to dub the video with narration. Good luck, incoherent

[incoherent]

BTW, I put a gauge on it and it is drawing about 5 kPa (even though it still has a few leaks). This works out to about .7PSI, or about 75lbs of force pressing the 9x12 inch frame to the glass, of which about 8lbs is evenly distributed over the 3x4 inch PCB.