Homemade PCB's. Share your methods / General Science and Electronics / Forums

Forums

4hv.org :: Forums :: General Science and Electronics

« Previous topic | Next topic »

Homemade PCB's. Share your methods

1 2 3 next

Move Thread

LAN_403

Physics Junkie

Fri Jan 18 2013, 04:37AM

Registered Member #7267 Joined: Tue Oct 16 2012, 12:16AM
Location: Detroit, Michigan
Posts: 407

Hi Ladies n' Gents! Recently I've been testing a few methods of pcb design and etchings. I'm trying to get the hang of it and find a good method. There is endless info out on the intraweb regarding this topic, but I've only read very few threads about it here on the forum. So I thought it might be interesting to hear from others. Please share all your details, methods, experiences, anything you want... everything, the type of chemicals you use, how much you dilute them, how do you dispose of them, precautions to take? Do you prefer to iron on the toner or use pre sensitized PCB? What type of printer paper do you use, temperature setting on the iron? Type of exposure lights do you prefer, including wavelength and how long you expose the board for. What software do you use to design? Do you tin, flux, soldermask your board? Any and all details you can share would be awesome

I have tried iron on and positive exposure, but only test strips, not an actual board yet.
Iron on details
I used eagle PCB freeware to mess around. I really like it because you can print in black and white, mirror your image, and export it to photoshop. My first design is going to be a challenge because it is double sided. For this i included alignment marks for top and bottom layers. I printed out the bottom layer and mirrored the top layer at max dpi setting. I dont have a laser printer, so I took my design to kinkos to get printed properly. I got several copies printed on transparencies and gloss paper. They were less than 75cents a print. If you look at the toner under a microscope, it is applied in tiny dots. This can mess up your design, So it is a good idea to get extra copies. Using the alignment marks, I carefully try to align my design on the board. I dont have a solid method for this yet. But i found that drilling a hole in your paper and board to align with helps a ton. Make sure your copper is shiny clean. I used scotchbrite pads to clean my board with. Once my design is aligned (toner side down of course) I lay a piece of plain paper between the design and my iron, to make for smooth ironing. I use the synth/silk setting on my iron, but this proved to not work well as the toner ended up smudging and merging traces. My temp setting was clearly wrong. Also i found that letting the iron sit still on the design worked better then moving back and forth, as that contributed to the smudging. It only took about two minutes of ironing on each side. Cooled the board in water, let it soak, pulled the design off. I found that some traces did not apply at all, due to temperature being wrong. For the etching, it seems that ferric chloride is the acid of choice. I purchased a 1L bottle by MG Chemicals from amazon for 10 bucks, free shipping for prime members. This part doesnt require much explaining. Poured FeCl into plastic dish, put the PCB in, agitate, rinse with water, done. All in all it worked just fine. Finding a good method to perfect the ironing process is what I need to do, as well as try different papers.

Positive pre sensitized PCB
This seemed to be the more ideal way to go, but was more tricky. Far more tricky. I bought a few MG Chemicals double sided 6x6 1/16 thick pre sensitized board to try (also from amazon $10bucks a board free shipping). Also MG Chemicals positive developer 500ml $8. What I was most unsure about was what lights to buy for exposure, more on this follows. This is what I used the transparencies for. Again, since the laser printer applies toner in dots, I doubled up on the transparencies and went over thin traces with a fine point sharpie. I got 8"x10" glass sheet 1/8" thickness from my local hardware shop for 2$ a piece for exposure purposes. Luckily having a photographer in the family helps with this process. So i had a patterson safelight available for use when opening up the packaged pcb. Using alignment marks, i made sure my design was perfectly aligned and taped it to the glass pieces toner side facing the pcb. Clamped together using a small vacuum vice, perfect for the job. Now this is where things got tricky. First I read that many people have success using common 13W CFL bulb because they emit small amounts of UVA (315-400nm). So i tried it exposed for 8 minutes. Using 1 part developer to 10parts water, I developed the board.... well not really. I either over exposed it, CFL's werent good enough, or made a miscalculation in my developer and it was over diluted with water. Even though I could see the developed pattern, absolutely no resist came off during development, which led the FeCl having no effect. I tried re developing, adding a little more developer, and then all of the resist slid off like jelly lol. So I did more reading, and I'm going to try exposing it to regular "party UV black lights". Wavelength is 365nm and I've read that they work well because they have that phospur filter to block the visible light emitted by mercury vapor in the tube. MG chemicals recommends using a 325nm light and exposing for no less than 5 minutes for their products. So using the black light, i will try again. What is your take on the kind of lights to use? I will take photos and post them later. I hope this helps anybody trying homemade PCB's. Please comment I would love the input, thanks all.

mikeselectricstuff

Fri Jan 18 2013, 09:29AM

Registered Member #311 Joined: Sun Mar 12 2006, 08:28PM
Location:
Posts: 253

Check out my page

The optimum tubes are the type used for insect killers - available cheaply from a good electrical supplier.
For artwork, laser print n heavy tracing paper.

Daedronus

Fri Jan 18 2013, 02:06PM

Registered Member #2329 Joined: Tue Sept 01 2009, 08:25AM
Location:
Posts: 370

UV leds work great, if you use 400 of them.

This is what I use, a PCB with about 400(and something) of them, in groups of 7 + one resistor in series all in parallel, powered directly with a 24V power supply

My exposure time on negative dry film laminate is about 20 seconds.

About the UV mask, a few lessons I learned the hard way:
1. laser printer toner is a good UV blocker, but
2. laser printers will wrap/deform the mask, also the black parts tend to have pinholes with un adhered toner.
3. inkjet printers work great for masks but, and this is a big but, dye based ink is not good at blocking UV, use only pigment ink based printers, or replace the original ink with one that is pigment based if the printer allows it. Epson piezzo print head can work with any kind of paint.

I had good results with INK Master EP230BK pigment ink in my Epson L100 printer. With the original dye ink I had a lot of exposed mask residues in the masked parts.

4. 1200/1440 dpi printers (never mind the other values quoted in specs) are actually 600/720 true dpi. This means a 720dpi picture will print pixel for dot on a 1440 dpi printer.
1440 inkjet printers will simply scale the picture to 720dpi discarding, yes discarding, no bi-linear no nothing, just nearest neighbor scale, every other pixel. 3/4 of a 1440 dpi picture is simply discarded.
5. more importantly then the dpi for inkjet printers is the drop size, usually given in picoliters. 4 pL will barely do 4/4 mils and it's good for 6/6mils

Linas

Fri Jan 18 2013, 02:14PM

Registered Member #1143 Joined: Sun Nov 25 2007, 04:55PM
Location: Vilnius, Lithuania
Posts: 721

best i could do at home with UV technology, please some one beat it

ConKbot of Doom

Fri Jan 18 2013, 05:46PM

Registered Member #509 Joined: Sat Feb 10 2007, 07:02AM
Location:
Posts: 329

1)make up my gerber,
2) open it up in a gerber viewer to make sure I'm not hosing anything up
3) send it to oshpark (or batch pcb) pay <$50
4) wait ~3 weeks and have my double sided, though hole plated, routed and gold plated PCB

worth it IMO because of the combination of time saved, additional features (PTH and solder mask), and not worrying about something goofing up and ruining an hour or two of labor + supplies.

Daedronus

Fri Jan 18 2013, 06:17PM

Registered Member #2329 Joined: Tue Sept 01 2009, 08:25AM
Location:
Posts: 370

try that with a 8 layer board :)
ahh, yes, diy-ers can't do that.....yet....

Linas

Fri Jan 18 2013, 07:18PM

Registered Member #1143 Joined: Sun Nov 25 2007, 04:55PM
Location: Vilnius, Lithuania
Posts: 721

i don't know how about you, but i can make pcb for 99% of my projects, that includes 2 layer with 0.1mm pitch, and if some on can't make layout that fit's to 2 layer board, and need lot of vias and even 4 layer, that means they don't know how to design stuff.

only exception pcb for BGA and similar packages

if i could get pcb made in week time with solder mask and reasonable cost, maybe i would use that service, but in my country only few companies can do that, and cost is astronomical (>100$ just to start making, but if you order like 1000x then is reasonable cost per board , and for ordering from china it will take at least 4 weeks

)

Daedronus

Fri Jan 18 2013, 07:33PM

Registered Member #2329 Joined: Tue Sept 01 2009, 08:25AM
Location:
Posts: 370

As a general rule you need a signal layer for 1 or in the best case 2 raws of bga balls

but my point was that ordering single unit PCBs online is not that great on the price, and if you want to use BGAs you are left high and dry (because the price becomes prohibitive)....not the other way around.

mikeselectricstuff

Fri Jan 18 2013, 10:39PM

Registered Member #311 Joined: Sun Mar 12 2006, 08:28PM
Location:
Posts: 253

wrote ...
2. laser printers will wrap/deform the mask, also the black parts tend to have pinholes with un adhered toner.

Tracing paper gives good adhesion. The thick stuff (90+ GSM) doesn't warp much.
Quality is good enough for 0.5mm pitch QFPs

Ben Solon

Sat Jan 19 2013, 12:10AM

Registered Member #3900 Joined: Thu May 19 2011, 08:28PM
Location:
Posts: 600

I use a mechanical etching method. I just throw a .127mm tip diameter v tip onto my cnc machine any get my avatar picture. I've done down to 16 mills with perfect accuracy over two layers, and even down to 10 mills to break out a qfn44 package with mixed results. with that trace width the pcb has to be so absolutely flat so that the angle of the v and the bit depth don't cause the trace width to vary. i'm in the process of making a universal pcb jig and topping off the surface with a end mill in order to get down to 8 mills.

1 2 3 next

Moderator(s): Chris Russell, Noelle, Alex, Tesladownunder, Dave Marshall, Dave Billington, Bjørn, Steve Conner, Wolfram, Kizmo, Mads Barnkob