HOWTO: Create your own PCBs on the cheap!

[MWP]

Originally Posted by tin

That's called the expensive way.

Nah, not so expensive for double-sided and accurate boards. They do take a while to make however.
The BJ transparency sheets are $1ea, and the presensitized PCBs can be found cheaply if you look around.

Two examples of double-sided boards ive made:




While etching the second one:


Protel has a "mirror" option in the print config.

[dakiller]

Originally Posted by MWP

Nah, not so expensive for double-sided and accurate boards.

Just wondering how you line up the 2 layers and how accurate you get it? I don't like the results I get with toner transfers

[MWP]

Originally Posted by dakiller

Just wondering how you line up the 2 layers and how accurate you get it? I don't like the results I get with toner transfers

Ok, this is what i do:
1. print the transparencies
2. cut them out leaving a good inch on each side
3. line the two transparencies layers on top of each other, then tape them together on 3 sides.
4. slip the PCB in the middle, and use a little tape to hold the transparencies in place
5. expose on one side under a fluro desklamp for 12mins, then flip it over and expose the other side.
6. develop
7. clean then etch
8. laquer
9. drill
easy

I get them accurate enough for very small vias which i make by soldering through a strand of wire.
You need a very sharp drill though to not push the pads off the other side of the PCB.

I think hand soldering all the SM ICs (esp new fine pitch ICs & 0603 resitors) takes more practice than producing the PCBs.

[dakiller]

Originally Posted by MWP

Ok, this is what i do:
1. print the transparencies
2. cut them out leaving a good inch on each side
3. line the two transparencies layers on top of each other, then tape them together on 3 sides.
4. slip the PCB in the middle, and use a little tape to hold the transparencies in place
5. expose on one side under a fluro desklamp for 12mins, then flip it over and expose the other side.
6. develop
7. clean then etch
8. laquer
9. drill
easy

I get them accurate enough for very small vias which i make by soldering through a strand of wire.
You need a very sharp drill though to not push the pads off the other side of the PCB.

I think hand soldering all the SM ICs (esp new fine pitch ICs & 0603 resitors) takes more practice than producing the PCBs.

Well that sounds real easy, next time I'll get the stuff needed to do photo resist.

I haven't and any problems in drilling the holes, I got a big drill press and put the board in the vice and it is very quick to drill 'em, they just center themselves after you've etched the boards. And soldering, I can solder SSOP chips with my eyes closed just about and I try and use 1206 throughout and only resort to 0805 when needed

[Goth]

Originally Posted by Stinger

Alright I can never get this method to work. Toner never sticks to board. I don't know for sure, but I suspect my iron is not hot enough. Could someone who does this successfully measure the temp of their iron (hopefully all you nerds have DMMs with thermocouple probes like me )

I've tried a few different printers/photocopiers so I dont think thats it and I've bought a few different types of high quality paper and tried transparencies too. I've roughed up the board with nylon pads and cleaned them very thoroughly with highish purity acetone.

The toner sticks initially, but while removing the paper chunks tend to fall off and other people seem to mention that you should be able to scrub quite hard.

I'm in the process of setting up the fuser roller from a laser printer as a motorized heating press. I just have to get around to making an AC chopper so I can actually apply a sensible amount of power to the fuser element. Connecting it directly up to mains makes the inside coating of the roller smoke

Get the paper nice and hot, and evenly and thoroughly iron it.


Then put it in a container of cold water and soak it, for several minutes at least. Don't try and remove the paper before it's thoroughly soaked.

You should use glossy inkjet photo printing paper.

After you've soaked for a while, the paper should start to soften and come apart - if you pick up the board, the paper will be soft, and it will come apart in flat layers - you'll be able to peel off the back-most layer(s), with the layers closest to the toner still there.

You might be able to peel off a little more - just white paper, without any toner in contact with it - then use your fingers to gently rub back the paper pulp on the layers closest to where the toner is, eventually leaving just the toner layer and a thin layer of paper stuff over the top of that.

I like this method of PCB fabrication, because it's really cheap, although it can't do uber fine traces.

Photolithography is more expensive, but makes it much easier to do double sided boards, or reproduce the same board over and over.

Press-n-Peel film is no better than using photo paper, in my opinion, once you start getting the hang of it, and it's heaps more expensive.

Originally Posted by dakiller

Well that sounds real easy, next time I'll get the stuff needed to do photo resist.

I haven't and any problems in drilling the holes, I got a big drill press and put the board in the vice and it is very quick to drill 'em, they just center themselves after you've etched the boards. And soldering, I can solder SSOP chips with my eyes closed just about and I try and use 1206 throughout and only resort to 0805 when needed

MWP -

What kind of photoresist coated boards do you use - positive or negative? What brand?

Where do you get the boards - and do you use fibreglass or phenolic?

It's the board stock that you're using that is the most expensive part of photolithographic processes - and the specific photoresist will determine how easily it can be exposed with ordinary fluroescent lights - a big factor in determining how practical it is to do.

Cheers.

[TRG.dOinK]

I learnt this stuff at TAFE, my friend left tafe and went to uni, and is doing java programming and some silly maths problems

Tafe ftw. We even used ISPLever i think it was called, and wrote ABEL/VHDL to program our Lattice ispLSI 1016e FPGA (fairly sure its a FPGA)

We used Protel to design our circuit boards, we then printed it off onto paper. You had to mirror the output so that writing/layout came out the right way.

Then we made up sheets used for overhead projectors. We would then cut it out, and lay it on top of the copper board, and put it into this light box for a certain time. This was done so you get your tracks and remove the top protective light sensitive layer (Im fairly sure).

Once down, basically the same steps as described in this post, it goes into the chemical processes (2 of them), one to develop, then you clean it, then it goes into another chemical to be etched which eats away the bare copper, then you wash it with water again and dry it, drill the holes and then spray laycor (spelling?) over the tracks to protect it from corroding which is an important step.

I might have to fiddle some day and make something i think, before i forget everything

Maybe I should make my class ab amplifier onto a made circuit board and make it stereo instead of mono. Hrm.

[MWP]

Originally Posted by Goth

What kind of photoresist coated boards do you use - positive or negative? What brand?

Where do you get the boards - and do you use fibreglass or phenolic?

It's the board stock that you're using that is the most expensive part of photolithographic processes - and the specific photoresist will determine how easily it can be exposed with ordinary fluroescent lights - a big factor in determining how practical it is to do.

The PCBs i use are called "Kinsten positive acting presensitized PCB".
The PCB type is FR4 glass-epoxy, but they also come in CEM3 glass-epoxy and phenolic in different thicknesses and copper weights.

I buy them from a company here in Adelaide that i cant remember the name of (ive had enough to last me a few years).
A quick search found this Aus company that sells Kinsten too http://computronics.com.au/kinsten/prices/ (decent prices too!)

[Goth]

Yeah, the Kinsten postitive stuff works really well, and is cheap.

Computronics has a minimum $20 postage or something, though.

I tried Riston negative, which is more expensive, but Jaycar has it, and found that it's not as easy to expose.

I tried DSE's spray on positive resist once, didn't work too well.

[lahmyaj]

great to see other ppl making their own pcb boards! Geez i haven't made one for a couple of years, been moving house too much to settle into it!

But another major problem of mine is getting my hands on a FREE schematic/pcb design program! Protel is like millions of dollars! lol so what do u use?

I'd like something that runs in XP, none of this dos crap or sumthan and i have to be able to make my own custom chips etc. There was a program i used to have that i was making layouts for 16x2 lcds etc. whatever i needed to add to my board pretty much.

cheers! etch on!

[MWP]

Originally Posted by lahmyaj

But another major problem of mine is getting my hands on a FREE schematic/pcb design program! Protel is like millions of dollars! lol so what do u use?

I use Protel

[Stinger]

I usually soak for at least 30 mins - hour and it doesn't seem to help much. I've tried some high quality inkjet printing paper - 80-90 gsm ~ $1 per sheet as well as some really shiny stuff. The really shiny paper (Celcast premium 190 gsm) didnt work too well as it has a plastic coating which doesnt break up and on my last run adhered itself so firmly to the copper that the board had to be chucked out.

One point is that the design I am using is rather sparse; might mean that the paper lifts at the edges of the tracks more readily, but the way that the toner comes off gives me the impression that a denser design wouldnt help that much/enough.

[GooSE]

Originally Posted by Stinger

I usually soak for at least 30 mins - hour and it doesn't seem to help much. I've tried some high quality inkjet printing paper - 80-90 gsm ~ $1 per sheet as well as some really shiny stuff. The really shiny paper (Celcast premium 190 gsm) didnt work too well as it has a plastic coating which doesnt break up and on my last run adhered itself so firmly to the copper that the board had to be chucked out.

One point is that the design I am using is rather sparse; might mean that the paper lifts at the edges of the tracks more readily, but the way that the toner comes off gives me the impression that a denser design wouldnt help that much/enough.

I guess maybe your iron isn't hot enough then?

Although it good be that your paper is too good... I tried Kodak Ultra Glossy 220 gsm once and it was rubbish. The OfficeOne paper that I have used successfully was $14 for 50 sheets.

[nux]

Originally Posted by lahmyaj

But another major problem of mine is getting my hands on a FREE schematic/pcb design program! Protel is like millions of dollars! lol so what do u use?

Eagle has a free version with only a few limitations.

[Goth]

Originally Posted by GooSE

I guess maybe your iron isn't hot enough then?

Although it good be that your paper is too good... I tried Kodak Ultra Glossy 220 gsm once and it was rubbish. The OfficeOne paper that I have used successfully was $14 for 50 sheets.

I use Kodak and it works fine, i'm not sure specifically what kind though.

I'll find out.

I couldn't find any cheaper stuff, and it's still by far the cheapest and easiest PCB manufacturing - for single sided boards without really really fine artwork - method i've used.

[Stinger]

Still, if anyone has a DMM with a temp probe and can whack it on an iron they have used to do toner transfer successfully (at the setting they use) I would be very appreciative.

Yeah I reckon we should dodgy up a page in the wiki for this (if someone hasn't already), with a list of paper types that are shown to have worked. I've seen a lot of sites on the net describing toner transfer, but most of them seem to be geared at American audiences, with their products listed (Staples paper is mentioned a lot). It'd be good to see some specific paper types/brands that work and approximate prices.