DIY PCBs-- worth the trouble?

[Veraxis]

Hello all,

I am still relatively new to the world of electronics, but lately I have felt that I am finally beginning to reach a point in my learning where I am able to build circuits with a reasonable degree of complexity, and, on more than one occasion, the thought of making customs PCBs for my projects has crossed my mind. For the most part, I have gotten by fairly well using ordinary perforated proto-board, but it seems to me that making my own PCBs would make it much easier to put together a circuit once a layout has been designed (not constantly referring to the schematic to figure out where to place a particular component), and would generally make everything look a lot cleaner and more compact. Naturally, I have looked at dozens of guides on the internet discussing various methods of making home-made PCBs. However, I have a number of reservations related to each of the methods I have read about, and I thought I would ask other members about their experiences with various techniques, as well as with the process as a whole.


With that long-winded introduction out of the way, I’ll get straight to the point about my concerns:

Ferric-chloride etching:
What is the relative failure rate of this process? Can failed etchings be minimized as long as one uses proper techniques?

With regards to buying special printers, toner, and photo paper, how high are the costs related to this, even after one has made the initial investment in a printer? Does using paper from magazines instead of photo paper greatly increase the failure rate or transfer unwanted marks to the board?

what are the relative advantages or disadvantages of using UV photgraphic transfer methods relative to heat transfer methods?

Apparently “reverse electroplating” can be used to speed up/ increase the efficiency of the process. Has anyone here tried this?


Copper chloride etching:
Is copper chloride etching significantly slower than ferric chloride etching?

If you constantly have to “regenerate” the solution with HCl, is it actually any cheaper than using ferric chloride?

Supposedly the standard regeneration method of mixing in HCl and hydrogen peroxide can create toxic fumes. Can this be easily avoided?

I saw a post somewhere about using copper acetate instead of copper chloride, and using vinegar to regenerate the solution. Has anyone ever heard of someone trying this?


CNC router milling:
How time-consuming is the process (programming and running the CNC) relative to chemical methods(transferring and etching)? For example, roughly how long would it take to mill a board that is, say, 4”x4” (about 10x10 cm)?

Supposedly using machines to mill PCBs is not quite as precise as using chemical methods. How severe are the limitations? Is it still possible to get at least a reasonable degree of precision using this method?

Apparently even the smallest CNCs available for PCB milling are often unjustifiably expensive even for more serious hobbyists. Is it possible to make a home-made CNC without having to program everything from scratch to work with your design?

In general:
Is there any truly practical way for hobbyists to make PCBs effectively on a small scale? Or is this really something that is only justifiable for large numbers of PCBs and/or highly intricate designs?

I understand that I am asking a lot of questions here, and I do not expect the members of this forum to be able to answer every one of them, but I think that it would be useful get a general sense of how practical this actually is before I spend my money on something that is not justifiable for the benefits it gives. Perhaps if we are able to gather together enough information on the topic, this could be used to make some kind of FAQ on PCB making?

Thank you all for taking the time to help me out,
JLNY

 

[HarveyH42]

Wow! Lots of considerations... Since you are just getting started, would suggest the simplest, cheapest method first, and see how it works out, and where you would like to see some improvement. You are going to get various views, some argument and debate, and suggestions that might not fit your finances or experience. I'll put my two cents in the pot early...

I use toner transfer, and have brought it down the quickest, cheapest, and most reliable. I use $100 Hp laser printer, Pulsar transfer paper(10/$15), an Ibico GL-4 Laminator($20 surplus), paper thin 1/2 ounce PCB(surplus), paper cutter, Acetone, Ferric Chloride, and a drill press. I use the laser printer for normal uses as well, also the paper cutter, laminator, and drill press, so not excessive purchases, all well worthwhile investments. I have printed a layout, transfer, etched, drilled, stuffed, and tested a few boards, even on with two ICs and discretes, in an hour or less. Kind of a rush job, usually take about 2 hours, and use more care.

You can mess around with magazine paper and a clothes iron, but you have to learn it by trial and error (many), which you simply clean the board, and start again. The temperature of the iron, and the quality of toner, time and pressure, right paper... It's a lot of variables to work out, if you got a lot of time and patience. Not me though, I just want to get the job done quickly, first try. The Pulsar paper works great, and isn't too expensive when cut to the size you need, instead of printing the whole sheet. The hot laminator I got surplus has worked great, unmodified. It only does a max of 4" wide, but I seldom need more than 2-3 inches. The light weight PCB stock cuts easy on the paper cutter, etches quickly, and drills without effort. A drill press will save you a bundle on tiny, brittle, tiny bits, that can break if you drop them on a hard floor. Not to mention, that it's much easier, when you have a couple dozen holes to precisely drill.

The only thing I use perfboard for these days, is for led arrays, or simple things that aren't really going to benefit from an etched board. I hate drilling holes... An etched board is so much easier to repair, or change a part on, than perfboard, which I tend to use the component leads instead of wires.

Anyway, this is how I do my boards, and feel it doesn't get much easier or cheaper. It works well for me, quick and reliable. I did start off with various common papers, a clothes iron, a dremel, all got the job done, eventually, but took a lot of time and re-work, to get less than thrilling results. The only place where dumb-luck came to play, was the Pulsar paper, and the laminator both arrived on the same day, so never really tried other papers with it. The Pulsar paper releases clean, no pulp to scrub and rub, not to mention only takes seconds to get it to let go, rather than soaking for minutes, and careful peeling.

 

[Dr_Doggy]

HI, when i do a board I use a one without copper etching, it just has 1mm holes spacing, great for mounting, then instead of etching I do a stich type thing with telephone wire i got(fits the hole perfect!), it gets a little messy at times, but much simpler than breaking out the chems.

IDK wats its called but its like a mini peg board!

 

[Lake9398]

It's called a perf board and the spacing is 0.1". It is point to point soldering and it is good for a quick prototype, but in the long run PCB's are much much better.

 

[Joe G]

i made my first so-called pcbs out of copper wire that i bent and covered in fiberglass resin, not great but my hotwire power supply is still working after 5+years, it don't
look great, but.... i started making my first real pcb w/ some paint and a toothpick, I penciled it on the bare board and then dabbled the paint to fill the traces in.
I've since been using a laser printer w/ peroxide and muriatic acid. the printer was needed for the family's use anyway and the best thing about the laser print is if the
kids get there homework wet the ink dosen't run.

 

[cachehiker]

Single sided boards? I'd say yes it's worth it if you do more than a few boards each year.

I've never used anything but Ferric Chloride. I have reasonable luck (>80%) with 30 mil traces, 1 oz. copper, toner transfer using Canon photo paper in a Laserjet 6P, and a 60 year old boat anchor of a clothes iron turned all the way up. I've never tried 1/2 oz. copper as I have more 1 oz. copper clad laminate than I think I'll ever use up. I etch the boards in a vertical tank with an aquarium heater and an aquarium pump blowing bubbles at both ends to circulate the etchant. The etching gets done in the center of the garage. There was enough vapor produced to slightly discolor a few square feet of linoleum in the basement but if you don't know it's there, it's not easy to see. It also made the house smell for a few days.

What I do now is primarily power circuits and things like LED arrays, always single sided and never more than 10-15 parts on the board. I did a few non-PTH, double sided boards in the past but the registration process (getting the pads on each side to line up) was a pain given my equipment so I prefer to send those out for fabrication. I can't justify the time and money to set up a better registration process for just one or two boards each year.

 

[bryan1]

I went and bought a heap of UV boards cheap but my laser printer wont print black enough images so I need to print 2 copies then lay them over each other precisely which is a pain. with the last boards I did the image didnt go on clean and I looked at a brochure I had, so I printed on that. Ironed it on and when the pcb image was clearly seen on the ironing side I knew it was enough. Just simply soak the board in water for 10 minutes then the paper will come off easily. For a first time go this method worked purfect and I am tempted just to expose all my uv boards for clean copper then stay with this method. I do have a head unit of of a laser printer which would be a purfect laminator but thats another job as the iron is so easy.

Cheers Bryan

 

[3v0]

The cost of TTF (toner transfer paper) is $15 a pack. That works out to 10 8.5x11 inch sheets or 935 sq inches for $15. 1500/935 id 1.6 cents pre square inch. The green TRF (toner reactive foil) add about another cent. In round number about 2.5 cents per square inch. But you need to cut the TTP a bit larger then the PCB so maybe round it up to 3 cents. In short it is not as cheap as free but still cheap.

There has been talk about a paper product with glue that is said to work similar to TTP but I did not see the product name mentioned.

TTP allows the paper to float away from the toner without any embedded fibers. Kudos to anyone who has good results with an iron. I prefer a laminator and most people do once they use one.

I have been etching with sodium persulphate for a while. It is less of a mess then ferric chloride but it seems to be more prone to pitting traces. The TRF helps with this to some degree. I may go back to ferric chloride now that I can etch in the shop instead of the house. With the TRF over the toner you can etch a board in just a few minutes with a sponge soaked with chloride. You wipe away the unmasked copper. You need very good venting.

I drill blank copper clad PCB stock with a CNC. Then I apply toner TRF and etch. To make DS boards I mask one side with box tape while working on the other.

 

[Wond3rboy]

Hi, bet the cheapest boards are made by me( )

A sheet of glace paper costing 6 cents does 6 boards, ferric chloride comes at 1$ a kilo(thats the way they sell it here). Printing on the paper is a pain though(gotta tape it to a normal paper as it gets stuck). Copper boards are available at 82 cents for a 6x6inch single sided piece, use an iron for the transfer(again a pain), etchings a messy business and i use a mini drill(those adapter ones). Single sided boards are easy to do, have done a few double sided boards this way too with the tape technique as written by 3v0(came out as a decent board).

In the end i would say with all the shortcomings... atleast i am getting boards made.

Also, 3v0, using a sponge for the etching, do you mean swiping the board with a ferric chloride dipped sponge? That would make etching quite tidy

PS: Getting the right kind of paper was a hell of a task.

 

[KeepItSimpleStupid]

Bryan:

If you can find some polyester paper, it works really well for exposure. I also use water to adhere the mask to the board during exposure.

 

[3v0]

Also, 3v0, using a sponge for the etching, do you mean swiping the board with a ferric chloride dipped sponge? That would make etching quite tidy

Exactly. Frank at pulsarPro came up with the idea as far as I know. But I recall that it may not work without the TRF foil over the toner.

I would like to know the name of the gummed/sticky paper that someone has found. It was mentioned in the thread about making you own quick release paper but I do not recall that the name was given.

 

[KeepItSimpleStupid]

I haven't used it, but it's been suggested on other forums that Disolvo paper works well. It disolves in water.

 

[bryan1]

I brought home about 12 pages of a bearing cattledog that is nice and heavy glossy paper which I reckon will be purfect for my next boards. As I only have double sided boards without the UV I'll be putting some UV in the developer to get a single sided board for my next project. I'm currently working on a 12f683 and ucn5408B for the drive for this laser toner setting head I ripped out of a Digital laser printer. It will be interesting to see whether this homebrew laminator will transfer glossy paper and if it does workout I'll try and organize a swap of my UV boards for uncoated boards. As my current laser printer won't do a dark enough black on transparencies but does a purfect black on this glossy paper I reckon it will be the way to go.

Cheers Bryan

 

[KeepItSimpleStupid]

Bryan:

Translucent Polyester paper works really well. It's translucent enough for exposure and prints dark. A drop of water adheres it to the board during exposure. I need to find a source. I found some lying around at work and it exceeded my expectations.

 

[DirtyLude]

Tried to make a nice video this weekend on my method for doing via's but it's all blurry. Need a macro lens for my poor little video camera.

 

[MRupp]

I use a blue paper to print my design to and than transfer it to the boards. Its called Press-N-Peel.

 

[Boncuk]

Hi JLNY,

whatever method you use to make PCBs they all require close watch from beginning to the final product.

I started out with printing a mirrored image, then poke holes through the paper print onto the copper clad (for component location) and used a fine brush and asphalt laquer for the pads and traces.

Painting a 2X2" board took me about 4 to 6 hours. Etching was no problem using ferric chloride while checking the progress regularly and keeping the etchant moving to wash off etched copper to increase etching speed.

Next to try was UV sensitive PCB material and rub-on traces and pads on clear foil. That method was a bit cleaner (no free hand painted traces and pads) but very time consuming as well as the asphalt paintings.

When I got a flat bed plotter I plotted the PCB image scaled 2:1 on white glossy paper and turned it into a photo shop to reduce to scale 1:1 and make a deep black copy of the board on clear film. That method worked well but was way too expensive.

Later I built an UV-exposure box and used plots scaled 1:1 on clear film. I used a light table (a box with two dimable CFLs using RF frequency) to check for printing errors (like interrupted ink flow at bends) and corrected them using a marker pen.

When it came to double sided boards I had no other choice than making vias off the footprint of ICs to connect bottom and top layer using thin silver wire to be soldered on top and on the bottom, well clear off components (requires lots of extra space).

Again later I purchased a through hole plating machine and from then on used Gerber plots which are the best to use for home made PCBs. I got the Gerber prints for a six pack leaving my Eagle software on the PC used in the print shop.

They offer deep black prints of the traces, pads and vias on a clear film scaled 100% accurately with sharp contours which a laser printer at 2,400 pixel can't offer. Mounting two prints on top of each other you must have a light table and have to work from the center outwards because of unequally sized (distorted) prints.

Moreover if your laser printer software doesn't allow to switch off toner saving mode the image won't be as black as it should be.

It's very important to have PCB material of equal UV-sensitive layer. Trying out no-name products you'll have to make a test including exposure and development time before making the "real thing".

I found out that BUNGARD (Bungard Elektronik) makes the best photosensitive PCB material with constant film thickness allowing for instant exposure without experiments.

Depending on your budget you'll find all kinds of machines making your PCBs with almost industrial quality there. (CNC drill presses are above industry standard offering 1,000th of an inch (0.0254mm) of accuracy. (Industry standard is 0.254mm).

Talking about toner transfer, that method will certainly work well after some pit falls, like the right paper, TRF foil and laminator.

My friends in Germany claim that catalogue paper of Reichelt Elektronik (Notebook, PC Zubeh) works best for toner transfer.

Regards

Boncuk

 

[Veraxis]

Guys,
Thanks for all of your input and your help on this, it really helped me out. I was finally able to etch some boards for my digital comptometer circuit (I still have not tested it fully, but if I get it all working I will definitely post up some schematics on another thread in the future). I used heat transfer with magazine paper, which I then put in boiling water before peeling off the paper and etching with ferric chloride. the boards are far from perfect, but I think that they came out reasonably well considering the complexity and size of the boards and that this was my first time trying this. I have a lot of drilling to do now, but this is certainly a step up from manually wiring things on perf board. also, if it is of any help to anyone here, I found that sharpie marker actually works reasonably well as a substitute for an etch resist pen, although it is not perfect.

 

[KeepItSimpleStupid]

Staedtler pens such as these **broken link removed** work much better. The lines are much finer.

You can erase lines with the solvent methanol.

 

[mvs sarma]

All said, i feel, that for complex designs , EDA based artwork and tone transfer method is still better

I do agree with 3v0 that while toner transfer, one can use press& peel or pulsar pro etc where they are available freely in their local markets.
pcb artwork of complex nature, using marker or etch-resist pens , would turn out a workable board perhaps. But doesn't look good and has its own issues.