When I was in third grade, back around 1959, they pulled me out of class and asked me if I would like to go to Mrs. Spencer's Workshop for part of each day. This was a special class for bored and restless kids where we could work on projects of our own.
I started out drawing maps of freeway interchanges. A new freeway had just come through town (this was a very new thing at the time) and I was fascinated by the interchanges. I made sketches of all the ones between Eugene and Portland, and then I looked for something new to do.
I had been playing with electrical circuits for a while, so naturally I wanted to make a printed circuit board for a science fair project.
I didn't have one of those fancy phenolic sheets with copper on one side, but I figured I could just use a sheet of copper and put electrical tape on both sides to act as resist for the traces, and then mount the whole thing on a wooden board when done.
It would be a circuit, on a board, etched like a real printed circuit. So I thought that counted as a printed circuit board.
I asked Mrs. Spencer if she could get me a copper sheet and some electrical tape, along with a tank of nitric acid. And she did!
I taped up the sheet, dunked it in the nitric acid, and watched the unprotected parts dissolve away. And then I pulled the remaining circuit out of the acid, rinsed it off, peeled off the tape, and mounted it on my board.
Of course we didn't need eye protection or anything fancy like that back in those days.
And the circuit worked the first time! It may have been the first and last circuit or program where I had such good luck.
I once did a science project in the school lab that dealt with sulfuric acid in an ehrlenmeyer flask. I completed the experiment, and went home. Only then I noticed that the whole front of my shirt had huge holes in it.
This kind of stuff is cool from a educational perspective but with services like OSHPark I've lost motivation to make diy circuit boards. The batch services are just so cheap and fast now days.
Scotty is doing a good job of catching everyone up on what is going on in China. Chinese manufacturers have been putting out 10x10x10 boards for 50 RMB ($7 USD) for many years now. Rush order (24 hours)? Another 50 RMB. Yellow solder mask? Also another 50 RMB. Once you start talking about 4 layer, flex, rgba solder mask or other customization the difference is even greater.
I used a cheap* Chinese service recently and the quality was excellent. They're a great option if the dimensions of your design make OSH Park's by-the-square-inch pricing uneconomical. In addition, the boards were delivered in under a week.
If you can squeeze your board down, 100 x 100mm seems to be the sweet spot with the cheap Chinese board makers that I've used [1].
10 boards under 100x100 are typically around $5-10, whereas a 10mm more pushes it to around $30
JLC are currently showing me some sort of promotional pricing of $2 for 10 x 100x100mm 2-layer boards (+$6 shipping - I'm close), which is ridiculously cheap.
Thanks for the links I didn't know about these services.
They use TG-130 or TG-140 which are lower temperature (130C/140C) FR4 board ratings. The surface may melt when de-soldering at high temperature. I usually prefer TG-170/TG-180 boards as I find them to be more reliable during rework.
I dumped my trusty UV box and stacks of chemicals earlier this year, developed some patience and get JLCPCB to do the honours. It’s cheaper for me to run a not too small 2 layer board there than pay for the substrate and consumables here and the boards are several orders of magnitude better.
Plus you really can’t spin a decent SMD board at home.
When I around 12 years old I attempted to etch my first circuit board. I took my copper clad board, placed it in a pie tin and then proceeded to pour in the etching acid. Within seconds, the acid reacted violently with tin and dissolved through the bottom of the pan and spilled all over the table saw and garage floor where I was working. Let's just say my dad was not a happy camper when I showed him my handiwork!
The pie 'tin' was probably aluminum - maybe sheet steel, if it was one of the robust types.
Aluminum and acids usually react fast, with evolution of hydrogen. I used to use that reaction to top off balloons (already partially inflated with air) with hydrogen. Tie off the balloon, touch a match - a satisfying kaboom!, with no shrapnel.
What acid did you use, when I etched my first board in 1969, I used Ferric chloride (that stuff stains everything!) and an aluminum pie plate, it worked fine, no reactions at all, other than slowly etching my copper circuit board.
That was my first and last PCB. Until yesterday that is. I bought one of these 100$ Chinese PCB milling machine. It "works", I guess. I mean, wont do BGA or any fine print anytime soon, but for personal projects, it beats a glorious mess of wires or ducktaped breadboards :P
I don't own one, but the cheapest on AliExpress for me right now is this: [1]. It lists as $130 for me at the moment. I did not do any exhaustive searching, perhaps there are cheaper machines to be found. I know nothing about this particular machine, I've never owned one just wanted to provide an example of what the OP might have.
If you're new to making circuit boards, I highly recommend giving KiCad a try, for drawing your schematic and board layout. It's open source, and while there are some rough edges, it seems to be built on a good foundation and there's been lots of good work being done on it lately. The learning curve isn't too steep, as these things go, but you're not likely to be limited it anytime soon, if ever.
Don't forget Fritzing! While Fritzing is most commonly used for making very professional-looking diagrams of breadboard layouts, it's also a complete and extremely capable EDA suite. I've done several boards in Fritzing, and taken a few of those to production (> 1000 qty).
Also, unlike Eagle and Orcad and circuit studio, Fritzing is free as in beer and speech. Same with KiCAD, but the learning curve on Fritzing is just so much easier.
While I think your assessment is correct, I do want to note that it's not like starting with Eagle, Altium, Orcad, etc is going to be easier. Just jumping into a complex, domain specific piece of software is tough, but totally worth it here.
Right - the problems attacked by these programs are inherently hard, I think. Some projects care a lot about things that are completely irrelevant in others (BOM management, creepages, RF concerns), etc. Any given component might have a single schematic symbol, multiple footprints, and several parameters which may or may not change with either of those (flash size/temperature rating/etc). Then, what's the right way to attach any library changes to the project where they are applied, how do you handle changes...
I use KiCAD when I am forced to. Of the two open source EDA packages that I am aware of, KiCAD certainly has an easier learning curve than gEDA. That said, I started with gEDA so I have surmounted the learning and libraries problems sufficiently for my needs. My beef with KiCAD is that it is weak sauce, I find it missing a number of the features that I take for granted with gEDA. I don’t use KiCAD for personal projects, the limitations frustrate me. OTOH Somebody just starting out and doing simple boards would probably find KiCAD more approachable than gEDA.
I too dragged myself up the seemingly insurmountable learning curve that was gEDA about a decade ago and used it exclusively for many years. I poked around in KiCad every now and then and didn't think much of it either.
Since the changes that CERN made, KiCad is a completely new tool. It is now my tool of choice, and unless gEDA gets a similarly drastic revamp, I'm probably never going back.
Out of interest, what KiCad limitations frustrate you? My main frustration is with pcbnew's (the layout program) apparent inability to store sets of design rules, so they need to be set for each run.
I've not got any skin in the game as such, but occasionally think about contributing to it - usually after getting irritated by random UI things, but so far it's never quite been enough to dive in and fix anything.
KiCAD is not bad but I definitely file under the category of software for people who like screwing around with software. If you’re making boards in your kitchen then the free edition of Eagle will be more than satisfactory, and it comes with decades worth of user-contributed libraries and scripts.
Depends on your taste. Personally, I hate Eagle and find it really painful to use. KiCAD has rough spots, but I like its workflow a lot more. Plus Eagle's free version restrictions are pretty tight, 80 cm^2 and 2 schematic sheets are quite restrictive even for a hobbyist. I don't begrudge people who use it, but I genuinely prefer KiCAD. And if you're going to pay for a license for your EDA software, I think there are better options as well (OrCAD).
Regarding libraries, in the hobby community Eagle has lost a lot of momentum in the past couple of years. KiCAD is now what you see a lot of makers releasing libraries and stuff for.
Yeah I’ve noticed a preference among the library authors for kicad-compatible releases. That’s fine with me. I consider it analogous to the situation between clang and GCC. All the compiler people are working on clang these days. But GCC is still the better compiler, through the sheer mass of accumulated contributions. Someday the new thing will surpass the old thing.
The iron-on toner transfer thing has been around since laser printers came out. It doesn't work very well. Note the mention of needing a few tries to get it right. "Touch up" with a conductive ink pen and X-acto knife is often required. The resolution of this process is not that good. There also tend to be registration problems, because laser printer roller feeds are not precision devices. If you're doing modern surface mount components, it may not be good enough. There also tend to be registration problems, because laser printer roller feeds are not precision devices. You don't get plated through holes. So through-board connections have to be made by soldering in little wires. There are special rivets and tools available for this.
Other than as a curiosity, or for very, very simple boards for large components, this is more trouble than it's worth. Quick-turn one-off boards are so cheap now.
If you want to make your first PCB for the simplest of projects this is overkill.
Instead:
+ Get a single sided copper board and draw your design with a marker pen
+ Etch it in ferric chloride in a plastic takeaway tub
+ Drill holes
If it is too hard to draw with marker pen, then it is time to get the boards made for you from one of the links in this thread. The method in this article just looks frustrating.
The stuff you use for etching your own boards tend to be pretty nasty. Enough to burn holes in clothes and give you a rash if you're careless. Like other said, it's more economical to order boards through services like OSHPark and Seeed. If you want to make your own board for educational purposes or super quick turnaround (<24 hours) I think you're better off milling it. More of a hassle but much safer. Get a cheap CNC like a 3020 or something.
The PCB tutorials by Chuck Adams(K7QO) on YT are pretty good and go over a unique PCB style that he calls MUPPET. Zero thru-hole mounting, but uses thru-hole parts (and SMD sometimes). His tutorials also go over creating the print transfer using a small lamination machine rather than a hot clothes iron. I recommend checking out the series.
I totally understand the hobbiest aspect here, but... I just paid $15 for 3x 4-layer boards that will be delivered to me in a week. IDK, cool, but I’d spend more than $15 making 2-layer with way worse tolerances.
People don’t really appreciate what a miracle the quick-turn electronics industry is. You can upload a Gerber file and get boards back in 24 hours. 2 layers, 4 layers, 6 layers whatever. Tiny little holes all over the board. Boards two feet square. If I order a huge box of stuff from Digi-key right now I’ll have it on my desk at 8am tomorrow. Or I can drop-ship it to the PCB place and get stuffed boards, inspected and reworked, in a couple days. It’s incredible, and all this has been up and running for at least 20 years.
I 100% agree that these services are incredible, but I think you're downplaying the changes in the last few years a bit much, considering the tooling improvements and plummeting prices.
As an example: last week at work, we had a situation where some product designer types and some engineers were having a hard time thrashing out aspects of a particular thing we've been talking about. It literally took me one working day and ~$300USD, to go from an email thread, to an order of custom PCBs and parts to make 10 physical widgets, which will be in our hands this week. In the meantime, I can nut out the firmware on a dev board from the pile of such things in my desk drawer, with free tools, easy-to-get docs, etc. That dev board costs about the same as two or three pints at the local pub.
I think in 1998, an engineer at my level would've spent an equal or greater amount of time on making a document and illustrations to explain the concept and tradeoffs that go in to it, then meetings to talk over that document, and possibly still no concrete conclusion. Instead, we get a small pile of working prototypes to fuss with and iterate on - if we go in this direction then we're already a substantial step toward the real product. And, our main office is in California, the factory is in China, and I'm in a small city in New Zealand.
Sure, this was something that could be done back then, but only if you were in the right place, and had the relevant parts pretty well figured out, and it would've cost a fair bit more money.
Back in 98 the company I was working for was paying three salaries for EDA software licenses and $2-3k for a 48 hour turnaround on a couple of eurocards.
I still have one of the first circuit boards I made, sometime around 1982. I plugged my soldering iron into a Variac to run it at a lower temperature, then used it to draw with beeswax onto a piece of bare board, correcting my mistakes with a knife.
Today I just use a quick turn service, or some form of "dead bug" breadboarding.
Cool! So do I, from 1969, in my high school electronics class and it still works too.
It's for a variable audio oscillator that rises and is re-set by a micro switch.
I tuned it with a precision trim pot, to sound "like" a then new, electronic Police Siren, it was a big hit at school with all the guys.
Unfortunately, I used the case from my first Transistor Radio, which was an xmas present in 1962, a beautiful red and creme, Jupiter, 6 transistor radio, I regret cannibalizing it for the case, to this day.
I started out drawing maps of freeway interchanges. A new freeway had just come through town (this was a very new thing at the time) and I was fascinated by the interchanges. I made sketches of all the ones between Eugene and Portland, and then I looked for something new to do.
I had been playing with electrical circuits for a while, so naturally I wanted to make a printed circuit board for a science fair project.
I didn't have one of those fancy phenolic sheets with copper on one side, but I figured I could just use a sheet of copper and put electrical tape on both sides to act as resist for the traces, and then mount the whole thing on a wooden board when done.
It would be a circuit, on a board, etched like a real printed circuit. So I thought that counted as a printed circuit board.
I asked Mrs. Spencer if she could get me a copper sheet and some electrical tape, along with a tank of nitric acid. And she did!
I taped up the sheet, dunked it in the nitric acid, and watched the unprotected parts dissolve away. And then I pulled the remaining circuit out of the acid, rinsed it off, peeled off the tape, and mounted it on my board.
Of course we didn't need eye protection or anything fancy like that back in those days.
And the circuit worked the first time! It may have been the first and last circuit or program where I had such good luck.
It was the most awesomest day at school ever!