DIY PCB STARTUP

[joergen14]

We’re a group of engineering students at SDU University in Demark

Right now, we’re building a DIY PCB machine to make electronics skills more accessible for students. Many universities, including ours, don’t have easy access to PCB equipment, and we think this machine could open up a lot of possibilities for students who want to experiment and build.

We’ve created a short survey (about 5 minutes), and your feedback would be incredibly valuable to us as we develop this project.

https://eu.jotform.com/form/242832076794061

Thank you for all the inspiration we’d love to hear what you think!

[SDU Engineering Student]

[Overspeed]

Hello

most challenging on DIY or small setup is the Vias as that request metallisation of the holes , except if you put rivets or copper nails

Other concern is the base material , as you shall stay in FR4 1.6 mm thickness basic specs , all other FR4 and also RF ''Rogers'' cost hell .

A nice process is laser engraving but you can also use your laser to burn paint as a mask and after make a etching , some US universities use laser burning to produce prototype in house

Good luck

OS

[janoc]

I don't want to spoil your student enthusiasm but how do you expect to compete with the PCB fabrication services like JLCPCB or PCBWay (or even Aisler here in the EU)? Did you consider that many places don't have their own PCB making equipment simply because it is not economical to do so? Is your machine going to be able to produce a board with soldermask, silkscreen, metalized vias and more than 2 layers?

That aside, why to reinvent the wheel?  If you want to manufacture your own PCBs on a small scale beyond the hobby DIY level like tone transfer, then a spray etching tank/machine, UV exposing unit and consumables cost very little and are certainly available even to a Danish university (apropos, greetings to SDU, I used to work at AAU in Esbjerg just across the street from the SDU campus, with many SDU colleagues). You still won't have metalized vias or soldermask unless you invest in additional process but that's true for pretty much any kind of "PCB machine" you are likely to build that would be still relevant for a student.

There is no need to invent a "PCB machine" (I guess some kind of CNC engraver/router?) here, IMO and probably very little market among cash strapped students for something like that.

Apropos, consider that in many cases the problem isn't the cost of purchasing the equipment but the recurring costs of operating it - e.g. required supervision if to be operated by students, consumables, handling the toxic waste, dealing with toxic dust (in case of using a CNC router) in a way that won't give the occupational safety authorities a heart attack and the operator a hefty fine. This adds up quickly to way more than the original cost of the equipment. E.g. my hackspace here down the street has gotten rid of their PCB production equipment for exactly these reasons, with the argument that it is much easier for everyone involved to order the boards online instead. There is a huge difference between etching or engraving a board at home - and operating (or selling) this equipment in a public space/institution/business where the laws and regulations are strict and the authorities are watching you.

[jduncan]

I'm pretty curious who will answer "No" to this question

[thm_w]

I'm pretty curious who will answer "No" to this question

I think what they meant to write was 2 layer PCB, not double sided load PCB.

[coppercone2]

you can do the metalization, the open process works, it just stingy with palladium. you need to add alot more. patents say 10x more. just because its black does not mean its good.

[tooki]

We’re a group of engineering students at SDU University in Demark

Right now, we’re building a DIY PCB machine to make electronics skills more accessible for students. Many universities, including ours, don’t have easy access to PCB equipment, and we think this machine could open up a lot of possibilities for students who want to experiment and build.

We’ve created a short survey (about 5 minutes), and your feedback would be incredibly valuable to us as we develop this project.

https://eu.jotform.com/form/242832076794061

Thank you for all the inspiration we’d love to hear what you think!

[SDU Engineering Student]

I would encourage you to look at prior efforts in this direction (like Voltera and pdi3d) and see why they have failed or just not caught on.

Everyone I know who has dabbled in PCB milling — including with CNC machines made specifically for that purpose — have given up on it because it’s too inconsistent, and lacks what are now basic niceties like solder mask and plated holes. (Yes, you can do those yourself but the effort is enormous.) Never mind multilayer (4+) boards, which we can simply dismiss as “impractical enough to be called impossible” to do yourself.


In the past, when commercially made PCBs cost a lot more, DIY PCBs were done by etching, and that’s still a decent approach from a technical standpoint, but the effort is completely disproportionate to the cost of just ordering from JLCPCB, PCBway, Aisler, OSHpark, etc. It just doesn’t make sense. And then you have chemical waste to deal with.


So what problem is it you aim to solve? Cost? You will lose on that. Capabilities? You will lose on that even more. Turnaround time? Well… maybe, if you have your process completely perfected. (Professionals just pay through the nose for express service from a local PCB manufacturer if they absolutely must have a board ASAP — as fast as 1 day.)

[tooki]

I don't want to spoil your student enthusiasm but how do you expect to compete with the PCB fabrication services like JLCPCB or PCBWay (or even Aisler here in the EU)? Did you consider that many places don't have their own PCB making equipment simply because it is not economical to do so?

Indeed, many universities, corporate prototyping labs, and the like had that equipment for years, and got rid of it in recent times because JLCPCB etc is cheaper, easier, and better.

At the university department where I work, single-sided PCBs in in-house devices were mostly DIY etched until around 1995, then little by little moved to commercial, and I haven’t seen a DIY etched PCB made after around 2000.

Nobody wants to go back to making them themselves, it’s not worth it.

[jduncan]

It seems like you may think assembly is the "secret sauce" that will make this worthwhile.

you're not the first to think of turning a 3d printer into a pick and place machine. however, what you'll find is that setup effort makes this less efficient than just hand placing components with tweezers until you're making 10+ boards.

[tom66]

Making your own PCBs is fraught with difficulties.

CNC engraved boards are very limited.  The resolution is restricting, and the boards lack soldermask which makes building up large boards (especially with SMDs) very painful.  The exposed copper can patina after a while which weakens it too.

3D printed boards are still very nascent, mostly still a toy idea rather than anything practical.

If you're talking about a traditional PCB process which uses photoimageable materials then it's even worse.  The chemicals required for industrial PCB processes are nasty.  They require hazardous material handling processes.  You absolutely CANNOT put the waste for a PCB process down the public sewer.  Even the rinse-water from the final part of the process is contaminated.  So you need either on site decontamination or pay for a professional company to take the waste liquid away.  Sometimes the etchants are flammable and usually they are corrosive so there are storage headaches too.

Fibreglass (FR-4) needs air filtration and safety precautions to cut.  You will need to cut the FR-4 because it comes on large rolls and board sizes will vary. 

The exposure process needs controlled lighting.  Most PCB fabs use orange or red LED lighting like the lamp in a photography lab.  This is usually carefully targeted to avoid the operating sensitivity of any photoresist. 

If you are building the boards yourself, then you will also need a reflow oven (requires fume extraction and 3ph power supply if it's a decent sized one) and pick and place machine.  You can probably get away without wave soldering, with hand soldering either done by your students or by staff, but make sure to work out the actual cost per joint as it can be surprisingly expensive (IIRC we were coming in around 5p per joint at the last company, which led to the next board revision having a lot of THT stuff removed).   The issue with pick and place machines is proper feeders for SMD reels are very expensive.  Open source projects to get a gantry moving X/Y/Z to place parts are common, but there are very few out there with the precise feeders required that can remove the parts from the reel.  Feeders need to be designed for the varying tape widths too as some parts take up 2/3/4 reel spots on a machine. 

If you wanted to get the cost down, buying a cheaper second hand machine that is otherwise obsolete and controlling it using modern software may be a better approach.  I remember that we had a PnP machine powered by a Win98 PC in an old job, and getting it to work with modern formats was a pain (even supporting some USB drives was tricky).  But the company that made that machine didn't support it any more so we were stuck using it.

[janoc]

Making your own PCBs is fraught with difficulties.

CNC engraved boards are very limited.  The resolution is restricting, and the boards lack soldermask which makes building up large boards (especially with SMDs) very painful.  The exposed copper can patina after a while which weakens it too.

3D printed boards are still very nascent, mostly still a toy idea rather than anything practical.

If you're talking about a traditional PCB process which uses photoimageable materials then it's even worse.  The chemicals required for industrial PCB processes are nasty.  They require hazardous material handling processes.  You absolutely CANNOT put the waste for a PCB process down the public sewer.  Even the rinse-water from the final part of the process is contaminated.  So you need either on site decontamination or pay for a professional company to take the waste liquid away.  Sometimes the etchants are flammable and usually they are corrosive so there are storage headaches too.

Fibreglass (FR-4) needs air filtration and safety precautions to cut.  You will need to cut the FR-4 because it comes on large rolls and board sizes will vary. 

The exposure process needs controlled lighting.  Most PCB fabs use orange or red LED lighting like the lamp in a photography lab.  This is usually carefully targeted to avoid the operating sensitivity of any photoresist. 

If you are building the boards yourself, then you will also need a reflow oven (requires fume extraction and 3ph power supply if it's a decent sized one) and pick and place machine.  Y
...

I don't believe the OP is after setting a PCB fab here but to make the technology somehow available to a typical student at home/in the dorm.

For a DIY board production you certainly don't need a reflow oven or a pick & place machine. Even a typical university lab or hackerspace doesn't need these. You are not building hundreds of identical boards in your kitchen, are you?

You also don't even need the yellow light or dark room - a typical resist is not sensitive to daylight, it works with UV so unless you open the package in direct sunlight or have a lot of flourescents in the room, you will be fine. Just don't leave an open package of dry resist or of pre-sensitized boards laying around. Fab work where they are after maximizing yield is a different ballgame and not relevant for a typical DIY setup.

Concerning waste - you neutralize it with baking soda and most places have a municipal waste collection facility that will take such chemical waste from you in small quantities. Universities and such typically  have a way to dispose of such waste because they have laboratories (I guess we are not making PCBs for a degree in philosophy or humanities in general). Those deal with way worse chemicals than hydrochloric acid, ferric chloride and copper compounds. Again, we are not setting up an industrial scale production facility here - that would be a different story.

For a typical DIY/hobbyist grade work a toner transfer or photolithography is totally doable and sufficient for a lot of projects already. I  personally wouldn't touch PCB milling because of the noise & dust (plus it is very limited & slow) but many people use that too. Whether it is economical/worth doing for you that is something everyone must decide for themselves, because there are many factors that come into it, not only board production/shipping cost or even board complexity. E.g. iteration time or deadlines ("I need a board for my semester project I must hand in next week - and it will take 3 weeks to get the board from China, what do I do now?!") play an equally important role.

[jduncan]

deadlines ("I need a board for my semester project I must hand in next week - and it will take 3 weeks to get the board from China, what do I do now?!") play an equally important role.

So sick of hearing this silly talking point. JLC will fab a 2-layer board in 24 hours, and for $20 you get delivery to the US on the other side of the world in 3 days more. You can click "order" then spend the hours you would waste DIYing the board (for a much worse result) on your other classes.

[dietert1]

I think in a future with nationalist governments everywhere and without free trade making PCBs of some complexity should belong to the skill set of EEs. Especially european countries will probably seek different solutions. Depending on China (PCBs) or on USA (Chips) appears more and more risky.

Regards, Dieter

[janoc]

deadlines ("I need a board for my semester project I must hand in next week - and it will take 3 weeks to get the board from China, what do I do now?!") play an equally important role.

So sick of hearing this silly talking point. JLC will fab a 2-layer board in 24 hours, and for $20 you get delivery to the US on the other side of the world in 3 days more. You can click "order" then spend the hours you would waste DIYing the board (for a much worse result) on your other classes.

Maybe check how much does the postage cost to Europe in such case. And how much a 24h turnaround costs vs the normal 3 days. Aisler has even 8 (!) working days standard.

We are talking students here, paying that out of their own pocket, likely with no income yet. And even the "3 day DHL shopping" is more like a more than a week to Germany (had that happen recently). The much better results is of no use if you don't get the board assembled, integrated (the project is rarely the board from China!) and delivered on time.

To paraphrase you - I am sick of hearing these talking points of people assuming everyone has the same needs, means and resources as they do.

[tooki]

For a DIY board production you certainly don't need a reflow oven or a pick & place machine. Even a typical university lab or hackerspace doesn't need these. You are not building hundreds of identical boards in your kitchen, are you?

A reflow oven certainly is basic equipment these days. It doesn’t need to be a super-expensive continuous conveyer system, after all. Decent ovens based on toaster ovens can be had for very little.

[janoc]

For a DIY board production you certainly don't need a reflow oven or a pick & place machine. Even a typical university lab or hackerspace doesn't need these. You are not building hundreds of identical boards in your kitchen, are you?

A reflow oven certainly is basic equipment these days. It doesn’t need to be a super-expensive continuous conveyer system, after all. Decent ovens based on toaster ovens can be had for very little.

A basic requirement for what, exactly? Esp. for a cash-strapped student?

If you said hot air, sure. One needs to be able to rework boards and it helps with some difficult components. But a reflow oven (plus keeping a stock of paste, etc)? Let's get real, shall we?  I think a student at home or in a dorm or shared flat somewhere has completely different priorities than getting a reflow oven.

A hackspace or a school lab can have a small oven, there it makes a bit more sense - but even there it is rarely used. E.g. we have one at the local hackspace - and it gets used only once in a blue moon when someone needs to pre-assemble a large amount of kits for some event. Otherwise for the one off stuff I see regularly being assembled there it is just faster to do it by hand (either using iron or hot air) than to mess with stencils, paste, setting the oven up, etc. And people are building even fairly large boards with a fair amount of SMD components there.

[tooki]

Who said anything about a student buying this for themselves? It’s obvious that the entire context of this discussion is as some kind of service for students.

Many modern components demand reflow soldering using stenciled paste, and handheld hot air cannot solder all such parts.

[janoc]

Who said anything about a student buying this for themselves? It’s obvious that the entire context of this discussion is as some kind of service for students.

Many modern components demand reflow soldering using stenciled paste, and handheld hot air cannot solder all such parts.

That's not at all from what OP has posted (and never came back) - the post & survey are about a "A DIY PCB machine" and not any kind of service.

I have also yet to encounter a component "demanding reflow soldering using stenciled paste" and that hot air cannot do. I am sure there are such components - but they are not something a DIY amateur/hobbyist/student is going to work with or that doesn't have an alternative in a more sensible package. Even more so given that we are discussing a DIY low cost boards (i.e. 2 layers max, nonmetalized vias, no solder mask, etc.).

[tooki]

Well, when you are talking about students who are too poor to buy a $60 toaster oven to reflow in, and the OP’s mention of makerspaces, it certainly suggests a device owned by a larger organization, with users using it themselves or the staff doing it for them. Either way, not the end user owning it.

[ebastler]

Maybe check how much does the postage cost to Europe in such case. And how much a 24h turnaround costs vs the normal 3 days. Aisler has even 8 (!) working days standard.

22€ for shipping, 6.80€ extra for 24 hour service. (That's for 5 pieces of a 100*160 mm² board from JLCPCB.) So what?

The fact that Aisler has 8 days standard processing time is disappointing and rather dumb. They are forfeiting the big benefit they have, namely being local for European customers. In most cases, JLCPCB will easily beat them on price including shipping, delivery time, and capabilities -- all at the same time of course.

[cgroen]

Maybe check how much does the postage cost to Europe in such case. And how much a 24h turnaround costs vs the normal 3 days. Aisler has even 8 (!) working days standard.

22€ for shipping, 6.80€ extra for 24 hour service. (That's for 5 pieces of a 100*160 mm² board from JLCPCB.) So what?

The fact that Aisler has 8 days standard processing time is disappointing and rather dumb. They are forfeiting the big benefit they have, namely being local for European customers. In most cases, JLCPCB will easily beat them on price including shipping, delivery time, and capabilities -- all at the same time of course.

Exactly. When I order from JLC, it takes almost every time exactly 7 days from Gerber until I have the boards in my hand using DHL Express (usually around Euro 23 to 27).
I have now done more than 470 separate orders with them, and they have never let me down (4 and 6 layers).
I'm also using their CNC machining for a lot of stuff as well as their 3D printing
(some of the Alu I have had CNC'ed from them: https://www.moonbounce.dk/hamradio/repam-remote-pa-monitor.html)

[PlainName]

Exactly. When I order from JLC, it takes almost every time exactly 7 days from Gerber until I have the boards in my hand using DHL Express (usually around Euro 23 to 27).

I use JLCPCB a lot, but waiting 7 days and spending $50 for the pleasure isn't in the same race as something I can have in my hands in half an hour for nothing.

Sure, there are caveats for this: only works with simple boards that are mostly single-sided (mostly because I can't be arsed dealing with double-sided alignment), and it's for prototypes that aren't going to need solder masks and all that stuff (but 0.5mm pitch SMD is not a problem).

In reality there aren't that many jobs like this that I want to do, and maybe instead I'll use protoboard and PTH components. But the point of this post is to refute the absolute certainty that waiting for shipping from China and paying steeply for that is always the best solution.

[tooki]

A half hour if you have your process totally nailed in, and don’t need to drill many holes. It doesn’t take much for it to take a lot longer than a half hour.

[PlainName]

If it took all day it would still be a winner if you don't want to wait or throw money away.

if you have your process totally nailed in

Nonsensical argument. Sending off the JLCPCB is only quick if you have you CAD and stuff 'totally nailed in' already. Start from scratch and you won't have sent anything off before you'd expect it to be back! Same applied to pretty much any process, and you have to start at the beginning some time. You could wend your way down to McDonalds and think "must be a better way, but it would take too long and be too much hassle to dial in that process of 'cooking', and you gotta have a cooker and raw materials... nah, I like burgers anyway."

[tooki]

If it took all day it would still be a winner if you don't want to wait or throw money away.

How is it sensible to spend an entire day making a simple one-sided PCB?

if you have your process totally nailed in

Nonsensical argument. Sending off the JLCPCB is only quick if you have you CAD and stuff 'totally nailed in' already. Start from scratch and you won't have sent anything off before you'd expect it to be back! Same applied to pretty much any process, and you have to start at the beginning some time. You could wend your way down to McDonalds and think "must be a better way, but it would take too long and be too much hassle to dial in that process of 'cooking', and you gotta have a cooker and raw materials... nah, I like burgers anyway."

I don’t think you understood what I am saying.

I am starting from the assumption that one has the PCB layout done, ready to send gerbers or to print out to use to etch at home. Doing the board layout is not part of perfecting your PCB manufacturing process — that is a separate step individual to each board.

All I was saying is that if you have your home PCB production process already perfected (that’s what I mean by “nailed in”) — you have already perfected your print settings and material selection, exposure times (for photosensitized boards) or transfer temperature (for toner transfer), and have perfected the etching time and temperature for your etchant — THEN you can quickly and reliably produce a PCB from your prepared PCB layout. As long as you don’t have dozens and dozens of holes to drill and don’t have complex board shapes to cut.


If you haven’t been doing enough DIY PCB manufacturing to have a repeatable, reliable process, then you may end up spending a lot more time, thanks to failed attempts.