Isn't it strange that you can make better PCBs at home than factories make?

[MaximRecoil]

With regard to two-layer boards at least. Normally it's the other way around.

Traces: Because factories use archaic Gerber files, your traces are limited to lines and arcs. If making a PCB at home, there are no shape limitations whatsoever. You could draw them in e.g., Adobe Illustrator, which supports bezier curves, so anything can be drawn. You could also draw them by hand, which, of course, has no shape limitations either.

Pads and vias: A through-plated pad/via is relatively delicate. It won't hold up to much heat/rework, and the plating in vias can crack over time resulting in an open circuit. At home you can use copper or brass rivets for this, which are made of sheet metal, which is infinitely tougher than mere plating:



https://youtu.be/ywfjknf6Vtg

You could rework a riveted through-hole until the cows come home and it wouldn't harm it in the least. It won't ever crack either. Additionally, it provides a mechanical fastener for the traces at each point that one is installed, which beats the mere chemical adhesive that is the only thing holding the copper to the fiberglass in a factory-made PCB.

Silkscreen: No limitations here; you can print anything at home up to and including full-color photographic images if you wanted to, because, again, you don't have the Gerber file limitation.

[T3sl4co1l]

I would trust rivets less than plating -- the FR-4 creeps over time, and the unusually meaty rivet will fatigue away from the board (delamination) or trace (cold solder joint).

Also, "infinite" is a lot, even for hyperbole.  A rivet will be a bit stronger than plating because of work hardening, and plated copper might not be very ductile (though it depends on the type of plating).  The material strength of plated copper will be about half; the material cross section determines the rest.  (I suppose a rivet with a 10 mil wall would be about 20 times stronger than typical PCB plating?)

I'd be quite interested to see suggested soldermask and silkscreen methods.  How would you do full color?  How do you maintain layer registration?  Where do you get the epoxy ink from?  (Toner doesn't stand up to soldering heat!)

Tim

[ataradov]

All of that makes no sense. Gerbers can be as detailed as you like or as equipment resolution will allow. The rest of the statements are unsubstantiated also. Show us how would you do high resolution silk screen at home, for example.

[Mr.B]

...Show us how would you do high resolution silk screen at home, for example.

That.

[MaximRecoil]

I would trust rivets less than plating -- the FR-4 creeps over time, and the unusually meaty rivet will fatigue away from the board (delamination) or trace (cold solder joint).

A copper rivet is going nowhere, provided it is set properly. It is the same idea as a crimped connection, like every single connection in the wiring harnesses of your car.

Also, "infinite" is a lot, even for hyperbole.  A rivet will be a bit stronger than plating because of work hardening, and plated copper might not be very ductile (though it depends on the type of plating).  The material strength of plated copper will be about half; the material cross section determines the rest.  (I suppose a rivet with a 10 mil wall would be about 20 times stronger than typical PCB plating?)

It will be a lot stronger, because it's far thicker. It is made from sheet metal, while the thickness of plating is measured in the millionths of an inch. You could build a structure using those rivets as fasteners, such as a sheet metal cabinet or attaching body panels to a car. You can't hold a substantial structure together with plating; not even close.

I'd be quite interested to see suggested soldermask and silkscreen methods.  How would you do full color?  How do you maintain layer registration?  Where do you get the epoxy ink from?  (Toner doesn't stand up to soldering heat!)

Those things are second nature to a silkscreen printer (which I am). Nazdar makes any type of screen printing ink you want, including epoxy. The entire process of making a PCB (aside from riveting) either is screen printing or is very analogous to it (i.e., photosensitive stencil-making emulsion type process). Full color photographic images can be screen printed; just look at a typical audio CD with album cover art printed on it. That is usually done with screen printing. The same goes for plenty of T-shirts. It is a CMYK halftone process, the same process that digital printers (e.g., inkjet, laser) use (simulated process can also be used).

All of that makes no sense. Gerbers can be as detailed as you like or as equipment resolution will allow.

No. As I said, Gerber files don't support bezier curves, which means you're limited to lines and arcs. Try to draw the Coca-Cola script logo in a Gerber file, for example.

Show us how would you do high resolution silk screen at home, for example.

See above. By the way, screen printing is fundamentally analog, i.e., there is no "resolution", or you might say that the "resolution" is infinite. If using a digital printer instead of a film-based "stat camera" to make the film positives, then the resolution is established by the digital print, but with a good printer, that's high enough to appear analog to the naked eye.

[BradC]

Show us how would you do high resolution silk screen at home, for example.

See above. By the way, screen printing is fundamentally analog, i.e., there is no "resolution", or you might say that the "resolution" is infinite. If using a digital printer instead of a film-based "stat camera" to make the film positives, then the resolution is established by the digital print, but with a good printer, that's high enough to appear analog to the naked eye.

Try again. He asked "Show us how you would do...". You said "Those things are second nature to a silkscreen printer (which I am).". Good on ya. Those things are basic to anyone with a grasp of physics and a bit of manual arts, but there is a massive chasm between knowing how it is done and actually doing it. I know how to, and have done screen printing. I aso know I would make a complete hash of a multi-colour screen print with the level of detail required on a PCB. Apparently you can do it, so *show* us how you do it. I assume given your confidence you actually *have* done it and therefore know all the little pitfalls that the rest of us would trip over.

While you are there, give us some shots of your bezier curve artwork, rivetted boards and let us know how they've held up over the years

Crimped connections rely on a correct "cold-weld" to attach the wire. How do you intend to achieve that with the plated copper without damaging the board? There's a reason crimped connectors are not soldered. Will it work, sure. Will it work long term? Probably not. Look at the DIY kits for rivetting plate through holes. There's a couple of very good reasons why they don't have flanges on them, and they are generally used for "prototypes".

[ataradov]

No. As I said, Gerber files don't support bezier curves, which means you're limited to lines and arcs. Try to draw the Coca-Cola script logo in a Gerber file, for example.

You can have rasterized pictures. The actual equipment does not support bezier curves either and rasterizes artwork anyway.

See above.

I don't see a picture anywhere.

By the way, screen printing is fundamentally analog,....appear analog to the naked eye.

So does it appear analog or analog? You have not shown a single example. I'd call this trolling and will stop waste my time until you show a picture of a board made at home that it better looking than a professional board. Don't forget to put some micro-vias there.

[Mr.B]

Silkscreen: No limitations here; you can print anything at home up to and including full-color photographic images if you wanted to, because, again, you don't have the Gerber file limitation.

This is what got me.
As @ataradov says... "Show us how".

If you are a screen printing expert, show us how to do this at home.
I would love to be able to do this on a PCB...

[ataradov]

BTW, most (all?) board houses will take vector files, they don't really care, since they import all this stuff into CAM software anyway.

I know at least one case when a board was made based on an Adobe Illustrator drawing (it was a front panel, so no actual functional PCB was required).

They will charge you more for this, since if you are asking, you either have no idea what you are doing, or know exactly what you are doing, and prepared to pay extra.

[sleemanj]

Making PCBs at home is great, a useful skill, and one that I regularly use for quick one-offs or prototypes. 

However, trying to make PCBs at a "factory quality" is a foolish waste of time, even without factoring in the economics.

double sided is a pain for all sorts of reasons
rivets take work to install,  are expensive (even if you buy them from me, or maybe especially if you buy them from me ;-)), and can loosen if you don't solder then top and bottom
soldermask is a pain unless you have dry film soldermask, which is not readily available out of china, and expensive - liquid soldermask is just a complete pain in the ass to work with in a diy setting for anything more than the tiniest of PCBs, even Vitrea 160 and toner masking is a more trouble to get right than it's worth (especially if you don't have an airbrush/spray gun)
silkscreen is virtually impossible unless you use toner transfer to do it, in which case it's black, or you can use that white toner reactive foil from Pulsar (I have never found a Chinese supplier of toner reactive foil), either way, it's a pain in the ass, and who in their right mind would go to the trouble of making an actual silkscreen for a PCB they are making at home
any pcb shape other than rectangular with round holes would be a massive pain to create unless you have a CNC router

That said, it's easy to get sucked into the challenge of making a real good looking PCB at home, but you'll spend a lot of time trying to perfect it, when for $10-20 USD you can get 10 of them shipped to you from China in a couple of weeks.




 
 

[Mr.B]

^ Absolutely

And even if I could get a photo realistic silk screen process perfected at home, I would still have my 2 layer PCBs made in China and shipped to me with no silk screen...

//troll me thinks
ignoreThread = true;

[MaximRecoil]

Try again.

No thanks. My reply is fine, as-is.

He asked "Show us how you would do...". You said "Those things are second nature to a silkscreen printer (which I am).". Good on ya. Those things are basic to anyone with a grasp of physics and a bit of manual arts, but there is a massive chasm between knowing how it is done and actually doing it. I know how to, and have done screen printing. I aso know I would make a complete hash of a multi-colour screen print with the level of detail required on a PCB. Apparently you can do it, so *show* us how you do it. I assume given your confidence you actually *have* done it and therefore know all the little pitfalls that the rest of us would trip over.

What pitfalls? Anyone who can do full-color process screen printing can do this. With regard to your text that I placed in bold, if that's true, the rest of your paragraph makes no sense, unless you don't know how to do halftones. To make the film positives for halftones, you need either a PostScript printer or some type of RIP software that can generate them in the printout. You also need to know how to do color separation in e.g., Photoshop.

Crimped connections rely on a correct "cold-weld" to attach the wire.

No, they don't. A proper cold weld is ideal, because it is air-tight, thus preventing corrosion of the wire within the crimp, and a specific degree of crimp ensures the connector won't detach from the wires when under strain, but it isn't necessary to attach the wire nor to achieve an electrical connection. 

How do you intend to achieve that with the plated copper without damaging the board?

You don't need to. A rivet isn't going anywhere because it's mechanically locked into a hole, so it doesn't need to be so tight as to cold-weld to the trace. The earth ground wire in a Metcal power supply is riveted to the housing by the way, and in my ~30-year-old one it hasn't gone anywhere, nor do I expect it ever will. Screw terminals work on the same principle, but unlike screws, rivets don't loosen, unless you stress them enough to physically deform them. That won't happen with a PCB.

There's a reason crimped connectors are not soldered.

Yes, because solder sucks for wire connections in an environment such as a car, which is subject to movement/vibrations and extremes of hot and cold temperatures. Solder is relatively brittle.

Will it work, sure. Will it work long term? Probably not.

It would work forever if the rivets are properly set. Have you ever worked with rivets before? You seem to think they are a lot weaker than they actually are.

Look at the DIY kits for rivetting plate through holes. There's a couple of very good reasons why they don't have flanges on them, and they are generally used for "prototypes".

I don't know what you're talking about. All rivets have flanges or are otherwise mushroomed/expanded on both ends once they are set; they wouldn't stay in the hole if they didn't, nor would they be called "rivets".

You can have rasterized pictures. The actual equipment does not support bezier curves either and rasterizes artwork anyway.

What "actual equipment" are you talking about? And which PCB manufacturers accept raster images (or vector formats such as SVG, EPS, AI, PDF, etc.) for the silkscreen layer and the copper layers (for the traces and pads)?

I don't see a picture anywhere.

This is a non sequitur.

So does it appear analog or analog?

I mentioned two different things there, so your question is another non sequitur. As I said, screen printing is fundamentally analog. It has been around for about a thousand years. But if you use a digital printout to burn the screen, you've obviously introduced a digital resolution to the process. Some digital images are high enough resolution to appear analog to the naked eye, and some are not.

You have not shown a single example. I'd call this trolling and will stop waste my time until you show a picture of a board made at home that it better looking than a professional board. Don't forget to put some micro-vias there.

I don't need to show an example for facts which are independently verifiable. I don't currently have any plans to make my own PCB at all, for any purpose; this thread is about the irony of being able to make better PCBs (within certain limitations, such as 2-layer) at home than factories normally make. As I said, it is usually the other way around, because factories tend to have huge resources, knowledge pools, labor force, and equipment that is too large and/or expensive for the average Joe to own, and that tends to result in a product that is hard to match at home, let alone exceed.

[MaximRecoil]

Making PCBs at home is great, a useful skill, and one that I regularly use for quick one-offs or prototypes. 

However, trying to make PCBs at a "factory quality" is a foolish waste of time, even without factoring in the economics.

double sided is a pain for all sorts of reasons

Of course it would be a pain, which is why I wouldn't do it. I wouldn't want to set hundreds of rivets for one PCB, nor would I want to make the screens/stencils for the printing and solder mask, nor would I want the clean up job (cleaning up after printing with epoxy ink is a lot more of a hassle than the plastisol ink I normally print with).

rivets take work to install,  are expensive (even if you buy them from me, or maybe especially if you buy them from me ;-)), and can loosen if you don't solder then top and bottom

What sort of stress is a PCB subject to that will cause a copper or brass rivet to physically deform enough to loose contact with the trace? Similar rivets are often used for applications that actually have significant stresses involved, such as attaching body panels to cars. In such cases they are supporting a significant amount of weight, and have to handle all of the flexing that a car body does under normal driving conditions.

[MaximRecoil]

If you are a screen printing expert, show us how to do this at home.
I would love to be able to do this on a PCB...

And even if I could get a photo realistic silk screen process perfected at home, I would still have my 2 layer PCBs made in China and shipped to me with no silk screen...

Then learn how to screen print and acquire the equipment/supplies you need. There are tons of resources pertaining to that online. I wrote a tutorial a couple of years ago on how to use Ghostscript (free) as a RIP to generate halftone printouts from Adobe Illustrator to a non-PostScript printer, which is a primary key to being able to do full-color process screen printing, so you can start there if you want to - http://www.t-shirtforums.com/p2630370-post1.html

//troll me thinks
ignoreThread = true;

You don't get to redefine the word "troll". Ironically, the most common example of "trolling" on the internet is falsely accusing someone of being a "troll".

[Monkeh]

rivets don't loosen, unless you stress them enough to physically deform them. That won't happen with a PCB.

No, it won't, because the PCB itself will deform first and make it loose, promptly followed by the solder cracking.

I have repaired many more failed rivets than plated holes. Proper plated through holes are mechanically excellent unless your name is Jeremy Clarkson. As an added bonus, you can plate through any shape cutout you can make in the PCB trivially. Where do I get 0.7x1.5mm slot rivets again?

[MaximRecoil]

rivets don't loosen, unless you stress them enough to physically deform them. That won't happen with a PCB.

No, it won't, because the PCB itself will deform first and make it loose, promptly followed by the solder cracking.

Why would the PCB deform? I have many large PCBs from the 1980s (arcade boardsets), and none of them are deformed, not even the bootleg PCBs I keep around for parts boards.

I have repaired many more failed rivets than plated holes.

Where are you finding all of those riveted PCBs to repair? Are there any PCB factories that produce or produced them?

Proper plated through holes are mechanically excellent unless your name is Jeremy Clarkson.

I don't get the Jeremy Clarkson reference, but I know you can damage a through-plated pad with nothing more than a soldering iron. You could hold a soldering iron to a copper or brass rivet all day long and it wouldn't damage it in the least. Additionally, because plating is very thin/weak, it can crack over time due to it thermally expanding at a different rate than the fiberglass. 

As an added bonus, you can plate through any shape cutout you can make in the PCB trivially. Where do I get 0.7x1.5mm slot rivets again?

Yes, as I said, "within certain limitations". Many, if not most, PCBs, don't have, or at least, don't require, anything other than round vias/through-holes.

[tautech]

We can learn a lot from what we think are processes of yesteryear.
Riveted I/O's were common decades ago and withstood much rework but of late they've been dropped in favour of cheaper processes. Are these better, IMO no.
Rather than dismiss such reborn procedures I'm all ears as to how they've been improved for todays use.

[Monkeh]

Why would the PCB deform?

Because you just crushed a piece of metal around it and then subjected it to intense thermal cycling.

Where are you finding all of those riveted PCBs to repair?

Old equipment. Once upon a time, through-plating was an expensive procedure.

I don't get the Jeremy Clarkson reference, but I know you can damage a through-plated pad with nothing more than a soldering iron.

With nothing more than excessive heat applied for a ridiculously long period of time, yes.. And I can destroy any PCB with nothing more than a screwdriver used incorrectly, I don't see the difference here.

Additionally, because plating is very thin/weak, it can crack over time due to it thermally expanding at a different rate than the fiberglass. 

Crap or faulty plating, yes, quality work, not so much. Decent modern PCBs are very durable, plus you get a component lead in the larger ones as support. Smaller ones can be plugged with solder, epoxy, or even fully plated.

You can do a lot with home manufacturing, and rivets aren't bad, but.. they're huge. I do home etching for prototyping - but I can't do 4-layer boards, .25mm vias, plated holes (yes, you do need them - again, can't do small holes and via-in-pad with rivets), getting mask between 0.4mm pitch QFNs just isn't happening with home gear no matter how experienced you are with the silkscreen process.. All of these are easily available almost as cheap (if not cheaper) than a home-made process, in quantity, with 100% testing. I'm finding it hard to call my laser printer and an iron 'better' based on the ability to do fancy graphics which don't impact function.

I, too, would love to see a guide on doing multi-colour silkscreening affordably at home, as it could be applied to a professionally manufactured PCB after the fact just as easily as a home etch..

[MaximRecoil]

Why would the PCB deform?

Because you just crushed a piece of metal around it and then subjected it to intense thermal cycling.

That's a one-time thing and very localized. Your assertion was with regard to PCB deformation that would cause a solder joint to crack, which would have to happen sometime later down the road. By the way, old solder joints often crack, and it has nothing to do with PCB deformation; a PCB isn't even always involved.

Old equipment. Once upon a time, through-plating was an expensive procedure.

Nice to know that factories used rivets once upon a time. In my experience, old equipment tends to be rather reliable. By the way, what exactly did you repair with these rivets? If it was cracked solder joints, that's not the rivet's fault. I've repaired lots of cracked solder joints; they are especially common on old CRT TV/monitor chassis, and none of them had riveted PCBs. I've never even seen a factory-riveted PCB before.

With nothing more than excessive heat applied for a ridiculously long period of time, yes.. And I can destroy any PCB with nothing more than a screwdriver used incorrectly, I don't see the difference here.

You don't need to apply heat for a "ridiculously long time" to damage a through-plated pad. Also, a soldering iron is something normally used on pads, either for the initial installation of a component or rework. Vandalizing pads with a screwdriver isn't normally done. I've come into possession of a lot of old PCBs that had damaged pads from previous owners' handiwork with a soldering iron. Had they been riveted through holes, they wouldn't have been damaged. I've yet to acquire one that someone mutilated with a screwdriver. 

Crap or faulty plating, yes, quality work, not so much. Decent modern PCBs are very durable, plus you get a component lead in the larger ones as support. Smaller ones can be plugged with solder, epoxy, or even fully plated.

That's true, but it would never be an issue with copper or brass rivets.

I, too, would love to see a guide on doing multi-colour silkscreening affordably at home, as it could be applied to a professionally manufactured PCB after the fact just as easily as a home etch..

It is a lot of work. For full-color process printing you need to make 4 screens, or more than 4 for some variants of simulated process, which tends to give better results than standard CMYK (4-color) process. As for affordable, it's affordable for me, because I'm already a screen printer, so it is just a matter of the cost of screen printing ink (and it takes very little of that for each print), waterproof inkjet film for making the film positives, and however much emulsion I'd use to make the screens. If you're not already set up for screen printing, there would be a pretty big startup cost.

The basic procedure for CMYK process printing is like this:

1. Separate your high-resolution artwork into CMYK. This can be done in Photoshop using channels, or there are specialty applications for it as well, such as "UltraSeps". 

2. Print your 4 film positives as halftones using a PostScript printer, or any printer combined with specialty RIP software. You can use [free] Ghostscript like in my tutorial that I linked to, or there are [expensive] dedicated software packages for it, such as Accurip (about $500).

3. Use your 4 film positives to burn 4 screens which you have already coated with emulsion and allowed to dry for at least 4 hours or so. You'll want relatively high mesh count screens, such 220.

4. Once you have your screens burned, washed out, and dried, you need to register them. Some people register to the film and others register to a test print. It's a matter of positioning each screen properly so that they all line up with each other and with where you want to print on the substrate, and locking them all down so that you can move/rotate them out and when you move/rotate them back in, they are still registered.

5. Print. Some people flash cure between each of the 4 screens and others print wet on wet. It depends on the effect you're going for and the type of ink you're using.

ETA: I'll add that a lot can be done with just one-color screen printing (a lot less work than 4-color process, no color separation or registration required, and the startup cost is far less, because you technically can use a pair of hinge clamps instead of an expensive multi-color screen printing press), from simple spot-color art such as the famous Che Guevara T-shirt to black & white photographic images, like you see every day in the newspaper (which are also printed with a one-color halftone process).

[T3sl4co1l]

I would trust rivets less than plating -- the FR-4 creeps over time, and the unusually meaty rivet will fatigue away from the board (delamination) or trace (cold solder joint).

A copper rivet is going nowhere, provided it is set properly. It is the same idea as a crimped connection, like every single connection in the wiring harnesses of your car.

You aren't even reading the words I'm writing.

I said the FR-4 creeps over time.

Does that not make sense to you?

Which therefore proves this entire thread is futile, and honestly, puts a lot of proof towards the "troll" suggestion.

It will be a lot stronger, because it's far thicker. It is made from sheet metal, while the thickness of plating is measured in the millionths of an inch. You could build a structure using those rivets as fasteners, such as a sheet metal cabinet or attaching body panels to a car. You can't hold a substantial structure together with plating; not even close.

You really are stupid, aren't you?

I normally try to avoid calling out peoples' ignorance, let alone stupidity, because I normally want to help, as frank as I am.

But this can only be willful stubbornness; trolling.

Electroforming (another term for electroplating) produces perfectly sound material, especially if some postprocessing is allowed.  Macroscopic thicknesses are standard practice.

They wouldn't be looking at it for rocket nozzles if it weren't legitimate.
https://books.google.com/books?id=BfR0LexeuREC&pg=PA564&lpg=PA564&dq=electroformed+rocket+nozzle&source=bl&ots=PdgrC57_wU&sig=T7E633GCz3Nz1Z4Wrm7snlXfDOE&hl=en&sa=X&ved=0ahUKEwi-_eWBtofRAhUlwYMKHc6zBIwQ6AEILzAH#v=onepage&q=electroformed%20rocket%20nozzle&f=false

The "foil" that a board starts with, is perfectly sound foil in the first place.  A thousand microinches of copper is a far sight stronger than your average roll of kitchen aluminum foil.

Anything can be measured in microinches, but that doesn't mean it's useful to do so.  Twenty thousand microinches is a lot of foil.

I'd like to see you go at this with a pair of pliers, trying to peel it apart!
https://www.eevblog.com/forum/eda/20-ounce-pcb!/


Tim

[EEVblog]

Making PCBs at home is great, a useful skill, and one that I regularly use for quick one-offs or prototypes. 
However, trying to make PCBs at a "factory quality" is a foolish waste of time, even without factoring in the economics.

This.

[MaximRecoil]

You aren't even reading the words I'm writing.

I read what you wrote perfectly well.

I said the FR-4 creeps over time.

Does that not make sense to you?

And I said the copper rivet is going nowhere (which is relative to its original installed position in the board), which means the amount of "creep" is insignificant.

Which therefore proves this entire thread is futile, and honestly, puts a lot of proof towards the "troll" suggestion.

False. "Troll" doesn't mean "someone who disagrees with me".

You really are stupid, aren't you?

Comical irony, from someone who established himself as an idiot by asserting that the fiberglass will "creep" out from under a copper or brass rivet which is mechanically locked into a hole.

I normally try to avoid calling out peoples' ignorance, let alone stupidity, because I normally want to help, as frank as I am.

But this can only be willful stubbornness; trolling.

Comical Irony Alert: Part II. And given that you didn't actually present an argument here, your tacit concession on the matter is noted.

Electroforming (another term for electroplating) produces perfectly sound material, especially if some postprocessing is allowed.  Macroscopic thicknesses are standard practice.

They wouldn't be looking at it for rocket nozzles if it weren't legitimate.
https://books.google.com/books?id=BfR0LexeuREC&pg=PA564&lpg=PA564&dq=electroformed+rocket+nozzle&source=bl&ots=PdgrC57_wU&sig=T7E633GCz3Nz1Z4Wrm7snlXfDOE&hl=en&sa=X&ved=0ahUKEwi-_eWBtofRAhUlwYMKHc6zBIwQ6AEILzAH#v=onepage&q=electroformed%20rocket%20nozzle&f=false

The "foil" that a board starts with, is perfectly sound foil in the first place.  A thousand microinches of copper is a far sight stronger than your average roll of kitchen aluminum foil.

Anything can be measured in microinches, but that doesn't mean it's useful to do so.  Twenty thousand microinches is a lot of foil.

Utterly irrelevant in this context. The plating in the hole of a via or pad of a typical PCB is incredibly thin and weak; it provides very little structural support; essentially zero in comparison to a sheet metal rivet. That is why it is so easy to lift a pad, even if you're not actually trying to do so. To rip the flange off a rivet requires considerable effort, and you'd need a pair of pliers to do so.

I'd like to see you go at this with a pair of pliers, trying to peel it apart!
https://www.eevblog.com/forum/eda/20-ounce-pcb!/


Tim

Given that that's far from a normal/typical PCB, it's a non sequitur.

[rollatorwieltje]

Just because the file format doesn't support the formal definition of a bezier curve doesn't mean you can't use it. Just approximate it with a shitload of straight lines, nobody cares. It's an editor feature. Your PCB will be rasterized anyway. I've seen several PCB tools that can do flowing lines, i.e. TopoR.

Besides, what is a "better" pcb? Try doing 6 thou traces at home, or solder mask for stuff like SSOP packages. I would need a substantial investment in tools for that, meanwhile services like Seeed and OSHPark can deliver perfect boards for almost nothing.

[MaximRecoil]

Just because the file format doesn't support the formal definition of a bezier curve doesn't mean you can't use it. Just approximate it with a shitload of straight lines, nobody cares.

It would take forever to approximate a bezier curve with straight lines. Even if each line were only one pixel long (~0.0139" at the standard monitor PPI of 72), it would end up looking "polygonish". For it to look smooth, you'd have to draw at least 300 straight lines for each ~inch of the bezier curve.

It's an editor feature. Your PCB will be rasterized anyway.

If you could just send them a raster image to begin with (or a standard graphic design-type vector file which can instantly be rasterized), that would work fine. Someone above said you can do that, though I've never heard of a PCB manufacturer that accepts raster or .AI, .SVG, etc., files.

I've seen several PCB tools that can do flowing lines, i.e. TopoR.

Does that program allow you to draw flowing lines manually with a bezier tool?

Besides, what is a "better" pcb?

Copper or brass rivets are more durable than a typical through-plated hole. Unlimited drawing is better than crude drawing, and that applies to both the copper layer and the silkscreen.

Try doing 6 thou traces at home, or solder mask for stuff like SSOP packages. I would need a substantial investment in tools for that, meanwhile services like Seeed and OSHPark can deliver perfect boards for almost nothing.

Illustrator and other vector graphics programs have no problem drawing ultra fine lines. But to effectively transfer them into the real world, you need a good printer/imagesetter and a good vacuum-based UV exposure unit.

[ruffy91]

But to effectively transfer them into the real world, you need a good printer/imagesetter and a good vacuum-based UV exposure unit.

When you're at this point your home IS a factory.

Also you would need a really good printer or a CtP printer.