I have been making PCBs like this since 2013, using a cheap CNC router. It took me over 3 months to work out the optimum machine settings, software settings, and tool sizes, during which time I nearly gave up. But I can now make really dense microprocessor boards reliably. I use 0.5mm bits for the vias and 0.8mm for the components. If any components won't fit (usually connectors) I enlarge the holes by hand with a 1mm drill. This is much quicker than changing the bit and setting up again to make another pass. It took me ages to realise that to stop a V-bit from breaking you need to increase the depth of cut, not reduce it. I use 3mm dowel pins to register the board. Screws don't make accurate register pins. Via connections can be made really quickly by taking a single length of wire and threading it backwards and forwards through all the via holes. Solder each one and then cut off all the spare wire links.
Genuinely expected this to be another formlabs resin printer used to expose board etch or something similar. Delightful to see such an old school method explained in such detail! Will definitely be checking out the rest of the channel :)
@@thatguythatdoesstuff7448 to use the vbit to pre drill center holes would definitely be a good idea, especially with those small drill bits. Could also speed the retrect movement out of the drilled holes. I would guess that the slow is only needed while pushing the drill bit down, not retracting it. If one want to save time. But yes, pre drill the holes with the vbit would be a drill bit saver. As he said in the end of the video.
You’re getting some really good quality results there, considering the bearings/belts/rails! I think the only advantages the professional machines have is they can be run with faster feed rates and a smoother workflow. Having machined several hundred PCBs on three different LPKF mills, I’ve universally drilled all the holes first, then run over the top surface of the PCB with very fine abrasive paper. This removes any small burrs. I’ve then run the insulation passes (smallest to largest). This consistently yields the best quality results. You then can flip the board and run the bottom insulation passes (if required). If you drill the holes last, you risk the pads being ripped up as the drill bit plunges. This becomes particularly important on special RF PCB materials (laminates) where the copper to dielectric adhesion is not as strong as it is with regular FR4.
I run the LPKF for my college's fabrication lab, and doing the the drills first to deal with losing pads is genius and I'm mad I didn't think of it. Have any other sage advice for these machines?
Another note for others, when you're designing a PCB (using something like Eagle) and you use the AutoRouter, I believe you can set some tolerances to ensure that the routes generated will work with this kind of process. But with SMD components, sometimes things just get too small. You can also set your minimum trace widths.
"This was a very high learning curve for me" - *THAT* is what this is all about - what you learn is something you cannot purchase from a PCB manufacturer Liked the way you used a ready made design so you could concentrate on what you wanted to achieve. Would be interesting to see if you *can* solder on the parts and fully complete a home made Nano.
This wont work, he has no way of plating the vias. If he changed the via holes to 0.9mm you can buy rivets. Or use chemicals, which you may as well use chemicals to etch it then.
> This wont work, he has no way of plating the vias. You can always solder the two sides of the via using a 30AWG wire. It's a lot of work but for one-off prototypes that works very well.
Amazing! Not only is this a great DIY solution it also prevents any unscrupulous third party board maker from stealing your designs. Maybe in a future video you can show how to make a screen and apply solder mask. Also, do you insert metal sleeves in the vias or just leave as is? Heck, with a little practice you can have a tidy little business building unique devices for people.
This is why I make my own PCBs, sure they are cheap from companies like jlpcb but I am not a production house that needs a ton of PCBs. I only make one-offs for my own projects and I can get them done in very short order. I use a "normal" CNC engraving machine though, so my controller is grbl based. Originally it came with a parallel port controller but I converted that to grbl and kept the original stepper drivers/power supplies. That was one of the first PCBs that I made, a shield that plugs into an UNO and breaks out the pins to a parallel cable to run the original controller without any additional mods. In fact the only modification that I made to the original controller was to run a wire from the new probe pin (from the arduino) to a banana jack on the case. For layout I use Diptrace, and for G-code I use CopperCAM. One thing I am going to try next is exporting to a DXF or similar format and try more advanced toolpaths in a CAD package. This will actually require a bit of extra time processing the file in CAD to make it suitable for a toolpath but I wanted more options that cannot be had from CopperCAM and other "engraving" software. I want to be able to ramp in and easily set multiple depths, these are things that I can do easily in my CAD software.
@@Tech2C No, thank you for sharing. Sometimes I think I would like to make videos showing all the projects that I do, but I'm far to busy to produce such content. Perhaps some day I spose.
Interesting method, and quite intense amount of detail for each operation. Thanks for taking the time to develop and demonstrate it. I have have very good success making prototype PCB using the laser etch of black paint (I use Eagle image export into Lightburn), then Muriatic acid (HCL) and hydrogen peroxide solution to etch away copper. Very clean and much faster
What a great result! I have been using relatively inexpensive pyramidal drill bits, of 60 degrees, and using a depth of 0.05mm and single pass. Also, as a hint, I was able to make my own vias by placing a single wire of copper of the same thickness as the via and 2mm long, and smashing it from both sides until it makes contact with the copper clad without soldering. The result can be good, just requires a bit of practice.
Are there 3d printers that can print copper on a board? Seems like much copper is wasted with this milling approach. Even through the copper board is not expensive, wouldn't printing copper be a good (possibly better) solution?
GREAT !! Thank You , Sir but HOW do we get the reference designators onto the PCB? That thick white "paint" we see on commercial PCBs is just TOO SMALL for DIY, .....or is there a WAY ? Thank You !
I mill PCBs at home before sending off to fab shop. Usually for anything complicated, mostly SMD I would just use toner transfer and etch the boards. It's way faster and less dusty. I'd add an enclosure or use mineral oil, the FR4 dust is extremely hazardous - takes decades, but it'll get you eventually.
LPKF machines makes marks centers on small holes with engraver bit so small drill bits under 0.8mm does not wiggle. you may use same technique if it is possible in flat cam.
Very good video ! Looks like a really good quality PCB ! Just one thing though, we can see two vias under what seems to be the processor. That will be quite difficult to solder them flatly enough, as there will be a chip on top of them ... Maybe you could find some small rivets to go through ?
@@Tech2C All those details are what makes your video stand out for good reasons and make it credible. Unlike Sanladerer’s which is mostly useless, greatly uninformed and only makes its way to the recommendations because of the channel’s overall profile.
This is really excellent :) ! Tank you so much ffor your work, and indeed I wasimpressed by the pcb quality level you reached, with clearly explained parameters. It might help me on project.
The tiny holes usually use VERY high speed drilling. 200K-250K RPM is normal and 300K is sometimes used. It takes high speed to get the cutting edges of the tiny bits to go any reasonable speed since they are so tiny.
@Tech2C 22:00 try replace 0.3mm drilling with etching with V-bit, but only to 0.1mm depth, it will make small holes to the copper, then you can drill with 0.3mm drill bit ;)
i think the bits that did not get cut out are because of the way flatcam decides where to put cutting paths. .... if the distance is too close it will simply skip it and leave a bridged area.
Oh i was watching that tiny drill bit, waiting for it to snap! - You can get drill bits that are a bit shorter, and as such less flexible so there's a small chance of it going dancing on the PCB. Regarding doing the "center drills" with the V-bit - that's a good habit of doing.
Was hoping to see him do the vias. Do you use solder wire that fits in the .3mm hole, stick it through and solder one side then go back and cut it down/solder other side? I don't think you could just solder it without some kind of wire going through, no way it will flow through the hole on its own.
When I was at school I had a holiday job one year hand soldering circuit boards for traffic light controllers, probably 40 years ago. The boards were double sided tinned but no solder mask and the vias were made using a roll of tinned wire poked through and soldered on both sides. I think the factory installed a wave solder machine not too long after that.
I'm glad you went into z probing. You really can't get good results milling copper clad without it, at least not without buying the expensive copper clad and using a fixture. I've got one of those cheap desktop CNC engravers and I simply couldn't get consistent results without doing z probing (using bCNC + grbl since it's not a converted printer). Also like the idea of using screws for alignment, I ended up making a 3d printed clamping jig but screws would probably be a lot easier. FlatCAM is a little quirky, but it's so much more capable than the other software like pcb2gcode. You might want to add a blurb about dealing with the dust, though. Especially fiberglass resin boards, the dust from milling is *extremely* dangerous and you really need to be careful about breathing it.
Yes I was wearing safety glasses, ear muffs and a breathing mask whilst in the room. I usually mention the saftey gear I wear but I didn't in this video. FlatCAM is amazing for free software. I can live with the quirks as editing the exported text files is easy enough. Cheers!
Tech2C i have a modified version of flatcam that works directly with marlin, have to polish it up and make an installer for it but that saves a lot of time ;) tell me if you want to beta test it.
Any updates which make it work better with Marlin is welcome. Will your changes be merged into the main FlatCAM branch? I'd like to see the M211 S0 included and F gcode fixed. If possible expand the feed rates to include plunge, retract and travel Z speed. And while I'm wishing for unrealistic changes, include auto leveling command G29 with the boards co-ordinates box in the gcode to be auto levelled... :-)
What if you drill the board underwater? wouldn't that hinder the dust from fusing back onto the board? Or maybe hook up a vacuum cleaner next to the drill bit to instantly get rid of all the dust. Just a thought.
I would have got rid of all the excess unused copper on both sides. May take longer but would be easier to look at the circuit for my eyes at least. Cool AF though! I'm torn in what to purchase as far as a printer or mill myself. I'm looking for a 3d printer with a large usable surface area with the ability to 3d print, but also mill too, and with the precision yours has. Excellent vids so far though. Very informative
The benefit with milling PCBs are saving time during prototyping. As soon as the design is defined I use a manufacturer. To mill PCBs as an alternative to manufacturing will normally never be done.
Why don’t you leave holes in the pad areas when the v-bit cuts the traces so the drill have a centering hole to start? Won’t that produce better aligned holes? Thanks.
In my experience with pcb's, I noticed that , as the drill makes contact with the surface, there is going to be a bit of pressure, as the z-axis keeps moving the built plate upwards, but the drill did not grab into the copper layer to drill the hole. At least, with regular drills. Also, since the spindle would run at around 20K rpm, it tends to run away across the surface, upon this first contact. Especially with such thin drills, or when the drill is not sharp enough. Might be a sign, the drill is not exactly perpendicular to the plate, or the spindle attachment is a bit flexible; it cannot handle strong forces. Spot drilling might help, or maybe you can fix an attachment for a push center pointer. Anyway, I think this printer did an amazing job. I did not expect that with a belt driven system. Nice.
Here's a thought. Laser cut Kapton tape with the negative image (ie kapton hole is where there is no copper). You could place the kapton tape on the pcb to start with. Once the lasering is done, chemically etch the exposed copper. potentially you could do both sides of the pcb at once assuming you lasered the kapton on both sides. Peel kapton off after etch and you'll have a nice set of traces left. Drill final holes and plate. You could also laser cut kapton with the soldermask and the stick that on the finished pcb.
just laser printing on photo paper then transfering the ink onto the double cladded copper pcb then etching with the ink as a mask would be more practical which is the standard way. The whole point of this video is to get rid of the need to use etching chemicals
Can you show me a link please. I used to do laser print onto acetate and the photo expose. The issue is that toner is not 100% light blocking so you get defects on the exposure. Direct cutting of Kapton is better in so many ways. But laser etching would be perfect.
Hello, I have been using your really amazing work for a long time. I built HyperCube and followed the same reasoning for the printer to have more functions. I also wanted to be able to create PCBs, but I discovered JLCPCB and the whole concept turned out to be completely unnecessary, just at the price JLCPCB offers and above all quality, creating your own PCBs is simply pointless ... . Only one can attach to the price for a shipment which is (in my case) 4x bigger than the price for the PCB itself, but still the price for the whole order is reasonable. Immediately I say that no one pays me for any advertising just yesterday I got my order and I am very impressed with the quality. Greetings from Poland.
I guess one has to pre-plan the alignment hole locations so that the homing action of the printer does not wipe out a drill bit if it passes over a screw head? Tip on finishing a PCB, use "liquid tin". Stuff is expensive but makes a world of difference when soldering. Good video, I might stick to etching my PCBs but use the Hypercube for drilling. Maybe speed up the drill and spot-drill as well to reduce the drill dancing? Cheers,
If you added some method of cooling like with water or air, you wouldn't need to do two passes because the copper wouldn't heat up enough to melt and move around into other traces.
I gave a thumbs-up already some while ago to this video. But your results are so gorgeous, that I need to comment as well. I have tried in the past to make doubled side boards on my cnc, and I would rather blame the screws for the misalignment (I used sewing pins). Also, if you start with a board, that is just big enough to have the alignment pins, you can easily flip it around without getting affected (so much) by the bed being misaligned, or a bit of dirt left on the bed
Looking awesome. I went the CNC DIY way because local post offices delays international packages by up to 6 months... Getting pcb made in China needs really a lot of patience.
would it be a simpler method to use a fine tip pcb pen mounted in a 3d printed holder and paint the traces then simply etch them? no mechanical stress, no vibrations for the printer. there are even markers with a tip of less than 0.1mm but I don't know if
@@Tech2C thanks for replying. i understand what your videos explain. but what I meant was to drill holes using the method in the video but instead of milling traces, I think painting them would extend the 3d printer life indefinitely and will allow you to make very fine traces.
''your doing it wrong'' - the 0.3mm drill bit should not be drilling into flat copper, the V bit should have made a centre punch hole for the bit, this is done when etching circuit boards by etching a ''hole'' for the bit..
seeing how the drill bit is vibrating at the begining of the hole drilling it seams like it's gliding a bit over the copper. That is confirmed seeing the end result (holes not centered on the pads). That's preaty tipycal, coper is tought to enter. To avoid that i would advice you to have the hole's center slightly marked with the V bit befor the actual drilling, then the drill will self aligne in the prehole. Just my 2cts...
Could you set your Z-axis Max Rate in the firmware to 15, and leave it at 300 in the g-code, thus the firmware will do everything at 300 except the firmware max rate will take priority over the maximum z speed? I think editing the g-code may be unnecessary.
Thats what i thought was going to happen by the title of the video. He would need to heat up the board and use a 0.2mm nozzle, even then it may not stick.
Thanks for sharing. My advice buy a 2500 mW laser. Paint your pcb board black and laser engrave your design on it. Through it in an etching solution and you will get 100 times finer PCBs. Based on your design, the process might be also faster.
Cheers. 100 times finer than the 0.2mm I achieve here is 0.002mm? I guess you still have to drill holes and cut out the board, so speed may not be any faster?
Wow I see so much success, now, tech-san, try make it better! Make a cool color protective layer before it begins. 1 automate the swapping tools and drill 2 automate magnetic check sensor detach (aligators) 3 automate (spin 180 degree) plates. 4 Protetive layer be scrapped by a tiny metal brush tool right at the end. 5 With tree fingers an automated selecting pcb compontents 6 Automate ball placing for ball, and paste placing for... 7 Automate set up tool for soldering with hotgun chips, and ball. 8 - Automate soldering! (wow that one should also be used to add adicional tin trails to already cut out pcbs) 9 - Nirvana. The initiation is done.
Nice video,, but at this point, I'm using various board houses these days. I don't mind the turn-around time and the price is certainly right for a fully-formed professionally-manufactured circuit board complete with solder mask.
How Ironic. I just soldered an Arduino nano from a bare PCB! I did not use a stencil, but I used some solder paste and a steady hand. I had just found a bunch of them in my school's advanced soldering room and felt like challenging myself.
Nice work. I've been thinking of doing something similar. I'm not going to try to solder mask the whole thing just the places under the ICs that aren't visible. So, the solder will not go there and cause a short.
You are doing this series of other interesting stuff to do with a 3d printer, I rebuild the Hypercube myself with the parts of my China i3 but I often was wondering if it is possible to 2,5D Scan of Objects with a 3D Printer, like a Mainboard other an other PCB to create a custom Heatsink for it, would be very interestet in stuff like that.
I was hoping this would be showing a new conductive material to 3D print boards. I’ve seen a couple of board 3D printers. Very good, but too expensive for my use (around $2,000 each for the basic models).
merhaba, ben 0.1mm 10 derece V kazıma ucu kullanıyorum. Fakat böyle işlemci veya 0.2 mm sinyal yolu kazıyacak iken ortada yol falan kalmıyor. Z cut 0.035 mm olarak ayarlıyorum. Sizce nerede yanlış yapıyorum. Teşekkürler
Yes, your result looks nice - but is it worth the effort. While I truly see the usefullness of CNC engraving, I really don't think it's a good choice for pcb making. You might argue that toner-transfer is mostly only working well for single-sided pcbs and that's true, but you can make a lot of designs single-sided, even using smd microcontrollers. The counter-argument of handling "nasty" chemicals is counter-acted by the huge amount of fine glass-fiber dust which milling creates. I like the automated drilling very much - and for toner-transfer I always include a drill spot in the design so that the drill (manually operated on a press) does not wobble but has a point to grab directly as you said yourself later in the video. If you had a harder sacrificial layer underneath most probably your standard drill bit would have been sufficient for drilling the 0.9mm holes.
Your video is sponsored by a PCB manufacturer, but you show how to make PCBs at home?
The irony is real!
Better quality, make the prototype at home then order 100s from them $$$$
The prototype is real, the the bulk order is real too!
@@vizionthing means you can test the prototype
@@jlebrech Ok Captain obvious?
Haha
I have been making PCBs like this since 2013, using a cheap CNC router. It took me over 3 months to work out the optimum machine settings, software settings, and tool sizes, during which time I nearly gave up. But I can now make really dense microprocessor boards reliably. I use 0.5mm bits for the vias and 0.8mm for the components. If any components won't fit (usually connectors) I enlarge the holes by hand with a 1mm drill. This is much quicker than changing the bit and setting up again to make another pass. It took me ages to realise that to stop a V-bit from breaking you need to increase the depth of cut, not reduce it. I use 3mm dowel pins to register the board. Screws don't make accurate register pins.
Via connections can be made really quickly by taking a single length of wire and threading it backwards and forwards through all the via holes. Solder each one and then cut off all the spare wire links.
great insight
Brilliant
Thanks for the tips
Interesting and thanks for sharing.
Genuinely expected this to be another formlabs resin printer used to expose board etch or something similar. Delightful to see such an old school method explained in such detail! Will definitely be checking out the rest of the channel :)
That 0.3mm drill sure scared me. I thought it was going to snap every time it started wobbling. Definitely spot drill first, next time.
Don't forget to wear you glasses.
I think the spindle speed is the problem. Those bits are made to run from 20 to 100k rpm.
I think a shorter bit would help too .
Agreed. A quick touch with that v bit would center that drill pretty quickly.
@@thatguythatdoesstuff7448 to use the vbit to pre drill center holes would definitely be a good idea, especially with those small drill bits.
Could also speed the retrect movement out of the drilled holes. I would guess that the slow is only needed while pushing the drill bit down, not retracting it. If one want to save time.
But yes, pre drill the holes with the vbit would be a drill bit saver. As he said in the end of the video.
You’re getting some really good quality results there, considering the bearings/belts/rails! I think the only advantages the professional machines have is they can be run with faster feed rates and a smoother workflow.
Having machined several hundred PCBs on three different LPKF mills, I’ve universally drilled all the holes first, then run over the top surface of the PCB with very fine abrasive paper. This removes any small burrs. I’ve then run the insulation passes (smallest to largest). This consistently yields the best quality results. You then can flip the board and run the bottom insulation passes (if required).
If you drill the holes last, you risk the pads being ripped up as the drill bit plunges. This becomes particularly important on special RF PCB materials (laminates) where the copper to dielectric adhesion is not as strong as it is with regular FR4.
I run the LPKF for my college's fabrication lab, and doing the the drills first to deal with losing pads is genius and I'm mad I didn't think of it. Have any other sage advice for these machines?
How do you plate the holes in that process? Like for vias?
My heart literally jumped when the 0.3mm bit was trembling on the surface, and all of a sudden went in like butter. What a moment :-D
hehe yep :P
@@Tech2C the end result was fine I must say! It turned out great. Just nearly had a heart attack :-p
i wish i had the time for a project like this. they need to just come out with an affordable machine that does this right out the gate. good work
first of drilling deep hole may do bit hole sixty-degress drillfreze. its be of centre deep hole and small drill not vibration or broken
Spot drilling ..... would have lowered my heart rate a lot during the drilling pass :)
Another note for others, when you're designing a PCB (using something like Eagle) and you use the AutoRouter, I believe you can set some tolerances to ensure that the routes generated will work with this kind of process. But with SMD components, sometimes things just get too small. You can also set your minimum trace widths.
"This was a very high learning curve for me" - *THAT* is what this is all about - what you learn is something you cannot purchase from a PCB manufacturer
Liked the way you used a ready made design so you could concentrate on what you wanted to achieve.
Would be interesting to see if you *can* solder on the parts and fully complete a home made Nano.
This wont work, he has no way of plating the vias.
If he changed the via holes to 0.9mm you can buy rivets.
Or use chemicals, which you may as well use chemicals to etch it then.
> This wont work, he has no way of plating the vias.
You can always solder the two sides of the via using a 30AWG wire. It's a lot of work but for one-off prototypes that works very well.
PCB looks professionally built, I tried making with cnc machine but couldn't get result as you build. Very nice and thank you for sharing.
Amazing! Not only is this a great DIY solution it also prevents any unscrupulous third party board maker from stealing your designs. Maybe in a future video you can show how to make a screen and apply solder mask. Also, do you insert metal sleeves in the vias or just leave as is? Heck, with a little practice you can have a tidy little business building unique devices for people.
This is why I make my own PCBs, sure they are cheap from companies like jlpcb but I am not a production house that needs a ton of PCBs. I only make one-offs for my own projects and I can get them done in very short order. I use a "normal" CNC engraving machine though, so my controller is grbl based. Originally it came with a parallel port controller but I converted that to grbl and kept the original stepper drivers/power supplies. That was one of the first PCBs that I made, a shield that plugs into an UNO and breaks out the pins to a parallel cable to run the original controller without any additional mods. In fact the only modification that I made to the original controller was to run a wire from the new probe pin (from the arduino) to a banana jack on the case.
For layout I use Diptrace, and for G-code I use CopperCAM. One thing I am going to try next is exporting to a DXF or similar format and try more advanced toolpaths in a CAD package. This will actually require a bit of extra time processing the file in CAD to make it suitable for a toolpath but I wanted more options that cannot be had from CopperCAM and other "engraving" software. I want to be able to ramp in and easily set multiple depths, these are things that I can do easily in my CAD software.
Great! Thanks for sharing!
@@Tech2C No, thank you for sharing. Sometimes I think I would like to make videos showing all the projects that I do, but I'm far to busy to produce such content. Perhaps some day I spose.
Hi, I was wondering is it possible to make a pcb for a 12kw inverter & a solar charger controller I'm trying to make my own
why you need to check heights for drilling? i think thats not necesary
Interesting method, and quite intense amount of detail for each operation. Thanks for taking the time to develop and demonstrate it. I have have very good success making prototype PCB using the laser etch of black paint (I use Eagle image export into Lightburn), then Muriatic acid (HCL) and hydrogen peroxide solution to etch away copper. Very clean and much faster
What a great result! I have been using relatively inexpensive pyramidal drill bits, of 60 degrees, and using a depth of 0.05mm and single pass. Also, as a hint, I was able to make my own vias by placing a single wire of copper of the same thickness as the via and 2mm long, and smashing it from both sides until it makes contact with the copper clad without soldering. The result can be good, just requires a bit of practice.
Good stuff! I mentioned at the end of this video to solder a wire through each via, but I like your approach!
Smashing?
@@wf2v I used pliers to press the both ends of the wire against the board.
Kudos for the time taken to put such clear content together!
Thank you!
Hi! My interest is large two-sided copper clad diy pc boards, with a mix of surface and through-hole. What have you got? With process.
Are there 3d printers that can print copper on a board? Seems like much copper is wasted with this milling approach. Even through the copper board is not expensive, wouldn't printing copper be a good (possibly better) solution?
GREAT !! Thank You , Sir
but HOW do we get the reference designators onto the PCB?
That thick white "paint" we see on commercial PCBs is just TOO SMALL for DIY, .....or is there a WAY ? Thank You !
Thank you for the review and additional knowledge. Another excellent review from Dave!!!!
Amazing video, but, how can I use flat cam to create pcbs with a ender 3 for example?
Great quality for home made machine !! Thank you for sharing your experience with us !
I mill PCBs at home before sending off to fab shop. Usually for anything complicated, mostly SMD I would just use toner transfer and etch the boards. It's way faster and less dusty. I'd add an enclosure or use mineral oil, the FR4 dust is extremely hazardous - takes decades, but it'll get you eventually.
There are also little copper rivets you can use to create through-hole connections
Sounds like David from the EEVBlog :D
Good Video by the way!
saw this comment everywhere :)
I can't even watch EEVBlog because of his voice, is like nails on a chalkboard. This video was much more soothing ;-)
@@tux1968 hes just got a really strong aussie accent,
i kinda like it (though im also an aussie)
agreed
its like Dave, but with a much more mellow voice.
LPKF machines makes marks centers on small holes with engraver bit so small drill bits under 0.8mm does not wiggle. you may use same technique if it is possible in flat cam.
Thanks for showing us what else our 3D printers are capable of! This is pretty cool.
Thanks SaberShip
Very good video ! Looks like a really good quality PCB !
Just one thing though, we can see two vias under what seems to be the processor. That will be quite difficult to solder them flatly enough, as there will be a chip on top of them ... Maybe you could find some small rivets to go through ?
Great video! Lots of great info and lots of care taken to present it all very clearly. Thanks! Please keep 'em coming!
I was concerned there would be too much info. 30mins video is a bit long but I'm glad it's done. Cheers!
@@Tech2C All those details are what makes your video stand out for good reasons and make it credible. Unlike Sanladerer’s which is mostly useless, greatly uninformed and only makes its way to the recommendations because of the channel’s overall profile.
Really enjoy the 3d printer as a swiss army knife content.
Great job, buddy! It works for me. Thank you!
This is really excellent :) ! Tank you so much ffor your work, and indeed I wasimpressed by the pcb quality level you reached, with clearly explained parameters. It might help me on project.
If you painted the board with solder mask then ran a compliant/spring loaded scuffing bit to remove the mask you would be in perfect shape.
The holes need plating though, and it's very hard to do at home :-)
The tiny holes usually use VERY high speed drilling. 200K-250K RPM is normal and 300K is sometimes used. It takes high speed to get the cutting edges of the tiny bits to go any reasonable speed since they are so tiny.
@Tech2C 22:00 try replace 0.3mm drilling with etching with V-bit, but only to 0.1mm depth, it will make small holes to the copper, then you can drill with 0.3mm drill bit ;)
@Tech2C You said it al 27:00, sorry for spam :D
Haha no problem :P
Thanks a bunch for your very informative video. This got me going with flatcam.
You just convinced me that sending the gerber to the fab is a far better/quicker way, even for my first prototype.
i think the bits that did not get cut out are because of the way flatcam decides where to put cutting paths. .... if the distance is too close it will simply skip it and leave a bridged area.
Why not create pilot holes with the v drill bit for when you want to drill the 0.3 holes?
he literally said he will, next time
Oh i was watching that tiny drill bit, waiting for it to snap! - You can get drill bits that are a bit shorter, and as such less flexible so there's a small chance of it going dancing on the PCB. Regarding doing the "center drills" with the V-bit - that's a good habit of doing.
Excellent video!
Nice work. I have so far failed with pcb milling on a modified T8 CNC and a Cyclone PCB Mill. In both cases, the spimdle runout was ridiculous.
Yes I had to change the flexible shaft I was using as the Z axis wasn't consistent. Such a pain but worth it once you sort it out.
Excellent video , very interesting to see what can be achieved.
The big problem with making your own double sided PCBs with a CNC is that you can't make vias, and will have to add wire jumpers between layers.
Was hoping to see him do the vias. Do you use solder wire that fits in the .3mm hole, stick it through and solder one side then go back and cut it down/solder other side? I don't think you could just solder it without some kind of wire going through, no way it will flow through the hole on its own.
When I was at school I had a holiday job one year hand soldering circuit boards for traffic light controllers, probably 40 years ago. The boards were double sided tinned but no solder mask and the vias were made using a roll of tinned wire poked through and soldered on both sides. I think the factory installed a wave solder machine not too long after that.
I'm glad you went into z probing. You really can't get good results milling copper clad without it, at least not without buying the expensive copper clad and using a fixture. I've got one of those cheap desktop CNC engravers and I simply couldn't get consistent results without doing z probing (using bCNC + grbl since it's not a converted printer). Also like the idea of using screws for alignment, I ended up making a 3d printed clamping jig but screws would probably be a lot easier.
FlatCAM is a little quirky, but it's so much more capable than the other software like pcb2gcode. You might want to add a blurb about dealing with the dust, though. Especially fiberglass resin boards, the dust from milling is *extremely* dangerous and you really need to be careful about breathing it.
Yes I was wearing safety glasses, ear muffs and a breathing mask whilst in the room. I usually mention the saftey gear I wear but I didn't in this video.
FlatCAM is amazing for free software. I can live with the quirks as editing the exported text files is easy enough. Cheers!
Tech2C i have a modified version of flatcam that works directly with marlin, have to polish it up and make an installer for it but that saves a lot of time ;) tell me if you want to beta test it.
Any updates which make it work better with Marlin is welcome. Will your changes be merged into the main FlatCAM branch? I'd like to see the M211 S0 included and F gcode fixed. If possible expand the feed rates to include plunge, retract and travel Z speed. And while I'm wishing for unrealistic changes, include auto leveling command G29 with the boards co-ordinates box in the gcode to be auto levelled... :-)
What if you drill the board underwater? wouldn't that hinder the dust from fusing back onto the board? Or maybe hook up a vacuum cleaner next to the drill bit to instantly get rid of all the dust. Just a thought.
I would have got rid of all the excess unused copper on both sides. May take longer but would be easier to look at the circuit for my eyes at least. Cool AF though! I'm torn in what to purchase as far as a printer or mill myself. I'm looking for a 3d printer with a large usable surface area with the ability to 3d print, but also mill too, and with the precision yours has. Excellent vids so far though. Very informative
The benefit with milling PCBs are saving time during prototyping. As soon as the design is defined I use a manufacturer. To mill PCBs as an alternative to manufacturing will normally never be done.
Why don’t you leave holes in the pad areas when the v-bit cuts the traces so the drill have a centering hole to start? Won’t that produce better aligned holes? Thanks.
In my experience with pcb's, I noticed that , as the drill makes contact with the surface,
there is going to be a bit of pressure, as the z-axis keeps moving the built plate upwards,
but the drill did not grab into the copper layer to drill the hole.
At least, with regular drills.
Also, since the spindle would run at around 20K rpm, it tends to run away across
the surface, upon this first contact. Especially with such thin drills, or when the drill
is not sharp enough. Might be a sign, the drill is not exactly perpendicular to the plate,
or the spindle attachment is a bit flexible; it cannot handle strong forces.
Spot drilling might help, or maybe you can fix an attachment for a push center pointer.
Anyway, I think this printer did an amazing job.
I did not expect that with a belt driven system.
Nice.
He did mention using the v-bit to spot drill for the holes.
Here's a thought. Laser cut Kapton tape with the negative image (ie kapton hole is where there is no copper). You could place the kapton tape on the pcb to start with. Once the lasering is done, chemically etch the exposed copper. potentially you could do both sides of the pcb at once assuming you lasered the kapton on both sides. Peel kapton off after etch and you'll have a nice set of traces left. Drill final holes and plate. You could also laser cut kapton with the soldermask and the stick that on the finished pcb.
just laser printing on photo paper then transfering the ink onto the double cladded copper pcb then etching with the ink as a mask would be more practical which is the standard way. The whole point of this video is to get rid of the need to use etching chemicals
Can you show me a link please. I used to do laser print onto acetate and the photo expose. The issue is that toner is not 100% light blocking so you get defects on the exposure. Direct cutting of Kapton is better in so many ways. But laser etching would be perfect.
I recomend making small pre-holes with V-bit. Then your small drills have longer life, and dont wobble like crazy.
that music was so chill I almost fell asleep
Hello,
I have been using your really amazing work for a long time. I built HyperCube and followed the same reasoning for the printer to have more functions. I also wanted to be able to create PCBs, but I discovered JLCPCB and the whole concept turned out to be completely unnecessary, just at the price JLCPCB offers and above all quality, creating your own PCBs is simply pointless ... . Only one can attach to the price for a shipment which is (in my case) 4x bigger than the price for the PCB itself, but still the price for the whole order is reasonable. Immediately I say that no one pays me for any advertising just yesterday I got my order and I am very impressed with the quality.
Greetings from Poland.
I guess one has to pre-plan the alignment hole locations so that the homing action of the printer does not wipe out a drill bit if it passes over a screw head?
Tip on finishing a PCB, use "liquid tin". Stuff is expensive but makes a world of difference when soldering.
Good video, I might stick to etching my PCBs but use the Hypercube for drilling. Maybe speed up the drill and spot-drill as well to reduce the drill dancing?
Cheers,
Marlin has a setting for homing travel minimum height for just this reason. The tool will lift up to that minimum height before homing back to 0,0.
For holes, make center holes with the same cutter that you cut the tracks with. Then the drill will not jump when drilling.
If you added some method of cooling like with water or air, you wouldn't need to do two passes because the copper wouldn't heat up enough to melt and move around into other traces.
Hi,
very impressive result and well presented video, do you have any information on the "modifications" that were done to the 3D printer?
bye
Have you seen the ANT pcb project?
I just watched a few videos on it. Looks good! It's basically a Core XY CNC machine with a short throw Z axis. I bet it could 3D print too :)
@@Tech2C someone is porting the machine to 2020 profile aluminum frame. files are available on thingiverse.
I gave a thumbs-up already some while ago to this video. But your results are so gorgeous, that I need to comment as well. I have tried in the past to make doubled side boards on my cnc, and I would rather blame the screws for the misalignment (I used sewing pins). Also, if you start with a board, that is just big enough to have the alignment pins, you can easily flip it around without getting affected (so much) by the bed being misaligned, or a bit of dirt left on the bed
Looking awesome. I went the CNC DIY way because local post offices delays international packages by up to 6 months... Getting pcb made in China needs really a lot of patience.
Well in that case this would make a great addition to your work bench.
Where did you get the bit that's present at 14:02?
Why does the drill need to be quite so long? It seems like you could avoid the issue by using a much shorter one.
would it be a simpler method to use a fine tip pcb pen mounted in a 3d printed holder and paint the traces then simply etch them? no mechanical stress, no vibrations for the printer. there are even markers with a tip of less than 0.1mm but I don't know if
That is an alternate method of making PCBs. You would still need to drill holes and cut out the board. Might as well mill the traces too.
@@Tech2C thanks for replying. i understand what your videos explain. but what I meant was to drill holes using the method in the video but instead of milling traces, I think painting them would extend the 3d printer life indefinitely and will allow you to make very fine traces.
Great video! Thank you! it's instructive as entertaining. Pretty good pedagogy
Great work, I love that you used an existing design (the nano). Next you need to do the mask & soldering :-D
Cheers!
@@Tech2C You need an automatic tool changer so you can mill and solder pcbs with the hypercube
awsome. hand solder is totally doable for such.
''your doing it wrong'' - the 0.3mm drill bit should not be drilling into flat copper, the V bit should have made a centre punch hole for the bit, this is done when etching circuit boards by etching a ''hole'' for the bit..
He said that after the finished drills
@@England91 i should have watched to the end lol {insert foot => mouth}
Great video!!!! Thanks!! love everything you did
seeing how the drill bit is vibrating at the begining of the hole drilling it seams like it's gliding a bit over the copper. That is confirmed seeing the end result (holes not centered on the pads). That's preaty tipycal, coper is tought to enter. To avoid that i would advice you to have the hole's center slightly marked with the V bit befor the actual drilling, then the drill will self aligne in the prehole.
Just my 2cts...
where did you by that spiral bit? all the v bit that i buy break after two uses
Hi, is it possible to demonstrate the printer as a sandwich press?
After the Cappuccino video.
Very high-quality video. Thx!
Could you set your Z-axis Max Rate in the firmware to 15, and leave it at 300 in the g-code, thus the firmware will do everything at 300 except the firmware max rate will take priority over the maximum z speed? I think editing the g-code may be unnecessary.
What about applying PLA to the copper then etching it in the classic way?
Thats what i thought was going to happen by the title of the video.
He would need to heat up the board and use a 0.2mm nozzle, even then it may not stick.
Great great great!! I have a CNC mill machine and your results are extremely good!!:)
Nicely done
Thanks for sharing.
My advice buy a 2500 mW laser. Paint your pcb board black and laser engrave your design on it. Through it in an etching solution and you will get 100 times finer PCBs. Based on your design, the process might be also faster.
Cheers. 100 times finer than the 0.2mm I achieve here is 0.002mm? I guess you still have to drill holes and cut out the board, so speed may not be any faster?
Careful when vacuuming the PCB with a endstop wire still attached to your Arduino, you can get a static discharge
Have you tried out ChiliPeppr and the Eagle BRD Import widget? Just drag n drop your Eable board and mill the PCB.
Does ChiliPeppr support Marlin firmware?
why plot area in flatcam is black
Wow I see so much success, now, tech-san, try make it better!
Make a cool color protective layer before it begins.
1 automate the swapping tools and drill
2 automate magnetic check sensor detach (aligators)
3 automate (spin 180 degree) plates.
4 Protetive layer be scrapped by a tiny metal brush tool right at the end.
5 With tree fingers an automated selecting pcb compontents
6 Automate ball placing for ball, and paste placing for...
7 Automate set up tool for soldering with hotgun chips, and ball.
8 - Automate soldering! (wow that one should also be used to add adicional tin trails to already cut out pcbs)
9 - Nirvana. The initiation is done.
Nice video,, but at this point, I'm using various board houses these days. I don't mind the turn-around time and the price is certainly right for a fully-formed professionally-manufactured circuit board complete with solder mask.
Yeah it's so cheap these days. If you need a board on same day there really isn't an option though.
I always thought you measured with a calliper at the notched part rather than the thick arm part.
Could you try to use the carbon fiber rods for x-axis for milling ?
How Ironic. I just soldered an Arduino nano from a bare PCB! I did not use a stencil, but I used some solder paste and a steady hand.
I had just found a bunch of them in my school's advanced soldering room and felt like challenging myself.
That is a coincidence. Well done with the soldering, I assume you've tested the board afterwards and it works?
Nice work. I've been thinking of doing something similar. I'm not going to try to solder mask the whole thing just the places under the ICs that aren't visible. So, the solder will not go there and cause a short.
You are doing this series of other interesting stuff to do with a 3d printer, I rebuild the Hypercube myself with the parts of my China i3 but I often was wondering if it is possible to 2,5D Scan of Objects with a 3D Printer, like a Mainboard other an other PCB to create a custom Heatsink for it, would be very interestet in stuff like that.
can i do this with 3018 pro cnc please
Could you please tell which exactly 0.1mm v bit you used for engraving. The link you mentioned in the description is not showing the same bit.
I was hoping this would be showing a new conductive material to 3D print boards. I’ve seen a couple of board 3D printers. Very good, but too expensive for my use (around $2,000 each for the basic models).
I'd like to give you two thumbs up!
Please i need all electronic part nessessary for a arcade joystick .ans where i can to bay et thank You.
merhaba, ben 0.1mm 10 derece V kazıma ucu kullanıyorum. Fakat böyle işlemci veya 0.2 mm sinyal yolu kazıyacak iken ortada yol falan kalmıyor. Z cut 0.035 mm olarak ayarlıyorum. Sizce nerede yanlış yapıyorum. Teşekkürler
Yes, your result looks nice - but is it worth the effort. While I truly see the usefullness of CNC engraving, I really don't think it's a good choice for pcb making. You might argue that toner-transfer is mostly only working well for single-sided pcbs and that's true, but you can make a lot of designs single-sided, even using smd microcontrollers. The counter-argument of handling "nasty" chemicals is counter-acted by the huge amount of fine glass-fiber dust which milling creates.
I like the automated drilling very much - and for toner-transfer I always include a drill spot in the design so that the drill (manually operated on a press) does not wobble but has a point to grab directly as you said yourself later in the video. If you had a harder sacrificial layer underneath most probably your standard drill bit would have been sufficient for drilling the 0.9mm holes.
Awesome video!! Do you have any link to the tool you use for traces engraving? Thanks a lot
why not vacuum while drilling?