UV Solder Mask: Not as easy as I thought

Solder Mask paint sticks to: 1. Bare copper or plated copper. 2. Fibreglass (FR4) 3. Polyimide or Polyester Film (Flex PCB) 4. Other solder mask (white/black Silkscreen) Anyone who's gotten enough cheap PCBs made will notice how they manage to put white component labels over Vias, pads, bare FR4, and the green Solder Mask. The correct way to apply this stuff is via silk screening, you don't need a machine, it's perfectly possible to do with a Squeegee + Silk Screen on frame (which you just custom order locally). UV solder mask would be no different to T shirt paint, so any general guides should do.

I've never had much luck with that UV solder mask liquid, either :) I've tried multiple times to add a solder mask to my home-etched PCBs using that liquid solder resist and overhead projector transparencies, but always had the resist stick to the transparency at least as well as to the board itself, exactly as you experienced here :( I was considering trying cling film at one point, just as you did, so at least you've saved me the time of trying that! :D I don't recall finding that it stuck to bare FR4, bare copper, or tinned copper (using "liquid tin"/stannous chloride) any better than you're seeing it stick to existing solder resist... The only success I've had so far is literally just painting it on with a fine brush, which is functional - it keeps the solder in place, at least to whatever accuracy I could manage with a paint brush - but it's not the most-attractive result. I've since bought some green dry film solder resist ("Dynamask") from a Dutch supplier (https://www.rembrandtelectronics.com/product/dynamask-solder-mask-starter-kit-5-pcs-15x10-cm/), but I've not had a chance to try it out yet. I wonder if some of the steps from their instructions would help with this liquid stuff? e.g. letting the board sit in a dark place for an hour after exposing it, or thoroughly cleaning the board with acetone before starting... As others have said, I think a more powerful UV lamp would help, and is why you had to leave it for ages, but my tests involved a much more powerful lamp and I still had the resist peel off the board even when what peeled off seemed to be totally cured! I get good pre-sensitised PCB results from a £10-15-on-eBay 36W "UV nail polish/gel lamp" that contains four small, 9W fluorescent tubes - the same thing that Big Clive's used a lot in the past. The newer LED-based nail lamps would probably work fine as well, and a lot of them at least claim to use LED packages that emit both 365nm and 405nm. For Pride in 2018 I designed and built what was intended to be an illuminated and animated Pride "badge" with six colours of 3D-printed diffuser stripe in front of 18 LEDs driven by an ISSI IS31FL3218 (used a lot by Pimoroni), controlled by a STM32, and powered by a Li-Ion/LiPo cell with the common Chinese combination of DW01+8205+TP4056+MT3608 for charge & boost, but it ended up being far too heavy to be worn - at least pinned to clothing; maybe it would be okay hung from a lanyard or something! https://imgur.com/a/gGJw4f1 One of the main weight issues was the 2000mAh LiPo cell I used. At the time I didn't know as much as I do now about energy-efficient electronics designs, and my health was such a disaster that I barely had time before Nottinghamshire Pride to get it built with the absolute minimum of functionality, let alone troubleshoot its high battery drain (even when switched off, which is impressive) or get any animations programmed into its STM32! I'd love to revisit that project some day :) Maybe I can figure out how to incorporate the black & brown stripes. You can't really get black or brown LEDs - not counting blown-up LEDs, that is! 😹 Some other points: 1) Ultraviolet is typically divided into 3 main bands - UVA (approx. 400-315nm, aka "black light"), UVB (315-280nm), and UVC (280-100nm) - and two more (overlapping) bands beyond that: "vacuum UV" (VUV, 200-10nm) and "extreme UV" (EUV, 121-10nm). Your UV tube will be UVA, as that's the UV band generally used to make things fluoresce, and for the majority of UV-curing/developing substances (photoresist, glue, nail polish, etc.). I've found that most such things specify light somewhere between 350-450nm, and it's often not super critical that you get exactly the right wavelength, e.g.: • Mega Electronics (RIP) dry film resist: "recommend lamps with peak emissions at 360-380nm"; • Mega Electronics Microtrak pre-sensitised PCB laminates: "spectral response: 350-450nm"; • Bungard pre-sensitised laminates: "maximum spectral response is in the range of 350-400nm" or "maximum spectral sensitivity of about 400nm" (depending on the web page/PDF you read :)) • The resin used in consumer resin-based (SLA, mSLA/LCD, DLP) 3D printers is nominally 405nm, which isn't even UV, technically... It's long been evident that most people are unaware of the UV subtypes and the effects of the different wavelengths, such as people suggesting that their resin-based 3D printer's ~405nm "UV" LEDs could give them a sun tan (nope) or sterilise surfaces (nope). However The Current Situation™ has made that a much bigger problem, as people have been attempting to sterilise things with fluorescent lamps or LEDs that are typically UVA at best, and of course there are endless listings on eBay and AliExpress for "germicidal" lamps that are actually just UVA (often 370-400nm or so), near-UV (around 400-410nm), or even just the "ice blue" colour of LED that seems to have emerged a couple of years ago. Those ice blue LEDs look really nice, but they ain't gonna sterilise anything! (They just happen to produce a colour that's close to the blue light that mercury vapour lamps emit alongside their actually-germicidal wavelength of 253.7nm...) I've no doubt that some of the eBay/AliExpress listings are from sellers that are well aware that what they're selling won't do what they claim, but I suspect that a lot of them are as ignorant of what theyr'e selling as their customers are... 2) Standard glass is good at blocking UVC, and some amount of UVB but to a lesser extent. Regular glass doesn't really block UVA significantly more than it blocks visible light. Wikipedia states, with 3 citations, that "common soda-lime glass such as window glass" passes "about 90% of the light above 350nm" (i.e. UVA) while "blocking over 90% of the light below 300nm" (UVB and UVC) [https://en.wikipedia.org/wiki/Ultraviolet#Blockers,_absorbers,_and_windows] Therefore using glass to hold things in place during processes like this will work fine, and is probably a good idea - it's what I've always done while making home-etched PCBs with pre-sensitised laminates and laser-printed transparencies (printing the designs onto numerous different materials while trying to find the best one). Glass is absolutely vital when you need the kind of precision required for developing boards prior to etching them, as even a small gap between the transparency and the PCB will blur the edges of tracks/pads. In fact, with higher-resolution designs you have to print the design mirrored and then use the transparency with the ink/toner layer against the board, as even the thickness of the transparency itself will give you blurred lines! (Obviously this is all to the best of my knowledge, alongside some brief research/fact-checking just now; I'm not an expert at any of this!) (And, as always, apologies for length - ADHD and medication for same is quite the experience... 😅

Nice video! I like getting to see problem solving in action, so many YouTube videos are just end results or omit failures so it's refreshing to see things go wrong. Looking forward to seeing the flags soldered up!

There's just never enough Watts is there David? It's a shame it didn't work out, but I'm really looking forward to seeing a finished PCB Pride flag!

Enjoyed that thanks David! Always good to see some trial and error. Have you checked the wavelength is right? "UV" technically covers everything from "Just outside of visible" to "Someone call Bruce Banner!".

I struggled with this stuff too, very tricky to use. Also all younger people asking what an OHP is 😅

Try using a nail curer. Big Clive uses them for doing PCB Transfers. I think the UV from the bulb you have isn't true UV you need one that has more UV.

so honest to show us the whole process. will you livestream the soldering ?

Only 4 Watts? Ya need more power captain! Maybe ask the other Watts for help :P

That’s too bad. The acrylic ones look fine though. Looking forward to the next video. Cheers

just mount the components and mask the LED's and spray with paint. use a latex mask from RS to cover the led's before painting and it will just peel off after. or you could dab a bit on the solder pads before putting the solder mask on. used these masks a lot when you are conformal coating PCB's and you need to stop it getting in to pot's and sockets etc.

If the solder mask is cured with a UV source >320nm light, then you should be able to use glass just fine or even FEP films used in resin 3d printers should work as the solder mask shouldn't stick to the polymer film. As all the products I viewed in my brief search just said UV light or the sun, my guess is 405nm light is enough, but as the painted versions looked fine, I'd say just go that route. But if you want to experiment in the future, glass and preferably glass+FEP film or any fluoropolymer film like teflon should work without the soldermask sticking to the film and/or possible the glass. And the coated copper boards are probably the ones your want to go with, maybe with a light pre-sanding to help the mask preferentially stick to the board.

Good morning (5 am here) from Oklahoma-ville. I haven't used any pcb manufacturers yet but I try to support the underdogs. It may stick to the board better if you rough it up with a fiberglass pen. I think the finish on those boards is too smooth.

You may want to try a squeegee to spread the stuff. A pvc card may even work

Get a nail polish curing light, they have way more power and the proper wavelength.

Hmm, sorry it didn't work out as expected :/ I was wondering just the other day if those Inkjet printers with CD tray could be used to print on a PCB 🤔 I had one for years, but it died last summer :/

Would't it be easier to solder the LEDs on first, then paint with acrylic, then carefully clean off the LEDs with a cotton bud.

Thanks David, atleast I don't have to go through the trial and error process waisting my own money.

Also did some experiments with uv solder masks, uv silk screen and photosensitive dry film. I even wrote a python program to generate gcode scripts from KiCad svg plots to control a cnc uv laser. Results are acceptable with the right settings, but it's a very slow process: https://github.com/chsc/laserg

The bank note checker is the reason for the poor result.