If you've ever bought a cheap "UV torch" to check conformal coating on a board, hunt down flux residue, or find a leak in a liquid-cooled PC, you've probably noticed something odd: the whole bench turns purple and you still can't see much.
That's usually not your eyes. It's the wavelength.
This post is a practical look at the two wavelengths you'll actually find on the market, 365nm and 395nm, and which one fits which job on an electronics bench.
The short version
| Job | Better choice | Why |
|---|---|---|
| Inspecting conformal coating coverage | Filtered 365nm | Less visible light, so the tracer glow stands out |
| Spotting rosin flux residue | Filtered 365nm | Same reason: contrast |
| Finding UV-dye coolant leaks | Either, 365nm is cleaner | 395nm works but washes out faint traces |
| Curing UV resin, UV glue or UV solder mask | 395nm (or 405nm) | Most of these materials are formulated for that range |
What's actually different
Both are UV-A. The difference is how close they sit to visible light.
395nm LEDs sit right at the edge of what we can see. A good chunk of their output spills over into visible violet. That's why a 395nm torch looks bright purple: you're seeing a lot of that light directly, and so is everything you point it at.
365nm LEDs sit further into the UV. They look dim to the eye, but that's partly the point: less of their output is visible light, so it doesn't drown out the fluorescence you're trying to see.
The other piece is the filter. Many 365nm torches put a dark glass filter (often called Wood's glass or ZWB2) in front of the LED. It blocks most of the remaining visible light and lets the UV through. A filtered 365nm torch looks almost dark when you shine it on a wall, but anything fluorescent lights up clearly against that dark background.
So the trade-off is roughly this:
- 395nm: cheap, bright to the eye, lots of visible purple, lower contrast.
- Unfiltered 365nm: better, but some visible light still leaks out.
- Filtered 365nm: lowest visible spill, highest contrast, usually more expensive.
Use case 1: Checking conformal coating
Most conformal coatings (acrylic, urethane, silicone) come with a UV tracer added so that you can check coverage after spraying or brushing. Under long-wave UV, coated areas glow and gaps don't.
With a 395nm torch, the purple light bounces off the solder mask, the silkscreen and the components, and thin or patchy areas are hard to tell apart from bare ones. With a filtered 365nm torch, the board itself stays dark and the coating glows, so you can see:
- areas that the spray didn't reach (often the shadowed sides of tall parts)
- coating that has wicked into connectors or onto test points where it shouldn't be
- bubbles and thin spots
Tip: inspect in a dim room and at a low angle. Shadows behind tall components are where gaps hide.
Use case 2: Finding flux residue
Many rosin-based fluxes fluoresce under UV. That makes a UV torch a quick way to see whether your cleaning step actually worked, especially under QFNs, around connector pins and between fine-pitch leads.
Two caveats:
- Not every flux fluoresces strongly. Some no-clean and synthetic formulations show very little, so "no glow" doesn't automatically mean "clean". Test the flux you actually use on a scrap board first.
- This tells you residue is present. It isn't a substitute for proper ionic contamination testing if you need that level of assurance.
Again, contrast is what matters here. Faint residue is easy to miss when everything around it is lit up purple.
Use case 3: Leak checks in liquid-cooled builds
If you run a custom water loop, UV-reactive coolant or a UV dye makes small leaks much easier to find: a drip behind a fitting or a weep at a radiator seam will glow.
A 395nm torch will show an obvious leak. For small weeps and old dried traces, a filtered 365nm torch is noticeably easier to read because the surrounding hardware isn't glowing purple too. The same idea applies to automotive and HVAC dye leak detection.
Use case 4: Curing UV resin and solder mask
This is where 395nm wins.
Most UV-curing resins, glues and UV-curable solder masks are formulated to cure in the 395–405nm range. A 395nm torch is usually the right (and cheaper) tool for spot-curing a resin fix or a solder mask repair. A filtered 365nm torch is built for seeing, not for pumping out curing power, and it may cure some materials slowly or unevenly.
If you do both inspection and curing, it's normal to own one of each.
A two-minute test for any UV torch
You don't need a spectrometer to get a feel for what a torch is doing:
- Wall test. In a dark room, shine it on a plain white wall. A big bright purple spot means lots of visible spill (typical 395nm). A dim, deep violet spot means less spill (typical filtered 365nm).
- Paper test. Most white printer paper contains optical brighteners. Under a good 365nm torch it glows a clean blue-white. Under 395nm you'll see the glow mixed with a lot of purple.
- Banknote test. Many banknotes have UV-reactive security features. A good 365nm torch makes them pop. Under a 395nm torch they're often harder to read.
- Don't trust your phone camera. Phone sensors render UV and violet light unpredictably, so photos often look more purple (or less) than what you actually see.
Also worth knowing: how bright it looks isn't how much UV it puts out. A 365nm torch can look dim and still emit plenty of UV, which is exactly why you need to be careful with it.
Safety (please don't skip this)
- Don't look into the beam, and don't shine it at anyone's eyes. The beam looks dim at 365nm, so it's easy to underestimate.
- Wear UV-blocking glasses for longer inspection sessions. Many safety glasses rated to block UV are fine; check the rating rather than assuming.
- Keep skin exposure short. Brief inspection is fine; holding your hand in the beam for minutes isn't a good habit.
- Keep it away from kids. It looks like a toy torch; it isn't one.
Wrapping up
If your goal is to see things (coating coverage, flux residue, dye leaks), a filtered 365nm torch gives you much better contrast than a 395nm one. If your goal is to cure things (resin, glue, solder mask), 395nm or 405nm is usually the better fit.
What do you use UV light for on your bench? I'm especially curious whether anyone has a reliable way to check no-clean flux residue, since UV doesn't always catch it.
Disclosure: I sell filtered 365nm torches in the UK, so I have an obvious interest in this topic. I've kept this post about the physics and the use cases rather than any product, and there are no product links here. I drafted this article with help from an AI writing assistant, then checked and edited it myself.
Top comments (0)