Scanner test success story - advice needed

Which state? Some have better material than others

So I finally got these polycarbonate states to a point where they feel pretty much spot on with the real ones. The weight, the slight flex when you bend 'em—it's uncanny how close it is now. I've been tweaking the lamination process for a while, and I think the key was getting the right layer stack and pressing temperature just right. But I'm still trying to dial in a few things and could really use some community advice on the finer security features.

Starting with the hologram overlay—the construction I'm using is a multi-layer film with a custom diffractive pattern. On the real ones, the hologram has this really defined depth when you tilt it under a light, almost like the image floats above the surface. I've managed to get the rainbow shift and the state-specific design elements to align, but the overlay edge is still a tiny bit thicker than I'd like. On the originals, you can barely feel a step where the overlay ends, and mine still has a slight ridge. Any tips on thinning that out without damaging the polycarbonate layers?

The ghost image clarity is another area I've been obsessing over. Real polycarbonate IDs have this ghost image that's subtly offset from the main photo—it's not just a shadow, it's a distinct secondary image that fades gradually. I've got the opacity right now so you can see it clearly when held at an angle, but the transition from the ghost to the background isn't as smooth as the real ones. It still looks a bit too crisp on the edges. I suspect it's a UV exposure time issue, but I haven't nailed the sweet spot yet.

Perforation pattern quality has been a beast. The micro-perf on the birth date fields—the real ones have tiny, perfectly round holes with clean edges and no burrs. I'm using a laser perforator, and the spacing is consistent, but the hole diameter varies slightly across the card. Under magnification, some holes look a little oval instead of round. I've tried slowing down the laser pass, but that seems to risk melting the edges. Anyone found a good frequency and power setting for polycarbonate that gives that crisp circle?

UV detail accuracy is probably where I'm most proud so far. Under a blacklight, the state seal and the guilloche patterns light up with the correct color—a slight off-white with a greenish tint. The fine line work is all there, and the UV ink doesn't bleed into the surrounding material. But the real ones have a hidden UV state flower that's only visible under a specific wavelength, and I'm pretty sure I'm using the wrong UV pigment for that. Mine shows up under standard 365nm, but the originals need 395nm to pop. Should I be mixing two different UV inks or using a dual-band approach?

As for the scanner test success story—I ran these through a couple of common scanners used in bars and airports, and they passed without any flags. The machine read the barcode and the magstripe data cleanly, and the OCR on the text fields was dead on. But I noticed that when I put it in the same scanner that flagged my previous attempt, the polycarbonate's near-infrared transparency was slightly different from a real one. That might be the laminate thickness causing a darker transmission reading. Is that something I should worry about, or do modern scanners have enough tolerance?

Overall, I'm really close, but I want to make sure the hologram edge, ghost image transition, and perforation uniformity are as close to OEM as possible before I consider these done. If anyone's cracked the code on those, I'd love to hear your methods. Also, if there's a specific UV response test I should run that I'm missing, let me know.