Weird crosshatch artifact

[brucek]

I have a weird artifact showing up with Helicon stacks on only one of my microscope objectives.
Optics:
Nikon D3
Nikon 200mm, f/2.8 manual focus, set at f/8 infinity focus (as a tube lens)
5X Mitutoyo Plan Apo Infinity Corrected Long WD - works great
10X Mitutoyo Plan Apo Infinity Corrected Long WD - works great
20X Mitutoyo Plan Apo Infinity Corrected Long WD - exhibits this crosshatch artifact in Helicon stacks most of the time
Homebuilt stacking stepper - set at 2.5 micron steps for 20x, 3.5 microns for 10x, 10 microns for 5x
Nikon flash through diffusion paper for illumination.

Input Images:
I'm taking photos of micro-crystals. I have separate dust maps for each objective, which I enable in the Helicon UI. The 5x and 10x objectives almost always produce great results with the default Depth Map stack parameters and a bit of retouching of the out-of-focus areas. With the 20x objective, though, most stacks show this cross-hatch pattern in the sharp areas (which is obvious in the image I've attached) but absent in the source images. It does not seem to matter what stacking method or parameters I select. The source images look fine to me with sharpness that appears where expected.

The attached image is a cropped section that shows the artifact clearly, not the entire image.

Questions:
Any idea what is causing this cross-hatch pattern?
Do I have a bad objective?
Maybe I am exceeding the capability of my optics?
Maybe there's some interference between my step size and the depth of field?
I'd really like to be able to use this objective for its increased magnification, but the results aren't usable unless I can figure out a way to get a cleaner result.

Thanks in advance for any ideas or guidance.

[Catherine]

No idea, but since there's no problem in the source images it's probably not related to the optics you're using. Does the same happen if you don't assign a dust map?

[brucek]

Thanks for the suggestion, Catherine. Ive added another Helicon stack image to this post without the dust map. The crosshatch is gone. Guess I'll have to clean my sensor.

[Catherine]

Please share one of the source images, and the dust map that you're using. We'll look into why this happens.

[brucek]

Attachments in this forum can't handle the size, so I put the source image and dust map into a RAR file here:
https://drive.google.com/file/d/1SVFjhH ... drive_link

[Catherine]

Thanks, downloaded. One thing that stands out is that you cannot have a TIFF map with raw images. The raw converter will crop images, so the pixels will no longer be aligned. The dust map must be in raw format as well. If you're pre-processing source images before stacking them, the dust map should be processed in exactly the same way.
I can't say for sure whether this is why it doesn't work, but it's possible.

[brucek]

Thanks Catherine, I really appreciate your help! I re-ran the stack with a matching DNG dust map and it the problem is gone. A suggestion for the program would be to pop up a warning if we try to mix file types this way.

[Catherine]

Indeed, it should be easy to add a warning. Happy to hear it works now, thanks for the confirmation!