Minolta 5400 Mark 1 - ZigZag artefacts

Yeah, that's what I noticed, huge variance between units. The 4 5400MK1 that I own all presents differents noise levels and patterns.

Archiloque said:

Meanwhile in my futile quest for the perfect scanner, I've received a 185 euros quote for the glass removal of the CCD sensor. I may go ahead with it, I just need to find a way to desolder it from its PCB as Eureca told me the procedure would be a lot riskier with a PCB. So, now, I need to get a desolder pump and hope for the best . I think i'll do it for completion sake.

Click to expand...

I know it's a crazy idea, but I'm always thinking about how to cool the sensor. A drop of 5-10C° can reduce the noise. I was thinking of a thin heat pipe or a peltier, but it's probably impossible to put it under the sensor without it moving out of place.
The PCB is screwed to a plate. Does this have a thermal conduction function or is it just a support frame?

If you remove the sensor, can you take some pictures?

Thank you for the interesting thread!
I am new to scanners, but am in the process of finding the optimal scanner to digitize my family's 35mm negatives.
I recently came across a project which seems relevant, as earlier @scarbantia mentioned (narrowband) trichromatic (RGB) light sources.
Here's a link:

Maybe someone would find this interesting and useful.

In order not to be completely off-topic, @Archiloque, did you manage to remove the glass of the CCD sensor?

I own three of these scanners (one stands out with much better noise performance compared to the other two. My guess is that the bulb is in much better shape on this unit).

I've read about this issue before, but hadn't really noticed it in my daily usage. However, after going through this thread, I decided to run some tests on a few very dense color negatives (I don’t usually shoot slides, which might explain why I hadn’t noticed it).

I discovered that I could easily trigger the zig-zag issue in very dense areas on my less capable scanners. When scanning color negatives as linear positives with auto exposure turned OFF, grain dissolver ON, the zig-zag pattern and some banding primarily appeared in the blue channel, with some hints in the green. The red channel remained clear. I found that by using the analog gain in Dimage to elevate the blue and green channels to match the red channel and get them out of the “gutter” on the histogram, most artifacts would vanish. However, this also made the scan times ridiculously long. Plus, this won’t help at all when scanning transparencies. But for color negatives, it mostly solved the zig-zag issue on my scanners. So, it seems that optimizing the signal-to-noise ratio is key.

I also found some information stating that diffusion filters, like the grain dissolver, diffuse light more in the blue spectrum than in the red, so the blue channel is already at a disadvantage by receiving less light.

Additionally, I found that if I let the scanner (and the lamp!) warm up for about 30 minutes and then recalibrated the scanner by pressing ctrl+shift+I, it tended to performed better. So, I suspect that a lot of the issues might stem from worn-out lamps, possibly in tandem with aging capacitors.

@Archiloque,
Your LED swapping experiment was genius! Could you share more specifics on the exact parts you used? I’ve also been curious if simply swapping out the bulb could fix many of the issues. I’ve seen that folks are replacing the bulb in the Multi Pro with new off the shelf parts on the Minolta scanner group on FB, and it seems to work pretty well. Do you know the dimensions of the bulb in the 5400?