Dichroic filter confusion

[Pau]

Do you think it would work okay to use it as emission filter, even though the Y52 dichroic mirror would have a somewhat higher cut-off wavelength? I am thinking that the relatively [edit: high] transmittance in the 350-450nm range would prevent me from using the 'yellowLP' filter as DM for blue-violet excitation.

Now I understand why you have tested the yellow BS at 0 deg. angle.
I've never used a BS as EM filter, although looking at your graph it would work on the right side of its spectrum to clean the emission of the BS but be aware of its high transmission in the UV and V range, not to be recommended without a UV cut complementary filter. As you know there are better alternatives for EM filters (and I have them).

I was thinking of using it as BS but for violet-deep blue excitation, in principle closer to its intended design, likely it also will work well for UV in the 380 range with the other filters you have. Just you can't profit the strong 365nm peak of the mercury lamp

[viktor j nilsson]

I'm now leaning toward using the 420ALP as DM in a UV filter cube, and the '470LP' as DM in a violet-blue filter cube.

It seems like a very general trend that the cutoff threshold of dichroic filters shifts to shorter wavelengths as they are tilted, here's some data from thorlabs.



It makes sense, but I hadn't thought of this at all, and I'm glad you pointed it out. This suggests to me that my 420ALP filter would likely work like a 390-400LP filter when used as a beam splitter, which would be pretty good.

[viktor j nilsson]

Just as an exercise, I ran the mislabeled "T400LP" filter in the home-made webcam spectrometer I made recently (total cost: less than $30).
First, I calibrated the peaks in Theremino with my 50W HBO lamp (the webcam or lens clearly transmits nothing below 400nm as the 365nm peak is completely missing):

I then recorded a reference spectrum using my 100w Halogen lamp, and then a spectrum with the T400LP filter in the lightpath, brought the data into R, and took the ratio between the measurements to calculate transmittance. Here is the result:

Actually not completely useless - I could have told you that this was a 495nm rather than 400nm filter based on this.

Sensitivity below 475nm is terrible, though. Based on the image of the HBO lamp, it looks like I should be able to get decent readings down to 400nm. In retrospect, it seems critical to align the halogen properly, if it is out of alignment, the 400-475 tail gets lost first. As an illustration, the only difference between these two spectra are the position of the lamphouse relative to the slit:

I might even try to change the lens to one that transmits UV, although I am not sure what light source I would use. Speaking of which: did you guys see the people who recently built 55W Xenon lamps for their microscopes using bulbs and ballasts made for cars? Pretty ingenious.
LINK:
https://www.microbehunter.com/microscop ... 12&t=14018

[Duke]

I think what Pau meant, is that you can still use this 475 nm beamsplitter in another cube, like very popular Alexa488 cube (aka Green Cube), implying you would source appropriate Exiter (i.e. 450BP20) and Emission (490LP) filters, and legit 400LP beamsplitter for your cube. Thankfully, this parts are not that rare or expensive - you can get those on eBay for ~50$ each, or if you are looking for cheaper options, there is a selection of filters on aliexpress.

As OmegaOptical stores on ebay closed, Ali became the main source of optical components for me. It's a pain, because how poorly optical stores are organized, most of the time you need to contact the store in DMs to get the part you need give them exact specifications, measurements, anyway... what are the options...

I unadvice you to repurpose the filters out of their intended use - it's a risky business, you may burn your eyes, especially if you are using broadband gas discharge sources - HBO, XBO etc.