You should be able to use the Sangaboard PWM channels for a higher power LED, depending on the diode forward voltage it will provide ~1A peak current. Just don’t run it continuously at high power without a heatsink on the current limiting resistor (a thermal pad on that area and a heatsink worked well for me with high power LEDs (which is where the heatsink kit came from) but I would still only run it at full power briefly. The plastic around the LED might also start melting quickly at full power).
3 Likes
The Emission filter graph looks like the signal is from about 510-650nm:
But the specs say it is full width at half maximum 575nm. So from 525-625nm. The Emission peak of GFP is 509nm and seems to fall of much sharper than the absorption?
Also how well focusses is the light at the objective. If you focus the microscope with a slide, remove the slide and replace with a piece of paper do you get a nice sharp spot?
1 Like
I think there may be a couple issues with the choice of filters described above. First, there are several types of GFP and the wtGFP has an excitation peak in the 390s. eGFP (an engineered version) has a peak at about 470nm. Thus using a Blue LED at 470nm will work just fine. Second, the beam splitter should be a dichroic beam splitter not just any beam splitter. The one in the link above doesn’t specify its cut-on nm. Ideally you want a dichroic beam splitter that is set to 500nm. Nest, the emission filter needs to be around 509nm to get most of the GFP’s emission wavelengths. The Stokes shift for eGFP is pretty narrow and the dichroic mirror needs to split the two peaks.
2 Likes
You might try this filter from the same company instead to get more of the emission peak.
I’ve had great photos from my DIY 3D printed microscope (not an OFM) with GFP and a 470nm Blue LED with the appropriate filters. I can provide more info if people are interested. It would be great to have an affordable set of filters as the ones I use come from a standard filter cube and cost $800USD which make it too expensive for many people.
3 Likes
Thanks everyone, this is all super helpful!
That would be great, I would be really interested to learn more and see some of your images! Thanks so much for sharing.
2 Likes
Jo, Here’s a link to a YouTube video showing the scope imaging GFP in a living nerve cord of a larval drosophila. It was taken with a Basler global shutter camera and shows Ca++ influx when neurons are excited along the nerve cord.
The paper describing the older version of the scope is at:
Ryan et al. (2020) Building Your Own Neuroscience Equipment: A Precision Micromanipulator and an Epi-fluorescence Microscope for Calcium Imaging. J Neuro Sci Educ 19(1):A134-140.
I’ll post some pictures of GFP taken as stills when I get a chance along with the GitHub link.
1 Like
@jorodeo It is also worth thinking about if we are photobleaching. It may be worth trying to align everything optically. Then add in the filter then turn on the illumination just before capturing? If that is the case then we will need a way to toggle the illumination.
I say all of this with no knowledge of how susceptible GFP is to photobleaching
1 Like
Ive just looked through old data from our 2020 paper It is worth noting that the original image:
Isn’t super bright:
We then take the raw bayer data and some calibration images to calculate the fluorescence image as:
The data is in fluorescence folder of the image. The calibrate_data.py script will throw an error unless you delete the calibrated images. Or change line 16 from:
data_paths = [os.path.abspath(d) for d in data_paths]
to
data_paths = [os.path.abspath(d) for d in data_paths if "calibrated" not in d]
1 Like
GFP will photo bleach if the light is left on too long. There are mounting mediums that reduce this for fixed specimens. Another thing to consider is the loss of the blue led light before it even hits the excitation filter. I’ve modified the OFM optics column to fit my led with the cylindrical heat sink if anyone is interested. I think this prevents the that light scattering and should produce a stronger image. I think it will also make it easier to swap out led lights as it just slides in and out. I haven’t mounted it in my OFM build yet and I don’t have the smaller filters sourced yet, but I’m working on it. I can post the stl file all the files for building and wiring a separate on/off/dimmer led switch.
2 Likes
This looks great. Being able to slide it in and out to swap the LEDs will make it much easier to use. It might be a few weeks before I can get back to testing but I do think the lowest hanging fruit to start optimizing is the LED set-up. Yes, if you could share the cylindrical heat sink and LED board you’re using that would be great. The on/off/dimmer is a a great addition! I’m sure @gerrit will also be interested in this.
@j.stirling Thanks for sharing the original manuscript images. I tracked down some similar samples from Lieder in Germany. We should have a Convallaria rhizome acridine orange stained slide coming in a few weeks. We thought it would be good to test a second sample along with the GFP slides. Lieder have a whole range of prepared slides for anyone that’s interested!
@jorodeo I forgot to put the link in to the raw data:
https://researchdata.bath.ac.uk/734/
1 Like
I finished building my IORodeo OFM kit last night. I added the new str for the optics module with filter cube and a cylindrical holder for the led heat sink I use. It’s all up and running. All I need now are some fluorescence filters for the cube. I had to slightly enlarge the hole in the black base to make it easier to slide the led heatsink in and out. The led control box runs on either a 9V battery or a 12V wall adapter and uses a potentiometer for dimming and (in this case) a momentary push button to activate the led. I also have a control box with a toggle switch to free up your hands for other things. I’ll post a wiring diagram and bill of materials for the led and control box when I get some time. I can post a link to the new optics module str file if anyone would like it.
3 Likes
That looks great, @JRyan ! Is it possible to swap out the LED and filter cube without disassembling the microscope? Is the filter cube attached to the LED?
Yes. The LED just slides out of the 3D printed tube and then you would have to reach in with something to remove the filter cube. I’m working on a way to do that with a 3D printed tool that would use something like small double-sided tape to stick to the 2 flanges on the cube on its right and left. I may also try to add a small extension on one side of the cube (modify the str file) to make it easy to grasp with plastic forceps. I’ll work on that.
I’ll post the bill of materials for the LED later. Here’s the wiring diagram for the LED with heatsink.
Note that you can swap out a 9V battery connector in place of the 12V wall adapter. The JST 2-wire connectors are so that you can use the same control box with different color LEDs. My undergraduate students and I use the same control box/LEDs for ontogenetic control of behavior in fruit flies so it can have multiple uses.
1 Like
1 Like
Thanks @JRyan , that’s really helpful!
Together with another collaborator, Peter, I’ve been working on using magnets to create a snap-fit mechanism for the filter cube and reflection illuminator. So while not compatible with @JRyan 's design above, it does allow for the original mechanism (as @jorodeo used above) to be swapped out without disassembling the microscope. I’m including the STL files if anyone wants to give that version a try.
fl_cube.stl (67.0 KB)
reflection_illuminator.stl (209.9 KB)
rms_optics_module.stl (419.9 KB)
Peter created the modifications in TinkerCAD, and I then recreated those modifications in OpenSCAD so that we can contribute it back to OpenFlexure in future. The STL files above are from the OpenSCAD design, and I’ve created a branch in my fork with the changes:
4 Likes
I tried to upload my .stl for the cylindrical LED optic carrier, but it says new users can’t upload files. I’ll try to find another way.
@gerrit’s suggestion of magnets is a great idea for locking the filter cube in place. I’ve used magnets a lot on other neuroscience DIY projects and my only concern would be getting the filter cube back out when you want to use a different fluorophore. It’s a very tight space and it will be hard to grasp the filter cube with enough force to break the magnetic bond without scratching the delicate filters.
My other suggestion for using the jorodeo version of the “reflection_illuminator” is that you will still be losing a lot of the light before it hits the dichroic mirror in the filter. It’s open on the sides and bottom and a lot of light will be lost unless it is moved all the way forward. You’ll also want to add a heatsink as these LEDs can get hot enough to soften and warp the PLA plastic.
One other reminder for people who have not done work much fluorescent filters is that you cannot touch the filter glass with your fingers - the oils on your fingers will degrade the thin coatings. So always wear gloves when working with the glass filters.
I’m glad to see that there are several others interested in an OFM with epic-fluorescence. It’s a great project. I’m happy to contribute in any way I can. I’ll try posting my modifications to my Github and linking that here if I can.
1 Like
I solved the problem if swapping in and out the LED and filter cube thanks to @gerrit’s suggestion. I modified the filter cube so that it fits into the outer rim of the LED heatsink and thus the LED and cube become one unit. This can simply live in and out of the modified optic housing. See photos below.
The stl files are on my GitHub at:
https://github.com/CrawFly/DIY-Neuroscience/tree/main/OFM%20%20LED%20parts
4 Likes
Thanks for the link @Gerrit. Can you open an MR on GitLab to master on OpenFlexure? If you click the “Draft” button, then it won’t be possible to merge and it will be clear that you are still working on it, but this way the OpenFlexure team will automatically see when changes are made and we can suggest other changes/discuss it in the MR interface.