Orthoplan lightning

[Leitzcycler]

As it was asked here viewtopic.php?f=16&t=9139&start=30 which lamp I use, I put the answer here as it is quite off topic in the place the question is asked.

I have home-made led conversions in both transmitted and epi-fluorescence lightning. The transmitted light led is 3w 9-12V 20 mm diameter 6000-6300K form Ebay. It has a thick brass-aluminium base for removal of heat. The power source is a two-channel voltage regulated transformer with voltage readings for both the transmitted and epi illuminators. The fluorescence illuminator has three 3w leds from Avonec, Ebay: 370-380nm, 490-495nm and 515-525nm. They are mounted in a shape of a triangel into a large heat sink. I had to remove the original base plates of the leds to get them fit. As you can see from the photos, the central tube is eccentric and can be turned around so that one of the leds can be selected to be in the center of the optical path - at least near The markings in the rings help to set the correct led into its position: black marks = led in a position, red marks = the number code for each led 1,2 and 3. The power for each led is selected with a switch. The condensor lens in front of the leds is purely experimental. The system is still a prototype and under develomental process. It works satisfactorily, however the intensity of the fluorescence leds is not very high. I have also tried to use a very high-power white light led. However, it generated too much heat and would have needed a better coolign system. And the result was not better with the fluorescence ligh intensity.

[Leitzcycler]

[Hobbyst46]

Impressive job ! wow !
About the illumination:
1. in general, a 3W LED is really not powerful for fluorescence. The question is, how you evaluate the efficiency. The best would be some standard sample. A common one, in labs, is stained polymer beads, such as polystyrene latex. These fluoresce at several wavelengths. And they come in narrow diameter ranges: 3um diameter, 10um etc, very uniform (and beautiful). A permanent, or semi-permanent slide of such beads is excellent for judging both the brightness and the uniformity of the field.

2. Another, maybe simpler option, is a slide of a concentrated fluorophore, for example Rodhamine B in glycerine (or other viscous, not-volatile solvent). If it is a saturated solution, the visible fluorescence is only from the surface and does not depend on the thickness of the layer. Can be excited with your 515nm LED (if I remember correctly). Rhodamine is relatively stable against photo-bleaching.

3. In epi-fluorescence, the Kohler method is valid, and the objective serves as condenser. If the Orthoplan has an epi- iris aperture, it should be used for the purpose. Also, the distance of the LED from the optics should be optimized, if feasible. This holds true even for modern fluorescence microscopes.

4. Although the 3W LED is perhaps less than optimal, a high power white LED is not necessarily better. The 3W intensity covers the specific excitation range that one needs to cause fluorescence, whereas the white LED covers a broad range of wavelengths. I mean, using the white LED along with excitation filter might prove to be less than the 3W specific excitation LED. A 15W specific wavelength LED would be best.

[Rorschach]

Seems pretty impressive to me, even if the output isn't quite up there yet!

[Leitzcycler]

Staining cells with DAPI and Acridine orange works. However, the intensity is not so high compared to what I remember it was at universitys Zeiss Axioplan. However, I did fluorescence microscopy only a very short time many years ago and only to check the function of some antibodies. I used them in immuno-EM and actually did much more electron microscopy than light microscopy. Unfortunately (?) I do not have access to that infra any more. So I have to put up my own systems. But it will be much better eventually
My present aim is to express some GFP fusion proteins in some non-model yeast species. If that is visible in my system, it is quite ok. Though I would need a much better camera which will be an investment somewhere in the future...

I didn't find any more powefull leds in Ebay. They might exist, however may be diffucult and expensive to get.

[Hobbyst46]

Staining cells with DAPI and Acridine orange works. However, the intensity is not so high compared to what I remember it was at universitys Zeiss Axioplan. However, I did fluorescence microscopy only a very short time many years ago and only to check the function of some antibodies. I used them in immuno-EM and actually did much more electron microscopy than light microscopy. Unfortunately (?) I do not have access to that infra any more. So I have to put up my own systems. But it will be much better eventually
My present aim is to express some GFP fusion proteins in some non-model yeast species. If that is visible in my system, it is quite ok. Though I would need a much better camera which will be an investment somewhere in the future...

I didn't find any more powefull leds in Ebay. They might exist, however may be diffucult and expensive to get.

The Zeiss Axioplan was probably equipped with a 100W mercury lamp or 75W Xenon, both of which provide powerful excitation at these wavelengths. So the LEDs face an ancient but strong competitor...
Commercial LEDs for fluorescence microscopy are very expensive, so any DIY investment is potentially of advantage.
BTW, it is possible to stain baking yeast cells with FITC.

[Leitzcycler]

The Zeiss Axioplan was probably equipped with a 100W mercury lamp or 75W Xenon, both of which provide powerful excitation at these wavelengths. So the LEDs face an ancient but strong competitor...
Commercial LEDs for fluorescence microscopy are very expensive, so any DIY investment is potentially of advantage.
BTW, it is possible to stain baking yeast cells with FITC.

Yes, and a strong heat generator I also need lightning that can be turned on and off without any limits.

Commercial systems are simply unreachable for me.

Bakers yeast has been studied too extensively. The research focus must be set to something more special

[fero]

Just found out this discussion, I am converting my Olympus BH2 RFL light source to LED, I was thinking about 6500k LED as I have only B filter, something from these:
https://www.tme.eu/en/katalog/white-pow ... 1802409%3B

I was just curious whether you made any progress with your impressive setup that could give me some inspiration.
Thank you,
Fero

[Leitzcycler]

Not so impressive... The led system for the transmitted light is working well. The system for epifluorescence doesn't as it would need more light intensity. I am currently working on it to improve the system. Have many other projects so the progress is slow. However, I think should get something ready during the next few weeks. I can post here if someone is interested.

[smollerthings]

I can post here if someone is interested.

Please do!

I found electronics components that can drive and dim 10W+ LED without flicker for $2 if you are interested.

[Leitzcycler]

You may find more information here:

viewtopic.php?f=11&t=13170&p=113413#p113413

I have used variable voltage regulation for dimming leds until I learned here it doesn't work.

The current project is to make a multi-purpose illuminator base/lamp house into which I can add either led or car bulb as discussed in the thread above. The reason why all this is so slow is that my real bottleneck is not here, however in the research I am doing in the lab. That should produce something to see under fluorescence.

I found electronics components that can drive and dim 10W+ LED without flicker for $2 if you are interested.

Please, add the link here if you have one.

[smollerthings]

So I went on my quest for cheap led driving and dimming. I found a solution that seems to work. I cascaded a constant current driver like this one:
https://fr.aliexpress.com/item/32580214850.html

to a pwm driver:
https://fr.aliexpress.com/item/33012725299.html

Both are common affordable components and have high commutation frequencies and can handle high current (see the spec sheets)

I haven't noticed any flickering on video. I think the commuting frequencies are high enough for video applications.

The PWM seems to cover the current range nicely (0-max current). In the video below, I have a 3 ish W LED, I put the max current to 300mA for the demonstration. However, even at 0 on the multimeter, there is some current passing through and the LED is somewhat illuminated.

12V wall wart -> constant current driver -> PWM generator -> LED



Packaged in a box:

[MichaelG.]

.

That seems to work nicely

Thanks for the demo.

MichaelG.

[Leitzcycler]

Thanks for the presentation! As I said I have used variable voltage. It works for me, however doesn't seem to work for others As I have systems ready, I don't thing I am going to start building everything from the start again at least in the near future. However, I will keep all this information in mind and perhaps will make some progress when (if) I have time.

I am not very convinced of those cheap Chinese circuit boards as I have ordered some and they worked just a couple of days. Maybe you have better luck with those...

[smollerthings]

Thanks for the presentation! As I said I have used variable voltage

How do you limit the current and not fry the LED?
But yes, the best solution is the one that works for you!

[Leitzcycler]

Long before the existence of the mighty Ebay I used to make some electronic devices myself. Not that I really understood anything about electronics, however construction is possible if you follow the circuit diagram. Here you may find the circuit I used in many applications for voltage regulation: https://www.componentsinfo.com/lm317-pinout-equivalent/

What comes to the Orthoplan lightning, I use 3W 9-12V led of 6000-6300K from Ebay regulated by the circuit described above. The useful voltage area is far below 12V. I just have a 10-turn potentiometer which is easy to operate and you don't easily turn it too much to blow up the led. I have also fixed V meter in the panel. You could of course fix the upper voltage limit by arranging the resistors, however I didn't do that. And using a bigger capacitor in the output end will abolish any flickering.

All the experts in the fields of electronics and optics are welcome to shoot down my construction ideas

[smollerthings]

Sounds like it is working fine. Keep on eye on the current, I was blowing 10W into my 3W led without noticing...

[Leitzcycler]

Sounds like it is working fine. Keep on eye on the current, I was blowing 10W into my 3W led without noticing...

If I ever happen to blow up the led, I change a spare led And here I just extended an old Olympus transformer for use of led as I think the original transformer is so nice and I want to keep it as interior decoration. viewtopic.php?f=24&t=11044&p=107208&hil ... ht#p107208
When I want to use led, I plug the rectifier element in. And I am still able to use the original bulbs. There is mechanical voltage limit knob.

BTW, no Paris-like weather any more. Snow and cross-country skiing

[Leitzcycler]

This is the version 2.0 of my fluorescence LED illuminator for Orthoplan. I am not a fan of 3D printing and such things or even know anything about it. Besides I like to do something with my hands. I bought leds of 30W 450nm and 470 nm. The first experiments are very promising….

Changing the leds is a bit tricky. I will also try a 30W white led and see if it can be used wit different wavelengths. One option is to arrange several leds into a slider which can be moved against the surface of the heat sink. However, this works with the FITCH-filter for now.

And Merry Christmas for all of you!

[Microscopy_is_fun]

Hi Leitzcycler,

interesting project! I am also looking into replacing a 100W Hg-lamp with LED for fluorescence. I think it is very important to put focus on the right collimation of the beam, and not so much onto maximum power of the LED. If you look at commercial systems, e.g.

https://www.leica-microsystems.com/prod ... ification/
https://www.thorlabs.com/newgrouppage9. ... p_ID=13597

you can see that the output power of a single wavelength band is typically less than 1W. If you need a 30W LED in your setup for sufficient fluorescence intensity, your collimation of the beam is probably not optimal. There is also a general rule in optics that states that the light intensity in the image plane can never be higher than the intenisty of the light source itself. Since the high-power LEDs are mostly relying on enlarging the emitting area of the LED array, this means that you probably lose a lot of light in your case, and only a small fraction ends up at the sample plane.

[dtsh]

This is the version 2.0 of my fluorescence LED illuminator for Orthoplan. I am not a fan of 3D printing and such things or even know anything about it. Besides I like to do something with my hands. I bought leds of 30W 450nm and 470 nm. The first experiments are very promising….

Changing the leds is a bit tricky. I will also try a 30W white led and see if it can be used wit different wavelengths. One option is to arrange several leds into a slider which can be moved against the surface of the heat sink. However, this works with the FITCH-filter for now.

And Merry Christmas for all of you!

It's nice to be able to cast parts, isn't it? Amazing what one can do with a makeshift foundry and a lathe, nice work.

[Scarodactyl]

That is really cool!

One thing I'd note is that one thing 3d printers are nice for is lost wax-style casting Obviously not something you needed for this, but they can even simplify the workflow for true hand work like this.

[Leitzcycler]

your collimation of the beam is probably not optimal.

Sure it is not optimal. However, now it is done so let it shine...

[Microscopy_is_fun]

However, now it is done so let it shine...

Certainly a very nice project, thanks for showing and putting up the links to the driver board suppliers!

[Leitzcycler]

My latest version.

Version 2.0.jpg (196.48 KiB) Viewed 1245 times

[jean-marc]

Hi,

Very good homemade system. I like the possibility of a quick change of your LEDs.
You could have put smaller LED, I guess a large part of the light is lost.

I used LEDs also for my episcopic light and fluo lights. The chips are not larger than 5x5mm. My blue 450nm LED is 3W and give more blue light than my 18W Cree XHP50 white LED.
I have UV Led, and they are not more than 1400mW with a very little chip.

IMG_5806.jpg (72.98 KiB) Viewed 1173 times

IMG_5807.jpg (89.94 KiB) Viewed 1173 times

IMG_5808.jpg (45.77 KiB) Viewed 1173 times

Regards

JM