A. pellucida - Blue LED

Hello Group,
I've done some experiments with my MICAM program.

First I added the possibility to get pictures twice the original size (I think Don asked for that extension). Then I added the possibility to capture four images and present the average image. This gives less noise.

I also used a blue LED (470 nm) for the illumination. Pictures were taken with the Euromex Novex B microscope 1.25 NA oil immersion (100X) and 1.25 NA condensor.

The image of the objective was directly focussed on the CCD of the ToUCam 840 webcam.

You can find the result in my photo album called "Microscope Pics":
 Link : http://tech.ph.groups.yahoo.com/group/Microscope/photos/...

Kind regards, Marien

Hello Marien,
Thank you for presenting some very interesting results with our old friend A.pellucida. But, all four images are from the same slide?

Could you tell us about your slide preparation?

You wrote:
> ....capture four images and present the average image. This gives less noise.

Yes indeed. It's somewhat surprising, but I think you may be the first microscopist to present Ap images using averaging. As the signal-to-noise can be expected to increase by the square root of the number of images averaged (approximately), did you happen to check the result with (for example) 16 or 64 captured images?

> The image of the objective was directly focussed on the CCD of the ToUCam 840 camera

Hmmm. As I recall, the ToUCcam is a 640x480 webcam similar to the Logitech used by Don. I'm assuming that you removed the standard lens from the webcam. So what was the optical setup at the eyepiece location?

Thank you again for sharing your work with us.

Regards, Keith

Hello again Marien,
In a previous post I wrote:
> I'm assuming that you removed the standard lens from the webcam.
> So what was the optical setup at the eyepiece location?

I was thinking pickup location for the eyepiece, but you may have been using the webcam on a trinoc tube?

Regards, Keith

"marienvanwesten" wrote:
> Pictures were taken with Euromex Novex B microscope 1.25 NA oil
> immersion (100X) and 1.25 NA condensor.

Hello Keith,
Thanks for your comment.

> Thank you for presenting some very interesting results with our old
> friend A.pellucida. But, all four images are from the same slide?

Yes they com from an 8 Forms slide by Klaus Kemp. (I saw that nr 22 is F. rhomboides with line distance of 330 (340) nm). The calculated distance of lines in A.pellucida is 270 a 280 nm.

> As the signal-to-noise can be expected to increase by the square root of the number
> of images averaged (approximately), did you happen to check the result with (for example)
> 16 or 64 captured images?

Not yet, I will delve into that subject later on. My goal was to see if I could get sgnificant improvement with stacking 4 captured frames.
It really does!
There is a lot of noise, because the LED I am using is only running at 6 mA, so the output power is only a few mW.

I removed the eyepiece. So the objective is the only optical element (and the condensor of course). You can see my setup at my website:
  Link: http://members.home.nl/mvanwesten/en_microscope.html

I tried photographing with the lens on the webcam, but at the back of the lens is a reflective filter for infrared and UV. this causes red spots in the pictures because of multiple reflection in the eyepiece. I read an article about this subject from someone who had the same problem.

Kind regards, Marien

Marien,
Thanks for the clarification. All the same, a nice image for F.rhomboides.

All the best for continued development of your MICAM program. Software development can be a very time consuming hobby - be it Delphi or MS Visual C++ !

Regards, Keith

"marienvanwesten" wrote:
>> But, all four images are from the same slide?
> Yes they come from an 8 Forms slide by Klaus Kemp. (I saw that nr 22 is F. rhomboides with
> line distance of 330 - 340 nm). The calculated distance of lines in A. pellucida is 270 - 280 nm.

Hi Marien, Very nice pics.

Does your scope have Koehler lighting, the site I looked at did not list it for this scope. The condenser on your scope, is it the standard Abbe 1.25 that comes with it . Did you use Oblique, Phase or other lighting procedure.

Thanks!
Ron

Hello Ron,
Euromex offers an additional Köhler illumination for this microscope, but I made my own light source with a blue LED and a small lens in front of the LED so the image of the LED was on the condensor diaphragm (that is Köhler I think). I did use the standard Abbe 1.25 condensor.

I must admit that the illumination is a bit oblique as well, because the lens can be moved above the LED to find the best result. My setup is rather simple and easy to construct, but you have to be a bit handy to arrange things properly.

I could show a picture of my setup if you like.

Regards, Marien

"marienvanwesten" wrote:
> I could show a picture of my setup if you like.

I would love to see it! Ron

Marien,
Right now there is quite a lot of interest in the use of a blue LED as an illuminator source. So pictures and any other information re the actual LED used, power supply, etc. will be of real interest.

I may have some input re using PWM to reliably increase the effective illuminator intensity.

Thank you for sharing your design with us.

Regards, Keith

Hello all,
> Right now there is quite a lot of interest in the use of a blue LED as
> an illuminator source. So pictures and any other information re the
> actual LED used, power supply, etc. will be of real interest.

I have put some photographs of my blue LED illumination in my "Microscope Pics" album:
 Link : http://tech.ph.groups.yahoo.com/group/Microscope/photos/...

I bought a plastic lid (the kind used by bricklayers to prevent concrete and plaster to enter open plastic piping. I hope you understand what I mean). The plastic lid fits nicely in the opening in the microscope house as you can see.
Then I made a round PCB (printed circuit board) with the blue LED and a series resitor of 470 ohm. I didn't etch the board, but I cut the tracks in the copper layer with a small drill with a round head, like dentists use.
The PCB is attached with three nuts and bolts. There is no need to use a PCB. You can use a piece of cardboard, pinch holes in it and attach the components, then solder the wires to the components.

Several years ago I ground and polished some small lenses (just to see if I was able to make a lens myself). These lenses have a focal distance of a few centimeter. I attached the lens in a piece of pipe (two pieces that were taped together in my case)To make sure the LED was focussed on the condensor diaphragm.

I had an old simple power supply that gives 5, 6, 7.5, 9 and 12 volts. Most of the time I switch it on 5 or 6 volts. The current drawn from the source is only a few mA, so any power source will do.

> I may have some input re using PWM to reliably increase the effective illuminator intensity.

Interesting, but doesn't that interfere with the webcam? Using the original lamp I get a flickering image on the computer screen, so I have to switch on the mechanism to prevent this. DC illumination doesn't have this effect of cause.

I hope this will illuminate you ;-)

Kind regards, Marien

"marienvanwesten" wrote:
> I have put some photographs of my blue LED illumination in my "Microscope Pics" album.

I like it, Marien.

This is very similar to one of the design choices I will try on my Lomo Multiscope. I have an advantage on the lomo, in that the bulb holder on the Lomo is desinged so I can unplug the original two pronged bulb from the holder and insert the LED right into the holder with no modifications. The original holder allows easy movement of the light. The resistor of choice can be added in-line at the plug-in on the variable transformer. The transformer supplies 3,4.5,6,7.5,9,12 volts DC.

I have removed the diffuser element in the condenser above the light and I believe all I need to do is put another lens element of proper focal length just above the condenser in a simple helical focus system to bring the filament/LED to focus at the iris.

This brings up this question, do you think that if the LED where placed at the iris of the upper condenser that Koehler will also be achieved? This was thought of only as a way to avoid the other lens element purchase.

Thanks for the Illuminated inspiration!

Ron

Hello Marien,
> I have put some photographs of my blue LED illumination in my "Microscope Pics" album.
Thanks. As always, a couple of pictures really help in describing a project such as this illuminator LED.

>> I may have some input re using PWM to reliably increase the effective illuminator intensity.
> Interesting, but doesn't that interfere with the webcam?
> Using the original lamp I get a flickering image on the computer screen, so I have to switch on the
> mechanism to prevent this. DC illumination doesn't have this effect of cause.
We can assume that the intensity of light output from the LED is proportional to the current flowing in the LED. If the maximum current is limited by a series resistor, the output can be lowered by using PWM with a duty factor less than 100%. This is a conventional variable intensity control. The output is actually a series of pulses of light, but if the base PWM frequency is say 200Hz the eye doesn't "see" these pulses.

I doubt that this would introduce a flicker via the camera - but I may be overlooking something here. Maybe another group member will have some input.

Now the interesting situation is when you remove the current limiting resistor and run with a much lower PWM duty factor - maybe 20% (at the same 200Hz). If the LED output is non-linear with increasing current, by increasing the supply voltage I think you can generate very high intensity light pulses - without burning up the LED. I think that the eye sees the high intensity pulses and "ignores" the blank period in the PWM scheme. The LED stays alive because the heat generated, by the average current, remains relatively low. Of course it's then a matter of playing with supply voltage and PWM duty factor range to get a reliable (smoke free), variable intensity output.

I seem to remember getting startling output from a red LED, but I would have to redo the experiment to make sure. Just what I need right now - yet another project on my bench!!

Have fun.
Regards, Keith

Hello Keith,
> I doubt that this would introduce a flicker via the camera - but I may be overlooking something here.
> Maybe another group member will have some input.

Yes, its absolutely impossible to have a PWM and use a High End Camera. Best is a variable current source from 10mA up to 1000mA for the RoyalBlue-LED. Normal max. is 700mA, but for a short time I use it up to 1000mA.
The UV-LED (385nm) from Roithner in Austria has a limit from 350mA !

So I never like to have a PWM power supply.
With the UV-LED you have to use a camera, you see nothing, or near nothing.

My results are in the photo folder Amphipleura_pellucida

Regards, Peter

Hi Marien,
First class pictures of AP, taken with simple apparatus.

You don't mention whether the condenser was immersed to the underside of the slide, but I assume not?

Your results are of particular interest, as I am trying to explain how microscopists before 1860 managed to image the stria on AP! Pre immersion lenses, pre homogenious immersion, and with a limiting NA of about 0.9! I have managed to duplicate the feat with a Lomo x85 water immersion, NA = 1, and with an immersed aplanatic condenser. The next stage will be with a non-immersed achromatic condenser, and a dry objective of NA = 0.95.

I think that the key point here is the form of the illumination. LED sources should give excellent longitudinal coherence, far better than any normal filament based source. Also, because the source is small and perfectly planar, oblique illumination can be achieved merely by tilting the source slightly, with respect to the central axis of the microscope.

The shorter wavelength of the blue light is useful, but I doubt it is the real reason for the good resolution.

Would your MICAM software work with the older Philips Vesta Pro? I have one which I still use because it is CCD based, not incredibly noisy CMOS! I missed out on buying a Toucam II - many astronomer friends bought them and can produce superb planetary pictures using image stacking.

I use a ribbon filament lamp for most of my work, but it looks as though it would be sensible to investigate LED sources as well.

Cheers, Merv

Hi Keith & Marien,
The flicker and rolling bars that gets into video from lighting is an alias from the frequency of the AC current to the light and the scan rate of the Video being out of sync. NTSC uses a 60 HZ scan rate and PAL uses 50 HZ. The AC lights in the US flicker on and off at twice the frequency of the AC current or 120 Hz. In the UK, EU an much of the rest of the world the flicker rate is 100 HZ. You can even measure it on incandesent bulbs.
To be sure there isn't a problem with the flicker from the lighting and scan rate of the camera most lighting I have seen is at least at 2,000 Hz for a 4,000 Hz flicker rate. I am sure that is great deal more than it needs to be to stop the aliasing problem but it was don't with florescent lighting and the high frequency allows audio transformers work instead of special built ones.
A NE555 will do the job at less than 20kHz see http://www.kpsec.freeuk.com/ 555timer.htm and search down to duty cycle. It it probably a bit more stable with two 555 timers one setting duty cycle and one setting the clock frequency. I expect there will be interaction among the setting particularly if there isn't a drive transistor or FET on the output of the 555 timer. Texas Instruments has one that will source up to 200 ma at 12 volts.
If you build a PWM solution I suggest that a current limiting resistor be used to set the maximum current draw with the duty cycle 100% on. At least on the first prototypes. Having a fail safe circuit will save a lot of LEDs in the beginning. Once things are working you can take the resistor out if you need to conserve batteries.

Gordon

> I like it, Marien. This is very similar to one of the design choices I will try on my Lomo Multiscope...

I suggest buying a simple magnifying glass like the one they sell for stamp collectors. My lens isn't corrected for any abberations and it also does work quite well.

> This brings up this question, do you think that if the LED where placed at the iris of the upper condenser that
> Koehler will also be achieved?
> This was thought of only as a way to avoid the other lens element purchase.

I did a simple test holding my LED richt under the condensor. Using the 60X (0.85 NA) objective I was able to see lines in F. rhomboidus. You have to be a bit handy holding the LED at the right spot and manipulating the condensor focus and microscope focus. I must say that the results were not as good as with the setup as I described, but it might work.

Success with your experiments

Regards, Marien

Hello Merv,
> First class pictures of AP, taken with simple apparatus. You don't mention whether
> the condenser was immersed to the underside of the slide, but I assume not?

Right, only the top side of the slide was oiled. I did try to oil the underside, but it was very messy and I didn't get a significant improvement, so I didn't try it any more.

> ...The shorter wavelength of the blue light is useful, but I doubt it is the real reason for the good resolution.

I think the absence of chromatic abberation is also important.

> Would your MICAM software work with the older Philips Vesta Pro?

My program should work with any kind of webcam. If it doesn't, drop me a note about your problem.

I did try my sons Logitech Quickcam Messenger. This webcam has a CMOS chip and causes much more noise. I think the Vesta Pro is a better choice because it has a CCD. The new Philips webcam SPC900NC is said to be also very good for astronomy, so will do for our work as well.

Kind Regards, Marien

Hello Keith, Gordon and Peter,
> The flicker and rolling bars that gets into video from lighting is an alias from the frequency of
> the AC current to the light and the scan rate of the Video being out of sync....

I also observe flickering with my webcam. That obviously has to do with the scanning of the chip. The driver of my ToUCam even has a provision, to deal with that.

> If you build a PWM solution I suggest that a current limiting resistor be used to set the maximum current
> draw with the duty cycle 100% on. At least on the first prototypes...

What I don't understand is that you folks speak about several hundreds of mA, while I use only 6 mA and I get enough light to see the structures visually. What is the reason to use so much light? Or do I miss the point somewhere?

Kind regards, Marien

Hello Peter,
You wrote:
> Yes, its absolutely impossible to have a PWM and use an High End Camera.

On thinking about the PWM scheme some more (and per Gordon's comments re a typical digital camera's raster scan) it's obvious that what our eye perceives is very different from a camera operating in a progressive scan mode.

> Best is a variable current source from 10mA up to 1000mA for the RoyalBlue-LED.
> Normal max. is 700mA, but for a short time I use it up to 1000mA.

Yes. A variable, constant current DC power supply is the most professional approach - but hobbyists enjoy finding out if there is an "economical" approach that gives similar results.

Is it similar to the Ortholux conversion as documented by Hugo Gonzalez Pineiro.

> My results are in the photo folder Amphipleura_pellucida

These images are impressive. Dr.Don acknowledges that these are the best images published with the group - to date.

Regards, Keith

Hi Merv,
Please excuse my butting in, but your comment:
> ....is the form of the illumination. LED sources should give excellent longitudinal coherence,
> far better than any normal filament based source. Also, because the source is small and perfectly planar,....
caught my attention.

I have wondered if the LED (probably on a die) could be considered to be close to ideal point source for light microscopy illumination. But from what I understand the most common, flat, surface emitting LEDs have a Lambertian output i.e. the light is distributed in a cone with a large included angle? I suppose this is why the LEDs are shipped encapsulated in a plastic 'case' which includes a lens to concentrate the light on a central axis perpendicular to the LED planar surface. However, it seems that this lensed output has a central lobe and two side lobes.

I would be interested in your comments as to if a more sophisticated optical component could be used with such a surface emitting LED (on a die) to construct an 'ideal' illuminator source. Or have the lads at companies such as Luxeon already figured this out?

Regards, Keith

PS. I seem to remember a European(?) company promoting such a source in conjunction with a fiber optic cable, for high resolution microscopy, but I don't remember a name right now.

One more question for you, Marien.

Have you tried to resolve the lines using the original "Critical" ilumination?

I am able to resolve the dots on F.Rhomboides with 47.5x n.a. 0,95 objective without blue filter or blue LED using "critical" and oblique illumination but, not any chance of lines with Amphipleura Pellucia. Is this similar to what you have found with "critical " on your scope.

Thanks for the lens tip, I think that Surplus Shed can fix my Koehler situation for about $5.

Ron

Keith & Gonzalez,
As long as the light is chopped fast enough that all the pixels are illuminated copping the chopping rate doesn't make an alias with the scan rate and the pixels charge is the same for the light from pixel to pixel PWM light should work. You can also put an inductor in the series with the LED and the PWM modulator and smooth out the pulses to almost ripple free DC. Off the top of my head an an RC circuit should work as well.

It is also not impossible to sync sync the camera and the light source together.

Again off the top of my head I would think with out looking anything up all the pixels should only illuminated once in the picture taking cycle or the pulse rate should be greater than the dynamic range of the sensor. In the case of 8 bits thats more than 256 puses during the period the pixel is cleared and before the camera reads it. That could be a very fast pulse rate.

The 2,000 HZ rate for florescent lights for video cameras was to keep from having aliasing problems with the ones with fast scans.

Gordon

Hi Gordon,
You wrote:
> As long as the light is chopped fast enough that all the pixels are illuminated [equally] the chopping rate
> doesn't make an alias with the scan rate and the pixels charge (CCD) is the same for the light from pixel to
> pixel PWM light should work.

This makes sense. There may be a Nyquist sampling constraint; but regardless, the key factor is "chopped fast enough". I think what Peter may have been referring to with respect to "high end cameras" was the higher resolution and corresponding pixel (sampling) rates. I have not run any numbers, but the required chopping rate may well be above 10Mhz!

> ....PWM modulator and smooth out the pulses to almost ripple free DC.
Again this makes sense, but I think it blows my basic idea out of the water re getting an increased effective intensity from the LED. But this concept was flawed as it only applies to the eyes perception of the intensity - I think.

While we are on the PWM topic, is it not correct that Luxeon uses PWM control to ensure that the LED does not self-destruct?

Regards, Keith

Dear Gordon
Integrating the output from the PWM to get smooth dc probably won't work well to reduce the LED brightness as the light output with varying dc voltage is unpredictable (and quickly drops to zero below a certain voltage). Anyway you could just use a variable in-line resistor (with a shunt resistor to ensure you don't go over-current) to get the same effect much more simply

PWM pulses the LED on and off very rapidly to produce the appearance of reduced brightness. Circuits abound on the internet:
http://superpositioned.com/articles/2006/02/25/diy-led-lighting-guide#PWM

Hello Ron,
It's possible to see the lines of A.p. with N.A. only 0.95. I made this picture last time. See in the photosection:
  Link : http://tech.ph.groups.yahoo.com/group/Microscope/photos/...

Regards, Peter

Hello Gordon,
My camera has a fast shutter with down to 0.2ms. The resolution is not 8bit as normal, this type has 12bit! So I have to pulse very very fast to get a constant brightness. 12bit means that I need 0.02% of constant light, or you see a flicker.

Its so easy to get a high end DC current source.

Best regards, Peter

Hello Ron,
> Have you tried to resolve the lines using the original "Critical" illumination?

That's a bit difficult. I can only make critical illumination by removing the diffuser and lens above the diffuser completely. I don't have a lens above the filament then. I didn't test that. I will try tomorrow.

> Thanks for the lens tip, I think that Surplus Shed can fix my Koehler situation for about $5.

Yes, it's not always necessary to use the most expensive solution.

Kind regards, Marien

Hi Peter,
Very clean pictures! Are these with Koehler lighting ? Can this be achieved with "critical " ?

Thanks! Ron

"scitech200" wrote:
> My understanding is that Marien's Ap image was a simple 'averaging'
> of static images, without re-alignment, with his MICAM software.

You are right Keith, In the experimental version of MICAM I use now I can average four consecutive captured frames (the only reason to do that is minimizing noise). I am working on a version with the ability to average more frames, to see were the optimum is. I think the optimum will be at 8 or 16 frames, Don also said averaging many frames doesn't give a much better result.

I wrote MICAM because I wanted to save a captured frame with additional information, just with one mouse click. I also want to incorporate features from my NAFIS program, were you can select a name from a database to give the image a meaningful name. That really saves a lot of typing.

When ready I will put this version with the averaging capabilities on my site.

Kind regards, Marien

Some members may not be aware that there is a simple statistical calculation for the reduction in noise achieved by averaging (arithmetic mean) frames. Taking the standard deviation as a measure of the spread of noise across one pixel in multiple frames the standard deviation of that pixel, in the average of N frames, is reduced to 1 divided by the square root of N of its original value, showing that as N increases pixel values get increasingly closer, on average, to the long term mean.

Thus the average of 4 images can be said to halve the noise, the average of 100 images to reduce it to a tenth, etc. It is apparent that diminishing returns soon set in.

Selwyn

Hello Ron,
the illumination is like an "Oblique Illumination":
  Link : http://micro.magnet.fsu.edu/primer/techniques/oblique/obliqueintro.html
The Slit or Sector Stop (like in the picture) is absolute necessary.

Only with this illumination I got the dots of A.p. with
N.A. > 1,3 , or the lines with N.A. >0,9

Regards, Peter

"Ron Jones" wrote:
> Very clean pictures! Are these with Koehler lighting? Can this be achieved with "critical"

Thanks Peter,
This will help.

Marien,
The problem with most achromatic objectives is that they don't perform very well in the blue. Therefore you find quite bad spherical aberration off axis, and increasing astigmatism. However, if the object is pretty close to the axis, and only a small part is in focus, the effect may be quite small.

The increase in resolution due to wavelength alone is only about 25%, and does not explain the very good contrast in your pictures, other than that by the Signal/Noise improvement due to stacking. I will down load your software and try it out with the Vesta Pro, will post some images.

Thanks, Merv

Hello Merv,
> ...The increase in resolution due to wavelength alone is only about 25%, and does
> not explain the very good contrast in your pictures, other than that by the Signal/Noise
> improvement due to stacking.

Yes, 430 nm is not so different from 560 nm. About 25 % indeed.

I can show you the influence of stacking. Take a look at my "microscope Pics" album.
"Captured AP" is a single frame captured. "4 images stacked" shows the result is of stacking 4 images. That is four different frames, not the same frames of course.

I noticed that moving the lens above my blue led just a bit caused a much more clear pattern. As the diffraction pattern of every "slit" has a sin(x)/x shape, isn't it possible that the patterns have some kind of constructive interference?

Kind regards, marien

Hello Marien,
you write:
> Yes, 430 nm is not so different from 560 nm. About 25 % indeed.

Right, but with a blue LED with 455nm I have had problems to get the dots of A.p.
With the UV-LED from Roithner with 385nm I have now no problems to get the dots.
May be that you have other results.

Best regards, Peter

Hi Marien,
It is very possible that constructive interference is taking place, in particular with such a small, planar light source.

If you move the lens above your LED, displacing its centre of curvature with respect to the optical axis of the microscope, then you no longer have a symmetrical illumination function, and the illumination becomes increasingly oblique. This will have a prefered axis with respect to the stria on AP - If you have a rotatable stage you can revolve the slide and watch the image contrast come and go. Ideally, the obliquity needs to be orientated along the length of AP, worst visibility occurs when the obliquity is orthogonal to the long axis.

To resolve the dots on Ap you need symmetrical illumination, such as that given by COL.

I had a look at your stacked images, they are first class, and you have certainly proved that you don't need excessive firepower to achieve excellent results. The other variable that will affect your results is the focal plane of the condenser, the old microscopists used to recommend racking up the condenser slightly to achieve the best image contrast. This cetainly makes a difference with an Abbe condenser, but not much difference with an Aplanatic or Achromatic condensers.

Cheers, Merv

Hi Keith,
That is a good point, and well worth investigating.

These LEDs must be very similar to the old 'Pointolite' which gave a very good approximation to a point source. This was a tiny ball of tungsten kept in a near molten state by an arc between two electrodes. The LEDs are small in size, perfectly planar, and therefore should perform even better than a pointolite. Also, the longitudinal coherence should be excellent. There is certainly an argument for exploiting this characteristic, coupled with a suitible optical chain.

My own experiments with ribbon filament lamps show the advantages of spatial coherence, and give much superior performance to a conventional folded tungsten filament. In fact, I have serious doubts regarding Kohler illumination, as although it gives the illusion of homogeneous illumination, I suspect that the lack of longitudinal coherence does still affect the relative phase across the aperture, and therefore the final result. The illuminating source's wavefronts carry their spatial distribution through the object on the stage, before it reaches the objective. Therefore, throwing the image of the filament into the back focal plane of the objective, although pleasing to the eye, can have little effect on the relative spatial coherence of the illumination - it remains that of the source. Hence you would not expect to see any difference in resolution between Kohler and critical.
But you do see a difference using an extended, planar source, and it is impossible to detct any difference between 'critical' illumination and Kohler using the ribbon filament lamp. This suggest that the form of the source is critical for the best results, rather than the position of the source image.

None of the modern Tungsten-halogen sources can be considered ideal, and what most people accept as 'Kohler' is a long way from that theoretical ideal. So high intensity LEDs could be a real answer. Psuedo point sources for metallurgical oblique illumination have been made using a TH lamp feeding a fibre optical - this simulates a source with good longitudinal coherence, and because of the large number of fibres in a bundle, gives poor lateral coherence. Which the theorists say is ideal!

Cheers, Merv

"scitech200" wrote:
> I have wondered if the LED (probably on a die) could be considered
> to be close to ideal point source for light microscopy illumination.