Help with Olympus BH2 (Reflected Light) LED retrofit

[siliconGary]

Hi, this looks like the right place to get some help on re-builiding a microscope illumination system.

I have a metallurgical Olympus BH2 (BHT model?) unfortunately the illumination lamp and tubing is missing broken off at the condenser aperture diaphragm.


I would like to restore and use the microscope, and figured retrofitting an high power LED light source would be interesting project and likely the most economical path. I have been researching this and have come unstuck as to what I actually need.

My initial plan is to use 1" SM-1 optical lens tubing and attach the LED to aluminum plate with heatsink, that is bolted to an optical cage like this one that will fit over the tube. My microscope has the field aperture, and bolted to it with a dovetail fitting is a lens with a thread, I am 90% sure this is an SM1 thread (1.035” diameter 40 TPI) but would appreciate if someone who knows about Olympus BH2 could confirm this (There is only ~5 thread pitches and even with a micrometer it is a bit hard to measure accurately.)

I am thinking of using a HP LED like this LumiLeds COB LED with 5000K colour temperature, 4542 luminous flux seems closer to a 100W halogen lamp brightness originally on this microscope. However one drawback is it operates at 900mA ~35V, and so far I have not seen many LED drivers that work up to 35V so I may build my own or choose another LED.

Now just whacking the LED on a tube will give me illumination that will let me use the microscope (I can already point a light source at the lens and resolve objects through all objectives). What I don’t know is how sub-optimal this will be.

I used to use reflected light microscopes frequently (Many years ago) and know optimizing the condenser aperture is key to maximum resolution. So I am thinking I probably need to add a iris diaphragm as a condenser aperture. This is doable with a lens tube system I can find two I could screw directly onto the lens element 1" variable aperture (I think a 12 mm maximum aperture will be sufficient as I measured the maximum field aperture as 8.6mm). Looking at diagrams I think the apertures on the BH2 are 1” apart, could someone confirm that. I would also like to know how critical the distance between those apertures is?

So I think I will add the condenser aperture back - now I come unstuck as to whether I need to put a lens in front of the LED to collimate and focus the light source. (The reflected light Kohler illumination information indicates you need to focus the light source are the plane of condenser aperture.) What I am looking for is with an LED light source would I see a difference/improvement if I were to have a proper collector lens set-up?

Now designing a collector lens system is getting out of my pay grade :-) ButI looked at lens and am thinking of using this aspherical lens https://www.thorlabs.com/thorproduct.cf ... =ACL25416U in combination with this bi-confex lens https://www.thorlabs.com/thorproduct.cf ... ber=LB1761 such that the bi-convex lens is ~25mm from the condenser iris (= focal length), and the back of aspheric lens is ~8mm from the LED (=back focal length). But I am getting way out of my depth here and would appreciate if this makes sense to someone who understands optics. The costs with the lens is now getting up there for me, and I would really like some feedback on whether it makes sense or not.


Bonus question: What does the spring loaded lever at the back of the trinocular head labelled "ND out” do?

[houstontx]

i think the ND stands for neutral density filter. Probably moves it in an out of the light path if present.

[enricosavazzi]

I can say nothing specific about this scope, but if you are using axial ilumination through the objective (not paraxial through a cowl around the objective), I have a similar setup in a Mitutoyo FS-60 equipped with Mitutoyo M Plan Apos up to 20x. With this scope, a 100W halogen fiberoptic source provides an exceedingly bright axial illumination of ordinary subjects like forams and sand samples. At the minimum intensity allowed by the illuminator (which is perhaps on the order of 10 W) intensity is just about right.

Based on this, I think you should try first a 5W or 10W single-chip LED, which should be enough for most subjects (I am aware that some metallurgical samples may require a much stronger illumination).

Single-chip LEDs are also far easier to collimate than a multi-chip arrays like those in high-power cobs. I seem to remember a recent thread where a transmitted microscope illumination using a multi-chip LED array failed to provide more light at the subject than a single-chip LED. I do not recall the details, but I think this was due to the difficulty of collimating the light from multiple chips mounted on a wide surface onto a single point. Basically, a conventional condenser only forms a smaller image of the array at the focal plane of the condenser, and most of the chips are cut off from contributing to illumination by the various apertures of the microscope, except for one or a few closest to the axial position. Avoiding this might require the use of individual optical fibers picking up light from each LED chip and combining it together by bundling the fibers together.

[Pau]

Enrico, I agree with all you say about LEDs for microscopes

Maybe you're referring to this recent thread:
http://www.photomacrography.net/forum/v ... hp?t=31950

or more likely to
http://www.photomacrography.net/forum/v ... hp?t=32012
where I linked to the materials I use:
http://www.photomacrography.net/forum/v ... 545#197545

Gary, welcome aboard!

It will be in function of the collimator lenses you use, with the complete optics designed for halogen or tungsten lamps, definitively you want a powerful single chip LED to mimic the lamp filament. I'm full satisfied with the Cree XLamp XML T6 Neutral White 4000k (10W), much better than COB LEDs already tested

[siliconGary]

Thanks for the feedback about the LED's. I will look at getting a high power single chip LED. I would like to be able to use dark field so I think it needs a good brightness level. It also has the benefit of 12V operation, so lots more choice for driver chips and modules.
In your comments you both refer to collimating the light, suggesting I do need to add lenses in front of the LED. So I am worried about what lens I need to get to collimate the light and focus on the condenser aperture. It would be a lot easier to retrofit an LED to an existing lamp housing with the filter tube, collimator lens (And condenser aperture) intact but I have not found anything for sale.

[enricosavazzi]

[...]
In your comments you both refer to collimating the light, suggesting I do need to add lenses in front of the LED. So I am worried about what lens I need to get to collimate the light and focus on the condenser aperture. It would be a lot easier to retrofit an LED to an existing lamp housing with the filter tube, collimator lens (And condenser aperture) intact but I have not found anything for sale.

Most lamp housings have a convex lens (often planoconvex with the flat side toward the lamp). Getting an old lamp housing may help you to obtain a suitable lens and a mechanical connection to the microscope, but you probably should discard the part of the housing that surrounds the lamp (a LED should be mounted onto a relatively large heatsink, which works best in open air).

It would be best to use a heatsink large enough that it does not need a fan. In general, heatsinks with fins work best with the fins vertical to create a natural air flow, so a cylindrical heatsink with longitudinal fins and a LED mounted at the base of the cylinder works best with the LED pointing downwards. A flat heatsink with fins on one side or on opposite sides works best with the LED pointing sideways. You also don't need the reflector usually mounted at the rear of the original lamp. Then mount the heatsink in the housing so that the LED sits at the same position where the lamp filament was originally located in the housing, and the chip surface points directly into the opening of the lamp housing and into the microscope.

A good idea is to build an aluminium bracket to hold the LED and its heatsink in the housing, make large screw holes in this bracket, and use screws with large washers to attach the bracket between the heatsink and the housing, so that you can adjust the position of the LED in three dimensions by a few mm. This allows you to center the LED.

If you don't have a lens from a lamp housing, try to get a LED parabolic lens not too large, since if it is larger than the exit of the lamp housing any light collected through the periphery of the lens cannot enter the microscope. This means choosing a lens with as low as possible focal length. You might even try to operate the lamp housing without a lens, since a low NA collimating lens (i.e. too small diameter and/or too high focal length) may make things worse than no lens at all. Mount the lens with its flat(ter) surface toward the LED, with the LED chip at its focus point (if you can see a magnified image of the LED chip approximately in focus through the lens from a distance of at least half a meter, the distance between LED and lens is approximately right, or slightly too much).

[siliconGary]

UPDATE:
I found an image of the RLA illuminator that was originally on my BH2 microscope



(Found at http://www.m2associates.com/wp-content/ ... inator.pdf)

Looking at it and reviewing an old BH2 brochure, I realized that this does not have a second aperture. All I am missing from mine is the filter tray, lamp housing containing a condenser lens. It also explains why I have a lens before the aperture diaphragm (I couldn't understand why there should be a lens between the two apertures in a kohler illumination.)

So this simplifies things quite a bit. I have ordered the parts. I went with the Cree XML LED that PAU referred to - and am going to add a Carlco LED optic (20mm narrow beam - I have ordered both clear and frosted pointing at an aspheric condenser lens (16mm Focal length) which I am going to mount in a lens tube.[/url]

[Pau]

Nice that you found the image.

To optimize the resolution of the lens close, the Aperture down 90%. You can visually see resolution improve as you close this Aperture down.

There is a mistake in the text, closing the aperture diaphragm does not increase resolution (in fact it reduces it), it increases contrast.
The Olympus brochure for the older BHM has it right:

Adjust the opening of the diaphragm to match the N.A. of the objective in use. If the specimen is low in contrast, it IS recommended to stop down to 70-80% of the objective N.A.

Dis you see the diagram here?:
http://www.alanwood.net/downloads/olymp ... ochure.pdf
It refers to the two diaphragms illuminator, but at pags. 8-12 you can also see your BF/DF single diaphragm illuminator

Likely you will need the lamp collector lens to be focusable.

Please keep we informed of the setup and results when done. At same point I will want to do something similar with a more incomplete Zeiss vertical illuminator I have (in some tests an external microscope illuminator seems to work well, but it is not very comfortable to use.

[siliconGary]

Dis you see the diagram here?:
http://www.alanwood.net/downloads/olymp ... ochure.pdf
It refers to the two diaphragms illuminator, but at pags. 10-12 you can also see your BF/DF single diaphragm illuminator

Likely you will need the lamp collector lens to be focusable.

Yes that is the brochure I saw, and confirmed for me I only have the one aperture.

The lens tube I am using, has internal threads that hold the aspheric lens with retaining rings, so I should be able to focus the lens during assembly. What I don't know is whether the distance between the Carclo lens/LED and the aspheric collector lens is important. I will be able to move the Carclo lens/LED assembly in an out by varying the screws I am holding the plate to the cage plate.

I will post images when it is done.

[siliconGary]

The last parts arrived today (The LED's were on a slow boat from China).

Here is the components laid out - From left to right Olympus lens with dovetail fitting and screw thread (Part of microscope) - 2x1" lens tube lens - 1" aspheric condenser lens - tube cage - frosted 20mm carclo lens and holder, cree xml led - 0.19" aluminum plate - heatsink


I was pretty worried before getting the lens tube that it would fit on the olympus lens. I thought it was a SM1 thread, but reading about SM1 threads they seem to be only used by Thorlabs and maybe proprietary. Anyway the tube does screw onto the lens!!

Here is the assembled lamp housing - I made two one for the vertical illuminator and one for the transmission illuminator

[/img]

I measured as best I could the back focal length of the olympus lens, and knowing the focal length of the aspheric lens I positioned the aspheric lens such that the two lens converge. That left the back of the aspheric lens ~6-8mm from the front of the carclo lens which I think is ok. I use the frosted carclo lens as a cheap diffuser. In all the parts were a little over US$100 which with addition of a little local hardware and aluminum with shipping costs and exchange rate was a little over CAD $200.

Here is the assembled lamp housing with the olympus dovetail fitting lens attached


And here is the microscope with the new lamps and controller



It is working - I don't have a condenser lens so cannot check transmission, but the reflected light is working fine and appears bright and with a good range of brightness and as far as I can tell uniform illumination.

I would post an image but I have a confession - I don't own a camera other then my iPhone5??! (I know this a photography forum - please don't kick me off ;-) Adding a DSLR is going to be stage 2 of the project.

FWIW I did try and capture an image down the eyepiece with my iPhone



I am thinking of getting a condenser and trying slide microscopy, oddly I have only ever used reflected light microscopes. In addition I need to look at the objectives on this microscope - I have an odd combination:-

5x & 20x Olympus NeoDPlan infinity/0
10x & 40x Nikon BDPlan 210/0 DIC

I think they are all ok/good objectives but the 5/20x are not par-focal with 10/40x. I am not sure if any of these were original to the microscope.

[Alan Wood]

5x & 20x Olympus NeoDPlan infinity/0

These 2 objectives were designed for use with your microscope.

Alan Wood

[Charles Krebs]

Nicely accomplished project!

As Alan pointed out, the two Olympus are the proper objectives to be used with the set-up you show. The two Nikon objectives are definitely not. Will you get an image with the Nikons?... likely yes. What will the image quality be...???

There are two basic types of microscope optical systems... finite and infinity. Your base Olympus microscope, with no accessories, is a 160mm finite tube length type. When you add that vertical illuminator it adds a 180mm "tube lens" and effectively converts it to an "infinity" system (note the infinity symbol and f=180 on the illuminator body.) The Olympus NeoDPlan infinity/0 objectives are "infinity" type, and designed to be used with no coverslip (the "0"). The Nikon objectives are "finite", and designed to be used with a finite, 210mm tube length microscope.

If this make sense, fine. If not, it's quite straightforward to understand, have a look at this page and similar ones online:
https://www.microscopyu.com/microscopy- ... microscopy