LED choice. Which one?

[Chris Dee]

Hi everyone. I'm planning to modify a used Chinese microscope. Its a trinocular finite model with hybrid kohler illumination (frosted glass diffuser in collector) currently with 6v 20w halogen illumination. Using the supplied 40x and 60x objectives is a dim experience, the lossy frosted diffuser and fixed photo port light-share ratio (via prism) being the culprits I suspect.

I want to replace the light source with an LED and have a couple in stock that have plenty of lumens for the job, a Cree XHP-50.2 (4500k 80CRI) and a Cree MK-R (6500k 70CRI). I plan on running these at ~13w max from an external 12v supply and a DIY variable current limiter. There's a noticeable difference in tone and I think I prefer the 6500k MK-R but the CRI is higher on the XHP.. Which one would you pick?

[MicroBob]

Hi Chris,
welcome to the forum!
The light colour you actually see depends also on the light path through the microscope. Some instruments are far from neutral in colour.
I have used a couple of XHP 50 4000K LEDs which give a pleasant view for me. The cold light LEDs tend to have a really deep valley in the blue-green region and a high peak in the blue region. This is more equal with the warmer tones.
It is also possible to use LEDs together with light filters: I just used a 2700K Osram LED because this was what was available right now. Together with a blue filter the light is just nice. Hobbyst46 did a test on filters last year or so.

You need a good cooling for strong LEDs. I use round aluminium rods, attatch LED with heat transfer paste and have a nice fit in the base to transfer as much heat as possible. On the outside end a couple of cooling fins and preferrably a thin coat of paint for better heat radiation.

Bob

[Chris Dee]

Thanks Bob, it didn't occur to me that I could use filters (lots to catch up on). As the guts of the present illumination system are coming out, I think I can shoehorn a substantial heatsink into the base for the LED and the semiconductors of the current shunt circuit. I have a spare XHP50 also, so probably the smart choice. Thanks for the info.

[MichaelG.]

Hello, Chris
The declared Colour Rendering Index will only be valid when running the LED at its rated current anyway ... So, taking Bob's advice into account, the right choice for you will be the one that looks right to you, using your 'scope.

The perceived/equivalent Colour Temperature and CRI might both change dramatically with reduction in current ... Experiments are called-for [and do please share your findings].

MichaelG.
.
P.S. ... there is some interesting ['though perhaps not specifically relevant] information here:
http://research.ng-london.org.uk/scient ... ?page=info

[Hobbyst46]

...

Hello Chris,
1. A 13W LED creates a lot of heat. If this heat is distributed by a heat sink WITHIN the base of the microscope, I would verify that the base is not harmed by the heat. Especially plastic parts, paint, nearby knobs etc.
Although 13W is less than 20W, heat distribution among the components of the illuminator is not the same in the halogen and LED.
2. A 6500K LED is very useful if you want intense blue illumination from time to time. Then, use a short-pass filter to remove all light from 460nm up. For the other sessions, when you want white or yellowish light, a warm-up amber filter, say K2 or KR12, will remove much (not all) of the blue component, although the camera will still sense it.
If intense blue light is not needed at all, I would choose the 4000K or 4500 K LED. Or even 3000K.
3. It is important that the LED die will be located at exactly the same point where the halogen bulb filament is now.

[Chris Dee]

Michael: Thanks and noted. I was hoping there would be an obvious choice but think I'll have to compare both LEDs. I'll do some tests time permitting. I assume a DSLR might be able to do some comparative shots for CRI? Custom WB?

Hobbyst46: Thanks also for the suggestions. The finned heatsink I've selected is 65x70x30mm with a 5mm thick mounting surface, it''l be open vented through the plastic base. I expect it to be sufficient, but if heat becomes an issue there's space for a small blower. I'm aware of positioning in relation to the bulb filament (gleaned from this forum), interestingly the alignment of the halogen filament is off-centre by ~4mm stock. I'm assuming this is slack production tolerances rather than intentional? Frosted diffusers being more forgiving than high-end Kohler systems.

[MicroBob]

3. It is important that the LED die will be located at exactly the same point where the halogen bulb filament is now.

For a true Köhler setup this is true, but many microscopes have a frosted lens somewhere in the light path that acts as the new light source. This lens would have to be illuminated evenly, from whatever position of the LED that does it. My Zeiss Jena NF offers the choice to use or not use the frosted disc. Zeiss Standard and Phomi have a fixed frosted lens surface. In my NF I replaced the coarse frosted glass with a disc I made from a special acrylic glas that offers less light loss and better smoothness. This material is unfortunately not readily available, but other materials are worth a try.

@Michael: Thank you for the link! You often come up with interesting sources of information!

Bob

[MichaelG.]

... In my NF I replaced the coarse frosted glass with a disc I made from a special acrylic glas that offers less light loss and better smoothness. This material is unfortunately not readily available, but other materials are worth a try.

I don't know if this is the acrylic that you used,Bob ... but it is very effective, and available at modest price from 'Southern Acrylics'
https://www.ebay.co.uk/itm/LIGHT-BOX-AC ... 2116646902
https://www.ebay.co.uk/itm/LIGHT-BOX-AC ... 2117844690

The laser-cut discs that I bought for my Nachet stereomicroscope were about 0.5mm undersize, so I will probably buy some squares and machine them myself ... but overall it's a great product at a bargain price.

MichaelG.

[MicroBob]

Hi Michael,
the stuff I have has a certain name and numer, it came from a german supplier, but it might very well be the same material. Where ground glass and white tinted glass have their drawback this material seems to have fine reflecting dust throughout the material. This leads to very nice properties.

Bob

[MichaelG.]

It's rather short on technical detail, Bob ... but here is the PERSPEX page:
https://perspex.com/product-ranges/pers ... ctrum-led/

The material I used is 'Colour 1TL2'
... it looks just like the material on X-Ray viewing boxes, but is allegedly optimised for white LED sources.

Edit:
I have just learned that, in the 380 - 790 nm range:
1TL1 has 37% light transmission
1TL2 has 51% light transmission

Ref: p34 of the 'Perspex Design Guide'

MichaelG.

[MichaelG.]

Michael: Thanks and noted. I was hoping there would be an obvious choice but think I'll have to compare both LEDs. I'll do some tests time permitting. I assume a DSLR might be able to do some comparative shots for CRI? Custom WB?

Here's some bedtime reading for you, Chris:
https://www.tandfonline.com/doi/full/10 ... 89802#_i32

Regrettably: I believe most [all?] DSLRs have only two 'sliders' to define a Custom WB
Amber <-> Blue
Green <-> Magenta
i.e. you are setting a point on an XY plane.
So I'm not sure how far that will get you.

The detailed assessment of Colour Rendering uses fourteen discrete bands, but eight of these are averaged to produce the CRI that is usually quoted [see the document above].

Please note: I have no real expertise in this subject ... it just fascinates me.

MichaelG.

[Chris Dee]

Thanks for the info' Michael. Conformal testing is beyond my resources. If simple comparative shots show anything worthwhile I'll post the results. Right now I'm busy experimenting with photo tube designs for the scope. I'll post a thread on the lighting mod once complete. Thanks for the helpful suggestions everyone.

[MichaelG.]

Conformal testing is beyond my resources. If simple comparative shots show anything worthwhile I'll post the results.

Thanks Chris ... That's what we need.

MichaelG.