De-doming (shaving) LED's

[Chris Dee]

A while ago I bought a few high power Cree LED's and tested them while refurbishing a Vickers M17 microscope. I found the Cree XHP35 to have the most even phosphorous distribution and light pattern. Its a single die 12v LED with a die size of 3.45 x 3.45mm which closely matched the original filament size of the 100w halogen in the M17 it replaced. More details in the restoration thread here: viewtopic.php?f=5&t=8975&start=30

The LED I really wanted for this project is the Cree XHP35 HI (high intensity) which has no dome and designed for use behind secondary optics (https://www.cree.com/led-components/med ... -XHP35.pdf). It has been suggested that domed LED's are not ideal for microscope illumination. Try as I might I've been unable source any for reasonable cost. Availability even for Chinese sellers has been limited due to the Covid-19 situation, and Cree has just recently discontinued the XHP35 line. Big brand stockists still carry them here in the UK, along with uneconomical delivery fees for hobby buyers :/ Then I stumbled up some posts on the candlepowerforums which documented DIY tests on LED's with the dome removed or shaved. Although these experiments were aimed more towards maximising the throw of flash lights it seemed like something to explore, and see what (if any) difference there was.

Slicing the dome off was very easy. I used two 1mm drill bits to form rails ensuring the cut was level when slicing with a wet razor blade.

XHP35

1.jpg (102.74 KiB) Viewed 3515 times

The domes are softish and cut cleanly and level with one pass of the blade at a 45 degree angle. In retrospect using 0.8mm drills may be a better choice but this seems to have worked fine and none of the domed area of the lens remains.

2.jpg (81.31 KiB) Viewed 3515 times

3.jpg (60.3 KiB) Viewed 3515 times

Needless to say anyone replicating this needs to be very careful handling razor blades. The damage they can inflict is no joke. I bare no responsibility for any injuries sustained!

Now for the results. When testing I tried to make things harder on the illumination by using achromatic objectives and higher contrast subjects for observation, hopefully making any differences more noticeable (objectives used were Vickers Microplan 4x, 25x and 40x). I took care to ensure the condenser and illumination variables (Kohler) were as consistent as possible across a range of test slides, and took a series of photo's using shaved and domed LED to examine. Rather than post a swathe of photo's I'll condense it to my observations. With higher power objectives I could not find any instance of variation in resolution, chromatic aberration, spherical aberration, or illumination evenness. However with the 4x objective, low NA illumination setting, condenser swing-top out, and no diffuser, there was a difference to observe.

Domed LED

4.jpg (234.84 KiB) Viewed 3515 times

Shaved LED

5.jpg (228.18 KiB) Viewed 3515 times

Hopefully You can see the improved illumination evenness of the shaved LED There is also slightly higher CA in some areas of the domed LED image, though I'm loathed to attribute this entirely to the illumination and could equally be a result of a slight variation in focus. A little disappointed there were no improvements to be had with higher power objectives but I suppose this was to be expected. Possibly something useful for members who have domed LED's and photograph subjects at low magnifications? Maybe, but I suspect the illumination systems of different microscopes models may also play a role. Done mainly to satisfy my own curiosity, but I thought I'd share the results.

[MichaelG.]

Neatly done, Chris
... and definitely looks worth the effort.

[not too sure about the ‘risk to life’ but you obviously survived intact]

MichaelG.

[Chris Dee]

As a side note I've recently been using UV-LOCA TP-2500F Glue for slide mounting subjects. This is a UV curing glue used for mounting screen covers on mobile phones and tablets. So far it is working very well. It doesn't harden fully so care needs to be taken with the amount used to avoid flooding of the cover slip edges. Shrinkage on drying is very low and bubbles are less of an issue than I've experienced with nail polish and PVA.

Also the Vickers Microplan achromats which came with the M17 scope are very capable objectives and resolve an impressive amount of detail. Well worth considering for the budget enthusiast given the low prices they generally go for on eBay. All photos are unedited, resized from Raw.

Vickers Microplan 40x

6.jpg (87.83 KiB) Viewed 3511 times

[Hobbyst46]

Thanks for posting. Amazing stuff!
It reminds me of a documented removal the domes of ordinary super-bright 5mm-diameter LEDs. These domes are made of hard transparent epoxy, and serve as lenses. It was done by a researcher about 30-40 years ago, to increase the illuminance of the LED. Removal was achieved by filing as much as possible without damage to the die. The surface was then ground and polished, to provide a normal surface for gluing the square-cut tip of a flexible optical fibre. Indeed, the process considerably improved the intensity of the light beam.

Is that photo of a pappus ?

[Chris Dee]

MichaelG, thanks. Yes my fingers survived without laceration, coffee drunk after the shaving

Hobbyst46 Yes its taken from a slide I made of Hypochaeris radicata pappus stained with Methyl Red. Thanks for the info on historical LED de-doming, no doubt where this all originated from.

[daruosha]

It was a very neat idea. I'll definitely try it next weekend.
Thanks for sharing your experience.

[MicroBob]

Hi Chris,
this is really an interesting topic! Also nice to see that your work resulted in a likely improvement.
I have read before of de-doming LEDs but had the impression that it were not modern anymore without knowing the pros and cons. Generally a domed LED is not the same as a flat bulb filament, this it clear. There are miroscopes with true Köhler-illumination whitout a frosted surface in the light path. Here the de-doming might have the strongest impact.

Bob

[Chris Dee]

Yes Bob this mod would have little benefit in illumination systems with a fixed diffuser in the light path, possibly a marginal increase in brightness. The example above is also a worst case scenario for the (full Kohler) illumination on my microscope, normally the diffuser would be deployed at 4x as recommended in the manual.