DIY microscope illuminator with LED?

Flashlight Modding and DIY Parts
1

As my first LED project, I’d like to make an illuminator for my microscope. I am totally inexperienced using LED’s. I have read this past post Microbrain tries to create a microscope light - #7 by drgentr which has a lot of ideas.

My initial thoughts are to use a single element LED (not COB), but I need it to provide as much light as possible. From my reading it seems I could use something similar to a Samsung LH351D COOL WHITE 5000K if I could find it already mounted on a board. Recommendations for something similar that is currently available?

To compare, some microscopes use a quartz-halogen lamp of 40-100W.

I will want to power it with 120VAC, so if I understand correctly I could use something similar to a Meanwell LPF-40D-12? I am confused by all the various offerings for drivers, so I also could use a recommendation for a driver that is currently available. It would need to be dimmable.

Would I need anything else other than a heatsink?

I will try to keep my budget for this frugal.

Please correct my ignorance. Thanks much for any guidance you can provide,
Keith
near Louisville KY, USA

2

I used a bench power supply - you can find them on Amazon or Aliexpress in any voltage/amperage rating a microscope light would require. This allows me to swap to different LEDs / LED arrays at my preference, using the same power supply.
I purchased one capable of 60V & 12A (Riden RD6012) since this would also allow me to power the BHS’ original halogen bulb from the bench power supply if I desired. Since it’s a switching power supply it has switching noise - you may want a linear power supply for less noise. They are 2x+ as expensive though.

Currently I use an array of Nichia E21A MCPCBs simply hot glued to simple aluminum heatsink roughly 12 inches by 6 inches (with thermal paste between.) Don’t worry about thermals, I’ve never had a problem there. You can acquire any desired color temperature by blending the correct mix of E21A emitters on an MCPCB. For blood stains I use an MCPCB covered with Nichia 219B SW45K R9080 emitters, because they are particularly pinkish in color, which is the one color you cannot get with E21As.

You can order MCPCBs pre-soldered with LEDs of your choice, from various amateur resellers (budgetlightforums and candlepowerforums are good resources to look at.) Or, you can buy bare MCPCBs for your specific LED’s footprint, and flow the desired emitters onto them yourself using solder paste and a heat gun, stove, or cigarette lighter. A good term to type into google is “Star PCB LED high CRI”

If you don’t want to bother with MCPCBs you can buy COB emitters. Sunlike and Bridgelux produce the highest-CRI COBs, but they are not quite as high CRI as smaller LEDs. An advantage of these giant COBs is you don’t have to worry about their focal plane very much.

This type setup is definitely superior to the original halogen, especially for high magnification darkfield with the oil condenser. It gives unprecedented control over the light entering the microscope. I can choose any color temperature, and any light intensity, which zero tradeoff between one and the other. I can use a reddish-tinted high-CRI set of LEDs (eg. nichia 219B) for examining blood. Really makes me wonder why modern microscope manufacturers don’t create more sophisticated lighting systems with multiple emitters like this.

Edit: btw, you should probably aim for 4000K or at the most 5000K color temperature, if not lower. Halogen is 5700K but most people use a filter that reduces it to more like 3000K anyway.
In LEDs, lower color temperatures (especially 4000K and below) have much less of a dip in the red color zones (r9 / r13) which are important for blood viewing. It’s also less blue and less harsh on the eyes for long viewing. 4000K is the golden zone imo. If you’re in an experimenting mood, you can de-dome existing LEDs to lower their color temperature in some cases.

Here’s an old pic from when I first started messing about with LEDs.

You can do all kinds of silly things like set paper specimens on fire, if you’re not careful.

image
image760×1012 46.8 KB

3

Thanks much drgentr!

OK to the 4000K color, was wondering about that.

That’s a pretty expensive power supply you have there… To my knowledge, bench supplies are neither constant current nor constant voltage… correct? I have a 12VDC 12A fixed linear supply- I would think that would be more than sufficient for a single emitter, but I’d have no way to dim control with that …

Confused by you using an array… for a microscope? Isn’t the goal for a microscope for the illuminator to be a point light source? That’s why I thought I should be looking for a single emitter rather than a cob? My scope is a vintage AO Spencer with a substage condenser (or substage miror) The substage space under my scope is pretty small- about 2X2x3 inches.

When you say you are using an array “MCPCB covered with Nichia 219B SW45K R9080 emitters” do you use some type of lens above them, or a diffuser?

Keith

4

Some of the arrays are used in focus, some out of it.
On my BH2 I found no bad effects from operating LEDs “outside kohler” (ie out of focus) though it could be different for your scope.

For using your fixed power supply you’d need to add a potentiometer and have different voltages of power supply for different MCPCBs, it’s more of a headache than I wanted to deal with, I’m no electrical engineer

Bench supplies can be either constant current or constant voltage depending on your setting. Well, some of them - depends on the features

You can prolly get a very cheap bench power supply as long as you don’t want it to be 60V for running the original halogen lamp.
I think on Amazon a decent one is like $50-$100

5

It’s a pleasure to have you onboard, keithostertag!

6

Here’s a Reddit discussion from someone with the same needs: https://www.reddit.com/r/flashlight/comments/18g3i7r/sources_for_nichia_e21a_219b_or_e17a_5000k_9080/

7

Thanks! Yes, that Reddit discussion sets out a lot of the same criteria! It’s months old- I’ll join that discussion to ask what how he’s discovered since.

Rather than a bench supply, which LED driver would work for this situation? There’s a bunch of cheap Chinese ones on eBay with questionable specs so I’d appreciate an recommendation for a specific available dependable one.

Here’s a different take on the problem: which flashlight might be useful? My microscope has a second option: a substage mirror that I could shine an appropriate flashlight on. The flashlight would need to meet most of the same criteria: tight beam, high lumens, high CRI, very low tint, single emitter point-light source, focus-able to about 8-10 inches… and enough energy to last maybe an hour of high output before charging? Should I post that query in a different section here on BLF?

8

Continuous dimmability is uncommon in flashlights, and you will want very granular control of light level. Many flashlights also have PWM which could give you eye strain over long durations

9

I have very little knowledge at designing lighting systems, just flashlights. I don’t have anything super concrete but will share some thoughts that might help you narrow preferences down:

  • I’m not sure how many lumens you need, but it’s probably lower than one would expect. People here use flashlights to light up large areas or long distances; the same flashlights would incinerate a specimen if used point-blank.

  • For tight beam at a short distance, look for a light with a small head diameter, which produces a beam that is divergent at long distances but very focused at short distances. I’d also look for a TIR-based flashlight instead of a reflector-based one, they give a beam that is more homogeneous in profile and color. A Convoy S2+ (with extra TIRs, running Nichia 519A LED) would be a decent starting point.

  • I’m not super convinced that low CCT is what you want. Lots of organic material is brown/yellow/red in color, and these colors become almost indistinguishable (even from white) under an excessively warm (yellow) light. A cooler light is much better for distinguishing warm colors. For context, overhead sunlight is around 5700K, and that’s the CCT I’d recommend.

  • I am curious, why the requirement for a point-like source? In this case look for emitters that have a small light-emitting surface. Simon at Convoy Flashlight Store (on Aliexpress) has Nichia 219F emitters in 5700K R9080, which are a bit smaller than 519A. Since you want high CRI, look for emitters ranked R9080, which means 90 CRI minimum (in practice they are 95+ typical) and 80 R9 (deep red) minimum.

  • Focusable will be hard, there aren’t many good zoomable lights. But these lights have so much output that maybe it’s ok to just increase the output instead of focusing to increase intensity, as long as the stray/wasted light doesn’t interfere with observation.

  • An hour on a high-ish output is a piece of cake for any light running 1x18650 or more/larger cells.

10

Like pretty much all single-die LEDs, that runs around 3V.

But that driver only maintains constant current down to 60% (7.2V), so no good. It’s meant for COB arrays, or strings of single-die LEDs wired in series, and will not work with one single-die LED. You could use it with multi-die LEDs, like a 12V XHP50 or XHP70, which use 4 LED dies placed next to each other to create one more-or-less continuous light emitting surface. (Picking your LED to fit the power supply would be a little backwards, of course; I’m just illustrating what it could be usable for. And if you did go that way, you’d probably want the LPF-40D-15 (for XHP50) or LPF-60D/90D-15 (for XHP70) instead – you’ll hit the 12V limit before you hit full current with either LED.)

In typical AC-powered lighting applications, it’s generally preferable to use several lower-powered LEDs instead of one big LED, and to connect those LEDs in series (more voltage) instead of parallel (more current), so most off-the-shelf LED drivers are set up that way. Then when you do want to drive a single, high-power, 3V LED? I’m sure there’s an off-the-shelf solution for that somewhere, but good luck finding it; I haven’t yet.

One option is a 5VDC power supply and a current limiting resistor. It’s pretty inefficient, but 5V power supplies are cheap and easy to find, and if I’m already screwing the LED board to a fan-powered heatsink, I can screw a power resistor (0.68 ohm 10W, for a 3A LED) right next to it. With a fixed power supply, there’s no dimming, but if your power supply is adjustable down to 2.5 or 3V (or you modify it to be), you get nonlinear dimming, which may be enough.

11

@QReciprocity42- Thank you so much for taking the time to provide such detailed info! I very much appreciate it, and I bet so will many other beginners who happen upon this thread in the future.

RE: point light source: most articles I have read about microscopes using a condenser mention that a point light source is best. I can’t explain it, but as one example the Wikipedia article on condensers begins with this statement: " A condenser is an optical lens that renders a light beam from a point light source into a parallel or converging beam to illuminate an object to be imaged."

My criteria for focusing comes from articles I am reading about Kohler lighting which apparently is best when the light source can be focused- when using filament bulbs the texts say to focus the lamp filament onto the substage aperture… but again, I am just repeating what I’ve read and am working on understanding… for instance it is still unclear to me how to apply that to other illumination sources. So it is not a matter of focusing with intent to increase intensity- the end goal is image resolution. See this if interested: Understanding the microscope. 3. Illumination. By Jeremy Sanderson

@BensonM- thank you for correcting me and providing me with more to consider. As drgentr suggested earlier in this thread perhaps a bench supply is easiest in this application- I have a bench supply that can provide several volts up to several amps. Maybe with a Nichia 219F or 519A-V1? as mentioned by Qreciprocity42 and which are reviewed elsewhere on this forum.

Keith

12

Thank you for explaining for me the need for a point source, and for linking the reference! I am still trying to understand it.

Now that I see the importance of having a point source, I’d suggest another LED to consider: the Luminus SST12 in 4000K, available at Kaidomain. It has a circular die even smaller than the 519A and 219F, and is also 95 CRI. It won’t be able to handle as much power but should still be more than enough. CCT at 4000K is lower than I’d like but still ok.

If the objective is to get light rays as close to parallel as possible, then a large lens/reflector would be needed, not a small one as I suggested before. But without knowing specific details of your implementation I should be cautious to make suggestions.

13

In this context of this microscope… the opening in the stage for light to pass through to the specimen is only about 22mm in diameter; any light outside that diameter is wasted and maybe contributes to glare. If using a substage illuminator, there is only about 2" or so to place the entire illuminator beneath that 22mm opening.

If using the mirror, for instance with a flashlight or spotlight etc, one still has just that 22mm opening. The mirror itself is about 2" in diameter which is partly why if using a flashlight it would need to pretty close, like 8-10 inches depending on lens/reflector/falloff etc. There are a bunch of older filament or halogen type illuminators out there: here’s one: OLYMPUS Microscope illuminator Model TL Tested Working Original TL (TL1) | eBay , but I think they are not powerful enough for my needs (with camera, filters, etc) though of course they would work for some things. This is my first microscope, so I have not had experience with any of them. This is the one I currently use: AO SPENCER VINTAGE LAMP ASSEMBLY ILLUMINATOR MICROSCOPE PART AS PICTURED R6-A-94 | eBay

Modern commercial LED microscope illuminators can cost a few hundred to a few thousand dollars.