Fixed Lighting - 10W Tungsten to LED Swap

Flashlight Modding and DIY Parts
1

The bulb in my microscope burnt out a couple of nights ago. It’s listed on their website for $10. I checked around on other sites and found similar items for $8.99. It seemed a bit pricey for a bulb that only lasts maybe 20 hours, so I tried to retrofit it with an LED using parts that I already had on hand.

Here are the parts that I chose to use along with the etched boards. The LED is a 6V Bridgelux with a 3000K color temp. I’ve had it for a few years and finally found a project to use it in. A 100uF tantalum was the largest value SMT capacitor that I could find in my parts bin. I was a bit hesitant to use it for filtering, but decided to go ahead with it anyways. More on that in a second. :wink:

I wanted the boards to be stacked on top each other with the bulb base at the bottom. Then it would all be held together with epoxy.

After it was all soldered together, I did a quick test on the bench with a 12V isolation transformer. I wish I took a photo of it, but once I flipped the switch the LED blinked and then suddenly the tantalum cap failed in spectacular fashion. It was so hot that it de-soldered itself from the board. The chard remains fell onto the plastic table below the microscope. I watched as the glowing red capacitor melted its way into the plastic.

When I initially tested this circuit on the bench with a breadboard, I was using a through-hole 100uF aluminum electrolytic capacitor. I imagine the ripple is what killed the tantalum. The ESR of this model wasn’t particularly low at 150 milliohm.

The design had to be changed a bit and this time I chose to squeeze on a 220uF electrolytic capacitor. When I went to change out the wires on the LED, I must have heated the board too much, or too many times and the positive contact pad lifted off the MCPCB. I tried to hold it down with epoxy with no luck. The LED had to be scrapped and a high CRI Nichia took its place.

Here’s a photo of the bulb holder and the tube that focuses and directs the light.

And finally, the finished light.

I plan to go back and adjust the current a bit now that a different LED is being used. It’s currently being driven at 100mA. With that small, thin MCPCB I can probably get away with 150 or 200 mA.

2

Nice, I think the Nichia has a much more useful colour temperature than the Bridgelux for most applications (even apart from the CRI) :-)

It might not be that hard to mount the led-board on a piece of aluminium that touches the side of the housing to improve the heatsinking a bit. But perhaps the 200mA is good enough and then you will not need it.

Very nice microscope by the way, looks very very useful :-)

3

Thanks, a 20mm star board should fit nicely into the tube. I’ll have to dig through some boxes and see if I have one with this footprint or maybe even glue a different one to back of this 16mm board with some thermal epoxy.

And I think you’re right. The Nichia is a much nicer LED.

4

Moar scope pictures! Here are the 2 in our shop. Make small things big! (Sorry for the blurry cam)

5

nice job! I’d love to upgrade all the scopes in our lab to LED, especially the metal halide mercury burner (to generate UV) which takes $400 2000h bulbs. The current big thing in microscopes is LED lighting, as you can be very precise with the wavelengths (eliminating excitation filters) and they can be switched on and off at will and at speed. It’s exciting stuff, even if they’re a little late to the party :slight_smile:

6

Nice scopes, is the one on the right a Keyence? I have its competitor, the HiRox KH7700 at work. The depth of field is outstanding.

Good choice though on using the high CRI LED, I think you would have found the colors shown with the Bridgelux to be rather flat.

What circuitry did you use to convert the incoming AC to ??v DC to power the driver?