Microbrain tries to create a microscope light

23 posts / 0 new
Last post
drgentr
Offline
Last seen: 1 week 4 days ago
Joined: 03/24/2020 - 21:03
Posts: 8
Microbrain tries to create a microscope light

Heyo. I recently lucked into possession of a vintage Olympus BHS microscope. These things are cool, except for the 100W halogen goiter hanging off their backside. That thing... definitely is not cool when it's turned on. Plus the bulbs are mad expensive these days.

So, having no experience with making lights before, I've sought to replace the halogen lamphouse with a dimmable high-CRI LED (or set of LEDs.) I currently am thinking of the SunLike SAWS1566A COB (25W), which has excellent CRI & R9, and more output than the original halogen bulb. I'm aware it will need a decently sized heatsink that won't fit in the original lamphouse, so I will probably cannibalize the lamphouse for the front lens housing that connects into the microscope and attach it to a big Cooliance heatsink.

I've been doing a LOT of reading in the past couple of days. Where I am out of my depth here is... okay, everywhere... but especially the electrical stuff. I understand that I need a constant-current driver that outputs 720mA current (700mA would also be fine, I suspect? just slightly lower output,) @ at least 35.2 volts, and I need a power supply that takes 120v AC power & outputs DC power within the accepted voltage range of the driver. That sounds straightforward. However, I've never built a light before and the sheer amount of options here is overwhelming. It seems there are about a billion brands of drivers and power supplies. I would appreciate any advice on what I oughtta buy.


Lastly comes the thing I understand the least: PWM dimming. How do I do this? I would like to have a dial dimmer, but my caveman brain knows very little about PWM except that higher HZ = better (less visible to the eye & to cameras.) Any literature on this subject, or else a dumbie-proof suggestion on what to buy, would be appreciated.

Edited by: drgentr on 03/24/2020 - 21:59
Lightbringer
Lightbringer's picture
Offline
Last seen: 4 hours 35 min ago
Joined: 08/30/2016 - 14:12
Posts: 11105
Location: nyc

No idea what fits with what, but a CPU cooler with fan can be attached to the COB, and it’d allow more cooling than just passive cooling like most lights. Then you can bolt on the lens and worry about the electronix later.

09 F9 11 02 9D 74 E3 5B D8 41 56 C5 63 56 88 C0

drgentr
Offline
Last seen: 1 week 4 days ago
Joined: 03/24/2020 - 21:03
Posts: 8

I did think about using an active CPU heatsink initially, but it turns out vibrations in microscopy (such as may be introduced by a moving part like a fan) are a very bad thing! Any minor movement becomes amplified by, well, about a thousand times.

The best heatsink option I’ve found so far is Cooliance’s stuff, which is specifically designed to passively cool COBs.

Hoop
Hoop's picture
Offline
Last seen: 15 min 7 sec ago
Joined: 12/20/2012 - 05:33
Posts: 999
Location: Spokane, WA

I would look at Meanwell LED power supplies. The dimmable ones. LPF-40D-42 might be ideal. 3-in-1 dimming function, so it can be dimmed by several methods as described in the data sheet.

moderator007
moderator007's picture
Offline
Last seen: 2 hours 46 min ago
Joined: 12/23/2012 - 04:47
Posts: 3361
Location: North Carolina

Hoop wrote:
I would look at Meanwell LED power supplies. The dimmable ones. LPF-40D-42 might be ideal.

The dimmable one looks like a 100k pot can be used to control dimming. That make it turn a knob for brightness.
The original bulb should have made around 1400 lumens and the COB the OP linked makes 2655 lumens at 720 ma.
Reducing the current to a little more than half (say around 450 to 500 ma) will make the led run more efficient with less heat and would be easier to cool with more light than the original. I’m no expert with microscopes, but if you get to much light shinny on something your trying to see you cant hardly see anything, it just looks white. It over saturates your eyes. Maybe that’s not the case here, but just seems like a lot of light to me.
drgentr
Offline
Last seen: 1 week 4 days ago
Joined: 03/24/2020 - 21:03
Posts: 8

Thanks for the suggestion. Isn’t the output too high for the COB, though? 40W vs 25W.

drgentr
Offline
Last seen: 1 week 4 days ago
Joined: 03/24/2020 - 21:03
Posts: 8
moderator007 wrote:
The dimmable one looks like a 100k pot can be used to control dimming. That make it turn a knob for brightness. The original bulb should have made around 1400 lumens and the COB the OP linked makes 2655 lumens at 720 ma. Reducing the current to a little more than half (say around 450 to 500 ma) will make the led run more efficient with less heat and would be easier to cool with more light than the original. I’m no expert with microscopes, but if you get to much light shinny on something your trying to see you cant hardly see anything, it just looks white. It over saturates your eyes. Maybe that’s not the case here, but just seems like a lot of light to me.

That’s the idea – you have a lot of light on tap for things like micro-photography (cameras have a hard time with low light, like microscope images) or microscopy techniques that consume a lot of light (like polarized microscopy – which uses a filter to polarize the light, and that filter eats a good % of the light.) Then for regular brightfield you have a dimmer to turn the emitter down to normal levels. As I understand it, dimming with PWM is very efficient.

moderator007
moderator007's picture
Offline
Last seen: 2 hours 46 min ago
Joined: 12/23/2012 - 04:47
Posts: 3361
Location: North Carolina

From the linked page it shows a 50k pot being used to get about 50% of the rated current. It looks to me as though the pot rating determines the output current.
Its in a graph about half way down that page.

alpg88
Offline
Last seen: 7 hours 30 min ago
Joined: 10/11/2013 - 12:35
Posts: 767
Location: usa

is the halogen light source focused in anyway? cuz if it is, most likely focusing system will not work with led, i tried converting lcd projector to led, did not really work out.
this is high precision optical instrument, i bet it is really expensive, i would not mess with it

moderator007
moderator007's picture
Offline
Last seen: 2 hours 46 min ago
Joined: 12/23/2012 - 04:47
Posts: 3361
Location: North Carolina
drgentr wrote:
moderator007 wrote:
The dimmable one looks like a 100k pot can be used to control dimming. That make it turn a knob for brightness. The original bulb should have made around 1400 lumens and the COB the OP linked makes 2655 lumens at 720 ma. Reducing the current to a little more than half (say around 450 to 500 ma) will make the led run more efficient with less heat and would be easier to cool with more light than the original. I’m no expert with microscopes, but if you get to much light shinny on something your trying to see you cant hardly see anything, it just looks white. It over saturates your eyes. Maybe that’s not the case here, but just seems like a lot of light to me.

That’s the idea – you have a lot of light on tap for things like micro-photography (cameras have a hard time with low light, like microscope images) or microscopy techniques that consume a lot of light (like polarized microscopy – which uses a filter to polarize the light, and that filter eats a good % of the light.) Then for regular brightfield you have a dimmer to turn the emitter down to normal levels. As I understand it, dimming with PWM is very efficient.


Understandable then. If you look right below the pot to current graph they also have a nice wiring diagram to show how it would be wired up. You might give them a call and ask a few questions before you decide.
drgentr
Offline
Last seen: 1 week 4 days ago
Joined: 03/24/2020 - 21:03
Posts: 8

is the halogen light source focused in anyway? cuz if it is, most likely focusing system will not work with led, i tried converting lcd projector to led, did not really work out.
this is high precision optical instrument, i bet it is really expensive, i would not mess with it



Here's how the housing looks inside.
 I am a little worried about properly collimating the COB's output, but it should be doable. 

I think I understand how potentiometer ("pot") dimming is accomplished now, thanks for the advice all. Am I wrong in thinking that PWM dimming is ideal compared to potentiometer dimming?

Lightbringer
Lightbringer's picture
Offline
Last seen: 4 hours 35 min ago
Joined: 08/30/2016 - 14:12
Posts: 11105
Location: nyc

Yeah, “point source” from the bulb, big surface area from the COB. Light’ll go in allllllllll sorts of directions, and most of it not in the direction you want it to.

09 F9 11 02 9D 74 E3 5B D8 41 56 C5 63 56 88 C0

alpg88
Offline
Last seen: 7 hours 30 min ago
Joined: 10/11/2013 - 12:35
Posts: 767
Location: usa

yea, definitely focused, no, i would not mess with it, unless you want to ruin it. to do it properly is a lot harder than it looks. entire optical system will have to be redone.
what bulbs does it use? those are bipin bulbs, maybe you can use a different bulbs, if you can’t find oem, what is the voltage and wattage? i have few bulbs that look just like that

p.s. i googled it, that is $3000+ instrument, are you gonna use it for intended purpose? if not sell it. it worth good money.,

moderator007
moderator007's picture
Offline
Last seen: 2 hours 46 min ago
Joined: 12/23/2012 - 04:47
Posts: 3361
Location: North Carolina

Is this not the same bulb?
https://www.lightbulbs.com/product/hikari-00023
Maybe the filament is lower.
Does this look the original?
https://www.interlightus.com/light-bulb/WW-3RCQ-8

alpg88
Offline
Last seen: 7 hours 30 min ago
Joined: 10/11/2013 - 12:35
Posts: 767
Location: usa

that bulb looks like 15v150w, or 24w 250w, but gold color pins tells me it maybe something more, a special purpose bulb, unless they just look like that on the pic.

drgentr
Offline
Last seen: 1 week 4 days ago
Joined: 03/24/2020 - 21:03
Posts: 8

The BHS bulbs can be had online (for not-very-good price) but they last a median of 200hrs and the tint & CRI characteristics are, well, not good. The power supply is also 70s-era stuff and breaks down with some regularity. I know it's possible to use some combination of lenses to use a COB for this microscope - there are existing LED lamphouse replacements that use 10w Epistar COBs (super-vomitous tint, CRI and output.)

The optical requirements here are not as nuanced and intricate as you would be led to believe from the picture of the lamphouse interior above. The light doesn't have to be perfectly collimated into a shaft when it leaves the lamphouse. It just has to get out of the lamphouse and into the collection tube. It goes down the tube, hits a prism that tosses it upwards, where it hits the condenser. The condenser is what collimates the light for focusing on the specimen.

alpg88
Offline
Last seen: 7 hours 30 min ago
Joined: 10/11/2013 - 12:35
Posts: 767
Location: usa

halogen bulbs have 100cri.

actually it does have to be perfectly collimated, it has a complicated prism\mirror system.
http://www.alanwood.net/downloads/olympus-bh-2-brochure.pdf

drgentr
Offline
Last seen: 1 week 4 days ago
Joined: 03/24/2020 - 21:03
Posts: 8

alpg88 wrote:
halogen bulbs have 100cri

Oh, I stand corrected. It still looks really bad though somehow compared to a lot of my lights. I guess tint is everything.
alpg88
Offline
Last seen: 7 hours 30 min ago
Joined: 10/11/2013 - 12:35
Posts: 767
Location: usa

pdf says the bulb is 12v 100w, but the flat winding of filament is like no other 12v100w i’ve seen before, has to be special bulb, unless the bulb on your pic is not original.

moderator007
moderator007's picture
Offline
Last seen: 2 hours 46 min ago
Joined: 12/23/2012 - 04:47
Posts: 3361
Location: North Carolina

If you are willing to try to retrofit to a COB led, then have at it. That’s how alot of us here got into this hobby. Anywhere there’s a incandescent bulb we try to figure out a way to change it to a more efficient led bulb. I just finished modding my headlights to led a few days ago. I purchased led headlight bulbs that used xhp50 leds, I ordered 8 xhp50.2 5000k 80 cri leds and set in changing them out for the better cri and higher output. Worked like a charm. More light and better color retention than the yellow incandescent it came with.
I say go for it, you may learn a lot and teach us something along the way. It will require some trial and error and lots of reading. Thumbs Up

Hoop
Hoop's picture
Offline
Last seen: 15 min 7 sec ago
Joined: 12/20/2012 - 05:33
Posts: 999
Location: Spokane, WA

drgentr wrote:
Thanks for the suggestion. Isn’t the output too high for the COB, though? 40W vs 25W.

You don’t have to worry about the wattage capability of the driver being higher. The aforementioned driver is a constant current driver, it regulates the current, and will provide up to 1 Amp to the LED. The digikey page for your SunLuck led says 1.44A max current. (The link to the datasheet bounces.) The drivers operating voltage range of 25.2 ~ 42V looks suitable for a 36v nominal LED.
djozz
djozz's picture
Online
Last seen: 6 min 5 sec ago
Joined: 09/07/2012 - 17:04
Posts: 16705
Location: Amsterdam

Because of the way microscope lighthing is designed (Kohler lighting), the light source must be fairly small. You may get away with a COB but for me it would make more sense to go for a small high brightness power led, i.e. a high CRI Samsung LH351D at 5A should get you a brightness comparable to the 100W halogen and will have a better chance to work for microscope lighting than a COB led.

drgentr
Offline
Last seen: 1 week 4 days ago
Joined: 03/24/2020 - 21:03
Posts: 8

I know it would be better to use a smaller emitter, but they aren't available at this output & at this CRI & R9 (at least not without chaining enough of them to make it pointless.)

Found these interesting "double-TIR" optics designed for COBs. I reckon I'll buy a bunch of lenses and reflectors, spraybottle test the beam of the lamphouse, and try to replicate it with some combo of optics.