Optical comparator LED Upgrade.

[whitmore]

[about modules with onboard current/thermal limits, "KTB-HLDL-06" ]

Do you know the LED die size on these? Might just order a few to see how they work out. (I have a couple of the JANSJO LED desk lamps, so might even have some around)

They're Luxeon Rebel model, specs here and seem to have 3 x 4.5mm dimensions in the 700 mA size
(which is about the current limit on the JANSJO lamps, two AMC7135 in parallel).

 

[dkemppai]

They're Luxeon Rebel model, specs here and seem to have 3 x 4.5mm dimensions in the 700 mA size
(which is about the current limit on the JANSJO lamps, two AMC7135 in parallel).

Quick check shows incandescent bulbs at around 12lm/w efficacy. I have no idea if the projector style bulb runs at this level, but it's probably a good guess. The stock bulb in this unit is 70W, or about 800lm give or take. From the datasheet those Luxeon Rebels are running about 170lm at 700mA.

Anyway, I did a quick test with a JANSJO light as a surface light last night. There just isn't enough power there, even with the light head right next to the part. The optics in the actual surface light may help, but based on how much light the XM-L3 puts out, I don't think the Rebel will work as a surface light. For the surface light, the XM-L is running at a bit north of 3A, or close to 1150lm light output. Even at that level, the surface light could be brighter. The rebel isn't up it, I think...

It might be OK as a contour light, though. Compared to the Rebel a 170lm, the Cree is running pretty close at ~140lm at the 250mA minimum current I chose. Although, I'll admit things look better at ~1A to 1.5A from the Cree (~475lm to ~680lm ).

Based on the above, I think in this unit it should be getting around 800lm for contour, and some of the smaller die Cree LED's would work well. The result might be a slightly sharper image. But even then the etching on viewing screen gives some 'granularity' to the image, so sharper image may get lost hitting the screen anyway.

For bigger units, you may need double the light. It seems some of the mitutoyo bulbs are 150W, perhaps 1600lm??? The XM-L3 at 5A should do ~1800lm, so would be a pretty good match.

Dan

 

[gerritv]

My comparator/shadowgraph uses 10v 70w bulbs, with no built in reflector. The contour bulb has variable voltage via a wirewound pot, the surface light is full on.
I was thinking of using a 110v dimmable driver and then replacing the wirewound pot with a regular dimmer (led capable). In conjunction with your suggested XM-L2 or XM-L3. My screen is 12" so more light is better. The 100x lens has a beam splitter internally and gobbles up a lot of light for surface viewing.
You can see the surface light behind the lenses, it requires using a front surface mirror to reflect light if using the 10x or 20x lenses. My unit is at eye level, on a wheeled cart.

 

[dkemppai]

My comparator/shadowgraph uses 10v 70w bulbs, with no built in reflector. The contour bulb has variable voltage via a wirewound pot, the surface light is full on.
I was thinking of using a 110v dimmable driver and then replacing the wirewound pot with a regular dimmer (led capable). In conjunction with your suggested XM-L2 or XM-L3. My screen is 12" so more light is better. The 100x lens has a beam splitter internally and gobbles up a lot of light for surface viewing.
You can see the surface light behind the lenses, it requires using a front surface mirror to reflect light if using the 10x or 20x lenses. My unit is at eye level, on a wheeled cart.

This looks like a pretty nice machine. Certainly the magnification from 10x to 100x could be handy at times. The X-Y stage sure looks like the one shinko was using. Is the elevation/focus changed by moving the table up/down with the nuts below the X-Y?

The surface light being full on is a good hint of needing as much light as possible! Makes me wonder if an XHP50.3 or XHP70.3 might even be good choice there. Depending on optics, the XHP50.3 might be the better choice. The surface light driver would be simplified as dimming wouldn't really be needed. Cooling the XHP70 might be harder, although the ~5500lm might make it worth the trouble!

 

[gerritv]

This looks like a pretty nice machine. Certainly the magnification from 10x to 100x could be handy at times. The X-Y stage sure looks like the one shinko was using. Is the elevation/focus changed by moving the table up/down with the nuts below the X-Y?

There is a handle/wheel on rh side near bottom front. It drives a bevel gear on the column that the stage mounts on. Inside the column is also the front surface mirror that brings the light for contour viewing. Needs to be replaced, when I get a break using the thing. Once you have one of these it comes into play for a lot of things, like deciphering 0.5mm diameter threaded items to determine pitch and thread angle. The usual fish tail isn't much use at that point.

What do you think about using a dimmable driver with household dimmer to avoid bad supplies?

 

[dkemppai]

Once you have one of these it comes into play for a lot of things, like deciphering 0.5mm diameter threaded items to determine pitch and thread angle. The usual fish tail isn't much use at that point.

What do you think about using a dimmable driver with household dimmer to avoid bad supplies?

Slow response. Sorry, buried in major truck repairs right now...

Yeah, I think this will be handy. Even at only 10x, it should cover a lot of things I do. I also picked up a measurement microscope a while back, same idea, X-Y table with Mitutoyo digital micrometer heads and a cross hair (That one is 90x zoom, so really nice). It's slightly handier for small parts, only because of the digital micrometers. It is a lot handier if you're trying to measure something through glass (Like a lightbulb filiment for example). The comparator will do angles that would be more difficult on the micrscope (although with four x,y and cad, it can be done.).

I've been thinking about the dimmable driver idea. If you had a simple transformer and diodes, going to a filter cap, and fed the led through a resistor it would probably work. Dimmers are (usually) phase angle controls, that is, they just cut off part of the sine wave cycle (think 0 to 90, or 180 degrees of the sine wave. This changes the waveform, which makes the average voltage fall (thus the dimming).

Basically you'd make an old school analog adjustable DC supply. But then, you might as well just look for an adjustable LED driver. I'm sure they're out there, for probably not a lot of $$$. I went the cheapskate route, and picked a 5$ supply to start with.

Honestly knowing what I do now, for the contour light I might have gone this route.
http://kaidomain.com/S028637 (No affiliation, and haven't actually bought one.)
Throw 12V into this board, and you get constant current drive to an LED. At 2.5A into an XM-L3 it ought to be similar or maybe just a bit more light than a 70W projector bulb.

Dan

 

[gerritv]

Thank you Dean, that helps narrow down my choices.
The first power supply you bought, and 'fixed', did it really need fixing for the average person? Or was it really really terrible. There are so many choices it has become paralyzing :-(

 

[dkemppai]

The first power supply you bought, and 'fixed', did it really need fixing for the average person? Or was it really really terrible. There are so many choices it has become paralyzing :-(

Full disclosure, I do switching power supply design professionally and for fun. The power supply I bought was absolutely horrible. There was no way to make it operate correctly out of the box. Short story is I picked it because it looked like it had all of the right parts in the right places. Upon close inspection after reverse engineering, it was absolutely unusable without rework.

Yes, there are a lot of options. However the biggest problem with finding an LED driver for a single LED, is the LED market right now. There are basically two segments, the flashlight/headlight group (small segment) and AC powered lighting group (big segment).

The AC lighting guys care about big lumens spread out over the area of a light bulb, T8 tube, or high bay down lighting fixture. They string A LOT of 3V LED's in series and run them at higher voltages (20V or higher) off an AC source. There you really don't want a 'point' source of light anyway, so many LED's is a good thing.

The flashlight guys on the other hand, run one really high powered LED in a reflector. They want one point source of light, so it can be focused as needed. The problem is they do that from one or two 3.7V Li-ion batteries, so the drivers are designed to run from low voltages (3V to 9V), and are designed for space constrained designs in a small flashlight tube. The boards are hard to hold on to, and hard to keep cool.

For a fixed instrument lighting source like this, running from 12V or 24V is nice. But there aren't many LED drivers for the 12V-24V input range designed for single LED use. That application really falls into the 'gap' between the market segments. Thus the reason I picked these supplies. If they operated as advertised, they would have been fine.

If you can find a fixed, single mode LED driver that runs from 12V, and drives one LED, that's the way to go. Otherwise, I think a resistor and 5V supply is the next easiest thing to try. Beyond that there are many other options, but none are neat/tidy or cheap.

Again, if you really want to do a conversion, let me know. I'll help where I can.

Dan

 

[gerritv]

Thanks Dan, that is an excellent explanation.

I think it best then if I just order a couple of XM-L3's and use my own variable supply for now. My brother also does analog stuff, from radar to high power servo drivers. (I'm great with digital only) Between the 2 of you I am sure there is a worakble solution. If the light output is good enough to replace my bulbs then I will go the next step, a driver.

The XM-L3 mounts with aluminum back plate is what I will go with. Machining a mount for that is not a problem for me.
Thank you for your input, it is greatly appreciated.

 

[dkemppai]

The XM-L3 mounts with aluminum back plate is what I will go with. Machining a mount for that is not a problem for me.

The XM-L3's I picked up were on a copper heat spreader, but same difference. Sounds like you have a current regulated bench supply, that's all you need to test out the concept.

Mounting and measuring the correct position is certainly one of the more important parts. Much more important for the contour light, less so for the surface light.

If you haven't ordered LED's yet. I'd grab an XHP-50.3 or two to test out for the surface light. I think the extra lumen output will be helpful with any magnification over 10x. The cross (or dead spot) between the dies shouldn't show up on the surface light optics.