Binocular Indirect Ophthalmoscope

Hi. I'm an eye doctor at the University of Virginia interested in developing my own equipment. In particular, I use a device called a binocular indirect ophthalmoscope. It sounds complicated, but is simply a glorified flashlight worn on the head. The problem is that the commercially available products are large, cumbersome, and cost 3 to 5 thousand dollars. There is obviously a lot of miniaturization that has occurred lately that could be capitalized upon.

I have made a prototype that weighs only a few ounces using a xenon bulb assembly from Princeton Tec Rage (only cost $3) and two CR123 batteries.

Its bright. very bright. However, the beam is too diffuse. I need it to focus to about a four inch spot at a distance of about half a meter. I have tried several converging lenses, but can't get it to look right.

Also, I'm sure that the light and power supply I chose could be improved upon. With the encyclopedic knowledge the users of this site seem to possess, I hope I can improve my design.

A couple stipulations:
1. The light source has to be similar in color to incandescent. I am using a xenon source and it is just right. LED just won't work.
2. I would prefer a rechargeable lithium ion battery. This is something I may use for up to two hours per day, and I can't spend a fortune on cR123 cells.
3. The entire assembly must be light enough to be supported by a spectacle frame.
4. I don't want to blow myself up.

Thanks in advance,

Matt

drneale said:

LED just won't work.

Click to expand...

Just want to make sure that statement is one with an understanding of the state of the art in LEDs. There are now high CRI LEDs that are completely different from what 99% of people have seen an LED do. That includes people here.

first, :welcome:
Next, I'll try to answer some of your questions. My first thought for this would be an LED build. I know I know, you said no LEDs but hold on. Have you heard of the warm white Crees or high CRI Seouls? These give an output similar and sometimes more pleasing than an incan bulb will and are much more efficient. That being said, the most simple idea I can come up with is making a headlamp using one of these LEDs and an aspherical lens to focus the light to a tight spot. Search on here for "aspherical" and you'll see what I mean. As for electronics, you could keep it simple but just direct driving the LED off of a single LiIon cell such as a 16430. Making it all look nice and fit on some glasses might be the issue. Maybe start with an existing headlamp with a reflector, remove the reflector and led and replace with the new led and aspherical lens. Just some ideas for you. Good luck!

What I would build:

1. Light source: Cree XR-E Neutral White or Warm White (This is an LED with color like incancescent while maintaining efficiency) behind an 8˚ optic. That optic will give you the light distro you want, and do it with very little spill light (however there will be a little). Based on my recent experience shining light into my eyes, you won't need more than about 70mA current to the LED. That's also good because that's very little heat to dissipate.

2. Power source: protected 18650. That will last a looong time running an LED at 70mA. Think like 35-40 hours between charges.

3. LED on star with optic weighs 6.2g. I don't see the weight of the assembly on the frame needing to exceed 10g. The converter, switch, and battery can mount on the belt or in a pocket.

4. Protected 18650s are pretty safe if you follow some precautions, like keeping the voltage above 3.5V - which a cheap multimeter will help you do. And you can recharge the cell before it's completley discharged, i.e. you could charge it every Monday morning and you'd never have to worry about over-discharge. You could also use 3 Eneloops in series for similar performance.

Construction note: an 18650 and the LED electronics and switch will fit nicely in a D-cell holder.

Thank you all for your input.

1. I am willing to try the warm LED. Is this one acceptable?
2. How about this collimator?
3. How about this battery?

drneale said:

Thank you all for your input.

1. I am willing to try the warm LED. Is this one acceptable?
2. How about this collimator?
3. How about this battery?

Click to expand...

Thats a "cool white" LED, not a "warm white" one.

http://www.cree.com/Products/pdf/XLamp7090XR-E_B&L.pdf

That PDF contains all the information for the XR-E binning codes, the graphs on page 7 show where it is in terms of colour.

Welcome. It is great to see a variety of professionals seeking advanced lighting technologies. Could you tell us what budget you wanted to work with? Even though I love DX/KD for their selection and low prices, I think that so you can truly be happy with the finished product without any future reservations from using LED products, you should spend a bit more.

Well, if that is truly a Cree brand optic (like the first that were available), then that seems to be a good choice. I have little experience with optics and Cree XR-E emitters (I worry about square-looking artifacts). So if you cannot find suitable optics, there are great reflectors out there that work well with the XR-E. I agree that asheric lenses, when slightly out of focus, would work swell for your application! They cast virtually no spill, and provide a very tight, but even beam. It may look ugly on a wall, but works great when actually being used. I believe saabluster would be the proper person to ask for more info on different sized aspheric lenses! DX does sell some lenses that are made of plastic and are lightweight. The quality of the plastic lenses will be pretty decent, so you probably will not need any made of glass.

As for the LED emitter, maybe you should shop around a bit first. Even though it is a XR-E Q5 WG emitter, you may find it to be too bluish compared to the warmer whites that you desired earlier. WH and WJ was the warmest of the cool white bins. I am not sure if you would enjoy the slow service (the shipping) of DX, either. There are several sources for other XR-E emitters around here at the CPF. Just ask (as I am not sure of the latest sources).

I remember reading a bit about neutral and warm white emitters, so I found those links for you...
Mudman CJ took some shots of an incandescent light source versus various high CRI emitters here:
https://www.candlepowerforums.com/threads/211911

saabluster took some shots comparing his WH cool white XR-E, his 5A XR-E, and his 6C MC-E emitters:
https://www.candlepowerforums.com/posts/2654418&postcount=269

To get an idea how the warmer of cool whites compare to neutral and warm whites, I stitched together two tint binning charts that DFiorentino created from X and Y data in this sticky thread. I hope he does not mind. They were scaled differently, and I did not have the original excel files, so I re-sized them and combined them using PSP8...

Notice that the WH/WJ tints overlap the 3A/3B tints slightly. With a neutral white (with a bin code from 3-5), you will be closer to that BBL line. The WH tint is pretty broad and can extend into the greens a bit.

Well, I am not sure where to get a neutral white or warm white (besides Cutter), so maybe you can check with mudman or saabluster as to where they got their warmer emitters.

As for batteries: get the best. I am weary about even using those protected li-ion cells from DX. The protection circuitry could be set incorrectly, which can be a bad thing. Those cheap, Chinease cells are also often over-rated, and you can never be sure if they will work at all sometimes. I reccomend AW's li-ion cells (found on the CPF). He sells a range of sizes, and they are of the highest quality! They are not that expensive, either. Plus he stands behind his cells confidently. The capacity rating of his cells are right on with what you will really get, I believe. If you have the external pack located on your hip, then you could use higher capacity cells. His largest cell is a C-sized cell that sports a 3300mAh capacity. On the other hand, if you want to stay away from li-ion, then buy Eneloop brand AA cells (3 in series should work -2 may work at low current levels). Walmart sells Duracell rechargeables that are the Hybrid types. Look for the ones with a white color surrounding the (+) button. Those are from Japan and have a better quality than ones with a black colored surrounding. There are other brands of high quality hybrid AA cells around here as well. Again, just ask around!

I would reccomend that you use some sort of driver circuit to regulate your current output. However, if you will run the emitter at below 100mA from a li-ion cell, you may be able to get away with DD with a resistor to reduce current. If the power is low enough, you may be able to use a linear variable resistor to dim the LED as needed. A cheap and simple linear voltage regulator would also work (using a potentiometer to dim the LED) since the operating efficiencies are not as dramatically important with these low current levels. Either way, you will get some great battery life if the currents are this low.

LEDs are very efficient, so the amount of power required is so much lower that what you may be used to with a light bulb. Well, due to the efficiencies, I think that with such a narrow beam, currents above 100mA will create enough light that can be uncomfortable for the patient. With several hundred milliamps, it is possible that you could do some damage as well . That would sort of go against your entire profession, wouldn't it? I say that once you get your LED and optic set up, experiment with current levels to see which provides the correct amount of lighting.

-Tony

Gryloc said:

Welcome. It is great to see a variety of professionals seeking advanced lighting technologies. Could you tell us what budget you wanted to work with? Even though I love DX/KD for their selection and low prices, I think that so you can truly be happy with the finished product without any future reservations from using LED products, you should spend a bit more.

Well, if that is truly a Cree brand optic (like the first that were available), then that seems to be a good choice. I have little experience with optics and Cree XR-E emitters (I worry about square-looking artifacts). So if you cannot find suitable optics, there are great reflectors out there that work well with the XR-E. I agree that asheric lenses, when slightly out of focus, would work swell for your application! They cast virtually no spill, and provide a very tight, but even beam. It may look ugly on a wall, but works great when actually being used. I believe saabluster would be the proper person to ask for more info on different sized aspheric lenses! DX does sell some lenses that are made of plastic and are lightweight. The quality of the plastic lenses will be pretty decent, so you probably will not need any made of glass.

Click to expand...

I popped a KD 8˚ on a Cree last night and it didn't have any artifacts at .5m. I don't know is an aspheric is called for here. It's heavy, it needs to be mounted just so, and it has chromatic abberation at the edges of the beam. And there's no guarantee that the aspheric lens will have an even projection of light at 8-12˚ and .5m. These optics don't have any of those problems.

drneale said:

Thank you all for your input.

1. I am willing to try the warm LED. Is this one acceptable?
2. How about this collimator?
3. How about this battery?

Click to expand...

1. Cutter.com.au sells warm Crees. DX/KD don't.
2. That's actually the one I meant to link to. I changed the link in my post. (the 10 pack is a much better deal per piece)
3. That would work, but these are four or five times more capacious in only a slightly larger volume. I don't own AW cells and I've had good results with the linked cells. I guess you could say there's two factions on CPF when it comes to li-ions (and everything else :nana. One group uses DX/KD cells, the other group looks at the first group like the first group is witches and uses only AW cells.

There's also LiFePO4 cells, which will be safer than any li-ion but have less capacity.

Gryloc said:

I would reccomend that you use some sort of driver circuit to regulate your current output. However, if you will run the emitter at below 100mA from a li-ion cell, you may be able to get away with DD with a resistor to reduce current. If the power is low enough, you may be able to use a linear variable resistor to dim the LED as needed. A cheap and simple linear voltage regulator would also work (using a potentiometer to dim the LED) since the operating efficiencies are not as dramatically important with these low current levels. Either way, you will get some great battery life if the currents are this low.

LEDs are very efficient, so the amount of power required is so much lower that what you may be used to with a light bulb. Well, due to the efficiencies, I think that with such a narrow beam, currents above 100mA will create enough light that can be uncomfortable for the patient. With several hundred milliamps, it is possible that you could do some damage as well . That would sort of go against your entire profession, wouldn't it? I say that once you get your LED and optic set up, experiment with current levels to see which provides the correct amount of lighting.

-Tony

Click to expand...

I have a limited understanding of electronics, so I was thinking about a 2-mode AMC7135 board running on low with 2 AMC chips removed to give 50mA. A resistor/pot might be better; I don't know if 50mA is enough and the increments you can get with that board aren't narrow.

KD does sell a warm white led (Cree P4), look here:
http://www.kaidomain.com/ProductDetails.aspx?ProductId=5568
If you are willing to wait, and wait, and wait
They claim 3000K but do not specify the color bin

clg0159 said:

KD does sell a warm white led (Cree P4), look here:
http://www.kaidomain.com/ProductDetails.aspx?ProductId=5568
If you are willing to wait, and wait, and wait
They claim 3000K but do not specify the color bin

Click to expand...

I had forgotten about that LED. Definitely worth checking out. The LED in the pictures has the warm white phosphor. But I can't be sure which bin it will be. 7C and 7D are 3000K, but I haven't seen a P4 yet with that tint. So it might be N2/P3 7C/7D or P4 7A/7B. But those bins are very similar. After all, they're all adjacent to each other on the binning chart.


The image below is a Q5 WC in series with a P4 7A (3200K) at 350mA. (P4 7A is on the left)

How about just a simple micropuck 350/400/500 ma
http://www.ledsupply.com/micropuck.php
and 2xAA eneloops with the KD warm white and optic. Simple, sturdy,no li-ion concerns, and should achieve more than enough runtime.

clg0159 said:

How about just a simple micropuck 350/400/500 ma
http://www.ledsupply.com/micropuck.php
and 2xAA eneloops with the KD warm white and optic. Simple, sturdy,no li-ion concerns, and should achieve more than enough runtime.

Click to expand...

Even 350mA is about 5 times the necessary current. I shined myself in the eyes with 150mA and it was way too much.

Okay. Thanks for your patience with my inexperience. I thought eye surgery was complicated--but flashlights are brutal!

1.I'm going to order the Cree MCE4WT-A2-0000-000HF8 Warm white, 280 min lumens, F8 (7C, 7D, 8A, 8B) from cutter.com.

Since most of the time I am using the light to differentiate between subtle shades of red, orange, and yellow, and shorter wavelengths are more prone to chromatic aberration, I am going with my intuition and selecting the yellowest light available. Thanks for that graphic with the spectra of each LED.

2. I'm going to focus the beam with:

120/147
6 Degree LED Collimator Lens. Designed to operate with Cree Xlamp 7090 High Power LED's High light collection efficiency of >85% Precision moulded in optical grade Polycarbonate for thermal stability and system durability Part of the Polymer Optics "Modular LED Optics"® range
Comes with the Optic Holder 147
I know it says it is designed for Xlamp, but I'm hoping it will work with MC-E as well. Is this a foolish assumption?

3. As far as the battery goes, I have been pushing toward smaller cells, because I want to fit everything on the spectacle frame. I would be really happy with a 2-4 hour run-time if a rechargeable could easily be swapped out. Its not like I'm out in the wilderness or anything; I could just keep a few spares in the desk drawer.

4. With respect to a driver circuit:

• I understand that I probably won't need high amperage; thus, I can probably get away with a variable resistor (10K?) without sacrificing too much efficiency.
• However, gear-lust is tempting me. If I DID go down that road, would this product (Boost and Buck Circuit For CREE XR-E Q5 LED) suit the application? It is tiny, light, and should provide about the right current off the shelf...right? Would the extra weight and expense return a significantly longer battery life and consistent illumination?

5. Budget. Since the commercially available products retail for $2000 minimum, I am comfortable spending a few extra bucks on a quality emitter. Plus, even the most compact devices have a belt clip battery. I think we can do better for far less money.

Again, your help has been invaluable.

Matt

Very interesting project, drneale - please keep us up to date as it progresses.

drneale said:

2. I'm going to focus the beam with:

Click to expand...

You do not want to use that with the MC-E as it will have four separate beams due to the fact that the MC-E is made up of four dies. Trust me I have those and have first hand experience. You either want to change to the XR-E which you can get in warm white from cutter as well or change to the polymer optics number 186 which is the same as the one you put a picture of up above except it has a diffuser on the face to smooth the four beams. My recommendation is to get the XR-E. As I understand it you will not need the higher output from the MC-E and you will probably have to wait on MC-E specific optics as they are not shipping right now.

Hi, I have the http://www.kaidomain.com/ProductDeta...ProductId=5568

It is really nice with very uniform tint. Best warm white I've had.

Can't tell what tint it is, as I have no other warmwhite to compare with, but probably around 3200k. Received the emitter in 12 days, much faster than DX.

I really cant see why you would like a MC-E. You will have problems finding really narrow optics for MC-E and the light output from a single XR-E will surely be more than enough.

gillestugan said:

Received the emitter in 12 days, much faster than DX.

Click to expand...

That would most certainly be an anomaly.

saabluster said:

That would most certainly be an anomaly.

Click to expand...

My KD orders take much less time than DX (and Cutter).

As for the MC-E, it's not the best choice. As saabluster mentioned, you will have a dark cross in the beam and you will have waaaay too much light which you won't be able to focus the way you would be able to with an XR-E.