LED Ringlight

[canonian]

This is the DIY LED Ringlight I used in this post .
8 powerleds are surrounding a plastic cup, giving a nice omnidirectional and dimmable diffused light on the subject.

.

It is bolted on a BHMJ microscope focusblock and has full travel of 3cm in 1µm increments.

[BugEZ]

Wow! That is one powerful ring light! How long are your exposures?

Keith

[ChrisR]

Indeed! And what's the output spec for the LEDs?

[canonian]

...How long are your exposures?...

EXIF-data tells me it's 1/5 sec@ISO100, 0,0 EV exposure compensation.

...Indeed! And what's the output spec for the LEDs?...

I have to (find and) check the specs, my guess they are 1 Watt and roughly 100 Lumen each , but I'll check this tonight.

Edit: Specs found: 3800Kelvin / 155lm-220lm / 2.3W at 700mA

[ChrisR]

I have four 300Lumen head torches waiting for batteries. I think, Fred, you've saved me the trouble of a test!

I've been drooling slightly over these
http://www.luxeonstar.com/Neutral-White ... -wn300.htm
which are 2000 Lumen ( for short periods).
Eight of those would be a significant investment though.

Thinking on the keyboard here:
On the face of it (probably naively) the exposure time could go from 200ms to 10ms.
Looking at Fred's arrangement, I can't recognise anything inside the "cup", so there may be potential to move the lights closer, or focus them more. Perhaps 1ms is achievable?

Is this a Good Thing?
It could certainly help with reducing vibration movement for some situations but I do see mirror/shutter induced movement with 1ms (flash). That goes with shorter durations (eg 1/32nd power is 0.031ms).
So we're perhaps an order of magnitude too long still, to be rid of those.

Mirror/shutter induced movements become insignificant if the exposure is much longer, say 2000ms (2 seconds), because they're strong in the first 50ms or so. *

With my particular situation - (wooden floor, table with isolation under a heavy slab, copy stand on that) I see vibrations going on forever, at low frequency. I think I still have to use flash.


*"Citation required"! I vaguely remember that from some German tests.

[BugEZ]

Canonian

FYI...

The much less luminant ring lights I have built use approximately 24-36 5mm LEDs (http://www.ledshoppe.com/Product/led/LE1019.htm). I don't drive them to the full continuous current as my power regulation circuit is quite simple and I don't individually regulate the flow of power to each bank of 3 LED's. Each LED is rated for 13000 mcd. An online calculator estimates that each can produce approximately .3 lumins. I probably drive them to 2/3 of this.

A better estimate for comparison to your brightness may be to compare the power dissipated . With approximately .020 A and 3.2 V, the power is .06W per 5mm LED. Thus I probably have .06 W/LED X 36 LED = 2.3 W total electrical power consumed by the LEDs. If each of your lights uses 2.3W (edit) then you have 8X more light.

I adjust the intensity to keep the exposure around 1-2 seconds (ISO 200) to help "de bounce" my rather rickety rig. My home built controller allows me to sequence the LEDs to turn on about 100 ms after the shutter opens to help reduce the shutter bounce. I have done comparisons with and without this delay (10X microscope objective lens), and it is difficult to detect any change in image quality, so I usually don't bother to cycle the lights.

On the ring lights I use with microscope objectives, I have shifted the plane where the lights are located to be on the camera side of the objective lens opening. The skirt of my diffuser extends perhaps an inch beyond the objective lens. My goal is to reduce how much light shines directly onto the objective lens face. This has helped a lot with "fogging" of the images, similar to taking a photo toward the sun with no lens hood. I am presently experimenting with some microscope objective lens hoods I taped together using black construction paper. My goal is to pull the illumination a bit closer to the subject to provide more natural side lighting yet to avoid the "fog".

Your setup looks very clean and neat. Once again many thanks for sharing.

Keith

[canonian]

Your setup looks very clean and neat. Once again many thanks for sharing.

Thank you, Keith. First of all I admire your DIY skills and makes me think that everything is possible and buildable.
You've shown this in the well-documented posts you have shared with us. I especially like the materials you use.

Now and then I use flash but I prefer constant light, using my Jansjo's or this new LED ringlight.
However when using flash the images come out much 'cleaner', less noise, less smear, less hot pixels and 'pixeltrails', and needs a lot less PP.
But I do like the control you get in lighting the subject with constant lighting.

The LED's I use are driven at 350mA.
I could go for 700mA but while protyping I blew up a couple of very expensive powerleds and like to stay within safe margins.
Safe to say I'm happy with the amount of light my 'octo-lite' now produces.

To reduce 'fogging' I use a conical M42 to C-mount adapter as a hood on my reversed enlarger lenses, which proves to be very effective.
The gain I get in better contrast weighs up to the loss in working distance, which is still very comfortable.
I have no experience with shading a microscope objective, but will be on my extensive next-thing- to-try list.

[canonian]

I've been drooling slightly over these
http://www.luxeonstar.com/Neutral-White ... -wn300.htm
which are 2000 Lumen ( for short periods).
Eight of those would be a significant investment though.

Chris, I've been playing with the idea of using those flash-LED's that you find in mobile phones.
Driving them at low current could be used as focus assist and driving them shortly at full burst would then act as 8 flashlites.
1ms could easily be reached I guess. Syncing it with the shutter would be a greater challenge.

[AndrewC]

... I've been playing with the idea of using those flash-LED's that you find in mobile phones.
Driving them at low current could be used as focus assist and driving them shortly at full burst would then act as 8 flashlites.
1ms could easily be reached I guess. ...

I believe that phone flash leds are usually actually "on" for a few hundred msecs, they are often used by dumping the entire charge of a low voltage supercap through them. Several companies make controller chips which will switch between a low current "torch" mode and a full charge dump strobe. Alas, they are usually SMD devices and a pia to play with at home.

[BugEZ]

Canonian wrote

To reduce 'fogging' I use a conical M42 to C-mount adapter as a hood on my reversed enlarger lenses

A nice trick! Very clean.

Regarding microscope lens hoods: My initial test with and without one using my favorite ring light shows no improvement. However, this light is constructed to keep light out of the lens with no hood. This weekend I hope to test with more circumferential lighting and that may be more interesting.

K

[Andrew®]

I've been drooling slightly over these
http://www.luxeonstar.com/Neutral-White ... -wn300.htm
which are 2000 Lumen ( for short periods).
Eight of those would be a significant investment though.

Chris, we use the blue derivative of these driven at max current for other purposes.
Just be aware, they generate significant heat and need to be mounted onto a heat sink with a cooling fan.
See: http://www.luxeonstar.com/v/vspfiles/do ... hermal.pdf

[ChrisR]

Yeah, I'd noted is as a key thing to watch.
With an exposure time of say 10ms the duty cycle would be low, but I'd start with short pulses at low voltages while watching the resistance of the thermistor, or it could get expensive!