Design of a new LED flash module.

fotoopa · Post by **fotoopa** » Sat Sep 06, 2014 12:36 pm

I completely agree Rik.

For the background, there is also the white LED part into the color led as 4300 K. I think it should work properly as you indicate. With this I would quite like to be able to set any background gadient color.
For the main exposure, I doubt even between 5000 K and 3000 K versions. In the diagram, the 3000 K have the best spectrum, but more expensive and less lumen. Hence my preference for 5000 K version.

rjlittlefield · Post by **rjlittlefield** » Sat Sep 06, 2014 12:51 pm

I don't know about the particular Cree chips you're looking at, but I was interested to discover that in the Cree replacements for incandescent bulbs, it is common for individual chips to have different colors. See image #4 at http://www.photomacrography.net/forum/v ... hp?t=22018 for illustration. If you have the option to guarantee that all of the same type LEDs you buy come from the same manufacturing lot, that would be a good thing to do.

--Rik

canonian · Post by **canonian** » Sat Sep 06, 2014 2:07 pm

fotoopa wrote:Now I am thinking to use a LED background....

In your new lighting setup there probably would be enough light to illuminate a colored card or anything else in the background.
You already use this method: in the photo of your rig I noticed you use a interchangeable background card.

RGB LEDs can combine nice subtractive- or complementary colors tuned to the color of the subject.
There is however a slight chance that bright background light hits the optics and cause unpredictable results like flare, colorcast or loss of contrast.

I have been thinking about various ways to integrating a nice gradient background as well.
Forummember Johan uses a field monitor which produces the background and I experimented using my mobile phone, displaying a gradient background that works evenly well as Johan's monitor.
In some cases I use cut-out images from a magazine of e.g. a forest that produces a nice OOF background.

These are just some random thoughs; I'm curious how the RGB LED background will work out.

fotoopa · Post by **fotoopa** » Sat Sep 06, 2014 11:57 pm

rjlittlefield wrote:I don't know about the particular Cree chips you're looking at, but I was interested to discover that in the Cree replacements for incandescent bulbs, it is common for individual chips to have different colors. See image #4 at http://www.photomacrography.net/forum/v ... hp?t=22018 for illustration. If you have the option to guarantee that all of the same type

I had already found and read your explanation a few days ago. The new color leds have internal 4 channels (R-G-B + W) just to arrange better the colors. On my drawing I say 4 channels, but this should in fact 4 chips x 4 channels. To drive this you need to have 16 channels or an 4x4 multiplexed array. This makes driving it somewhat more extensive but just a simple task for my FPGA module.

If you have the option to guarantee that all of the same type LEDs you buy come from the same manufacturing lot, that would be a good thing to do.

This is an advantage to buy the leds from an authorized dealer. Their LEDs are divided into groups with Bin and Order Codes. They have a specific range with limited guaranteed abnormalities. That makes also the price difference if you choose certain bins.

canonian wrote:In your new lighting setup there probably would be enough light to illuminate a colored card or anything else in the background.

Yes that's right, Earlier there was the light of the main light and I had a limited range of color plates. Arranging the background was more difficult precisely, this I wanted to make it better.

Forummember Johan uses a field monitor which produces the background.

Correct! I had this also read. With a monitor I've just trouble to have the control of it. I want the background but limited and active in a well defined exposure periode. I use a delay of 400 msec after shutter full open before turning the LEDs on. With a monitor that is almost impossible. That must remain enabled and if you do not have an EFCS function in camera you add the vibraton time to the background. If you use longerexposure times then there is less influence. My exposure times are now about 15 msec to 25 msec. Within this time, I would also want to control the background. Now my total camera open time is 500 msec. 400 msec for antivibration, and max 100 msec for lichting (forground and background).

You already use this method: in the photo of your rig I noticed you use a interchangeable background card

Yes and I can also put here a transparent card between object and background diffuser. This should lead to beautiful effects and is easy to control.

Frans.

fotoopa · Post by **fotoopa** » Fri Oct 10, 2014 7:10 am

A small update.
The led-flash module is now under construction. Here a more correct drawing of the module:

And here the outer casing on the lathe:

All the components are delivered:

There is still much work left. The object (insect) attached to a turntable can rotate form - 10 to + 190 degrees. The nodal point stay into the centre of the lens. This part is not show here on the drawing.

Update 11 okt 2014:
Drawing updated, softball is 12 inch version! pingpong ball now marked as reflector! Full resolution versions and more pictures stay on my Flickr web: https://www.flickr.com/photos/fotoopa_hs/

Frans.

TheLostVertex · Post by **TheLostVertex** » Fri Oct 10, 2014 9:00 am

If the pingpong ball covers the lens of the objective that much and is translucent, wouldn't it create flare and contrast problems? Wouldn't the pingpong ball serve better inverted to what is shown in the diagram?

fotoopa · Post by **fotoopa** » Fri Oct 10, 2014 9:35 am

TheLostVertex wrote:If the pingpong ball covers the lens of the objective that much and is translucent, wouldn't it create flare and contrast problems?

The aim is to make the inside black. This should I test. The outside should give a white surface so I hope for better lighting in the tip of the insects eyes. If it does not work then I can set up a conical metal snoot inside the pingpong ball in order to prevent flare.
Revering the pingpong ball would interfere to much with the direct light path from the other ball diffuser. I have yet to test this, it may be that I'm wrong.

rjlittlefield · Post by **rjlittlefield** » Fri Oct 10, 2014 10:38 am

fotoopa wrote:
TheLostVertex wrote:If the pingpong ball covers the lens of the objective that much and is translucent, wouldn't it create flare and contrast problems?
The aim is to make the inside black. This should I test. The outside should give a white surface so I hope for better lighting in the tip of the insects eyes.

So, if I understand your idea correctly, the pingpong ball will be acting as a reflector instead of a diffuser. No light will be going through it, only bouncing off it.

Do I have this correct?

--Rik

fotoopa · Post by **fotoopa** » Fri Oct 10, 2014 10:57 am

rjlittlefield wrote: So, if I understand your idea correctly, the pingpong ball will be acting as a reflector instead of a diffuser. No light will be going through it, only bouncing off it.
Do I have this correct?

Yes, that was my intention. So I'd like a little extra light in the center. Would be less reflected flash-rings into the eyes (I hope).

Drawing updated 11 okt 2014.

Frans.

johan · Post by **johan** » Sat Oct 11, 2014 1:52 am

fotoopa wrote: Correct! I had this also read. With a monitor I've just trouble to have the control of it. I want the background but limited and active in a well defined exposure periode. I use a delay of 400 msec after shutter full open before turning the LEDs on. With a monitor that is almost impossible. That must remain enabled and if you do not have an EFCS function in camera you add the vibraton time to the background. If you use longerexposure times then there is less influence. My exposure times are now about 15 msec to 25 msec. Within this time, I would also want to control the background. Now my total camera open time is 500 msec. 400 msec for antivibration, and max 100 msec for lichting (forground and background).

Good call, using a field monitor does require exposure times of approx. 1s - 3s. I use flash though rather than continuous, which does a good job of freezing the foreground subject as long as it's rear sync (to avoid ghosting).

On the main light diffusion diagram, not sure if it's to scale but I have in the past found it an advantage to have the diffusion material be able to extend beyond the subject as well, to give light further back on the subject. It might be an advantage to build in some flexibility with that. Also, I've found that hoods over the objective do need to have quite a generous radius relative to the front element, otherwise you'll have degradation/vignetting. Looking good!

fotoopa · Post by **fotoopa** » Sat Oct 11, 2014 4:14 am

johan wrote:On the main light diffusion diagram, not sure if it's to scale but I have in the past found it an advantage to have the diffusion material be able to extend beyond the subject as well, to give light further back on the subject.

Here I have to make a compromise. So I wanted to use larger objects that are wider than the LED flash module. I chose to stay there at least 12mm from the focal point. The object holder I wanted to pivot from -10 to +190 degrees, and into the other direction 360 degrees. The third axis of rotation is the camera itself via the tube lens bracket. So I obtain three axes for the rotation possibilities. For the linear displacements I have also the three axes: X-Y and Z

The backlight comes inline with the lens at about 100mm (range 50-250mm) and also moves on rails. I would use the other half of the softball which is 95 mm in diameter. There is also a regulation in the X and Y direction. Backlight contains 4 high power RGBW LEDs, this time will also be very short and under full control of my FPGA hardware.

There come still details once it is finished a bit further. Today I received my bearings for the rotary table and a power supply 24V 70W for the main led-module. Oh there's so much work.

Frans.

fotoopa · Post by **fotoopa** » Sun Oct 12, 2014 8:29 am

Today, the diffuser is machined to 86 mm in diameter. I have made an extra tool with bearing bal to hold the diffuser in position. Now, the diffuser clips in the outer casing. Lots of extra work but the result is quite good.

.

Now I can finish off the outer casing. There is a gap of 38mm for the Mitutoyo lenses. Then I can turn the rest of the mechanical part followed by the PCB for LEDS. Step by step, we are on the right track.

Update 13 okt 2014:
Added the finished outer casing with diffuser and reflector area. Possibly I'm going to paint the outside aluminium part black.

Frans.

fotoopa · Post by **fotoopa** » Tue Nov 18, 2014 12:15 pm

A little update after several weeks of work:

End of this week I hope to take pictures with the LED modules to see how much exposure time is needed. The led drivers and the leds are already tested. Then I have yet to make the objects holder. Extra thanks to Chris S for the help with the diffusers. Results will follow soon. Slightly higher resolution images see on my Flickr pages.

Frans

Update: confusing-- ChrisR corrected to chris S.

g4lab · Post by **g4lab** » Tue Nov 18, 2014 12:40 pm

Beautiful work as always.

GemBro · Post by **GemBro** » Tue Nov 18, 2014 12:50 pm

That update is singing to me ... love it ... as with all of your projects ... just wished you lived up the road ... so I could borrow your lathe

Perfect design & build ...

Regarding the background side ... I was thinking of using an LCD Photo Frame gadget, with SD card slot and load up some nice backgrounds, made in Photoshop, of gradients, Guassian Blurred images, etc ... then if needed, to control the gadget, rip it apart and find the CFC (background light) 'select' line to turn that on/off ... then interface accordingly ...

Gem