Choice of color temperature for the power LED flash module.

Have questions about the equipment used for macro- or micro- photography? Post those questions in this forum.

Moderators: Chris S., Pau, rjlittlefield, ChrisR

fotoopa
Posts: 194
Joined: Sat Dec 14, 2013 3:14 am
Location: Belgium
Contact:

Post by fotoopa »

Chris S. wrote:Sounds like an excellent choice--elegant, and potentially a nice step forward that the rest of us can copy. I enthusiastically look forward to your reports.
Well, I made a lot of extra tests. I have read many documents relating LEDs, color, brightness, bins, CRI, color spectrum etc. All very difficult terms to understand and thereafter to make choices.

I did a test with my existing 6-channel LED version. LEDs were types with 5000 K and a CRI above 90. I took out of the diffuser in the first test and set the WB via the D7100 as preset1. I put the gray card, the Mit 5X lens + the 200mm macro tube lens and take a picture. Via View-NX-i all the information is available. After this test I put the normal diffuser into my unit and take a second picture with the same P1 preset into the D7100. Due to the duffuser, the WB is no more correct, colors are shifted. Lets see an overview:

Image

The camera has a color temp of 4646 K (WB without diffuser as P1). Using now the same preset with the diffuser you have to increase the brightness to have the same exposure. The flash time goes from 700 to 2400 usec, the blue channel has less light, the red more. The picture is no more the gray value.

To have an idee how much change there was, I have done a new WB with the diffuser included and store into the camera as preset P2. lets now look at this picture:

Image

The color temp is now 3609 K and the flashtime 2550 usec to have the same brightness. After this WB, ballans is again perfect with the diffuser. My conclusion is that the color temperature shifts downward. What remains of the CRI value, I can not really measure. Everything is very relative and the question is whether you need the very expensive versions.

My proposal to use two color temperatures will be totally wrong. As long as you simultaneously using only one kind of group it will works, but no more if you change selectieve the 10 channels. The WB will by wrong.

Just now I prefer the 4000 K versions with a min CRI 90 : CXA1830-0000-000N0US240H 4000 K 2780 lm.
http://www.mouser.be/ProductDetail/Cree ... 1gtA%3d%3d
They are also less expensive.

mjkzz
Posts: 1237
Joined: Wed Jul 01, 2015 3:38 pm
Location: California/Shenzhen
Contact:

Post by mjkzz »

But I thought the reason to get two different color temp LEDs is NOT if we can WB a picture, we can always do perfect WB, the reason was to shift it up towards colder color that Chris S. has mentioned, so that when shooting bluish subject, we can get adequate saturation.

fotoopa
Posts: 194
Joined: Sat Dec 14, 2013 3:14 am
Location: Belgium
Contact:

Post by fotoopa »

mjkzz wrote:But I thought the reason to get two different color temp LEDs is NOT if we can WB a picture, we can always do perfect WB, the reason was to shift it up towards colder color that Chris S. has mentioned, so that when shooting bluish subject, we can get adequate saturation.
Yes this is correct about the WB. I wanted to show that the blue portion just reduced by the diffuser which is an advantage. Generally, LEDs have too much blue, especially those with low CRI and high K temperature. If this result will give a better color rendition, I do not know. I can not make a spectra graph.

mjkzz
Posts: 1237
Joined: Wed Jul 01, 2015 3:38 pm
Location: California/Shenzhen
Contact:

Post by mjkzz »

I see, so the diffuser is reducing overall color temp, this could be disadvantage, making blue channel noisier.

fotoopa
Posts: 194
Joined: Sat Dec 14, 2013 3:14 am
Location: Belgium
Contact:

Post by fotoopa »

mjkzz wrote:I see, so the diffuser is reducing overall color temp, this could be disadvantage, making blue channel noisier.
Not when blue light is already too strong.

I did the order of 4000K LEDs today. Monday they'll arrive. I've ordered 12 pieces so that I can make further tests and / or selection.
My goal remains:
- Very short led flash times range 300 to 800 usec.
- Freezing of small vibrations.
- using a doka space, no ambient light
- No high temperatures, duty cycle is extreme small, max 1msec/2.7 sec stacking cyclus.
- No batteries, everything is powered from the mains supply.
- No light changes during the cycle stack.
- LED Flash with diffuser always exactly the same setup / distance to object.
- WB only be measured once via gray card for each Mitutoyo lens.
- Focus distance to an object is always 33.5 mm, a fixed setup even for the 2x, 5x and 10x lens and WB is stored into the D7100 camera.
- Quick changing of lenses, focus point position exact the same with another lens.
- LED Flash does not move while making a stack cycle, only the camera moves.
- Live view on the 27 inch PC monitor and also the file transfert.
- Auto process all pictures via Capture NX2 batch file.
- Zerene stacking software for the final results 2D and 3D pictures
- 3D pictures at the end process via StereoPhotoMaker.
- Stage unit has 1 um resolution for a full step on the motor.
- Encoder into the stage unit give me the exact positiuon even if you turn manual on the dial with step controller off. Readout 0 to 250.000 mm on the display.

mjkzz
Posts: 1237
Joined: Wed Jul 01, 2015 3:38 pm
Location: California/Shenzhen
Contact:

Post by mjkzz »

As shown by your result, after diffuser, the color temp dropped, so some blueish light are absorbed by diffuser.

Nice lighting goals, and I think with LEDs, these goals can be achieved "easily" compared to xeon tubs. Very excited to see your results :-)
Last edited by mjkzz on Fri Jul 08, 2016 5:51 pm, edited 1 time in total.

mjkzz
Posts: 1237
Joined: Wed Jul 01, 2015 3:38 pm
Location: California/Shenzhen
Contact:

Post by mjkzz »

Stage unit has 1 um resolution for a full step on the motor.
I do not know what kind of stage you are using, but being able to do 1um step is not enough, axial stability is another factor, particularly for high magnification work, in my opinion.

fotoopa
Posts: 194
Joined: Sat Dec 14, 2013 3:14 am
Location: Belgium
Contact:

Post by fotoopa »

mjkzz wrote:I do not know what kind of stage you are using, but being able to do 1um step is not enough, axial stability is another factor, particularly for high magnification work, in my opinion.
Do not worry, this was the guaranteed basic accuracy. Both spindle as guides run on ball bearings. No backslash is also present. This stage unit comes from an industrial precision measurement system. I use microstepping (1/16), that 1 micron step could be much smaller if necessary. The Mit 10X lens has sufficient with increments to 3 um. The Mit 20x lens I have not yet! The axial stability is perfect due the laterally steel ball bearings in a V groove over a long distance.

mjkzz
Posts: 1237
Joined: Wed Jul 01, 2015 3:38 pm
Location: California/Shenzhen
Contact:

Post by mjkzz »

From what you described and your work, I am sure your rail is a high performance one.

fotoopa
Posts: 194
Joined: Sat Dec 14, 2013 3:14 am
Location: Belgium
Contact:

Post by fotoopa »

Here are the new set of high power LEDs. A total of 12, there are two spare for additional testing or selection. Each LED has to 54W power. The final version will have 10 LEDs, more than 500W.
There will be a 52V power supply for driving the LEDs and a total of 40000 uF/63V elco's to supply the full current over minimum time of 40 msec.
Controller module stay at 5 m distance, RS422 differential line drivers will by used to drive the led module inside the doka space.

Image
P7116035 by Frans, on Flickr

Here you see my old radio shack workplace ( ex ON5DK). There is no more radio or any transmitter but the electronica remains. I use a old labtop for the PicoScope and Logic analyser and also a copy of the FPGA module as controller. Once the test are ready, the logic is transferred to the current existing step controller. This will take several weeks.

Image
P7116056 by Frans, on Flickr

To mount the LEDs I need a support plate. This is also cooling. A solid block of aluminum will be used for this purpose. The photo shows the rough block. Later the finish follows.

Image
P7116050 by Frans, on Flickr

PCB layout for the 10 power leds. The central part is drilled out to 40mm in order to let through the Mitutoyo lenses. There are 10 channels and at each channel, the current can be set.

Image
cree_10led_module_v3_layout_pos_filter by Frans, on Flickr

RS 422 line drivers placed into the basic controller at 5m distance from the doka space. twisted pair lines will transfer the signals to the reveiver module into the doka space. Times are adjustable from 1 to 40000 usec.

Image
rs422_driver_v1_schema by Frans, on Flickr

There is a lot of work for the coming weeks. This week, the rest of the parts ordered will be delivered as: Fets, twisted pair cable, connectors, 52V power supply, SMD components etc. In the end if everything is tested the software in the FPGA controller must also be updated.
To be continued....

mjkzz
Posts: 1237
Joined: Wed Jul 01, 2015 3:38 pm
Location: California/Shenzhen
Contact:

Post by mjkzz »

OMG! This is awesoooome!

At 1ms, amount of energy is about the same as a Canon 580EX set at 1/128 level of power, but since this can be placed so close to subject, it is probably equivalent to 16 of Canon 580EX (assuming it can be placed 4 times closer than the 580EX, probably much closer).

The best part is consistent output level.

Thank you for sharing!

fotoopa
Posts: 194
Joined: Sat Dec 14, 2013 3:14 am
Location: Belgium
Contact:

Post by fotoopa »

mjkzz wrote:The best part is consistent output level.
Yes correct, I hardly need to change the flash times. The distance always remains the same to the object. My FPGA controller has default settings, these are usually correct. Given the 10 channels I can also set the direction of the lighting.
Thanks to the help of my wife, today all PCB have etchings. As with black and white photography, there is exposed first (here UV light during 3.75 min), followed by developing ( 40 sec), and final etching (5 to 10 min). Then, the drilling of the printed circuit boards can begin.

Image
P7126082a by Frans, on Flickr

The PCB for the LED is preliminary drilled. The final finish is done together with the mechanical holder. So all the holes fit together correctly.

Frans

fotoopa
Posts: 194
Joined: Sat Dec 14, 2013 3:14 am
Location: Belgium
Contact:

Post by fotoopa »

There has been some work done on the 500W power LED flash module. For example, the printed circuit boards are ready for the test.
Here the RS422 receiver 10 channels module:


Image
P7156131 by Frans, on Flickr

And here the transmitter 10 channels RS422 module:

Image
P7166134 by Frans, on Flickr

A view on the power led module basis alu plate holder during milling and drilling:

Image
P7266255 by Frans, on Flickr

Rear view of the 500W power led module. View of the 4 limiters for the Mit lens and separate mounting on the rail system. Wiring must be finished, in total 20 wires for 14 A / 38V on 10 channels.

Image
P7306307 by Frans, on Flickr

Another view of the mounting of the LED module with the camera setup

Image
P7306331 by Frans, on Flickr

Time to connect the Led module PCB on the power lines. 20 wires 0.5mm² are used to connect the leds.:

Image
P8016370 by Frans, on Flickr

The result after positioning all the LEDs and the test of the proper connections through the 20 brass contact screws M2:

Image
P8016395 by Frans, on Flickr

All LEDs work correctly, the light is so strong that you definitely can not directly look at the LEDs! The total power can go up to 540W. Now should be done the measurements for temperature and all the timing controls for flash mode and live view mode. This will be done through a temporary test setup FPGA unit. If everything is measured, the whole hardware can be transferred to the existing controller. This will take a few weeks.

To be continued...
Frans.

ChrisR
Site Admin
Posts: 8521
Joined: Sat Mar 14, 2009 3:58 am
Location: Near London, UK

Post by ChrisR »

W :shock: :shock: :shock: :shock: :shock: :shock: W
Chris R

g4lab
Posts: 1434
Joined: Fri May 23, 2008 11:07 am

Post by g4lab »

Amen!! :D

Post Reply