Electronics > Projects, Designs, and Technical Stuff
DIY Focus Stacking for Macro Photography
mawyatt:
--- Quote from: RoGeorge on July 02, 2020, 08:51:55 am ---
--- Quote from: mawyatt on July 02, 2020, 12:33:10 am ---The P601 and P603 PI Piezo Stages can handle moderately heavy loads in Closed Loop, this was the weak link of the speaker VCM. There was no practical way for feedback, so it operated Open Loop and suffered from position errors.
The THK KR rail types are very well designed and use a temp compensating base metal, probably the best available in screw type linear rails. With the Piezo stage operating in closed loop we created a setup at high magnification (800X I recall) and watched the image as a hair dryer on hot setting blew hot air across the piezo stage. It hardly moved, the control voltage was all over the place to keep the loop closed but the stage was rock steady. When this was done to the THK KR20 the image quickly drifted off screen, direct consequence of open loop operation, even with the superb THK KR20!!
The setup we use for high magnifications utilize the piezo stage and the THK KR20 rail. The KR20 is used for coarse camera/lens positioning and the Piezo stage for fine moving the subject. Many times I don't bother powering the THK rail up, just use it as a manual positioner by turning the motor shaft. Since we know the motor steps (400) and the KR pitch (1mm/turn) we can make movements by hand just counting motor cogs steps at 2.5um/step. If we need something between steps, the the KR20 must be powered to get fractional steps. The Trinamic controllers have 256 micro-step increments, so theoretically you could get 1mm/400/256, or 9.765nm increments. Of course this is ridiculous, a reasonable estimate would be 10 to 20 times that.
--- End quote ---
Indeed, load weight for a CD optical head is not much, probably a gram or so at most.
Another type of closed loop actuator, even easier to use and totally free for DIY experiments will be the actuator from an older hard disk. HDD heads already heave the control loop, are trivial to position with nothing but software and an old PC.
I've played a few times with opened HDD running. They can work in open air for a very long time, for many days if not weeks, in a normal living room. The drawback is that the absolute position of the actuator is read from one of the spinning disks, so the disk must spin for the original control loop to keep the actuator in place, therefore it will come with some vibrations from the spinning plates. Maybe synchronizing the camera trigger with the angular position of the spinning disks, or maybe keeping the objective open for infinite exposure time and just flashing some LED lights in sync with the disk RPM might help catching the vibrations always in the same position. I don't know.
However, once locked on a track, the HDD heads are very stable, it was not easy to unlock their position by hand. I didn't measured the force, but I'll guess the actuator could drive an extra load of tens to hundreds of grams. That should be enough for a small stage with a chip die or an insect.
I imagine the HDD to stage adapter like a big loop of fishing wire that is anchored to the arm of the HDD actuator. The fishing wire loop is bigger than the HDD, and leaving the HDD body through small holes in the sides of the HDD frame, so to not keep the HDD plates spinning it open air. Small ball bearings in the corners of the closed loop of the fishing wire might help. Seems too complicated already.
Maybe just a small aluminium stage anchored with 2 screws on the actuator's arm would be enough. Far from ideal, but good for insects and free to try, with only a weekend long, or so, of tinkering. Or, it might not work at all because the PID coefficients might not cope with a much heavier load... ;D
Speaking of insects and macro shots and focus stacking, a trick that I've discovered by serendipity:
To arrange the photo posture of a live insect, and to make it stay still in the given position for a few minutes, put the insect in a refrigerator, 40° F (4° C) for half an hour, or so. :)
This won't hurt the insect. To them, those cold temperatures might easily happen in the wild during a cold night. It will take to an insect a couple of minutes before its temperature will raise enough to start moving, even when put in the direct sunlight for a better photo. That should be enough time for a quick focus stacking shooting session.
Back to that gorgeous photo you posted before, "Iridium- DMap-Obj-1a-Edit.jpg", a stupid question, but please keep in mind that I know nothing about laying a circuit in a die, so here it is:
Are those spheres the soldering balls? Asking because they seem incredibly close to various die structures. Some of them look like they have no dedicated pad area underneath. Isn't there a danger for the melted soldering ball to cover or get too close to the die structures? Then, there are some circular structures in the most left corner, I assume are some microwave coils. Wouldn't those coils be too close to the PCB traces, or to the PCB copper planes that the PCB designer might put underneath the coils area, and thus ruin the coil's Q factor by the Eddy currents induced into the PCB's copper?
--- End quote ---
Yes those are tiny (100um diameter) solder balls for flip-chip mounting. Many years of material engineering went into the metal alloy, size, shape and wetting surface design of these solder balls so they make reliable connections without shorting to other structure.
Those solder balls are difficult to work with (photograph) since they are like tiny spherical mirrors and reflect everything around including the light sources. So highly diffused uniform lighting is required to keep the specular artifacts in check.
Great question, the coils are small inductors utilized for impedance matching networks. All the effects due to Eddy currents, silicon die subsurface and PCB placement are taken into account. The EM fields are mostly confined to much less than the physical distance between the die surface when flip mounted and the PCB surface due to the high dielectric constant of silicon (~11), so the PCB has little effect on the inductor.
Best,
mawyatt:
--- Quote from: fcb on July 02, 2020, 11:31:07 am ---Gorgeous images!! - any chance you can post these in high res so I can get one made into a poster for my lab wall. :)
--- End quote ---
Thanks. Sure I can provide some higher resolution images, but they are limited in size for posting here I believe. If you look at DP Review some are available at higher resolution. Do a search for Chip Images and more will show up, the author is mawyatt2002, I've posted a few there.
https://www.dpreview.com/forums/post/60270789
Another resource is at Nikon Image Space mentioned in one of the above threads, where you can download the images in higher resolution.
I also can send some directly but my email is limited to just 10MB.
Best
mawyatt:
--- Quote from: Rerouter on July 02, 2020, 12:26:09 pm ---I feel I have missed it, but are there any free focus stacking programs anyone can recommend to give it a try?
--- End quote ---
I use Zerene, but it's not free. You can download a free trial though.
Best,
Mechatrommer:
--- Quote from: Rerouter on July 02, 2020, 12:26:09 pm ---I feel I have missed it, but are there any free focus stacking programs anyone can recommend to give it a try?
--- End quote ---
there are few in the list... i used Enfuse as one of the tools to do HDR stitching, it seems now to support focus stacking as well...
https://en.wikipedia.org/wiki/Focus_stacking
ps: for keen programmers, you can start with edge detection algorithm https://en.wikipedia.org/wiki/Edge_detection. where more edges information in the picture (higher spectral content or less compressible a particular portion of the image), thats where the focus should be, thats where to give more weight (or to put less mask) on the image, i guess this is how classic software based "contrast detection" auto focus in digital cameras work. blur the mask to avoid nasty result and then overlap them all. it should be not that difficult for a simple stitcher.
--- Quote from: fcb on July 02, 2020, 11:31:07 am ---Gorgeous images!! - any chance you can post these in high res so I can get one made into a poster for my lab wall. :)
--- End quote ---
few replies up is the link in nikonimages.....
Dave:
--- Quote from: mawyatt on July 01, 2020, 08:41:35 pm ---The effective aperture (EA) generally follows as Lens Aperture, or LA, as LM(1+M), where lens aperture is the normal "reading" on most cameras (Nikon shows the EA, not LA). When you get an EA that's beyond about F20 or so, diffraction begins to eat your lunch if your images are viewed at larger sizes. Normal lenses this doesn't matter, since shooting a bird at 100 feet or a portrait at 10 feet, M is very small, so M+1 ~ 1 and EA ~ LA. However when shooting macro this becomes a problem, example shooting at 1X the EA is twice the LA, so one would begin to see diffraction effects around LA of just f11 (in very high quality work this is below f8!!). So "stopping down" the aperture to increase the depth of focus has a limited range.
[...]
--- End quote ---
Thanks for the explanation, knew there had to be some sort of catch behind it.
Navigation
[0] Message Index
[#] Next page
[*] Previous page
Go to full version