Desktop automated focus stacking setup

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

Moderators: rjlittlefield, ChrisR, Chris S., Pau

elf
Posts: 1416
Joined: Sun Nov 18, 2007 12:10 pm

Desktop automated focus stacking setup

Post by elf »

It's not really done yet as it seems there will always be something more to add :)

This is based on previous setup detailed in these threads: Major changes are
  • 12"X18" 80 pound granite base
    Heavier aluminum extrusions
    XYZ and rotation subject stage
    Rotary table base for the camera side
    Bellows draw increased to 300mm
    Microprocessor control of the bellows draw
Image

Image

Image

Image

Image

Image

The granite block is resting on 4 small rubber balls which eventually will be replaced with sorbothane hemispheres to better isolate it from external vibrations. It does move pretty freely on the rubber balls, so touching the setup while taking an image is not advisable.

The X, Y, and Z axes are controlled with 5mm X .8 leadscrews. A split nut is used on each axis so each can be rapidly moved manually. The rotation axis is currently manual. The stage is only used for initial positioning of the subject. The focus steps are done by changing the bellows draw.

With the addition of the stage, it was no longer necessary to have a tilting camera base. The rotary table was added to give a little smoother control over the rotation of the camera base. Unfortunately, it has quite a bit of backlash in it, but this is fairly well controlled with a bungee cord. The rotary table is scheduled to be replaced with a stepper motor controlled rotating base which should take care of the backlash problem.

The camera arm can be rotated down about 30 degrees and rotated up 90 degrees. I've removed the capability for camera swings and tilts as I found I wasn't using them for macros. The biggest problem was the small Live View screen just didn't have enough resolution to get the focus perfect. Focus stacking, on the other hand, is more tolerant of focusing errors. You just need to make sure the start and end points cover the focus range desired.

The stepper motor is controlled by an Arduino Duemilanove microprocessor board and an EasyDriver stepper driver board.
The stepper motor does 1600 microsteps per revolution and is further reduced with the belt drive and pulleys to 3200 steps per revolution. With the .8mm pitch of the leadscrew, each step is .00025mm, but it's unlikely to be that accurate due to stiction and friction. I'm also using the Arduino to trigger the camera shutter via IR. The main reason for this, is the Olypus SDK doesn't allow saving images to the camera. The download time for each image is 20+ seconds. By using IR to trigger the shutter, images can be taken every 8 seconds.

I've seen a big reduction in vibration problems by using the granite block and hands-off with the Arduino and stepper motor. Next on the agenda is fitting a Nikon BD Plan 10 microscope objective.

ChrisR
Site Admin
Posts: 8668
Joined: Sat Mar 14, 2009 3:58 am
Location: Near London, UK

Post by ChrisR »

A monstrance!

I don't follow why you don't save the images to the camera, as it's so much faster, and would help with droopy flowers?
Download is quite quick with UDMA cards and Firewire.

I expect your camera has video out, to connect to a monitor? You need a composite to RGB converter to use an old computer monitor which is my intention.

BD lenses - I have some of those. Small rectangular LEDs slip easily into the annulus, but they aren't super bright. I'm on the lookout for the kind of xenon flash devices which are used in cell phones, but haven't seen any yet. The lighting does at least mean you can manage the small WD, as long as the subject doesn't have too much depth.

Sounds like you might want 3D dampers on the granite block support? Once it starts wobbling , it'll take some stopping. You have to get their rate right - but you knew that!

AndrewC
Posts: 1436
Joined: Thu Feb 14, 2008 10:05 am
Location: Belgium
Contact:

Re: Desktop automated focus stacking setup

Post by AndrewC »

elf wrote: The stepper motor does 1600 microsteps per revolution and is further reduced with the belt drive and pulleys to 3200 steps per revolution. With the .8mm pitch of the leadscrew, each step is .00025mm, but it's unlikely to be that accurate due to stiction and friction.
Precision will also be governed by the thread on your leadscrews and backlash on the nut.. Check the specs on what you have and I doubt they'll approach what you are trying to achieve. This is basically why advanced applications use linear motors rather than screw/nut drives. The advantage of microstepping is to supply smoother movement rather than just more precise movement. 8 microsteps is gentler than one big step. Downside is that you lose power and so sometimes miss steps - it is why you find that a lot of precision movement systems also use shaft encoders to verify that what you expect to happen actually does happen. Really precise stages use laser interferometry for closed loop positioning but that is a whole other ball game.
The granite block is resting on 4 small rubber balls which eventually will be replaced with sorbothane hemispheres to better isolate it from external vibrations. It does move pretty freely on the rubber balls, so touching the setup while taking an image is not advisable.


Make sure you size the sorbothane hemispheres correctly - with the right loading they do a tremendous job of soaking up and isolating movement in all axes.

Anyway, you probably knew all that already :)

Andrew

DaveW
Posts: 1702
Joined: Fri Aug 04, 2006 4:29 am
Location: Nottingham, UK

Post by DaveW »

Looks like one of the old Oxford Scientific Films set-ups. Put a movie camera on it and use small models and you could tender for the close-up shots in the next edition of Micronauts as they did! :D

Great set up, nice when you have the engineering facilities to do it, cant wait to see the results.

DaveW

AndrewC
Posts: 1436
Joined: Thu Feb 14, 2008 10:05 am
Location: Belgium
Contact:

Post by AndrewC »

Another thought, if you are going to try a 10x lens. All those aluminum extrusions have the potential to ring like a bell. Load up their central cores with sand and it will dampen it significantly. You can keep the sand in place with a silicone or epoxy plug.

Andrew

DaveW
Posts: 1702
Joined: Fri Aug 04, 2006 4:29 am
Location: Nottingham, UK

Post by DaveW »

Another alternative to stop "ringing" is to use one of those aerosol expanding foam containers to fill the hollow aluminium sections, usually sold by builders merchants for fixing door frames and windows, if you can get them in the US?

http://www.tooled-up.com/Product.asp?PID=2522

Don't go this far though:-

http://boards.fool.co.uk/Message.asp?mi ... mendations

DaveW

AndrewC
Posts: 1436
Joined: Thu Feb 14, 2008 10:05 am
Location: Belgium
Contact:

Post by AndrewC »

There is a very good paper on positioning accuracy on this page:

http://www.dovermotion.com/WhitePapersPage.aspx

elf
Posts: 1416
Joined: Sun Nov 18, 2007 12:10 pm

Post by elf »

ChrisR wrote:A monstrance!

I don't follow why you don't save the images to the camera, as it's so much faster, and would help with droopy flowers?
Download is quite quick with UDMA cards and Firewire.

I expect your camera has video out, to connect to a monitor? You need a composite to RGB converter to use an old computer monitor which is my intention.

BD lenses - I have some of those. Small rectangular LEDs slip easily into the annulus, but they aren't super bright. I'm on the lookout for the kind of xenon flash devices which are used in cell phones, but haven't seen any yet. The lighting does at least mean you can manage the small WD, as long as the subject doesn't have too much depth.

Sounds like you might want 3D dampers on the granite block support? Once it starts wobbling , it'll take some stopping. You have to get their rate right - but you knew that!
Olympus doesn't allow saving to the camera when it's tethered and no other operations can be done on the camera until the image has downloaded. Changing camera brands would cure this problem as Canon and Nikon don't have this restriction.

I like the xenon flash idea. Have you looked in Mouser or Digikey?

I was hoping to find a less expensive set of Sorbothane feet than Edmund Scientific but surplus unhappy balls are hard to find:)
AndrewC wrote:
Precision will also be governed by the thread on your leadscrews and backlash on the nut.. Check the specs on what you have and I doubt they'll approach what you are trying to achieve. This is basically why advanced applications use linear motors rather than screw/nut drives. The advantage of microstepping is to supply smoother movement rather than just more precise movement. 8 microsteps is gentler than one big step. Downside is that you lose power and so sometimes miss steps - it is why you find that a lot of precision movement systems also use shaft encoders to verify that what you expect to happen actually does happen. Really precise stages use laser interferometry for closed loop positioning but that is a whole other ball game.

Andrew
One of the more interesting drive methods I've seen uses a fixed belt with the stepper motor moving along the belt and with idler pulleys to hold the belts together except where it goes over the motor pulley. Claimed accuracy was sub micron over a meter distance.
DaveW wrote:Looks like one of the old Oxford Scientific Films set-ups. Put a movie camera on it and use small models and you could tender for the close-up shots in the next edition of Micronauts as they did!

Great set up, nice when you have the engineering facilities to do it, cant wait to see the results.
A lot of the code examples I looked at for the Arduino were for animation and other camera movements like panning and moving along a rail. I think it would be kind of fun to do an animated macro video.

The images I've posted over the last couple of weeks were done with this setup. It took a while to get everything stabilized and looking good enough to post pictures of it.

AndrewC wrote: Another thought, if you are going to try a 10x lens. All those aluminum extrusions have the potential to ring like a bell. Load up their central cores with sand and it will dampen it significantly. You can keep the sand in place with a silicone or epoxy plug.
Excellent idea, I may even have to borrow some of NU's sand bags :)



DaveW wrote:
Another alternative to stop "ringing" is to use one of those aerosol expanding foam containers to fill the hollow aluminium sections, usually sold by builders merchants for fixing door frames and windows, if you can get them in the US?

http://www.tooled-up.com/Product.asp?PID=2522

Don't go this far though:-

http://boards.fool.co.uk/Message.asp?mi ... mendations

DaveW
I actually have several cans of this in the shop. It's quite useful for creating temporary bearings for small spindle turning on a wood lathe. The sand is perhaps a little safer to use as it's not as messy and easily reversable :)

ChrisR
Site Admin
Posts: 8668
Joined: Sat Mar 14, 2009 3:58 am
Location: Near London, UK

Post by ChrisR »

I think the foam's low mass is against it. How about filling your limbs with shotgun pellets? Add filings if you want to get the volume occupancy up.
Density of sand is about 2.8, = only about the same as aluminium. Lead is about 11. It's good to get the mass near the scene of the accident.
Have you ever used a "dead blow" hammer - remarkable things. Different use but they don't bounce!

I'm not familiar with "tethering" the camera. I assume that's what you're using to set the number of shots, but I'm not sure what else. The Arduino could do that of course. I have one of the Arduino LCD display "shields", which has a few buttons on it too, which you could use to set it up.
It's only 2 rows by 80(?) characters, but adequate I think. There's some examples of people using larger displays on the net, but more time and effort..!

AndrewC
Posts: 1436
Joined: Thu Feb 14, 2008 10:05 am
Location: Belgium
Contact:

Post by AndrewC »

I was hoping to find a less expensive set of Sorbothane feet than Edmund Scientific but surplus unhappy balls are hard to find:)
I've purchased from this guy several times:

http://cgi.ebay.co.uk/4-SORBOTHANE-SORB ... 5632a694d2

Andrew
rgds, Andrew

"Is that an accurate dictionary ? Charlie Eppes

AndrewC
Posts: 1436
Joined: Thu Feb 14, 2008 10:05 am
Location: Belgium
Contact:

Post by AndrewC »

ChrisR wrote: I'm not familiar with "tethering" the camera. I assume that's what you're using to set the number of shots, but I'm not sure what else.
Varies from brand to brand but generically means having the camera connected to a computer through a USB or other connection. The camera recognises it has been taken over and some of its functions dissappear - you can't format a card in a tethered Nikon for example.

Seems strange that Olypmus don't let you transfer images when tethered - I tether my Nikon simply for that. Takes just a few seconds to transfer an image. I don't have Liveview but approx 3secs after taking a picture I can see it full size on my computer monitor plus I can watch the stack being collected in realtime without risking touching the setup. If,as and when Rik creates an ingester version for Zerene I'll just be able to sit back and watch the stack create from scratch :)

Andrew

g4lab
Posts: 1437
Joined: Fri May 23, 2008 11:07 am

Post by g4lab »


elf
Posts: 1416
Joined: Sun Nov 18, 2007 12:10 pm

Post by elf »

AndrewC wrote:
ChrisR wrote: I'm not familiar with "tethering" the camera. I assume that's what you're using to set the number of shots, but I'm not sure what else.
Varies from brand to brand but generically means having the camera connected to a computer through a USB or other connection. The camera recognises it has been taken over and some of its functions dissappear - you can't format a card in a tethered Nikon for example.

Seems strange that Olypmus don't let you transfer images when tethered - I tether my Nikon simply for that. Takes just a few seconds to transfer an image. I don't have Liveview but approx 3secs after taking a picture I can see it full size on my computer monitor plus I can watch the stack being collected in realtime without risking touching the setup. If,as and when Rik creates an ingester version for Zerene I'll just be able to sit back and watch the stack create from scratch :)

Andrew
I think ZS Ingestor would be a fine name for it :)

Olympus is the opposite. It requires the image to be transferred to the computer and there is no option to save to the camera. On the newest cameras this isn't as big of deal but on my e330 it takes 20+ seconds to transfer the image. The e330 only has a slow USB 1.0 port.

I'm waiting to see if Olympus will introduce a better mirrorless camera before switching to another brand.

Another reason for using the Arduino is that I can adapt it to the older pano head and use it in the field. I have another Arduino board on its way that plugs directly into the back of a 2 line LCD (It's a prototype and not generally available yet).

I've been to this site before. There's some pretty interesting uses for sorbothane. The people using it for audio equipment are generally using the lower weight rated dampeners. I'll need something that handles a little over a 100 pounds.

Aynia
Posts: 724
Joined: Thu May 01, 2008 7:42 am
Location: Europe somewhere
Contact:

Post by Aynia »

DaveW wrote:Looks like one of the old Oxford Scientific Films set-ups.
I agree. Amazing looking set up. :D I'm not even going to try and understand the technicalities of what's going on! I'm not ready for that yet!

ChrisR
Site Admin
Posts: 8668
Joined: Sat Mar 14, 2009 3:58 am
Location: Near London, UK

Post by ChrisR »

I can watch the stack being collected in realtime without risking touching the setup
Yeah but, no but, I mean, why would you want to alter camera settings mid-stack?
I'm still missing the point here. What's the point of "tethering"?
OK it saves taking out the card and reading the data to the PC, but is that all?

--
I have another Arduino board on its way that plugs directly into the back of a 2 line LCD (It's a prototype and not generally available yet).

You'll find a board that does that on ebay. Roboduino perhaps. WHat's special about the one you have coming?

Post Reply Previous topicNext topic