My Budget 1x-10x Stacking Setup

A forum to ask questions, post setups, and generally discuss anything having to do with photomacrography and photomicroscopy.

Moderators: rjlittlefield, ChrisR, Chris S., Pau

apt403
Posts: 37
Joined: Mon May 06, 2019 5:29 pm
Location: Yelm, WA

My Budget 1x-10x Stacking Setup

Post by apt403 »

Burned way too much time last night trying to capture sharp images at 2:1 reproduction on the tripod/ballhead combo that's been marginal yet sufficient @ 1:1. Spent a few hours drawing up a CAD design for the longer term solution - a 2" thick reinforced concrete table w/ a welded steel frame, with a piece of MIC 6 tooling plate bolted to linear rails and a ballscrew. Being the impatient type that I am, I decided today to try and come up with a super cheap setup using bits and pieces I already had lying around. Hopefully now I can get back to finishing up the imaging tests I should have completed yesterday!

Image

I'm sure we've all seen these budget X-Y tables before - Mine is bolted through the board and a 123 block I grabbed from my welding fixturing gear, using a 1/4-20 bolt.

A 25lb weight is bolted to the MDF in an attempt to add mass, damp vibration, and push up the resonant frequency of the system. This sits on 3, 5lb weights I'm using as feet, with approx 1/16" (~1.6mm) thick adhesive-backed neoprene rubber sandwiched in between. Everything sits on a heavy oak desk.

Image

In order to monitor linear travel, a surface grinding vice holds a Starrett 25-131 dial indicator with .0005" (12.7µ) graduations. Also experimenting with a dial test indicator - Because the probe on a DTI is repositionable, cosine error can be used to one's advantage - With the probe at approx. 75°, a .0005" reading on the dial will correspond to approx. .00012" (~3µ) of actual travel. The mass of the vice seems sufficient to prevent unwanted movement during use.

I'd say the max positioning resolution easily achievable with these tables is somewhere around .00025" (6.35µ), adequate (barely) for the DOF of a 10x NA .25 objective. Beyond that, hysteresis of the dovetail ways becomes the limiting factor. I imagine if I were to recreate the gibs out of a more appropriate material, perhaps PTFE, accuracy could be increased somewhat. Regardless, the rack and pinion system employed creates too much backlash to warrant further attempts at improvement in my eyes.

Image

A Starrett machinist's jack is placed under the lens, and run up just until it makes contact with the body of the teleconverter. Sagging has been both problematic for framing shots, and cause for some mild concern for the health of the composite DSLR body.

Image

A scrap piece of PVC rod was turned down on one side to fit into the recess on the knob, and fixed in place using quick setting epoxy. The other end is cross drilled for a piece of 6mm TGP shafting, if a larger effective diameter is desired. I imagine this glue joint will fail after protracted use - After sealing, anodized aluminium is a very poor surface to bond to. When that happens I'll remove the knob and sandblast it. Perhaps even re-anodize it. Prior to sealing, anodizing makes a great surface for adhesive bonding!

Image

Imaging subjects are setup on an old surface plate scribe, fixed using small binder clips. Not the most precise setup, but functional. Eventually I'd like to acquire or build a precise positioning rig that will be held fixed relative to the camera setup.

I use the in-camera live view to frame the shot and take test exposures, then switch to Nikon's Snapbridge app on my phone for remote triggering. The app itself comes with a live view, but the resolution is appallingly low and inadequate for finding focus.

I know this whole conglomeration absolutely pails in comparison to some of the other amazing setups posted here, but hopefully it'll prove helpful to someone just getting started, to show you don't need much gear to get good results.

Lou Jost
Posts: 5948
Joined: Fri Sep 04, 2015 7:03 am
Location: Ecuador
Contact:

Post by Lou Jost »

Super-important point-- the subject holder needs to be on the same white slab as the camera. There shouldn't be neoprene padding between camera and subject. The idea is to rigidly connect the camera and subject so they move together when vibrating.

ray_parkhurst
Posts: 3417
Joined: Sat Nov 20, 2010 10:40 am
Location: Santa Clara, CA, USA
Contact:

Post by ray_parkhurst »

Lou Jost wrote:Super-important point-- the subject holder needs to be on the same white slab as the camera. There shouldn't be neoprene padding between camera and subject. The idea is to rigidly connect the camera and subject so they move together when vibrating.
+1...this is more important than anything else in the system. I really like the machinist's jack, though I would move it as far out on the TC as possible. Do all parts of the lens move? If not then the jack would be even better placed on the farthest out fixed part of the lens, though I'm not clear on what happens when the camera moves side to side.

apt403
Posts: 37
Joined: Mon May 06, 2019 5:29 pm
Location: Yelm, WA

Post by apt403 »

Ah, very good point! A few test shots prior to reading this post drove the point home nicely, too. I've since rebuilt the setup - Now there's a steel plate screwed to the board, to which I can affix a mag. base holder. Also tripled the amount of weight, and have the whole deal resting on a 6" (150mm) thick piece of foam I've had kicking around for years. Additionally, I sewed up some bags of shot from an old pair of jeans - Every time the shutter fired, I would get a few microns worth of movement. Weighing down the camera seems to have done the trick:

Image

Further, I've got to revise my previous numbers - 5x is about the best I can hope for w/ the X-Y table. It's possible, but exceedingly tedious to make adjustments finer than that. 2x is really the sweet spot. For anyone w/ one of these tables, I'd highly recommend affixing the X-axis w/ some sort of adhesive - As the photo below shows, it's held on by a single 1/4-20 bolt, with no provision made to prevent side to side movement. There's enough slop to move the camera more than a whole frame @ 10x:

Image

Here's a quick test shot of a Claytonia flower @ 2x with the revised setup:

Image

Playing around with my new Olympus LMPLFL10X, I've come to realize that
doing manual stack and stitch (finally making some progress with stitching!) is extremely time consuming. So, I shook off the laze and build a quick DSLR mount for my CNC router:

Image

I was considering picking up an automated macro rail, but that wouldn't solve the problem of panning. These machines are about the same price (~$300-$400) but can move in all three axes w/ at least .0001" (2.54 micron) precision, as checked with a dial indicator. I've got the steppers set to 1/16 microstepping right now, but can go up to 1/64, for a theoretical minimum step size of .4 microns on the 5mm pitch ballscrews (5mm/(200 steps/rev * 64)). I doubt these rolled screws are accurate enough to hit that number repeatedly. Still tweaking the acceleration settings in LinuxCNC - What's fine for machining plastic and wood isn't so got for smooth movement @ 10x.

I'm browsing Amazon now for an appropriate remote trigger - I plan to disassemble the unit, replace the switch with a relay tied into one of the outputs on the control board. With some simple gcode, I should be able to automate the entire process of capturing the individual stacks for stitching. It'll be a welcome relief to program the machine, and walk away while it captures multirow panos for me.

Lou Jost
Posts: 5948
Joined: Fri Sep 04, 2015 7:03 am
Location: Ecuador
Contact:

Post by Lou Jost »

That's a brilliant solution!

Post Reply Previous topicNext topic