Micro stage movement

[LordV]

I tried using my old ( almost antique) novoflex macro rail as a specimen holder when stacking at 5X and 10X. It turned out to be just about adequate at 5X but too coarse for 10X where oof striping was evident in the stack.
Just wondering whether anyone has tried using a lab microscope slide stage for this (my only other on tap source of contollable small movement). I'm guessing the movement control should be fine enough ?

Brian v.

[lauriek]

Brian quite a few of us use microscope focus stages - I use an old Olympus BH base&focus block (with the head sawn off) . I assume you're talking about the stage focus mechanism and not the stage X-Y movements, I doubt the latter would be fine enough for 10:1 stacking but I could be wrong.

[LordV]

Brian quite a few of us use microscope focus stages - I use an old Olympus BH base&focus block (with the head sawn off) . I assume you're talking about the stage focus mechanism and not the stage X-Y movements, I doubt the latter would be fine enough for 10:1 stacking but I could be wrong.

Hi,
Thanks for the reply,
afraid I was thinking of the slide x-y stage movement not the coarse/fine focus objective movement - i could try this (and probably will) without taking the microscope to pieces.

Brian V.

[NikonUser]

I can get 0.1mm increments of my x-ymicroscope stage, with extreme care perhaps 0.05mm. I think you may need something like 0.01mm or finer for 10x.
The fine adjusment (z-axis) on a good scope will give at least 0.002mm increments.

[rjlittlefield]

You can use the microscope focus mechanism without taking the scope to pieces. Just rig a little side extension on the stage. Nothing elaborate is needed, just a piece of wood fastened on with small clamps or even well taped.

--Rik

[LordV]

Thanks all for the comments and ideas - will have to have another play during the next spell of bad weather.
Brian v.

[PaulFurman]

If you can work with moving the subject (very lightweight), it's pretty easy to find parts, moving the whole camera and or bellows sounds like it needs a big focus block or machinist's table.

I've got a second matching top part left over from this micropositioner thing which moves about 2mm in 10 comfortable half-turns. Quarter turns are manageable but very precise with the tiny handle unless you can reach it to inspect & add markings so you have to over-sample to avoid a few ambitious turns messing up a huge stack by losing the eyes or something. It's easy to slip into a different rhythm & start doing larger/smaller steps. And if it's not lined up with the lens, you get wasted pixels on the edges. But it's small & easy. I just did a 10mm deep stack with 166 frames, which was overkill for 2.5x but my next option is the bellows & that's more like .5mm per slight nudge and easy to mess up. It will travel about 15mm, which is good. The other 2 below only move about half that distance.

And another larger positioning slide from that with marked micrometer knob (although you'd need to hunch over it and squint to comfortably move single units (hundredths of a mm or .01mm) and it's marked 0, 10, 20, then ends on 25 so awkward counting if you wanted to squint & land on full turns or 5 units. 10 full turns moves 2.5mm and half or quarter turns are pretty comfortable. Maybe marking the zero more boldly would be useful and the halfway point in another color. Double turns would be a bit tedious, you'd need 4 full turns to move a tenth of a mm.

The third one has a slightly larger micrometer dial with 5-unit indexes and is also 25 units (again units are hundredth of a mm or .01mm) for a full turn but also only moves about 7mm total. The plate it mounts to has a coarse adjustment slider with 50mm of play similar to the bellows where you can barely squeak out .5mm movements unreliably.

These are greased metal dovetails, some have a spring and a finely threaded screw pushes them to compress the spring. The smallest top one is hard-threaded with no spring and is most useful because of the longer travel and faster adjustment.

[PaulFurman]

I can get 0.1mm increments of my x-ymicroscope stage, with extreme care perhaps 0.05mm. I think you may need something like 0.01mm or finer for 10x.
The fine adjusment (z-axis) on a good scope will give at least 0.002mm increments.

I'm just clarifying this for myself as I learn, not to be critical. Yesterday I did a stack of 166 frames at about 0.05mm at about 2.5x which was overkill but maybe not by much for a fairly deep 10mm subject. Quadruple the scale for 10x and that's about your 0.01 estimate though like I said that's more than plenty and my 0.05 steps would be hard to produce reliably for 10x work; miss just one at the eyes and a couple hours are wasted.

0.01mm would mean a stack of 100 frames for a 1mm deep subject.
0.001mm = 1 micron and gives 1000 frames for a 1mm deep subject

edit: the underlined comment below was in error, I lost a zero and exaggerated the situation.

10x on a full frame SLR holds a 3.6mm long subject and if it's roughly the same depth and you wanted a stack of 50, you would only need a 0.07mm movement so it should be just barely possible with extreme care on the microscope stage you describe. The 1 micron increments in the sliding positioners I described where a full turn of the knob is 25 microns needs 4 turns to go 0.01mm and 20 turns to go that 0.05mm step described for a stack of 50, which is not comfortable or convenient. Probably indispensable for working at 50x. I have not used a microscope block, probably more convenient than these micrometers.

[PaulFurman]

0.01mm would mean a stack of 100 frames for a 1mm deep subject.
0.001mm = 1 micron and gives 1000 frames for a 1mm deep subject

...The 1 micron increments in the sliding positioners I described where a full turn of the knob is 25 microns needs 4 turns to go 0.01mm and 20 turns to go that 0.05mm step described for a stack of 50, which is not comfortable or convenient...

Here's one with a little crank handle that has 0.025mm ticks vs the 0.001mm of these micrometers: Re: Persian speedwell (modified Unislide) which sounds ideal to me. It's probably not hard to do half-tics. I believe that's 1000's of an inch, 25x bigger than microns. And it appears sturdy enough to hold a camera, although probably not with bellows & large-ish lens also.

It's basically just a 20/TPI A250 Velmex Unislide. ...The scale reads in approximately .025mm graduations...

[ChrisR]

The 1 micron increments in the sliding positioners I described where a full turn of the knob is 25 microns needs 4 turns to go 0.01mm and 20 turns to go that 0.05mm step described for a stack of 50

It's very easy to lose a zero!
1 micron = 0.001mm so 4 turns of 25 microns = 0.1mm

I've found the step sizes which work for me for scope lenses (Others will disagree - it does depend eg to a point on the resolution of your sensor) are approximately


If your knob has 100 tick marks, or you make a disc which can be used, such as
http://photomacrography.net/forum/viewt ... 8186#48186
and your thread pitch is 0.5mm (e.g. M6)
then each tick mark will be 5 microns, you can reasonably go to 10x or 20x
I don't have one but I believe the Proxxon milling table mentioned in the forum several times (initially by Betty ) works at about this level.
Beyond that, you probably need a microscope focusing stage.

[rjlittlefield]

Chris, thanks for the table. The numbers you give are consistently about 80% of the DOF calculated by microscopyu for e = 8 µm. This fits nicely into the concept of "a little overlap", as well as providing a crosscheck on microscopyu's numbers. Most appreciated.

--Rik

[PaulFurman]

The 1 micron increments in the sliding positioners I described where a full turn of the knob is 25 microns needs 4 turns to go 0.01mm and 20 turns to go that 0.05mm step described for a stack of 50

It's very easy to lose a zero!
1 micron = 0.001mm so 4 turns of 25 microns = 0.1mm

Oops! that's much more reasonable, I'll add a correction to my text above to minimize frustration.

I've found the step sizes which work for me for scope lenses (Others will disagree - it does depend eg to a point on the resolution of your sensor) are approximately

[ 4x - move 40 microns per step]
[10x - move 10 microns per step]

Another way of looking at it if I transcribed your table correctly:

[at 4x - a stack of 50 covers a 2mm deep subject]
[at 10x - a stack of 50 covers a half-mm deep subject]

So my stack of 166 the other day at 2.5x for a 10mm deep subject... let's just say I did it at 4x, those 150 steps would get me 6mm with your numbers. Hmm, wow, that is in the ballpark. It would take about 1-1/2 turns on the micrometers per step. As it was I made about the smallest turn I could on the small thumbscrew for my miniature positioner, so that thing won't quite handle careful 10x work and I wasn't really oversampling much for 2.5x.

Other factors as you said, are the sensor resolution (I'm on 12MP full frame) and also whether using macro lenses or microscope objectives which I understand are really much 'faster' than macro lenses, although the explanation eludes me.

Right now I've got a Canon 35/2.8 that I use at f/4 which is great corner to corner at low mag of 2.5x but doesn't gain much real resolution when I push it to 7x. I've got an Olympus 20mm f/2 for 6x-12x coming in the mail which should be usable corner to corner wide open. That's why I'm paying attention to this topic.

[ChrisR]

At 12MP on FX (our) D700s have approx 8micron pixels, which are biggish. Scope lenses are only "meant" to cover approx the DX sensor size, so we effectively have lower resolution sensors than say a D300, for microscopy applications.
Eg, resolution is given as 0.275/NA, so with a typical 10x we get about 1 micron. Multiply by 10 gives 10 microns on the sensor, which you'd need smaller pixels than 8 microns to see. So "e" depends on the camera.
It is said that 2 to 4 or so pixels are required to properly record a "line". Quite how resolution as measured in "e" and a "line" would be equated I'm not sure, but "e" of 8 microns would be too small for a D700.
So I'm overlapping steps by more than 20%.

The scope objectives I've tried vary in their coverage of the FX frame. One 2x gives a circular image. An "infinity" 10x produces a curved field, which is masked to some extent by stacking. ( Intend to try it with a longer tube, or a relay lens).

microscope objectives which I understand are really much 'faster' than macro lenses, although the explanation eludes me.

Perhaps we could do with a FAQ topic to include standard little formulae.. It's quite hard to find them in previous posts. I have that one somewhere.
F number is given by 1/(2*NA) * m/(m+1), with a fudge factor of a pupillary magnification factor which you'd have to know lens parameters to factor-in.
However, a 10x 0.30, which has a focal length 16mm works out to an F number
1/(2*0.3) * 10/11 which is f1.5

A (40mm) 4x 0.2 would be
1/(2*0.2) * 4/5 which is f2
but the 4x 0.13 would be f3
and the 4x 0.1 would be f4

[PaulFurman]

Thanks Chris. I'm hopeless with formulas. But if I understood, after spending some time in one of those pupillary magnification threads (huge if) is that the aperture is figured for the 160mm extension or whatever standard it's spec'ed for. A macro lens' specified aperture is presumably spec'ed for infinity focus with no extension. If true, that 40mm 4x f/2 objective compares with an f/4 macro lens that ends up with an effective aperture more like f/17.

That's using this formula: "multiply the f-number on the lens ring by magnification plus one", I used a different one before...

[Charles Krebs]

Chris,

About week ago Rik and I talked briefly about DOF and steps.

The "NikonU" formula seems to be the best one to use. I put it into a spreadsheet for the objectives I use and the results (as Rik mentioned) correlate nicely. I'm using twice the pixel size for "e". Here's what it works out for with a 4.6 and 8 micron pixel sensor.