in what situations do we need steps as small as 0.3 micron .

[Yawns]

Currently on the market there are automatic rails that declares minimum step values ​​of the order of fractions of microns .. like 0.3 microns ..

I also started seeing in some groups "exifs" declaring steps with an accuracy of 1/100 micron, such as step = 0.59 micron (??) .. this is very ]accurate and I'm confused ...

If λ / Na ^ 2 is valid for the entire range from 0 to 1.0 (I don't know if it is ...) The step starts to be <1 after Na is> 0.75, and this is a very high Na. ..

I was checking Mitutoyo lenses (usually with very large WD) and only one 50x was 0.75 Na. Above that, only the 100x were having 0.75 or 0.90, but the working distances were not that great .. as 1.3mm and 5mm

in what situations do we need steps as small as 0.3 micron ...?

[rjlittlefield]

λ / Na ^ 2 is a pretty good approximation for DOF; a better one is shown at https://www.photomacrography.net/forum/ ... 955#215955 .

However, this DOF reflects only part of the problem. It captures the issue that the MTF of a diffraction limited lens will sag as a result of defocus, while viewing a planar subject that gives uniform utilization of the lens aperture.

However, at high magnifications, there is another problem that is often more important than MTF sagging. That problem is that subject features can appear to move laterally as focus is changed, and the direction of movement is different for different features. This results in features sort of "squirming around" as focus is changed. For an example see the sunset moth scales at https://www.photomacrography.net/forum/ ... 187#149187 .

The amount of lateral movement can be significant even when the amount of sharpness loss is not. Particularly for linear structures where misalignment is more visible and the lateral shifting may be worse, you can sometimes get better results by using a step size that is several times smaller than the DOF calculation. If I were to reshoot those sunset moth scales, I would think seriously about dialing back to 0.5 or 0.25 microns instead of the 1 micron that didn't work very well, despite that nominal DOF for that objective is 1.8 microns.

--Rik

[Yawns]

Thank you for your explanation Rick.. In my limited knowledge I could not find a reason... anyway I just go up to small magnification like 10X and with simple achromatic and I don't need such small steps, but I was curious ...

António.

EDIT: I saw that complex formula before and I must have a print screen somewhere in my drive... I will put it a Excel sheet and compare the results with the simple one.

[rjlittlefield]

I saw that complex formula before and I must have a print screen somewhere in my drive... I will put it a Excel sheet and compare the results with the simple one.

When checking the numbers, remember that λ / Na ^ 2 gives the two-sided DOF = total distance from front to back of the in-focus slab, while the more accurate formula gives only the one-sided DOF, maximum distance from perfect focus.

To get comparable numbers, you have to multiply the fancy number by 2.

According to my spreadsheet, the more accurate formula predicts about 4% less DOF at NA 0.40, 10% at NA 0.60, and 20% at 0.80, all for objectives in air. For what I do, these differences are not worth worrying about because the effect is small compared to the "squirming around" problem.

--Rik

[Yawns]

I saw that complex formula before and I must have a print screen somewhere in my drive... I will put it a Excel sheet and compare the results with the simple one.

When checking the numbers, remember that λ / Na ^ 2 gives the two-sided DOF = total distance from front to back of the in-focus slab, while the more accurate formula gives only the one-sided DOF, maximum distance from perfect focus.

To get comparable numbers, you have to multiply the fancy number by 2.

According to my spreadsheet, the more accurate formula predicts about 4% less DOF at NA 0.40, 10% at NA 0.60, and 20% at 0.80, all for objectives in air. For what I do, these differences are not worth worrying about because the effect is small compared to the "squirming around" problem.

--Rik

I was suspecting that .. because I immediately noticed the old formula was returning values near to 2x bigger the new one .. 2,5 after Na >0,80.
to much of a coincidence .... and I suspected the "error" couldn't be that big.
Even verified if I wrote well the formula on excell ...
Thank you.

Screenshot 2019-10-15 at 00.03.01 by antonio caseiro, on Flickr

[rjlittlefield]

I also started seeing in some groups "exifs" declaring steps with an accuracy of 1/100 micron, such as step = 0.59 micron (??) .. this is very ]accurate and I'm confused ...

This sort of number sounds like what happens when one uses a screw that is threaded in turns per inch, then converts to metric, or uses a metric screw but with some complex gear ratio like 32:90 between the motor and the screw. Without more information, I would expect that the indicated precision means nothing more than that's how the computation worked out.

For example the StackShot rail is 16 turns per inch, which computes out to be 0.496 microns per microstep of its motor. But we know from careful measurement that the physical increment per microstep only has that number as its average value. I've measured the increment per individual microstep as large as 1.38 microns, even in High Precision mode.

--Rik

[mawyatt]

For example the StackShot rail is 16 turns per inch, which computes out to be 0.496 microns per microstep of its motor. But we know from careful measurement that the physical increment per microstep only has that number as its average value. I've measured the increment per individual microstep as large as 1.38 microns, even in High Precision mode.

--Rik[/quote]



Rik,

When I measured the actual Stackshot motor current Sine and Cosine waveforms from the Toshiba (I think) driver chip, the waveforms didn't look very good. Others had reported issues with various Toshiba and TI chips. These waveforms also explain the annoying "buzz" from the rail.

The motor magnetic field effectively integrates the waveforms, but with a waveform like these I'm not surprised at that kind of variation you measured.

https://www.photomacrography.net/forum/ ... c&start=15

The Stackshot Motor and Rail are quite good, the controller is the weak link IMO. I've run the Stackshot Motor & Rail with the Pololu and Trinamic based controllers, dramatic improvement without the noise

Best,

[Adalbert]

Hi Yawns,
If you wanted to do 3 pictures a DOF for the lens with the NA=0.75 you would need the step-size = 0.33 micron.
https://www.photomacrography.net/forum/ ... 7&start=30

But what about more than 3 pictures a DOF and NA=0.95 ?
https://www.photomacrography.net/forum/ ... highlight=
BR, ADi

[Pau]

I do not need so small steps for macro because I don't use high NA objectives but for microscope work in some cases there is a small but clear improvement. Some time ago I performed a test with a 25/0.65 objective in transmitted light and 0.35micron steps worked better than 1.05micron, very likely because of the "squirming around" effect explained by Rik

Of course I don't need a precision rail as I couple the stepper motor to the microscope drive

[Macro_Cosmos]

With the Mit 50x M Plan and HR version, you'd need that kind of steps. Depends on the device. Lead screw mechanism has inherent backlash and variable precision, even though the Mit 50x has a DOF of 0.9, I'd usually recommend goings in steps half of the DOF of the objective, in this case, 0.45 is good.

Most shouldn't be concerned about the camera's shutter, it's a tool after all. Most offerings nowadays even have fully electronic shutters which theoretically won't impact the lifetime of the shutter mechanism. I'd say in general, more is better. That's why my stacks go beyond the required DOF, it's better to delete 100 of them than to lack 20, IMO.


Just my 2 cents, some say their stacks are successful at 1um for the 50x Mit.