Microns …

[Kiwi BlokeParticipant @kiwibloke62605]

I stand by what I said: "That non-temperature-controlled apparatus thinks it can resolve 1nm repeatably enough to justify the scale's resolution? Delusional, surely. It reminds me of a friend tickled pink by his small bench-top oriental CNC mill. He discovered that his CAD software could work to microns, and sincerely told me that he could now machine things 'accurate to a few microns'."

I can accept that the piezo gizmo can, in effect, change its length by 1nm, repeatably and accurately, but I'm talking about the apparatus as a whole. I'm not clear whether the piezo gizmo is a positioner or a displacement sensor, and if a positioner, whether it's working 'open loop' or 'closed loop'. In any case, the gizmo seems to be attached to a complex mechanical assembly, with a complex, cantilevered shape, which presumably has built-in clearances, etc., and the whole lot is subject to temperature effects. Is the apparatus to which the gizmo is attached dimensionally stable to within 1nm? The chain can only be as strong/accurate as its weakest link. Hence my comment about the CNC mill.

Edited By Kiwi Bloke on 08/07/2019 06:43:20

[Michael GilliganParticipant @michaelgilligan61133]

Posted by Kiwi Bloke on 08/07/2019 06:41:36:

I can accept that the piezo gizmo can, in effect, change its length by 1nm, repeatably and accurately, but I'm talking about the apparatus as a whole.

.

But why ?

The whole instrument comprises a standard optical microscope [of the inverted variety] with two Marzhauser stages attached and a digital camera system.

The action of the Z stage can be considered pretty much in isolation, but; as its resolution is sub wavelength-of-light, it may be easier to consider the XY stage … This has a much coarser resolution, and larger movements, so it might be conceptually more familiar.

Here's another photo of the 'interface'

The diatom, imaged on the screen, is about 20 microns across

The XY control [via the joystick] is 'pixel perfect' on that screen image … You can use the cross-hairs on the screen to select any point, and the stage will traverse to put that point centre-screen.

This is not 'science fiction' … these are standard tools used in labs all over the world.

MichaelG.

.

https://www.marzhauser.com/en/pim/produktdetail-popup.html?view=details&pimid=a368&no_cache=1&m=null&p=null

Edited By Michael Gilligan on 08/07/2019 08:16:48

[Michael GilliganParticipant @michaelgilligan61133]

Posted by Kiwi Bloke on 08/07/2019 06:41:36:

… He discovered that his CAD software could work to microns, and sincerely told me that he could now machine things 'accurate to a few microns'.

.

… and that [apparently popular] fundamental misunderstanding is why I carefully chose the subtitle to my thread:

[quote] It all depends on the context … [/quote]

Your friend's sincerity, and your dismissal of it, are 'two sides of the same coin'

Within the context of something tens of microns in size, it is easy [and neccessary] to work to tenths of a micron.

… and so, ad infinitum

MichaelG.

.

Edit: Although it's a different species of diatom; this SEM imaging is worth a look

http://www.scielo.br/scielo.php?script=sci_arttext&pid=S0102-33062013000400007

Edited By Michael Gilligan on 08/07/2019 08:56:15

[Kiwi BlokeParticipant @kiwibloke62605]

Michael, we seem to crossing swords today – not my intention at all. I admire the nm-resolution (and, presumably, accuracy) of the gizmo, but was suggesting that, unless the rest of the apparatus was remarkably stable, such resolution was wasted, and the achievable performance would be determined by things other than the gizmo. A bit like putting a micrometer head on a micrometer frame made from rubber? Incidentally, could the system be used to check gauge blocks – simpler than optical interferometry, I'd imagine?

[Michael GilliganParticipant @michaelgilligan61133]

Posted by Kiwi Bloke on 08/07/2019 10:22:38:

Michael, we seem to crossing swords today – not my intention at all. I admire the nm-resolution (and, presumably, accuracy) of the gizmo, but was suggesting that, unless the rest of the apparatus was remarkably stable, such resolution was wasted, and the achievable performance would be determined by things other than the gizmo. A bit like putting a micrometer head on a micrometer frame made from rubber? Incidentally, could the system be used to check gauge blocks – simpler than optical interferometry, I'd imagine?

.

Again we agree … No intention of crossing swords

The answer to your incidental question is probably "only if they are very small ones"

I seem to be struggling to get the point across [not only to your good self], but

the context is all-important.

MichaelG.

.

Edit: This would probably be the most suitable of their stages to use for checking gauge blocks; but obviously would not compete with interferometry:

https://www.marzhauser.com/en/pim/produktdetail-popup.html?view=details&pimid=a272&no_cache=1&m=null&p=null

Edited By Michael Gilligan on 08/07/2019 10:51:31

[old martParticipant @oldmart]

That's an impressive image of the diatom, the details within it must be right at the limits of optical resolution.

[Michael GilliganParticipant @michaelgilligan61133]

Posted by old mart on 08/07/2019 14:36:54:

That's an impressive image of the diatom, the details within it must be right at the limits of optical resolution.

.

Diatoms have long been used by microscopists, as a reference 'grid' to test objectives.

Klaus Kemp still supplies an 8 form Test Plate for this purpose: **LINK**

http://www.diatoms.co.uk/pg.htm

… and the best 'light microscope' images I have seen from it are by Osama Oku: **LINK**

https://micro.sakura.ne.jp/bod/typeslide.htm

Even these are completely blown-away by Scanning Electron Microscope images, of course; due to the difference in wavelength.

MichaelG.

[Neil WyattModerator @neilwyatt]

It's possible to work to absurd levels of accuracy very easily if you incorporate a feedback loop.

My telescope tracks starts in real time, on a good night with an accuracy of better than a second of arc. It even compensates for the speeding and allowing of stars caused by diffraction as the move higher or lower in the sky – the two clever calculations are (1) finding the centre of the image of a 'guide star' to sub-pixel accuracy and (2) the various algorithms for translating that into a movement of a mechanical mount whilst accounting for backlash and periodic errors in gearing etc.

Neil

[Andrew EvansParticipant @andrewevans67134]

I spent a lot of time looking at microscopic fossils, down to about 5 microns – they start getting fuzzy at that size due to the wavelength of light being about half a micron. That was using a fairly standard optical microscope. I got some amazing photos as well.

Neil – I am impressed that you can work so accurately!

[old martParticipant @oldmart]

I still have, amongst many others, about half a dozen slides of diatoms bought years ago from Northern Biological Supplies to use with my Russian Biolam R25. I gradually acquired most of the accessories for it. Its a shame, I never get around to use it anymore.

[Michael GilliganParticipant @michaelgilligan61133]

Posted by Neil Wyatt on 08/07/2019 16:02:36:

My telescope tracks starts in real time, on a good night with an accuracy of better than a second of arc.

.

I'm not questioning your claim, Neil, but I would be very interested to know how you demonstrate that.

… are you saying that your 'GoTo' aiming at start-up is accurate within a second of arc ?

MichaelG.

.

Edit: Is the method revealed by the solution to Problem 2, here:

https://spacemath.gsfc.nasa.gov/weekly/10Page38.pdf

… or am I missing a trick ?

Edited By Michael Gilligan on 08/07/2019 21:37:34

[Neil WyattModerator @neilwyatt]

Posted by Michael Gilligan on 08/07/2019 21:16:13:

Posted by Neil Wyatt on 08/07/2019 16:02:36:

My telescope tracks starts in real time, on a good night with an accuracy of better than a second of arc.

.

I'm not questioning your claim, Neil, but I would be very interested to know how you demonstrate that.

… are you saying that your 'GoTo' aiming at start-up is accurate within a second of arc ?

MichaelG.

No, the GOTO capability is to within about a minute of arc, the basic limit on this is the accuracy of polar alignment. I can do this manually with assistance from a plate-solving routing to better than 1 arc-minute. A multi-star alignment can then make this even more accurate, but in practice once your GOTO is getting large stars within the centre tenth of the display there's no great benefit in refining it further. You can use plate solving and computer control to line up on target to almost any arbitrary degree of accuracy.

Tracking is 'guided' using a separate scope taking 1 to 3 second images, in which a 'guide star' is identified. Because even a small star covers several pixels the 'centroid' can be calculated to an accuracy of a fraction of a pixel, so the guidescope can be much smaller and rather faster than the main scope (which might be making 5-10 minute exposures). 2-2 seconds is ideal as it's long enough that distortions caused by poor seeing are evened out, but short enough to prevent excessive drift between frames.

Because the pixel scale and focal length of the guidescope/camera combination are know the actual movement between frames can be calculated and a 'root mean square' figure. On a good night this can be a little as 0.7" or even better, on a poor one nearer to 2.0". ) 6-0.7" is about the limit for my mount as it's limited by the resolution of the drive.

The guiding software (PHD2) constantly monitors and logs performance, you can view the logs in a viewer. Here are 40 minutes of pretty much perfect guiding. The graph looks terrible, but actually virtually the whole trace is within +/1 1 arc-second. This is with my longer guidescope that has a smaller image scale which I use with a 1200mm scope which needs better guiding than my usual scope.

Ideally you want it to be no more than pixel scale of the scope, typically 1.7" per pixel with my usual setup.

You can do a lot of optimisiation if you really want, but there are more variables than you can shake a stick at from optical resolution of different scopes to the seeing quality.

Neil

[Michael GilliganParticipant @michaelgilligan61133]

Thanks for the clarification, Neil … That makes sense.

My misunderstanding … I latched onto your phrase 'starts in real time'

MichaelG.

[Author]

Posts