I understood the problem to be more about consistent distances of step rather than totally missing (slipped) steps. For sure, torque is a key to missing steps. I remember an industrial contoller where we started with a motor about 6cm diameter and 6cm long. It worked, but lost a few steps. We ended up with a motor about 10cm diameter and 30cm long, though it still slipped the odd step here and there! That was 30+ years ago, I believe things have moved on, in the control of the current.
What I'm not clear about is your vertical axis. How does that relate to your "2.5um steps"?
metallurgical objective vs biological objective
Moderators: rjlittlefield, ChrisR, Chris S., Pau
Hi Chris, wow, I have never seen those motors you mentioned.
Currently, I think slipage is easy to detect and somehow easy to fix. It is very difficult (for me) to determine what is the cause of those zigzag thing in the displacement graph.
I have my eyes on a Suruga motorized stage PMG615-R05AR, seems to be impressive thing with resolution of 1 micron at half step. Maybe I should use this as reference -- there are many components in the system -- the stage (ie, the lead screw), the motor, the driver, the motor mounting plates, the ZS. This thing can eliminate some of the uncertainties -- the lead screw, the mounting plate, etc.
My vertical motorized slider uses a no-brand translation stage, I do not think it has the precision of XEG60. But I think it is good enough for some low mag work, worst come to worst, I can use it as positioning slider. :-)
I bought two used XEG60s and motorized them.
Currently, I think slipage is easy to detect and somehow easy to fix. It is very difficult (for me) to determine what is the cause of those zigzag thing in the displacement graph.
I have my eyes on a Suruga motorized stage PMG615-R05AR, seems to be impressive thing with resolution of 1 micron at half step. Maybe I should use this as reference -- there are many components in the system -- the stage (ie, the lead screw), the motor, the driver, the motor mounting plates, the ZS. This thing can eliminate some of the uncertainties -- the lead screw, the mounting plate, etc.
My vertical motorized slider uses a no-brand translation stage, I do not think it has the precision of XEG60. But I think it is good enough for some low mag work, worst come to worst, I can use it as positioning slider. :-)
I bought two used XEG60s and motorized them.
Armed with this 20X objective, the tube lens is set to 300mm at f/5.6, so final mag factor is 33.33X (300/180*20). Here are results for 1.25um step.
I am not experienced (very little over 5x magnification level in fact) stacker and not knowing how ZS performs stacking, I think this is not bad, we can throw away "abnormal" images, right?
Setup: Nikon D5200, 70-300 Zoom at 300mm focused at infinity and f/5.6. Objective is Olympus ULWD 20X 0.4 Infinity WD=11mm, one small flash at 1/128 power level.
I am not experienced (very little over 5x magnification level in fact) stacker and not knowing how ZS performs stacking, I think this is not bad, we can throw away "abnormal" images, right?
Setup: Nikon D5200, 70-300 Zoom at 300mm focused at infinity and f/5.6. Objective is Olympus ULWD 20X 0.4 Infinity WD=11mm, one small flash at 1/128 power level.
Here are results when stepping 2.5um each sample. With 33.33X magnifying factor, the pink line gets pushed up and the "springiness" zigzag line stays within about the same band.
The new setup is more springy (springier?) than the old because the 70-300 zoom lens is not supported/fixed on anything while the old setup has the bellow bolted down. With new setup, if I set zoom on live view to max (10x), the image in LCD monitor would shift as I walk towards the setup (floor is solid concret!!!), so this probably explains spikes and valleys in the second graph with the new setup under 33.33x mag and denser sensor (6000 for Nikon D5200 vs 5184 for Canon 550D)
But, one thing is, since the pink line gets pushed up and the zigzag band stays roughly the same, it seems to measure smaller steps, such as 0.625um or even 0.3125um step, higher magnification is needed to have meaningful result. For example, with 100x objective or setup, a 0.625um step means 62.5um displacement on sensor, that is roughly 17 pixels, that should be enough to set the pink line (the average) higher above zero line and IF the zigzag band stays the same, the graph should look much better.
[edit] What it means is that the motorized XEG60 might perform much better than I thought, it is my testing setup that is not good enough for the purpose of testing. If the zigzag band is treated as noise, then we can raise the "signal" level up by using higher magnifying objective and get better signal to noise ratio (I have left college for too long, pardon my use of terms). [end of edit]
Geez, this means I need another objective
The new setup is more springy (springier?) than the old because the 70-300 zoom lens is not supported/fixed on anything while the old setup has the bellow bolted down. With new setup, if I set zoom on live view to max (10x), the image in LCD monitor would shift as I walk towards the setup (floor is solid concret!!!), so this probably explains spikes and valleys in the second graph with the new setup under 33.33x mag and denser sensor (6000 for Nikon D5200 vs 5184 for Canon 550D)
But, one thing is, since the pink line gets pushed up and the zigzag band stays roughly the same, it seems to measure smaller steps, such as 0.625um or even 0.3125um step, higher magnification is needed to have meaningful result. For example, with 100x objective or setup, a 0.625um step means 62.5um displacement on sensor, that is roughly 17 pixels, that should be enough to set the pink line (the average) higher above zero line and IF the zigzag band stays the same, the graph should look much better.
[edit] What it means is that the motorized XEG60 might perform much better than I thought, it is my testing setup that is not good enough for the purpose of testing. If the zigzag band is treated as noise, then we can raise the "signal" level up by using higher magnifying objective and get better signal to noise ratio (I have left college for too long, pardon my use of terms). [end of edit]
Geez, this means I need another objective
-
rjlittlefield
- Site Admin
- Posts: 23621
- Joined: Tue Aug 01, 2006 8:34 am
- Location: Richland, Washington State, USA
- Contact:
True for the most part. The stacking process will not be bothered by moves that are too small or too large, unless a move gets large enough that some plane is never captured in focus.mjkzz wrote: I think this is not bad, we can throw away "abnormal" images, right?
When some plane is not captured in focus, in the stacked image you'll see a line of blur along the contour line where that plane intersects the subject.
As you get near that point, making smaller and smaller steps, the problem with a positioner that makes uneven steps is that it will make you shoot more frames than you would need with a more precise one. Using a microscope focus block under motor control, I have no qualms about setting a 2 micron step when I really need 2 microns. Using my StackShot rail in that same situation, I would have to set a 1 micron step to be sure of getting no more than 2 microns, as described at http://www.photomacrography.net/forum/v ... hp?t=16323.
I think I read someplace that things will move more than expected.With new setup, if I set zoom on live view to max (10x), the image in LCD monitor would shift as I walk towards the setup (floor is solid concret!!!)
That sounds right. So, for the sake of measurement, you might think about making some sort of rigid support for the objective. A simple V-block, shoved firmly against its underside and clamped in place, should kill the springiness quite nicely. One way to check is to take a series of exposures with the block under test clamped so that it cannot move. If the instrumentation still says that it is moving, then something's wrong with the instrumentation.The new setup is more springy (springier?) than the old because the 70-300 zoom lens is not supported/fixed on anything while the old setup has the bellow bolted down.
--Rik
Thanks Rik.
I sort of reduced "springiness" by replacing the 70-200mm lens with simple M42 extension tubes (3 sets of 49mm each). I do get an image, but a bit blurred around the edges, very sharp in center region. Now I can adjust the subject (a peice of paper with gradient pattern printed by inkjet printer) with little shifts even when camera LCD is set to 10X, good enough for me to adjust focus continuously. I hope this is OK for the purpose of alignments.
The result is much better, the zigzag band is now confined to +-0.5um. I think I can further improve this by using a larger step motor, unfortunately, I have to wait two weeks for it.
Lesson learned here (for those who are interested) is that reducing vibration is a key factor.
Objective shown here is not the one used in test.
I sort of reduced "springiness" by replacing the 70-200mm lens with simple M42 extension tubes (3 sets of 49mm each). I do get an image, but a bit blurred around the edges, very sharp in center region. Now I can adjust the subject (a peice of paper with gradient pattern printed by inkjet printer) with little shifts even when camera LCD is set to 10X, good enough for me to adjust focus continuously. I hope this is OK for the purpose of alignments.
The result is much better, the zigzag band is now confined to +-0.5um. I think I can further improve this by using a larger step motor, unfortunately, I have to wait two weeks for it.
Lesson learned here (for those who are interested) is that reducing vibration is a key factor.
Objective shown here is not the one used in test.
Just to conclude (for those interested), with larger motor (3 times as powerful as last one used), just a little improvement -- much less big spikes and valleys (most likely caused by lesser powerful motor and not so precise motor mount), but the zigzag band is still within -0.5um to +0.5um.
I also tested ZS software and I am amazed by its capability, so I think at this level of precision, professional setup is needed to accurately pinpoint where the zigzag comes from -- the XEG60 itself, testing setup, or both, or more factores, I do not know.
But for now, it is good enough for me
I also tested ZS software and I am amazed by its capability, so I think at this level of precision, professional setup is needed to accurately pinpoint where the zigzag comes from -- the XEG60 itself, testing setup, or both, or more factores, I do not know.
But for now, it is good enough for me
Finally decided to give it a run again with much secure/stable setup. The result is very good, even with 1.25um step.
The whole setup is using 15mm rods rail often used in film making gigs, by adding a lens support on the 15mm rail, and bolting down the whole thing to one piece of aluminum extrusion plate, it is much better. In maximum zoomed live view mode, touching the camera no longer cause violent image shifts, well, touching the middle of lens still cause some jittery but much less and manageable.
In following graph, the unit is in micron (um), formula for it is 1000*SQRT(A1*A1+B1*B1)*23.5/33 where 23.5 is the width of sensor on D5200, 33 is the magnification factor (20*300/180)
0.001269769 3.4036251E-05
0.002861171 -2.6956585E-05
0.003890311 1.7528347E-05
0.005781626 -6.2319654E-05
0.007479830 -2.7553478E-05
0.009235298 -2.4051931E-05
0.010546076 -4.5175393E-05
0.012402408 -4.5589706E-05
0.014121618 -7.3476737E-06
0.015943923 -2.9292730E-05
0.017252449 5.0958817E-05
0.019318031 -9.7586117E-05
0.020972855 -3.7674428E-05
0.022517064 -1.6146390E-05
0.023961137 -9.5904841E-06
0.025854671 -1.7096625E-05
0.027479054 1.2748039E-05
0.029357117 -9.8652469E-05
0.030663587 -1.7518897E-05
0.032756240 -2.3454860E-05
0.034376536 2.9147099E-05
0.035943877 1.4738306E-05
0.037383169 9.3522909E-05
0.039136232 8.6168627E-05
0.040965830 9.7954217E-05
0.042713277 2.1378608E-04
0.044121627 5.1052026E-05
0.046141289 3.3194156E-05
0.047826911 7.4250074E-05
0.049545101 7.0466256E-06
0.051101971 -1.0622609E-04
0.053027180 5.9960817E-06
0.054792017 1.2595670E-05
0.056735440 2.1157598E-05
0.058375451 -1.6943997E-04
0.060382243 -1.6300811E-04
0.062105041 -1.1080598E-04
0.063952759 -1.9761482E-04
0.065561970 -3.7287227E-05
0.067564770 -2.4232405E-05
0.069442669 2.2828421E-06
0.071361765 6.5821907E-06
0.072976796 9.9166191E-05
0.075069963 1.5202981E-04
0.076582213 3.1968051E-04
0.078662523 2.6273432E-04
0.079969224 3.2184660E-04
0.081813500 1.9790991E-04
0.083616613 3.8317384E-04
0.085442950 3.8801384E-04
0.087047749 2.9197365E-04
0.089114562 2.5009182E-04
0.090836881 1.7842569E-04
0.092571673 2.5003608E-04
0.093850796 4.7586513E-04
0.095862500 2.6684737E-04
0.097454152 3.7283849E-04
0.099470783 2.5057792E-04
0.100875706 3.5023085E-04
0.102976067 3.9898120E-04
0.104738945 2.9143183E-04
0.106716492 2.6837605E-04
0.108123905 2.0833526E-04
0.110145101 1.4769863E-04
0.111790774 2.6612944E-04
0.113847074 8.3038937E-05
0.115316341 2.2684026E-04
0.117387987 1.5970713E-04
0.119070319 2.4293702E-04
0.120844080 1.1970177E-04
0.122257865 1.9645859E-04
0.124035351 2.2838205E-04
0.125812838 2.6030551E-04
0.127703991 1.8179151E-04
0.129081973 2.3206542E-04
0.131217354 1.5803798E-04
0.132783640 1.9095337E-04
0.134765784 2.1180944E-04
0.136166484 2.4782992E-04
0.137932431 2.9565809E-04
0.139571895 1.9140716E-04
0.141419678 1.7127579E-04
0.142911334 2.3837897E-04
0.145044480 2.3034433E-04
0.146807969 1.2131011E-04
0.148780388 -5.9754973E-05
0.150183256 4.2199765E-05
0.151986360 8.5094410E-05
0.153839583 -3.4548550E-06
0.155761004 2.3269841E-05
0.157190376 2.3990702E-05
0.159055730 4.0377129E-05
0.160815998 -1.7728126E-06
0.162725385 -6.0684269E-05
0.164012065 -8.8567797E-05
0.165897113 -1.6772193E-04
0.167762057 -4.5143623E-05
0.169403391 -4.9219218E-05
0.170924714 -1.0913128E-04
The whole setup is using 15mm rods rail often used in film making gigs, by adding a lens support on the 15mm rail, and bolting down the whole thing to one piece of aluminum extrusion plate, it is much better. In maximum zoomed live view mode, touching the camera no longer cause violent image shifts, well, touching the middle of lens still cause some jittery but much less and manageable.
In following graph, the unit is in micron (um), formula for it is 1000*SQRT(A1*A1+B1*B1)*23.5/33 where 23.5 is the width of sensor on D5200, 33 is the magnification factor (20*300/180)
0.001269769 3.4036251E-05
0.002861171 -2.6956585E-05
0.003890311 1.7528347E-05
0.005781626 -6.2319654E-05
0.007479830 -2.7553478E-05
0.009235298 -2.4051931E-05
0.010546076 -4.5175393E-05
0.012402408 -4.5589706E-05
0.014121618 -7.3476737E-06
0.015943923 -2.9292730E-05
0.017252449 5.0958817E-05
0.019318031 -9.7586117E-05
0.020972855 -3.7674428E-05
0.022517064 -1.6146390E-05
0.023961137 -9.5904841E-06
0.025854671 -1.7096625E-05
0.027479054 1.2748039E-05
0.029357117 -9.8652469E-05
0.030663587 -1.7518897E-05
0.032756240 -2.3454860E-05
0.034376536 2.9147099E-05
0.035943877 1.4738306E-05
0.037383169 9.3522909E-05
0.039136232 8.6168627E-05
0.040965830 9.7954217E-05
0.042713277 2.1378608E-04
0.044121627 5.1052026E-05
0.046141289 3.3194156E-05
0.047826911 7.4250074E-05
0.049545101 7.0466256E-06
0.051101971 -1.0622609E-04
0.053027180 5.9960817E-06
0.054792017 1.2595670E-05
0.056735440 2.1157598E-05
0.058375451 -1.6943997E-04
0.060382243 -1.6300811E-04
0.062105041 -1.1080598E-04
0.063952759 -1.9761482E-04
0.065561970 -3.7287227E-05
0.067564770 -2.4232405E-05
0.069442669 2.2828421E-06
0.071361765 6.5821907E-06
0.072976796 9.9166191E-05
0.075069963 1.5202981E-04
0.076582213 3.1968051E-04
0.078662523 2.6273432E-04
0.079969224 3.2184660E-04
0.081813500 1.9790991E-04
0.083616613 3.8317384E-04
0.085442950 3.8801384E-04
0.087047749 2.9197365E-04
0.089114562 2.5009182E-04
0.090836881 1.7842569E-04
0.092571673 2.5003608E-04
0.093850796 4.7586513E-04
0.095862500 2.6684737E-04
0.097454152 3.7283849E-04
0.099470783 2.5057792E-04
0.100875706 3.5023085E-04
0.102976067 3.9898120E-04
0.104738945 2.9143183E-04
0.106716492 2.6837605E-04
0.108123905 2.0833526E-04
0.110145101 1.4769863E-04
0.111790774 2.6612944E-04
0.113847074 8.3038937E-05
0.115316341 2.2684026E-04
0.117387987 1.5970713E-04
0.119070319 2.4293702E-04
0.120844080 1.1970177E-04
0.122257865 1.9645859E-04
0.124035351 2.2838205E-04
0.125812838 2.6030551E-04
0.127703991 1.8179151E-04
0.129081973 2.3206542E-04
0.131217354 1.5803798E-04
0.132783640 1.9095337E-04
0.134765784 2.1180944E-04
0.136166484 2.4782992E-04
0.137932431 2.9565809E-04
0.139571895 1.9140716E-04
0.141419678 1.7127579E-04
0.142911334 2.3837897E-04
0.145044480 2.3034433E-04
0.146807969 1.2131011E-04
0.148780388 -5.9754973E-05
0.150183256 4.2199765E-05
0.151986360 8.5094410E-05
0.153839583 -3.4548550E-06
0.155761004 2.3269841E-05
0.157190376 2.3990702E-05
0.159055730 4.0377129E-05
0.160815998 -1.7728126E-06
0.162725385 -6.0684269E-05
0.164012065 -8.8567797E-05
0.165897113 -1.6772193E-04
0.167762057 -4.5143623E-05
0.169403391 -4.9219218E-05
0.170924714 -1.0913128E-04
Interesting how your "rig" has progressed 
.
I don't think you've mentioned the stepper motor drive mode used by the Stackshot they call "High Precision"?
I can't remember details of exactly what it does (apart from making the motor sing), but it makes a big difference to the movements.
Something about the current decay IIRC.
I/we can find details here I'm sure - if you haven't covered it?
.I don't think you've mentioned the stepper motor drive mode used by the Stackshot they call "High Precision"?
I can't remember details of exactly what it does (apart from making the motor sing), but it makes a big difference to the movements.
Something about the current decay IIRC.
I/we can find details here I'm sure - if you haven't covered it?
Hi Chris,
hahaha, well, I just happen to have a set of 15mm rod rail system (used for film making rig), then bought some more and made it very stable. I am surprised that even if I touch the tube lens, it does not jitter, only shifts, at 30x optical and 10x digital (live view). Someday, I will post a compete picture of the setup. I find it very usefull as it has low profile, easy to change (you can move camera back and forth on the 15mm rail and lock it), there are commercial made lens support plates, etc.
I believe now the zigzag in the graph still has a lot of vibration components in it as I did similar analysis with still captures -- nothing moves, just take 100 pictures and let Zerene align them, get the output from Zerene's log. The still captures shows similar zigzag, meaning, even if there is nothing moving, the zigzag are there. I suspect that somewhere, some parts of camera bend and cause lens shifts (at micron level)
I do not have Stackshot, so I can not speak of it. But from what Rik did long back (3 years back), I suspect that they still need to do something about that cyclic "bump".
hahaha, well, I just happen to have a set of 15mm rod rail system (used for film making rig), then bought some more and made it very stable. I am surprised that even if I touch the tube lens, it does not jitter, only shifts, at 30x optical and 10x digital (live view). Someday, I will post a compete picture of the setup. I find it very usefull as it has low profile, easy to change (you can move camera back and forth on the 15mm rail and lock it), there are commercial made lens support plates, etc.
I believe now the zigzag in the graph still has a lot of vibration components in it as I did similar analysis with still captures -- nothing moves, just take 100 pictures and let Zerene align them, get the output from Zerene's log. The still captures shows similar zigzag, meaning, even if there is nothing moving, the zigzag are there. I suspect that somewhere, some parts of camera bend and cause lens shifts (at micron level)
I do not have Stackshot, so I can not speak of it. But from what Rik did long back (3 years back), I suspect that they still need to do something about that cyclic "bump".
metallurgical objective vs biological objective
Maybe it is worth trying adding a little retention spring on the carrier in order to put all the "natural gaps" on the same side to reduce the amplitude of them like if you are putting the rig in vertical position?
Regards
Pierre
Pierre