Steppa motor and rail question

[UncleChip]

Still learning so please be paciente with me,

I have a rail that travels 1mm with each rotation of the screw, I have a stepping motor that is 1.8degree,
Am I correct in my workings?
360 % 1.8 = 200 per turn
1mm or 1000 % 200 = 5um

So 5um would be the minimum distance this combo would do if I set it up with a controller

On the MJkZZ website where I am looking at a controller, their prebuilt rails have a 2mm screw yet boast a distance of 0.078125UM (not mm) I am wondering where the discrepancy is,

[Pau]

This is because micro-stepping, you only made the calculation with full steps

[chris_ma]

I have a rail that travels 1mm with each rotation of the screw, I have a stepping motor that is 1.8degree,
Am I correct in my workings?
360 % 1.8 = 200 per turn
1mm or 1000 % 200 = 5um

So 5um would be the minimum distance this combo would do if I set it up with a controller

On the MJkZZ website where I am looking at a controller, their prebuilt rails have a 2mm screw yet boast a distance of 0.078125UM (not mm) I am wondering where the discrepancy is,

yes, your calculations are correct.
the MJkZZ website numbers must been calculated with microstepping on the stepper motor. either a 1.8deg motor with 128 microsteps or a 0.9deg motor with 64microsteps.
from my research it seems that microstepping only brings you so far (depending on the motor and the driver), you definitely will not get to 0.078125um accuracy (the 6 decimals are a clear indication that it's a mathematical concept and not a physical measurement). in fact it's nearly certain that you'll not get 0.1um accuracy with this setup. something like 1um could be possible though.

I wanted to do a real world measurement about microstepping with my setup for a long time, but always delayed it because of other matters were of higher priority.

[enricosavazzi]

At the same amount of steps or microsteps, microscope focusers with fine focus (typically 100 or 200 μm per turn) have an obvious precision advantage over linear rails with 1,000 μm travel per turn. If you need reliable and repeatable 1 μm (or slightly less) travel increments, microscope stands/focusers are a better choice than a synchronous motor controller with a high number of microsteps.

[UncleChip]

Thanks for this, didn’t know about microstepping, ( did say I was a noob)

At the moment I am only up to 5x, my next step will be 10x so I think this combo should do fine,

[UncleChip]

So
If I changed the motor to a 400step I would have 2.5um without microstepping?
Looking at the objectives I can’t see I would need to go smaller,

Is there any reason why the motor is directly connected and nobody uses cogs and timing belts? With my existing motor I could use different size cogs to achieve a smaller UM,

[lothman]

So
If I changed the motor to a 400step I would have 2.5um without microstepping?
Looking at the objectives I can’t see I would need to go smaller,

correct

Is there any reason why the motor is directly connected and nobody uses cogs and timing belts? With my existing motor I could use different size cogs to achieve a smaller UM,

this would add elasticity in the drive train and may lead to a jumping slider. So no consecutive 5x 0.2µm but 0/0/0/0/1 µm. That is the reason why drive units with repeatable steps below 1µm are so expensive (needs high end mechanics) and not only a high transmission between motor and spindle.

[UncleChip]

I presume the intended application will make a difference to the required accuracy, a hobbyist focus stacking rail for doing bugs needs enough accuracy to produce accurate stacks, I presume there are many areas that require much higher repeatability
The rail I got looks the same as the ones mkjzz are using on their ultra series, they claim up to 100x,I will get 2 more of these for the other axis

[rjlittlefield]

The rail I got looks the same as the ones mkjzz are using on their ultra series, they claim up to 100x,I will get 2 more of these for the other axis

I have no trouble believing 100X. That probably means with Mitutoyo M Plan Apo 100X NA 0.70, for which the best available formula says 0.96 microns DOF. (The number 0.6 microns is frequently quoted, but that is a single-sided value calculated with a less accurate formula.) Even at NA 0.80, the DOF is still about 0.69 microns. These are diffraction-limited numbers, so if the rail can do even this well, no focus banding will be evident in the stacked image.

--Rik

[Adalbert]

The rail I got looks the same as the ones mkjzz are using on their ultra series, they claim up to 100x,I will get 2 more of these for the other axis

This is probably the THK KR.
I also use one of them, the KR 2001A.
It runs in my setup with a belt drive and a ratio of 5:1.
That gives quite small steps e.g. 0.00033 mm.
viewtopic.php?p=252257#p252257
The deviations are still within limits :-)

[UncleChip]

Yes it’s the THK KR, it was listed as 20A but looks identical,
The belt drive was my consideration for the Z axes as I wanted to minimize the vertical height so it doesn’t interfere with the lights, I am also considering using 2 of these actuators for the Z axis, as I want a bigger specimen platform, the belt seams a logical way to turn both from the same motor, so good to see others using it,

[chris_ma]

It runs in my setup with a belt drive and a ratio of 5:1.
That gives quite small steps e.g. 0.00033 mm.
viewtopic.php?p=252257#p252257
The deviations are still within limits :-)

so that's 0.33um +/- 0.1um, or if you'd mount a laser you could differentiate the wavelegnt of a red and blue laser - quite impressive!

what I don't quite get is how you arive at 0.33um?
with a 1mm lead screw, a 1.8deg motor, 256microsteps and a 5:1 transmission I get
1000 / 200 / 256 / 5
= 0,00390625

[Adalbert]

It runs in my setup with a belt drive and a ratio of 5:1.
That gives quite small steps e.g. 0.00033 mm.
viewtopic.php?p=252257#p252257
The deviations are still within limits :-)

so that's 0.33um +/- 0.1um, or if you'd mount a laser you could differentiate the wavelegnt of a red and blue laser - quite impressive!

what I don't quite get is how you arive at 0.33um?
with a 1mm lead screw, a 1.8deg motor, 256microsteps and a 5:1 transmission I get
1000 / 200 / 256 / 5
= 0,00390625

You have already calculated the smallest step in microns, so 1000 instead of 1 on the left.
By the way, I use NEMA 17 0.9 degrees => 400 instead of 200

And I need 0.33um for the NA=0.75 and 3 shots per DOF, therefore the test and example for this step size.

[chris_ma]

what I don't quite get is how you arive at 0.33um?
with a 1mm lead screw, a 1.8deg motor, 256microsteps and a 5:1 transmission I get
1000 / 200 / 256 / 5
= 0,00390625

You have already calculated the smallest step in microns, so 1000 instead of 1 on the left.

sorry, what I meant that I don't see how to arrive at 0,33um with a single step.

but I see now that

1mm / 400steps / 256microsteps / 5 pulley
= 0,000001953125mm = 0.001953125um

means that you're running 170microsteps to get one 0.33um step between images, right?

that would also give a good estimate on the accuracy of microstepping, since you have about 0,1um of jitter, it would mean that microstepping is useful to about 32 or 64 microsteps

[Adalbert]

Hello Chris,
Yes, the deviations depend on the number of microsteps.
I have executed such tests where you could see
the periodic behaviour very well.
Best, ADi