CNC resolution

[sinha_nsit]

I am thoroughly confused by various terms with perhaps the same meaning.
Pl let me know the diference between (if any):

• Resolution of a CNC machine
• Control resolution of a CNC machine
• Mechanical resolution of a CNC machine

Finally, are these same as the minimum possible displacement of an axis, i.e., the movement caused by one electrical pulse?

[Robin Hewitt]

You've got it, resolution is how far it moves with one step of a stepper motor, one pulse from a linear or rotary encoder, number of decimal places on a digital display.

Not to be confused with accuracy of course, that's completely different and much harder to achieve

[sinha_nsit]

You've got it, resolution is how far it moves with one step of a stepper motor, one pulse from a linear or rotary encoder, number of decimal places on a digital display.

Not to be confused with accuracy of course, that's completely different and much harder to achieve

So, are these terms (resolution, control resolution and mechanical resolution) refer to the same thing - RESOLUTION of the machine.

[LeeWay]

I'm not really sure that there is a difference in those terms. Perhaps though. Mechanical resolution could be how many turns it takes to rotate 1".
Control could be how many steps it takes to get this to turn enough revolutions to get 1". Could be stated in metric as well.

Together you would have the resolution of the machine.

[Al_The_Man]

Resolution of the machine is usually the smallest resolution the machine is capable of.
Control resolution, is the smallest commanded increment that can be entered.
Mechanical resolution is normally tied to the first.
The encoder or feedback resolution may be much smaller than the smallest commanded resolution possible.
Al.

[jetpig1]

So, are these terms (resolution, control resolution and mechanical resolution) refer to the same thing - RESOLUTION of the machine.

I believe this is over simplified.

Resolution: ACTUAL smallest unit of displacement effectively generated for a cnc from code to hardware.
Control resolution:
smallest unit signal possible for the control to effectively generate for intended physical movement (often not the same as actual)
Mechanical resolution:the achieveable positioning or displacement as a result of mechanical advantage or the physical drive components towards intended position or displacement.
Since gears and steppers vary in steps/rev, revs per inch,etc, mismatches show up where Control and Mechanical resolution yield different results.

Peace,
-Geo

[sinha_nsit]

These terms are really confusing and not properly used. I have seen the attached write-up which discusses these things. Taking a cue from that, and from some other sources, I conclude that the mechanical resolution is the movement of the slide per pulse, which is also referred to as the basic length unit (BLU). The control resolution depends on the bit storage capacity of the control and the total distance to be travelled by the slide. And the actual resolution of the machine would be the larger of the two values. Usually, the control resolution (based on, say, 32 bits) would be pretty fine. So, it is the BLU which is the actual resolution of the machine.

Another point:
There are two increment systems IS-B and IS-C on Fanuc 0i control. IS-C has finer resolution (0.0001 mm or 0.00001 inch), but shorter stroke length (one tenth). The manual talks about least input increment and least command increment.These two are different if the input is in mm and the machine tool has inch system. This has confused me. Does it mean that if we give an input of 0.001 (least input increment), the slide will move by 0.0001 inch (rediculous statement, but what is least command increment?).
Since the increment system talks about resolution as well as stroke length, it has got something to do with the bit storage capacity of the control. Things are not very clear. Will somebody please throw more light on it.

[Robin Hewitt]

All well and good but resolution doesn't really count for a hill of beans next to the accuracy figure.

Flip a switch on your stepper controller and you can double the resolution. Whoopy do. But as soon as resolution goes beyond positional accuracy for the slide you haven't improved things one bean, let alone a hill.

[Al_The_Man]

The least input increment is the smallest amount the control will accept as a command.
In other words, if the least input increment goes to 3 places of decimals in " then the least input increment is .001".
As an example, I recently converted a simple tube cut-off lathe where the requirement was a least input increment of 2 decimal places, (.01").
The accuracy desired was with ± .001".
The servo encoder with reduction gave a resolution of 64,000 counts/inch, the control could position the servo within 1 count, so my potential resolution would be .0000156.
However, due to mechanical error that consisted of ±.0005", my mechanical tolerance was within the required 1 thou" customers spec.
My least input increment was 10thou"
Al.

[sinha_nsit]

So, control resolution will depend on encoder resolution AND its bit storage capacity - the larger of the two. This is however theoretical resolution which will usually not be practically attainable. Here comes the mechanical resolution in picture, which depends on the quality of mechanical systems. Then, in the file I attached with the previous reply, the distance between the addressable points should be called mechanical resolution. It cannot be control resolution. Otherwise also, repeatability of control system is perfect, i.e., 100%. It will always position the servo, for a particular input, in the same position, everytime. But, even standard text books call this distance control resolution. In fact, many also use the term control points in place of addressable points.

It is because of mechanical resolution that the control gives two options for the least input increment. Select the one which is practically possible on a particular machine. The control must be using some internal conversion to use a round figure for least input increment (somebody may throw more light on it).

Finally, what is important is the mechanical resolution, because that is the actual output. The control resolution is likely to be much finer, but of no use.

This is what I conclude. Please correct me if I am wrong.

[Robin Hewitt]

what is important is the mechanical resolution, because that is the actual output. The control resolution is likely to be much finer, but of no use.

It depends on what you use for positioning. If you depend on accurately turning a screw for position then positional accuracy depends on the precision of your screw and eliminating flex.

You can fit linear encoders to the slides and position to them. they will tell you where you are regardless of screw accuracy. Suddenly any old screw will do, so long as the tool pressure can't move the table.

I'm working on a box of tricks to position using cheap chinese digital scales, they resolve 20480 lines per inch but 3 counts of jitter means dumping the bottom two bits giving a useable 5120, about 0.2 thou. Their big problem is they only report about 50 times a second and the data becomes stale very quickly. The machine inevitably gets ahead of the scale and depends on screw accuracy between reports. However the error is not cumulative so I think it is worth doing.