View Full Version : Question on Micro step resolution?

23-03-2010, 08:39 PM
Hi. not sure if this is the best place to ask but here goes.

I've just started setting up Mach3 with the PM542 drivers and I'm a bit confused. mach recommends to use 10 micro steps (2000 steps per motor revolution) which gives a nice smooth drive and seems to be stronger than 1/2 or 1/4 steps. But this seems to go against everthing I've read and always thought microsteps reduced power?

The other thing is that Mach talks about using micro steps to change resolution and this seems to be true as the more microsteps i use the speed drops whilst jogging (using the same velocity settings) I'm sure I read that mircostepping dosnt affect resolution as the driver just uses the microsteps to transition smoothly from one step to the other.

So im not sure what to do as its tempting to use a higher micro step to keep it running smooth but if this is using up more steps then is there is the risk of hitting the max processor speed and therfore reduce the top speed?

Is there a standard-ish settup that everyone uses? :smile:

16-04-2010, 09:06 PM
I am also using Mach 3 and as far as can tell the resolution is set by the by the microsteps, i.e. the motor can be stopped at any of the fractional steps as set on your driver board not just whole steps. So if you select 16 microsteps and you have a 1.8 degree stepper your resolution will be 3200 steps per rev or 0.1125 degrees and you can stop on any one of these. Note there will be a tolerance on these steps in practice. If you have a 2.5 mm pitch lead screw the resolution per step or linear resolution will be 0.78125 um per step, i.e. very fine. In mach 3 you can configure the pulse rate as 25000, 35000 and 45000 pps and typically for a 1 GHz processor I believe you can get away with 35000 pulses per second which is one of he setting options. Assuming your stepper driver can handle this pulse rate then your maximum linear speed will be 27.34 mm per second, i.e just over 1" per second. If you have a fast PC processor you may be able to get away with 45000 pps and hence increase your linear speed to 35mm/s. I would start off at the lower pulse rate and work up. to see when you start to lose pulses. Clearly if you reduce your microsteps to 8 then you will halve the step resolution but double the maximum speed. The main reason for fine microstepping is to get smoother operation and stop judder not necessarily to achieve higher resolution accuracy.

John S
16-04-2010, 09:39 PM
Going above 8 or 10 micro-steps is a wast of time - literally, as you just reduce speed and torque available.
Micro-stepping does increase smoothness but there are limits.
Most machines just cannot get this resolution mechanically anyway and all that happens is if you command 1 microstep nothing happens, 2, 3 4 and possibly on the later ones it will jump to make up as it overcomes stiction.

Unless you have a really well built machine with absolutly no backlash and are in a temperature controlled enviroment trying to achive fine tolerances just will not happen.

Welcome to the real world.

17-04-2010, 09:06 PM
microstepping will only give you smoother motion and less noise and will NOT increase accuracy.
A stepper motor has a tolerance of +/-5% of 1.8 deg (on a 1.8 degree motor) per step and this is only on the full step position and microstepping between steps will in most cases not be linear, so dont expect to get good accuracy or repeatability between the full step positions.

Also higher microstepping resolutions will also help to eliminate the zero position glitch that is very prominent with drivers that are limited to lower resolutions or morph to a full step driver after a certain speed.
While this problem will not really effect applications like CNC machines it is a big problem in applications where very smooth motion is required.

12-02-2013, 03:08 AM
Hi Gary ( Zapp ) Will ex Emach here,