Stepper Drive Rotary Table Results Using Mode 1 of the Peter Norberg Consulting
Bill Chennault
Posts: 1,198
All-
In the original test (Stepper Drive Rotary Table Results), I used the Peter Norberg Consulting BS0710-USB in Mode 0. The manual for the stepper board says this about Mode 0 . . .
Mode 0 and Mode 1 Definitions
'0: Single winding full step mode: Exactly one winding will be on at at time, and will be on at the selected current for the motor. The "real" physical motor position (in full step units) therefore only updates once every 8 microsteps; thus the "full step" location will be in the (microstep location)/8, dropping the fractional part.'
And the manual has this to say about Mode 1 . . .
'1: Half step mode: Alternates between having one and two windings on at a time, thus causing the torque to vary at the half-step locations. The "real" physical locations will be at the half-step values, and hence the motor will "move" once every 3 microsteps. The "full step" location will be the (microstep location/8), with fractions·of 0 to 3/8 mapping into fractional location 0, and 4/8 through 7/8 mapping into fractional location 0.5.'
The data reported by·the·Mode 1·test, which is EXACTLY the same test as the previous test in hardware configuration and program--the only thing being changed is from Mode 0 to Mode 1 via a SEROUT command argument change--produced different results, as you might expect. However, I do not know enough about stepper controllers and stepper motors to explain why. I am hoping one of you will tell me.
Here are the two tests side by side for ease of comparison purposes . . .
Mode 0············ Mode 1
n:········ 100····· ·100
Mean:····· 0.026"··· 0.012"
Median:··· 0.025"··· 0.009"
Mode:····· 0.026···· 0.010
Extreme
Variation: 0.01"···· 0.074"
To me, it looks like the·Mode 1 test has a mean repeatibility about twice as good as the Mode 0 test. Does the explanation above regarding the differences between the two modes explain this difference?
I have no explanation for the huge difference in the median score between the Mode 0 and Mode 1 tests.
And, again, to me, it looks like the·Mode 1 test has a Mode about·two and half times·"better" (?) than that of·the Mode 0 test. In other words, more of the scores in the Mode 1 test were the same than the scores in the Mode 0 test. (I think this is good.) Does the explanation above regarding the differences between the two modes explain this difference?
If you have time, I would like to know what you think?
Thanks.
--Bill
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
You are what you write.
In the original test (Stepper Drive Rotary Table Results), I used the Peter Norberg Consulting BS0710-USB in Mode 0. The manual for the stepper board says this about Mode 0 . . .
Mode 0 and Mode 1 Definitions
'0: Single winding full step mode: Exactly one winding will be on at at time, and will be on at the selected current for the motor. The "real" physical motor position (in full step units) therefore only updates once every 8 microsteps; thus the "full step" location will be in the (microstep location)/8, dropping the fractional part.'
And the manual has this to say about Mode 1 . . .
'1: Half step mode: Alternates between having one and two windings on at a time, thus causing the torque to vary at the half-step locations. The "real" physical locations will be at the half-step values, and hence the motor will "move" once every 3 microsteps. The "full step" location will be the (microstep location/8), with fractions·of 0 to 3/8 mapping into fractional location 0, and 4/8 through 7/8 mapping into fractional location 0.5.'
The data reported by·the·Mode 1·test, which is EXACTLY the same test as the previous test in hardware configuration and program--the only thing being changed is from Mode 0 to Mode 1 via a SEROUT command argument change--produced different results, as you might expect. However, I do not know enough about stepper controllers and stepper motors to explain why. I am hoping one of you will tell me.
Here are the two tests side by side for ease of comparison purposes . . .
Mode 0············ Mode 1
n:········ 100····· ·100
Mean:····· 0.026"··· 0.012"
Median:··· 0.025"··· 0.009"
Mode:····· 0.026···· 0.010
Extreme
Variation: 0.01"···· 0.074"
To me, it looks like the·Mode 1 test has a mean repeatibility about twice as good as the Mode 0 test. Does the explanation above regarding the differences between the two modes explain this difference?
I have no explanation for the huge difference in the median score between the Mode 0 and Mode 1 tests.
And, again, to me, it looks like the·Mode 1 test has a Mode about·two and half times·"better" (?) than that of·the Mode 0 test. In other words, more of the scores in the Mode 1 test were the same than the scores in the Mode 0 test. (I think this is good.) Does the explanation above regarding the differences between the two modes explain this difference?
If you have time, I would like to know what you think?
Thanks.
--Bill
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
You are what you write.
Comments
I understood from previous reading thah microstepping was not as accurate as full stepping especiallly under load conditions.
I wonder how much load you have on the table as well. A stepper requires a load to be accurate as well and I would think you need to test at 50% and 100% loads to see if you have the tolerance you require.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Think Inside the box first and if that doesn't work..
Re-arrange what's inside the box then...
Think outside the BOX!
Here's a link to a page about a new line of Slo-Syn stepper motors...
www.motionandcontrols.com/electrosales/warner/New_Products/kml.htm
Slo-Syn are excellent, highly regarded stepper motors. If you look at the specs you'll see this line of info:
+/-2% typical step accuracy, noncumulative
Yes, that's right. The motor itself is only guaranteed to be accurate to plus or minus two percent of a single, full step. Micro-stepping doesn't increase accuracy, just resolution. Without closed loop feedback (an encoder typically), you won't get any better positioning accuracy than what the motor is capable of.
This link explains it in a bit more detail:
www.faulhaber-group.com/n391003/n.html
I assume you're running open loop, but I may be wrong. If you are, then your accuracy is only guaranteed to be plus or minus 2% of the motors full step, and 3% is actually a lot of common. The cheaper the motor, the worse the accuracy. Micro-stepping has many benefits, but increasing accuracy is not one of them.
Thanks,
PeterM
Thanks for the link and the advice. I have been coming to some of your same conclusions the empirical way (read, the hard way). Since my lengthy experiments have little variation, I believe the cheap stepper is simply ignoring a lot of the commands.
The mass it is moving is fairly negligble. It is just my magnetic clamp dial indicator. However, the results I am getting tell me that with the proper gear-train (I am using 3:1), even my cheap stepper might be useable in a dual drive configuration in a small robot. Especially, if every once in a while you stopped and took care of accumulated error . . . somehow.
I am nowhere near finished pla . . . testing! [noparse]:)[/noparse]
--Bill
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
You are what you write.