Machsupport Forum
Mach Discussion => General Mach Discussion => Topic started by: pgsierra on October 06, 2008, 11:24:44 AM
-
Hi Folks,
I have two different types of steppers on my machine.
Two are identical ( Motion Technologies N20 3451) and the literature says they have 200 Steps per revolution. The third Motor I bought off of EBAY ( it is a RHT34 - 1200 - Rex Technologies) and am not aware of the steps per revolution, but have sent the manufacturer an email requesting it.
I am using 2 start 1/2" - 10 lead screws, (5 turns per inch) with Dumpster CNC anti backlash nuts, and 201 Geckos ( which I bought new so I assume they were factory set at 10 Micro step) .
If I use the utility in Mach to determine steps for an inch, it says I need to set the steps to 9753.147828 for the two MT motors. If I measure the distance on the table, It does seem close, but It's not exact.
The numbers seem huge but If I read other posts right, I should be able to calculate it on my own like this: 200( steps per rev) X 10(Drive Microsteps) = 2000 X 5 ( to get to the inch) = 10,000... right?
The other motor seems to be around 1024... If anyone has any data on that motor it would be appreciated!
Thanks in Advance,
Pete
-
Pete,
The correct steps/unit is 10,000/
Regards,
Ray L.
-
You were right in your calculations.
Steps per inch should be calculated, not measured, unless you have a completely bastard leadscrew (bastard used iin it's proper context).
10,000 is not a lot - mine is 60,000 (3 to 1 gearing and a 10 tpi leadscrew) so don't worry, Mach can handle it.
Are you saying your third motor needs 1024 pulses per rev?? Seems a bit over the top. Motors are fairly standard and 200 is the common one.
You can try by connecting your motor and put in 200 as your steps per inch. Fasten some sort of marker on your shaft and tell mach to move one inch - see if the motor turns one rev. Get your speed fairly high to do this. Or use 2000 steps per inch and see if it turns 10 times.
-
Although 10,000 is the calculated amount, you may need to adjust the number up or down a little bit. Typically, the acme screws you're using have a tolerance of ±.009 per foot, so are rarely actually 10,000 steps per inch. Also, the steps per inch can vary along the length of the screw, so you need to check in different areas of the machine and pick a number that works best overall.
-
You are joking?? I can only say that if these screws are sold for use in lathes etc, then someone needs prosecuting. I will accept some tolerance, but as for saying the steps per inch vary along the length of the screw that is rediculous.
You cannot build a machine like that. If a screw has 10 threads per inch, it has 10 threads per inch. There will be play between the threads, but it is not the job of the steps per inch to take this up. For that you need backlash compensation. I will accept that backlash can changes, eg between a worn part of the thread and a relatively unused part.
If you do not set your machine up as calculated, then you are setting up an inherent defect into the system, one for which you cannot then compensate, and which will affect ALL movements on the machine, not just a few.
-
Mach use to have it and may still, not sure. Most all industrial machines I work on do. It's screw mapping. This is for the very reason Ger21 mentioned. Most screws, even the high end high precision ground ones have slight variations along the length of the screw. If Mach still has screw mapping, I have no idea how well it works.
Brett
-
Although 10,000 is the calculated amount, you may need to adjust the number up or down a little bit. Typically, the acme screws you're using have a tolerance of ±.009 per foot, so are rarely actually 10,000 steps per inch. Also, the steps per inch can vary along the length of the screw, so you need to check in different areas of the machine and pick a number that works best overall.
That seems to me like really bad advice. *Every* screw, even the best ground ballscrews, have some lead error. The only difference is the magnitude of that error. Rolled ballscrews are typically 0.004"/inch maximum error. Ground ballscrews can get down to 0.0005"/inch. However, the *average* pitch should be very, very close to the spec over any reasonable distance. If you adjust steps/unit, then you're forcing the error to be less *only* at the one place on the screw where you measured it, which means you've *increased* the error pretty much everywhere else. Not a good idea. The only time you should compensate by fudging steps/unit is if you have a screw that isn't really what you think it is. For example, perhaps your "10 tpi" screw is really a 2.5mm lead metric screw, in which case you'd need to adjust steps/unit by 25/25.4 to compensate. And if you have Acme screws, this is all a little academic anyway, since your biggest problem will be backlash, not lead accuracy.
Regards,
Ray L.
-
Ooops! Those numbers should be 0.004"/FOOT, and 0.0005"/FOOT. BTW - Nook has some REALLY nice rolled screws with 0.001"/foot lead error, for barely more money than the 0.004"/foot ones.
-
If you only use the calculated value, then why does Mach3 have a function to calculate your steps/unit based on distance moved?
If the screws specs are ±.009, then it could be off by up to .009 per foot. Most likely it'll be better, but imo it's a good idea to check that your getting the actual distance your trying to get. If not, you adjust your steps/unit.
-
Screw maping is still in Mach under "FUNCTION CONFIG" but don't remember seeing any info on it or a posting about it.
If you have a "sweet spot" along the axis, and, you do a lot work in that vicinity, it may make sense to adjust steps / unit
for that area and live with the tolerence outside of it.
As far as the +- spec for the ball screws, you MAY find that it will gain or decrease ie; for +-.003"/ft. ball screw on my mills X axis it will actually gain very accurately and linear in fashion .003 in a foot and will also linearly return to the start point, and it will vary differently in the other direction from that start point.....never mover than .003" over any 1 foot section of it. Spec's don't always say what you think they say.
So would you "fudge the steps", base it on the 12" run in one direction, or about the some point of the overall profile.
Most users will never really know what they have.
I totaly agree with Jimpinder that they should be calculated. I also totaly agree with Gerry on the little refinement part
of the steps.
In Neutral Territory ;D
RICH
-
Thanks for the help everyone... that seems to do the trick.
10,000 is 1 inch on the nose.
Is there a reason why after the steps per inch are saved ( in this case 10,000) Mach would default back to 2000 when the application is started again?
Have done it several times and am confident that I am saving each axis.
Pete
-
Make sure your xml is not read only, shouldnt be but can happen if you transfer the xml from a CD.
If thats not the case then try the Save Config option from the config menu.
Hood
-
Thanks Hood!
Took a look at the XML file and it doesn't seem to be read only in the properties tab. I'll be sure to select save settings ( near the bottom) from the configuration drop down menu, when I set this up again.
Pete
-
Yep thats the one, save settings, not save config, been a while since I looked :(
Hood
-
Remember, when a screw has a spec of .004" per foot it means the real accuracy is LESS than that amount. If it was more it would be out of spec and presumably the manufacturer would not ship it.
I am curious what instrument you have in a shop that can measure to an accuracy of .004 per foot? maybe a 10" micrometer? How would you set it up to make a table movement measurement?
My point is guys get all hung up over screw accuracy, when lots of other factors have as big, or bigger influence.
The 'calculate steps per in' button was added because so many people seem unable to understand the math involved in the calculation.
-
r
I am curious what instrument you have in a shop that can measure to an accuracy of .004 per foot? maybe a 10" micrometer? How would you set it up to make a table movement measurement?
Engineers use slip gauges for things like that with a dti. I could easy do 10" with my gauges. ;D
Phil_H
-
Ron,
I use either a jig transit, alignment telescope, optical level, which can focus as close as 7" ( 20x and better) from the target or out to infinity. The instruments use an optical micrometer and give direct repeatable readout down to .0002" inches at 12". Perpendicularity to the bed axis is done using autoreflection techniques. Mirrors used are optically flat.The target is a "calibrated" scale which when profiled / calibrated for each spacing have a very accurate profile. Thus i can easily see backlash immediately and measure it or even profile a ball screw. For that matter you can measure spindle run out. On a lathe you can check the carriage travel relationship to the spindle or set a very precise angle.......and on and on. But being practical about things you are absolutely right and sometimes we are happy the stuff just bolts together when done.
Could probably measure a hair from your head 20 to 50 feet away and i am just an amateur. This stuff is not cheap....the receipt for the Kolmorgan alignment telescope when bought 15 years ago was $14.5K, no i won't tell you what i paid for it!
I thought everybody had stuff like this. ???
RICH
-
Could probably measure a hair from your head 20 to 50 feet away and i am just an amateur.
Must be a really good tool if you could do that for a hair on my head ;D
Hood
-
When man was made, God put hair on only on the ugly ones, a few didn't get any !;D ;D ;) ;D
RICH