Machsupport Forum
Mach Discussion => General Mach Discussion => Topic started by: SpookyDad on June 01, 2010, 05:25:49 PM
-
I am extremely frustrated! I am running a job that I have done several times before. I upgraded the computer and replaced the Xylotex controller since running this code. I tuned the motors by measurement and it is off by a couple of thou but that is as close as I can measure.
This is what is happening:
(http://i13.photobucket.com/albums/a273/SpookyDad/pictures%20for%20forums/runoff1.jpg)
(http://i13.photobucket.com/albums/a273/SpookyDad/pictures%20for%20forums/runoff2.jpg)
It has happened on another job over an hour into the run. It seems to happen 7/8 ths of the way through the run. THis run has 842 lines of code but the other one has 4500. I thought it might be a memory problem but this run is small.
I am running XP Service pak 2
-
Do the DROs and Toolpath show the position of the code or the tool when it goes wrong?
Hood
-
Is your xylotex getting hot? Looks like one axis may be stopping?
-
I put the Xylotex in a separate case with a large fan blowing directly on to it.
I didn't pay attention to the exact line where it crashed but when I send it back to 0,0 it is several inches off
Neil
-
Definitely sounds like an axis stalling then so as Ger says its either the drive or maybe the axis binding in some way.
Hood
-
If it has been up and down the table over 20 times in this run, is it likely that it would be a stalling axis?
I should really optimize the code but I just wanted to run the same job again.
Neil
-
I just realized it has occured when doing a circular cut out or pocket.
The circles are never perfectly round.
Neil
-
IF you will notice there was a total SHIFT in the offset as when it shifted it shifted on the Center point of an arc/circle. AND stayed on the exact center of that arc/circle. AND still cut a good arc/circle I have seen it shift IN and I have seen it shift OUT.
THAT is not lost steps or NOISE as if it was the arc would have been offset by the amount of the missing steps AND it would have still cut the same radius as the code here it did none of the above.
-
And did your DROs show that Terry?
Hood
-
They certainly did(;-) it was following MACH direction in the dros.
BUT 'll keep quite and let yall handle that, sorry
-
IF you will notice there was a total SHIFT in the offset as when it shifted it shifted on the Center point of an arc/circle. AND stayed on the exact center of that arc/circle. AND still cut a good arc/circle I have seen it shift IN and I have seen it shift OUT.
THAT is not lost steps or NOISE as if it was the arc would have been offset by the amount of the missing steps AND it would have still cut the same radius as the code here it did none of the above.
I realize the WORDS are English but this make no sense at all to me. Could I get a "G-Code for Dummies" explanation of this.
-
Do you reckon this the same problem as you were always talking about on Yahoo? Do you still get it?
Hood
-
IF you will notice there was a total SHIFT in the offset as when it shifted it shifted on the Center point of an arc/circle. AND stayed on the exact center of that arc/circle. AND still cut a good arc/circle I have seen it shift IN and I have seen it shift OUT.
THAT is not lost steps or NOISE as if it was the arc would have been offset by the amount of the missing steps AND it would have still cut the same radius as the code here it did none of the above.
I realize the WORDS are English but this make no sense at all to me. Could I get a "G-Code for Dummies" explanation of this.
What Terry is saying is if you look at the pics you see the axis moves out one way then back the other, if it was lost steps then it would tend to just go one way, well that is if it was exactly on the apex.
Hood
-
How does that help me with my problem?
Is it some sort of computer setting?
The computer is a dedicated machine with no internet or network connections. I stripped out all the extra windows crap. There is no anti-virus software running.
Neil
-
If its the same problem that Terry gets then I dont think its ever been found and if I remember the only way he could get things back up and running was a fresh install of Mach, but Terry will advise.
If however the DROs dont show this movement then it is not likely to be the same and although Terry says it cant be lost steps because it goes one way then the next that would only be true if it was at the apex, if it was not then it could well be lost steps.
So do you think the line goes off directly at the apex, from your pic it looks to me that it is before. Also looks to me like the line going up seems to be at the max height that the circle would be, so that does suggest lost steps to me.
Hood
-
So do you think the line goes off directly at the apex, from your pic it looks to me that it is before. Also looks to me like the line going up seems to be at the max height that the circle would be, so that does suggest lost steps to me.
Hood
It does appear that it starts to go wrong just before the apex. The Y axis seems to be functioning but the X axis is falling behind.
I just did the driver test and it is flat line up through 75mHz. At 100 it gets some minor noise. Is there any advantage to upping the Kernal speed?
Neil
-
Neil,
I haven't had that happen to me in a long time. My cure back whenever was to slow the velocity down. Conservative here at 30 to 40 % of max velocity.
When steps are missed the steppers may just go off with a mind of their own. If the motors are getting hot then they can start to act funny and do strange things, same for a drive that is very close to it's max amp rating for the stepper being used.
Sorry, i don't have an exact answer for you.
FWIW...
RICH
-
It looks like the X axis practically stopped. Do you have any resonance on your machine, or dampers on the motors? My xylotex was practically unusable until I made dampers for the motors. Now it's rock solid.
As for binding, it can occur at any time. Do you lubricate your screws? They can get dry and start to bind. Also, if you see this lowering the acceleration 25% or so, and see if it helps.
-
The max speed is set to 60 ipm on Xylotex 425 motors. The spindle doesn't appear to be moving all that fast when it goes nuts.
I don't have any dampers on my motors but have been planning on building some. I cut back the speed when I was having resonance issues before. I have also added an optp-isolater board to cut out noise.
Every time I fix one issue, another pops up and I am beginning to hate this machine.
Neil
-
What kind of screws do you have, and how fast are you going when it happened? The Xylotex 425 oz motors will not spin very fast. Coupled with a poor choice in screws, it could be that you're trying to go faster than the machine is capable of, resulting in a stalled axis.
-
I have 1/2-10 Acme screws and Xylotex 269 motors. (I made a mistake on the 425's)
I am not sure how fast I was going but the max speed is set to 60 ipm
Neil
-
I've been thinking about this, and I'm betting it is stalling, but it may or may not be for mechanical reasons. I will say though, 60ipm with 10tpi screws is going to be turning those motors at the upper end of their torque curve, and slowing down may be your only option. That aside, I do have a couple observations.
If Y is vertical, and X is horizontal;
The little bump at the top would be where X stopped moving, and Y attempted to continue moving out to the edge of the circle. The lower mark would be where X started moving again, attempting to finish the right side of the circle. You can see that X was *trying* to move, as it does angle a bit to the right. Once it reaches the lower mark, the motor started moving again.
I know how frustrating this can be. It can seem like everything is sailing along smoothly, and out of the blue, stalling happens. I'd like to refer you to my stalling post from months ago, there may be some things in there for you to try, and there were lots of helpful posts from folks trying to help me. If you want to see what worked for me, check page 4, but don't skip what was said on the previous pages.
http://www.machsupport.com/forum/index.php/topic,9943.0.html (http://www.machsupport.com/forum/index.php/topic,9943.0.html)
Good luck!
(http://i26.photobucket.com/albums/c137/Guitar_Guy0101/help2.jpg)
-
that looks like one axis just quits for a while then starts back.you might want to check your wires from the motors back to the drivers.
everything might be ok except for one motor is not getting the signal all the time.
bill
-
that looks like one axis just quits for a while then starts back.you might want to check your wires from the motors back to the drivers.
everything might be ok except for one motor is not getting the signal all the time.
Good call, Bill. Especially considering the stopped axis looks like it might have been trying to move a little bit. I'd bet a short could do just that.
-
I have 1/2-10 Acme screws and Xylotex 269 motors. (I made a mistake on the 425's)
I am not sure how fast I was going but the max speed is set to 60 ipm
Neil
60ipm is probably borderline for that system. In motor tuning, try a velocity of 45 and an accel of 5, and see if it still does it. I think you're just pushing it a bit too hard.
-
I cleaned and lubricated the lead screw with Boeshield t-9 this morning. I cleaned the screw by spraying Blade & Bit cleaner while traversing the X axis. At full speed I could stall the motor with the rag with only moderate pressure. After lubrication the screw wouldn't stall unless I applied a fair bit of pressure so that helped some.
Last night I made one of the captive slug type dampers. It is crude because I cut it out on the router and the holes aren't perfectly round. My Harbor Freight 7x10 mini lathe is horribly sloppy but I did eventually get a flange that is round.
Installing the damper on the Y axis did little or nothing. The Y runs quiet and smooth all the time. Putting it on the X axis however reduced the noise by at least 50%. It will now rapid at upwards of 90 IPM without stalling. That isn't cutting anything though. It only stalled right in the middle of the travel. The rod is 1/2" x 10 and is nearly 5 ft long. I fear that the high speed whipping may have bent it slightly.
I am going to try running it at 50 IPM for a while and see if there is any problems.
I have a question about the numbers on my motor tuning but I will post that separately.
Neil
-
I have struggled with the *exact* same problems :-\
1/2" screws at 5' are really hard to get straight. Even with a "straight" screw, by nature they will begin to whip at high speeds. Over the 5' span, gravity alone causes the little 1/2" screw to sag a little. I switched to 8tpi 1/2" screws on one machine, and it's better since they don't have to turn as fast, but I still think I'd have been better off with 5/8" 6tpi. The problem with that is, that requires all new bearings and motor couplers. I was also concerned that the motors may have a hard time dealing with the extra mass on fast, tiny movements. Maybe someone else has experience with this.
I also made some of those dampers, and aside from making a LOT of noise, they did little, if anything to help. I took them off after a few hours of messing with them. They may help for other problems like resonance, but not for my problems.
On the machine that I'm talking about, I'm using those plastic anti-backlash nuts, and those things can be finicky. I could feel some small bumps / burrs on the acme screw, so I took some fine sandpaper and ran the screws back and forth with the sandpaper wrapped around it until things felt a bit smoother. After that I cleaned everything up really well, lubricated it again, and it's been better. I still run it slower than I'd like though. I decided I'd rather have my projects come out decent rather than wrecking every other thing I cut. Steppers have a lot of torque when they turn slower, but it quickly fades as they increase in RPM.
-
Neil,
Consider the following:
I would guess that the screw is not even straight within 1/16" ( .062") in 5 feet. So chances are that you will bind unless there is enough
motor torque to deflect the screw and keep it turning unless there is additional slop. I would adjust the screw position at each end
and see what happens in the middle.You may need to comprimise some.
Then do some rapids away from the ends, and end to end, and see what the max velocity you can get till it skips. Then cut the velocity in half.
Get a piece a scrap material ( worst case stuff you will be using) and now do some cuts at the ends and also the middle of the travel.
The depth should be the deepest you will cut. You will quickly come to some velocity and acceleration value where you can have confidence
with your machine.
Speed is nice, but is not worth crap, unless in the end the part can be machined. You have what you have and don't compare to something else
which is different.
Now if you are not satisfied with the results then change the machine to suite the desired operating conditions.
So consider experimenting to see what works for you.
FWIW....
RICH
-
I have a Xylotex with 250 oz motors. My X and Y axis use 1/2-8 2 start. I can get 180ipm on the 40" Y screw, but the two 60" X screws are limited to about 150 ipm, due to severe whipping.
My Z axis, however, has 1/2-10 acme, and it's only good for 55ipm with an accel of 5. On 2D parts, I can up the accel to 7 or 8, but it'll stall on 3D carving.
And I've found that a light oil is a much better lubricant than the dry lubricant sprays. I use pneumatic tool oil, and it works great.
-
I am considering replacing the X axis leadscrew. I have seen 1/2 - 10 2 start and 5 start. These would'nt require any sort of bearing or coupler change.
I have also considered going up to a 5/8 - 8 2 start. This would help with the whipping problem.
What are the pros and cons of these options?
As I understand it, multi-start will reduce the precision but this is a wood router so 1/32" is more than enough.
Neil
-
The only con is the price, although I think the Xylotex is not powerful enough to use the 5 start.
Also, 1/2-8 2 start is quite a bit cheaper than 1/2-10 2 start. About $20 cheaper for a 6' screw.
5/8 has more inertia, so it may hurt your accel a little. And I use Dumpster nuts, which you can't get in 5/8".
-
Gerry,
I've never used a screw with more starts than one. Does using a 2 start screw with the same TPI double your linear speed at the same rpm?
-
Ok I redid the job. The first time I had a glitch. During a straight X cut, the Y had a little freak out. The funny thing is it went back to the correct position on its' own. I didn't even notice it until I was removing the work. The second time it ran fine.
It is the cut between the circles. The code calls for it to cut out the panels last so it is going around the rectangular perimeter.
(http://i13.photobucket.com/albums/a273/SpookyDad/pictures%20for%20forums/runoff3.jpg)