May not be pretty, but it's certainly more effective than nonexistant PBC parts, eh? Nice job!
If (or when) I do it again, I'll do a few things differently....but in the end the PBC parts are nice, look good, and work pretty well overall.
Well, I'm starting this thread to quit jacking Winegars build thread.
I have been following winegars build almost to the tee in addition to his personal help. For about a month now all I've been lacking are the rails. Since then been learning the software and even bolted the Z axis to a 4x4 to perform some center drills.
I finally lost patience with a particular manufacturer and distributor while waiting for my aluminum rails ordered May 4th. I spent less than $40 on some angle iron and a 2x2 length of tubing and built my own rails. Now I know the quality will have its limits but I now have a fully functional table to progress on. If I ever receive the aluminum extrusion rails I will decide whether or not the improvements will be worth their installation. I was actually quite pleased how straight I could keep the angle iron and strap metal rails. Posted are the very first cuts I made. The dimensions for 1" entities were right on.
Gotta go buy some steel now....
Similar Threads:
May not be pretty, but it's certainly more effective than nonexistant PBC parts, eh? Nice job!
If (or when) I do it again, I'll do a few things differently....but in the end the PBC parts are nice, look good, and work pretty well overall.
Carl
Hey, received my gantry rail today from my favorite manufacturer. Weird that it was ONLY the gantry rail I received. Still waiting then...
On another note, made my first aluminum cuts. WOW, didn't realize how clean and smooth it would be. No discoloration, straight, almost looks like it was a (straight) cut with a hack saw.
Working on my home/limit switches now. Attaching the switch to the moving axis which means activating by a ramp on each end. Any tips on the Z with TNG? Seems like I have to re-zero Z after each cut and raise it to move the gantry.
What post processor are you using in TNG? I use the MP1000....or something similar to that, I can't remember now. It has the correct code for doing the torch touch-off before the cut, as well as raising the Z to the "safe" height between each cut.
Carl
Finally received my rails.
The videos links I'm posting are to my second real cutting.
Please analyze the shuttering/vibration I'm getting while cutting these fonts. The drawings are shown as zillions of line segments. 16ga steel. The bottom first attempt is slow at 50ipm, .020 cut height with the 220674 shield. Thought the slowing down was helping the shaking but started dropping arcs.
The second attempt above the first was at 90 IPM (by increasing the feed rates by 150%) feed rate and 40/50 cut height. It actually finished all the cuts but still a lot of jagged edges. I wonder what the cause is because even the outside border I added has the shakes and jagged edges.
Experts, let me have it! What am I missing?
[nomedia="http://www.youtube.com/watch?v=1RfgX-vUhsY"]‪M2U00278‬‏ - YouTube[/nomedia]
[nomedia="http://www.youtube.com/watch?v=Z9qXU6iJBio"]‪M2U00279‬‏ - YouTube[/nomedia]
[nomedia="http://www.youtube.com/watch?v=PXtvL0NrzcE"]‪Start of WoodBros logo - second ever cut on new table‬‏ - YouTube[/nomedia]
Thx,
Jerry
Boise
A couple ideas--first, you're cutting pretty slowly for 16ga. Second, you can try increasing the "lookahead" in Mach3 so the computer can "see" farther into the future--or redo the drawings so it's arcs rather than lots of little lines. Third, you need to get a different shield for your torch. The one you're using is for hand cutting, so it's made to drag on the material--it gives you the .06 cut height while dragging, where the machine shield gives the clearance you need. The part number is in the Hypertherm manual.
With my Powermax 45, I've found the cutting charts in the book to be really close to what is needed--at least a great starting point for amperage/voltage/cut speeds.
Carl
You are using (as Ilean stated) the wrong shield....switch to the mechanized shield and then cut at .06" off the plate just as the Hypertherm manual states. For best cut quality on 16 gauge (no dross, no warpage, smooth edges) the manual says to go 350 inches per minute. Your machine probably is not cabable of that speed....so I would suggest determining the maximum speed that your machine can handle....then reduce the current (amps) to 35 or so....maintain the .06" cut height.
Jim Colt Hypertherm
Thanks guys for the great info. I am waiting for the 220673 shield now.
My concern is over the jerkyness in the motors. Is that due to the slow speed? At higher speed the jerking seemed worse. Or is it because of the zillions of segments in the fonts? It didn't occur in my small test cuts of one inch circle and squares or the small stars we made.
Here is a question (and I thought I had changed this), my Y axis master motor is set at step=1361.7 velocity=343.9 acceleration=36.6 for the NEMA 425oz stepper. Does this sound right? My slave motor is not the same at step=2000 velocity 19.9 acceleration=10. I thought I changed that but apparently didn't save it.
Thanks,
Jerry
Boise
Oooooohhhhh, so I'm cutting .060 too high using the floating Z axis. And if I set it correct, then the tip will drag the material with that shield. Thanks! Took me a minute.
The machine shield is .060" shorter...and is smooth on the face so blowback slag will not stick as easily.
As far as the motors.....I would work with the motot/drive supplier to tune them to your machine.....obviously any anomalies in the motion will show up in the cut.
jim
It's mostly due to the many small segments. That's where the lookahead might help you, but cleaning up the drawing will be the biggest change.
You'll want to set your slave axis settings the same as your master. It will work most of the time the way you have it, simply because the A is slaved--it's working on the Y axis' settings. If you do a "Ref Y", then each uses it's own settings, and you'll end up with a progressively more crooked gantry.Here is a question (and I thought I had changed this), my Y axis master motor is set at step=1361.7 velocity=343.9 acceleration=36.6 for the NEMA 425oz stepper. Does this sound right? My slave motor is not the same at step=2000 velocity 19.9 acceleration=10. I thought I changed that but apparently didn't save it.
Thanks,
Jerry
Boise
You got it!
Carl
The jerking is most likely caused by the arcs being a zillion short line segments. What program are you using to draw? Some software will turn arcs into that. But as said earlier in the thread, fixing your drawing will help the most.
Also, there is a setting in sheetcam, under Application options > Drawing import > arc fitting tolerance. Increasing that setting, to my understanding, will smooth the arcs some.
If you increase it too much, you may get some faceting in your arcs, but with plasma cutting, you will probably not notice much.
Your motor settings look as though they should work fine, other than the slaved motor not being the same.
Hey All!
Table has been doing pretty good but its time to snuff things up and fix a couple issues. Most issues are probably just part of my learning curve.
Anyway...
My first question has to do with the motor tunings. X,Y, and A are all driven by NEMA 23 425oz motors on a 20 pitch rack and pinion system. Since these are all the same, shouldn't the tuning numbers be the same for the three (X,Y,A) axises(sp)? I used a dial indicator at first to tune on the mach3 setting page. This looked to be right on at first until I started cutting larger files. The first inch was perfect but for each 10 inches out the X axis I went, I was getting an additional .01" error. So a cut 20" out was actually 20.02" or so. I re-calibrated with just a straight edge on 10 and 20" movements and things are a lot better. So how can you get a good calibration with only an inch movement from a dial indicator?
Second, is there a good process to determine backlash? Is it typical for everyone to have backlash values configured in mach3? How can you correctly tune the motors if you have backlash in your configuration?
Thirdly, I've attached pics of drawing circles. I rigged a spring loaded pen holder and use it to verify designs before cutting into good metal. As you can see the circles lean alittle to the left or are flat on a 45 degree angle. Any ideas how to get them better? Is it the backlash in the table? or tuning parameters between X and Y? The pen plotter has some slop to it but the same shape happens in circles cut by the plasma also.
My home referencing seems to work well and keeps the gantry pretty square so I have ruled that out.
Any critiques or ideas would be greatly appreciated. Thanks!
Jerry
Boise
When I tuned mine, rather than using a dial indicator over a 1" distance, I just used a tape measure over the full length of the table. Less accurate over a short distance, but no cumulative error.
Carl
I'm a little confused... I was reading your thread while planning a build and came across a couple of points of personal interest.
What control software are you using?
Does your control software have feed ramping? I run a Fadal cnc mill that uses feed ramping (G08 is off and G09 is on) to slow the machine at the end of each line and increase at the begining to increase accuracy and reduce jerking.
When you refer to Backlash you are asking about a slop issue? If your machine control allows you to use optional stops (fadal is M01) or single line commands you should be able to measure backlash using a dial indicator. If you run the axis in the plus direction say 4 inches and stop, set your 1"dial indicator to zero at .5 (approx half of the indicator travel). Then move back a specified distance say minus .3 the difference between that number (.3) and what the dial says is your backlash.
When you figured your pitch number did you use the spur gears pitch diameter to figure movement per revolution for your motors? The 10.01 to 20.01 sounds proportional.
Jerry, it's been a while, but I too had the same problem. I worked on it for a week or so till someone set me straight. The jerkiness has to do with the CV control. I've cut letters with more arcs and segments than you can count and I have no jerkiness...... Smooth cuts.
Do a little experimenting with the cv control and you'll get it going. E-mail me and I 'll check my set up if you like and make a screen shot of my setup. ajclay@cox.net
aj
Last edited by ajclay; 03-28-2012 at 06:58 AM.
(Since these are all the same, shouldn't the tuning numbers be the same for the three (X,Y,A) axises(sp)?
Sure, same motors, same pinion, same gear rack... no mystery there.
( I used a dial indicator at first to tune on the mach3 setting page. This looked to be right on at first until I started cutting larger files. The first inch was perfect but for each 10 inches out the X axis I went, I was getting an additional .01" error. So a cut 20" out was actually 20.02" or so. I re-calibrated with just a straight edge on 10 and 20" movements and things are a lot better. So how can you get a good calibration with only an inch movement from a dial indicator?
Second, is there a good process to determine backlash? Is it typical for everyone to have backlash values configured in mach3? How can you correctly tune the motors if you have backlash in your configuration? )
Backlash doesn't come into play if you never change directions. When you are calibrating the machine for accuracy, make all the movements in the same direction. move it in +X direction, put the dial indicator on and zero it, now move it in the same direction one inch. There is no backlash and you'll be able to set the steps or count correctly. ( don't get this confused with "tuning" the motors. Two different things. )
Once you get it moving exactly one inch "1.000" and showing one inch on screen you got it. Your machine is only as good as the gears. Stack-up comes from gear irregularities, usually. Find a machinist with a Starrett long ruler to check stack-up or long measurements. Hang a indicating device on the torch an move the full length of ruler. If you can't see any irregularities you're closer than you think.... .010 or better.
Backlash is just that. Moving in one direction, stop, zero indicator, and start moving .0005 increments at a time in the opposite direction. That will tell you the backlash when things actually start moving.
( My home referencing seems to work well and keeps the gantry pretty square so I have ruled that out. )
Making a machine move exactly square isn't easy because so many things come into play. The smaller the part you're cutting the smaller the error in squareness / roundness......
Aj