Machsupport Forum
Mach Discussion => General Mach Discussion => Topic started by: Timbob on November 03, 2013, 01:31:51 PM
-
OK, it's hair pulling time.
If I use a wizard, any wizard, to thread, the threads come out with the wrong pitch. Not just slightly, but massively. Example, set for 16tpi, it turns out at about 26tpi. Next attempt, it'll be about 24tpi.The threads are consistent in the same run, and track correctly from beginning to end, spindle speed remains constant. These threads would be useable except for the fact that I'd have to make custom nuts for them. Not too many 3/4x26 nuts out there. This has been an ongoing problem. Originally, the threads would get wiped out at the end of the cut during the spring passes. That is, I was threading a 4.5" length and all was ok until the spring passes, at which point the last 2.5" would get wipe off. So upon researching I found that the version I was using was not up to date, 3.040.??? I decided to update to 043.057, but now I can't get past the wrong pitch to even do a spring pass.
I have:
-Updated to 3.043.057
-motor tuning and steps set correctly. 1" commanded = 1" move, 4" command = 4" move. It is not losing steps. Returns to correct start point and ends at exact same spot. Like I said, useable thread, just no known standard.
-Set to exact stop
-Backlash is not enabled yet. No need to muddy the waters.
-Diameter mode, tried radius for *********s and giggles
-Debounce to 0. Have tried various other settings. Mach sees the spindle and feeds consistently, just wrong.
-Spindle indexing, feedback, closed loop and averaging set. Indexing/RPM is working correctly. The RPM reads correctly, but the IPR DRO indicates that it is feeding at the bizarre rate for 26TPI. IE, it should read ~.0625 for 16TPI, but it reads ~.02 to .03 . It is like it just pulls an IPR setting out of its ass.
-Tried all infeed methods and angles.
-Tried my origignal xml. It does the same thing, but the spindle response is more accurate. The new xml slowly adjusts to the correct speed, where the old is just there right away. This doesn't seem to affect the threading,a s Mach compensates for the speed difference. Like I said, very consistent thread, beginning to end.
My next thought is to completely uninstall Mach and start fresh. Are there nasty files that don't get replaced when doing an update? If so, how do I kill them off for a totally fresh install?
Anybody have any thoughts?
Please???
Tim
-
Please attach your xml.
Hood
-
XML attached.
-
Turn of the Turn Diagnostic Plugin and restart Mach and see if that helps.
Hood
-
Turned it off. Exactly the same.
-
Are you using a VFD? Dont see any relays set up in Mach.
Is your ratio correct? you have it at 0.67
Hood
-
I'm using a VFD controlled through mach on a G540, through LPT. Ratio is correct at .67 . I might have to post a pic of the threads. They are really quite lovely! Just useless :'(
-
Any way you can confirm the reading in Mach is the same as the spindle, maybe a mechanical tach?
Have you tried with closed loop disabled?
Hood
-
OK, here's an interesting bit: If I change the thread t exactly 2/3's of what I want, they work perfectly. 16TPI, Clean and correct.
What the heck?
-
seems too much of a coincidence your spindle ratio is set to 2/3 ;)
Hood
-
That is, if I set the pitch to 10.666667 TPI, .0938", the threads come out at 16TPI, .0625" .
-
seems too much of a coincidence your spindle ratio is set to 2/3 ;)
Hood
Dear God, if it really is that simple, I'll never live it down. I'll try it right now.
-
Dont worry about it being that simple, if it works just be happy :)
Hood
-
why would you have set the spindle pulley to .67 ?????
-
Looks to have been the problem. Why would it have worked in the older version with a .67 ratio? Granted, I did have threads being wiped off for the last 2.5" of a 4.5 cut, but it was good up until the spring passes. Also, why is the spindle response slower in the xml i posted earlier vs the one I attached just now? This one nails the speed within a second and holds. The other starts slow and then gains and levels out.
Thank you so much for the help. Whenever there is a problem, you are usually the first to help. This board is lucky to have you.
Tim
-
why would you have set the spindle pulley to .67 ?????
A combination of what my manual said, and that .67 is the motor to spindle ratio. It made sense in my mind, but I can see how it made a problem. Again, it DID work before at .67
-
Difference is you have averaging on now and didnt before.
Regarding the pulleys ratio, you would only use a ratio if the actual sensor was on the motor side rather than directly on the spindle.
Hood
-
sounds odd, but great it's now sorted.
One thing I always add to the threading wizard, is a pause for a couple of seconds on the first pass to let mach see the correct speed. You can set this in spindle setup if you like.
-
FWIW,
Mach compensates for changes in spindle rpm, slow down only, and adjusts the next pathing based on what the real time ( ie; it is not manipulating some reading and the shared time with windows ) rpm was, and averages it if spindle speed averaging is selected and as such then you have two compensations fighting each other, namely the vfd trying to maintain an rpm and mach trying to correct for a slow down of rpm. So the fix was to set rpm manualy and not have the problem. Frankly, good threading relies on enough spindle power being delivered to maintain the rpm and as you deviate from that threading results / accuracy will suffer. Frankly, if your spindle speed is solid spindle speed averaging dosen't do much for you.
Not sure if things have changed in later versions of MACH as i still use an older version......
RICH