Machsupport Forum
Mach Discussion => General Mach Discussion => Topic started by: swarfboy on December 29, 2009, 04:13:42 PM
-
I upgraded from version 0.20 to 0.033 and find that 0.033 is overfeeding by around +25% on Jog and G0 and G1 moves.
-
are you meaning distance or velocity?
Hood
-
Velocity definitely and possibly acceleration too but its hard to perceive if its over accelerating.
Simple test.
My motor tuning is set to 1000mm/min for X.
Jog X and the velocity reaches approx 1250mm/min. It shouldn't go over 1000mm/min and doesn't in older Mach versions such as 0.020.
I upgraded from 0.020 to 0.033 so I don't know what version this bug as introduced.
-
Not heard of that one before, can you attach your xml and I will see if I can replicate it.
Hood
-
I just loaded Version 033 and checked the lathe axis and the feedrate in joging steps or G01 is spot on. ???
Did you do any motor tuning and not go out and back into Mach?
RICH
-
Didn't change any of my settings. Copied myscreen.set, machlic.dat, mach3_4axis.xml, my macro folder, my image folder, my machmill shortcut that loads my profile. Uninstalled mach3 0.020, deleted c:\mach3r, installed mach 0.033, rebooted, copied my files back, started mach.
I've done this 2x now and both times mach overfeeds velocity by about 25% on all jog and feed moves.
Attached is my .xml
-
Just ran you xml and see the tuning is set the same for all axis's...not 1000 but 850 and the vel dro is right on when rapid jogging.
Double check your tuning.
Hood will diagnose it better in the am.
RC
dang...forgot the pic again
-
I compensated the motor tuning to get it to run closer to my desired max velocity, but this is a workaround not a fix.
-
Oh..OK, I see.
Cant make it happen here, I'll be watching for the solution.
Good luck to ya,
RC
-
Working fine here also. Do you use any Plugins or devices such as the shuttle or PoKeys? might be worth disabling them to see if you can isolate the problem.
Hood
-
No. I have a 4 axis mill and controller that has gecko 201 drives, charge pump, 3 home switches and a probe, nema23 250oz-in 0.9deg steppers, and a couple 110VAC relays and my computer is more than ample to run mach, using XP pro. I've been through many versions of mach over the past 4 years and all have worked fine for the most part.
I went back to ver 0.020 and velocity behavior goes back to normal. Tried version 0.033 again after that and overfeed happens so there is definitely a software issue. I do use my own screen set, I'm going to try to use the default install, screen, and xml to see if maybe something from my screen set iis causing this. I have 2 machines so I will also try this on my other machine.
-
Doesnt seem to be anything there, I did notice you had a custom screen but I very much doubt it will have anything to do with that, it is more likely to have something to do with the threading code that Art has done as it seems to have caused other problems for the select few, you should feel privileged you have joined an elite group ;D
Will be good to hear how you get on with the other machine, presume its a different computer?
Hood
-
I downloaded and test all versions of Mach between 0.20 and 0.33 excluding 0.32 as I didn't see it on the server.
The velocity calculation changed began significantly overfeeding in version 0.30. With motor velocity max set to 850mm/min the machine jogs at 1115mm/min. Please look into this. I can reproduce this on 2 machines.
-
What CPU do both of these machines have? Is it actually an overfeed or is it the DRO reporting wrongly?
Hood
-
AMD Athlon 2000+/2100+
FSB 133
1.0GB RAM
The machine is traveling at a faster velocity, I can see and hear it, and the DRO is reading faster.
Some change the occurred in 0.30 caused this.
-
Strange, have seen reports of weird spindle speeds being seen with these latest revisions and its always been faster CPUs I think, yours is the first with the overfeed.
I think 030 was the version that the new driver for the threading was added but not sure, will ask Art and see if he can look in on this post.
Is there a specific reason you need 030 onwards?
Hood
-
Immediately no, but I don't want to see the bug persist into future versions as I'd like to be able to install newer releases and enjoy the feature additions and bug fixes.
It's odd that I'm the only person experiencing this when I have had it working fine since early Mach3 1.x versions and now it changes in v0.030.
My spindle is controlled by a potentiometer on the spindle controller not through Mach, so I can not verify that issue.
-
Rev 4 is a major upgrade, in fact its a total overhaul so unlikely bugs will be passed on from earlier revisions but even if there are bugs Rev4 will make them easier to sort as it will be in modules and fixing one thing will not break something else which is a problem at the moment from what I gather.
You are certainly the first report I have seen with this problem, thats not to say you are the only one, just no one else has reported it yet.
It was the spindle speed readout I think that seems to be messed up but I am just thinking it could be related as it was about that rev that the new driver code was added.
Hood
-
I have been seeing this also. By a lesser margin than Swarfboy is getting.
Check what the Pulse Frequency is on the diagnostics page. I run 35k and am getting 35317 reported.
Seems to have arrived with the latest driver upgrades.
I thought it was rather good, as my pulse frequency has always been lower than kernal speed, on every computer I have used since I started with Mach. :)
Greg
-
There is a new revision just posted (034) try that and see if it helps.
Hood
-
There is a new revision just posted (034) try that and see if it helps.
Hood
G'day Hood
Yeah I just tried 034, and it is all screwed up for me.
Feedrates are oscillating up and down.
Don't know yet if it is just me or a problem with the new driver update.
Reverted to 033 and all is good again. Or at least like it was before. :)
Greg
-
Oh well hopefully things will be fine in Rev4.
Hood
-
There is a new revision just posted (034) try that and see if it helps.
Hood
Actually I am very happy with 033.
The long standing rotary axis feedrate bug, that I have been complaining about for over a year, has been 90% fixed.
Happy as pig in mud. ;D
Greg
-
Lucky you are not in Scotland at the moment then as the pig would be frozen into the mud and not too happy ;D
Hood
-
version 034 has corrected the feed issue. Thank you very much.
My motor max velocity is 850.2mm/min (for some reason the 0.2 gets added when i enter 850.
The machine now jogs and accelerates properly at 849.9mm/min.
Close enough for me. Thanks.
-
:) glad to hear that, hopefully this kind of thing will not happen with Rev4.
Hood
-
Hello
Excuse me, I have a problem, may be like this (I don`t know).
For example:
Position "0" (to look black labels)
(http://forum.rcdesign.ru/attachment.php?attachmentid=322178&stc=1&d=1266926622)
%
G90 G21
F 5000
M98 P1234 L30
M30
O1234
G1 A5
G1 A0
M99
%
(http://forum.rcdesign.ru/attachment.php?attachmentid=322179&stc=1&d=1266922954)
%
G90 G21
F 5000
M98 P1234 L60
M30
O1234
G1 A5
G1 A0
M99
%
(http://forum.rcdesign.ru/attachment.php?attachmentid=322180&stc=1&d=1266922954)
Is it bug of Mach?
Help me please
-
Has forgotten to tell - Steps for A-axis 8.8888 (200*16/360)
And not understand:
type G01 A10, but in window: "+10.0126"
G01 A360 - "+360.0036"
G01 A720 - "+720.0072"...Etc.
-
Pavlll,
I think what you are seeing is because of your steps per degree.
If you move 10 degrees than the number of steps would be rounded to the lower or higher
of the whole step. Thus you get some parts of a degree like .0126 deg in this case, but at certain angles it will be
a displayed even degree in the dro.
"I think " you would need a number divisible by 360 and 10.8 reduction to get rid of that.
AT 10 degrees you have .0126 degrees = .756 min = 45 seconds which would be about 0.0025" in a foot.
You can ignore it or accept it for what it is.
So it's a resolution thing and good resolution will come from mechanical for a rotary.
I don't have any dro numbers like that for any angle,but , my steps per degree are 750.
Hope this makes sense,
RICH
-
Thank you RICH
I enter 10 (in steps) today... nothing :(
Result like yesterday.