Machsupport Forum
Mach Discussion => General Mach Discussion => Topic started by: killrmodz on November 05, 2024, 04:53:09 PM
-
Circles aren't circles...
After a long hiatus, i'm back in the saddle trying to churn out awesome creations. So i've upgraded to clearpaths, a new BOB for my ess and updated from VCon wheels to HGR20 linear rails and guides. Everything was seemless transitioning over.
So here is my dilemma... My circles are not circles, but something oblong... LOL I will take pictures and upload them along with my gcode tonight, I just wanted to put this out there to see if anyone else has experienced the same.
I use Fusion360 for my CAD/CAM and noticed that a lot of updates have been applied from the last time I created gcode for a project.
I'll post up more info and some pics when I get home and take pics of my mach configs.
Thanks!
Ismael
-
Here are some configs and a recent job.
-
If the circles are displayed as being correct in the Mach3 Table Display screen then it is most likely a mechanical issue with your machine.
If, however, your circles are not displayed correctly then you need to check your CAD/CAM settings etc.
Also, post your problem Gcode.
Tweakie.
-
Here you go Tweakie!
-
I ran your Gcode here and the circles are comming out round & OK but I had to reduce your F1200 feedrate as my machine would not cut small circles at that speed (certainly not in acrylic) so perhaps that could be your issue also ?
Another thing is that it is not a good idea to have a G90 and G91 on the same line of Gcode.
Tweakie.
-
My clearpaths are capable of that feedrate for what i was cutting but will try and lower the feed rate. They are configured for the x and y access for 6400 pulses per rev. Not sure what is machs limitation is any when running clearpaths.
When you imply g90 and g91 shouldn’t be on the same line, im assuming i think line 3 is what you are referencing, correct?
Should the rest of the g90 that are on their own lines be be changed or just line 3?
Ill change some things and give it a try.
Thanks again Tweakie!!
-
Changed some configs after reviewing doc on ess from warp. After that redid some gcode.
(1001)
(T5 D=2. CR=0. - ZMIN=-1. - FLAT END MILL)
G90 G94 G91.1 G40 G49 G17
G21
G28 G91 Z0.
G90
(2D CONTOUR1)
S16000 M3
G17 G90 G94
G54
G0 X11. Y6.
G43 Z15. H5
G0 Z5.
G1 Z-0.5 F250.
G3 X1. I-5. J0. F1000.
X11. I5. J0.
G1 Z-1. F250.
G3 X1. I-5. J0. F1000.
X11. I5. J0.
G0 Z5.
X6.566 Y15.674
G1 Z-0.5 F250.
G3 X5.566 I-0.5 J0. F1000.
X6.566 I0.5 J0.
G1 Z-1. F250.
G3 X5.566 I-0.5 J0. F1000.
X6.566 I0.5 J0.
G0 Z15.
M5
G28 G91 Z0.
G90
G28 G91 X0. Y0.
G90
M30
That attached pics were the results. Checked my physical machine and no issues found. Use AVID CNC r&p system with hgr20 linear guides/rails. Might tuned down the clearpaths as i'm not sure what else to do.
-
Very hard to make sense of your JPGs. May be a setting somewhere? I get huge pics which would fill the screen several times over. Not enlightening for me.
Anyhow, there are other ways to mill holes, especially using a helix. The following has been extracted from a much-used program, and may have some excess. It make a single hole centered on 0,0. Most parameters can be changed as you wish. A cleaner bottom is easy to do.
#1=10.00 % diameter of hole
#2=4.00 % diameter of cutter
#3=[#1/2-#2/2] % RADIUS for helix
#4=10.00 % Z thickness of material
#5=2.00 % Z cut depth per cycle
#6=fup[#4/#5] % N(z) cycles to do
#7=[#5/#6] % dZ increment per cycle
#8=0 % Z start on surface
g0 z40 % Z safe
g0 x#3 y0
g0 z#8 % touch surface
m98 p10 L#6 % do N(z) full helix circles
g0 x0 y0 % go to middle for safe exit
g0 z40 % exit straight up
m30
% Helix subroutine
o10
#8=[#8-#7/2] % down
g3 x[0-#3] y0 z#8 r#3
#8=[#8-#7/2] % down
g3 x#3 y0 z#8 r#3
m99
Cheers
Roger
-
I have resized the images so we can all see them :)
Tweakie.
-
Hi Tweakie
Thanks.
To me, it could look like a bit of backlash: jigglies at two opposite points on the circle.
Or maybe ??? something to do with the way the job is programmed: go down a bit, then go around a bit. But even so, it argues for backlash or bad flex.
Cheers
Roger
-
I agree with Roger.
Your Gcode works but is not good Gcode.
As the circles are created as two semi-circles Mach3 makes a slight pause between each and flex (maybe in the machine, the spindle, the endmill, etc.) is creating two knibs in the machined toolpath. I think this could be improved, as would the roundness of your holes if you slowed the federate down further.
Tweakie.
-
Checked gcode from previous runs on my old machine setup. Absolutely identical… the only things changed are the linear guides and clearpaths…
-
Fusion360 has gone quite a bit of updates… but the output gcode is relatively the same.
-
One thing I noticed when doing test cuts (larger circle in the pics) is that the little bumps are the same where ever I moved on the piece of acrylic. It's like it's cutting half arcs with little bumps enter and exit to the contour base it's direction in the routine. It's does a bump when it begins to travel from left to right and then again right to left on the x axises. I've double checked all of the mecahinals of my machine and nothing was loose nor did the rack and pinions have backlash... So stumped. Next steps is to redo my tuning of the clearpaths. Thanks you two for helping me through this while I try to find the root cause!
-
Look at this edited pic. This is what you referenced as well tweakie... which occured when I reduced my feeds from 1200mm to 500mm.
-
> nor did the rack and pinions have backlash...
With all due respect, rack and pinion always has backlash. Sorry, but that is how it is.
That is why we use ball-screws. Mind you, even a ground ball-screw with double ball-nuts will have backlash, but it will be sub-micron and beyond my eyes.
Cheers
Roger
-
Fair enough… im just going by my experience with what i have as a machine and similar gcode used with no issues previously. Im engaging in the avid forums as well on clearpath tuning and it possibly could be that and how there could be a contention between the clearpath servos and mach3 controlling the motors.
Thanks!
-
Something that I omitted that I want to get your opinion on. My Y axis is running two Nema 34 servo and my X is running a single new 23 servo. Prior I was running all three motors, two on the y and one in the X, the same. Could not having the X axis motor match the Y motors cause this issue? Thanks.
-
Not enough detail. What sort of Nema 34 servo: stepper or 'analog'?
Do the motors have integrated gearboxes?
What sort of (electronic) drivers?
Yes, lots of detail needed IF you want high performance.
Cheers
Roger
-
I just want my circles and arcs to not be rounded squares.
The setup:
MM is my default units
Ess with c82 bob
Clearpath sdsk nema 34 x 2 for Y
Clearpath sdsk nem 23 form X
Both have internal drivers
Im using avidcnc’s rack and pinion system.
Nema 23 has a 3:1 gearing reduction
Nema 34 has a 3.2:1 gearing reduction
Both servos have no microstepping defaulted to 800 steps (units) per revolution.
Both gearing are 20 tooth per 25.4 mm (1”) rotation
Hope this helps.
Thanks!
-
This gets a little complex - sorry.
The Clearpath servos do have encoders, but I don't know where the encoders are. If they are on the motor itself (most likely), then you can also have backlash in the 3:1 gearbox which is added to the motor. From an engineering perspective, you must have some backlash in the gears to avoid having them lock up and jam.
That said, I suspect that the rack&pinion is far more likely to be the source of any backlash.
Rack&pinion is good for large routers processing large wood panles: you would probably never notice any backlash. But for machining harder materials - from hard plastic to steel, industry uses ball-screws and ball-rails these days. To be sure, older CNC mills might have used ACME threads and dovetails in the distant past, but they always had backlash and wobble. That made getting precision difficult, and these days ACME and dovetails are only used on low-end manual mills - it's cheaper you see.
Cheers
Roger
-
Thanks for the information. My previous version using the same rack and pinion didn’t have this issue but the machine isn’t the same anymore. So with all thats been said and all that ive tested changed and verified, it is the only thing that makes sense at the moment. So backlash compensation is what ill be adjusting and testing next. Thanks Roger!
-
My pleasure. Anytime.
Cheers
Roger
-
You can easily measure backlash if you have a dial indicator - method is shown in the Mach 3 manual if it isn't obvious. I have a very small amount of backlash on my mill - ~0.01mm IIRC - and compensation works well. On my lathe however, which has an acme Z drive with more like 0.3 mm, I have not had much luck as the Mach3 BC algorithm doesn't seem to work very well with larger amounts, and the setup isn't very well documented. On a lathe Z axis you can usually avoid backlash problems by always cutting towards the headstock just like you do with manual machining.
Also whilst you can compensate from a purely metrical point of view, BC cannot prevent other effects such as the cutting force pulling the work into the cutter as it isn't constrained.
-
Latest update. To keep all things even, i've ordered the same servos that are on the Y axis for the X (considering a load issue and undersized the servo). Tuned with the MSP software from Teknic. The oblong shape is gone when cutting but definitely seeing, the best way to explain is stuttering. I'm leaning towards a tuning issue with the clearpaths as i've taken things apart to assure things are tight and straight (assuming some sort of binding). Then I noted something when homing. There slight stutters going to it's home position. Not for the Y axis but for the X. Once homed i brought the spindle/carraige closer to me and reduced the jogging to 5%. This made it a bit more noticeable as the carriage is closer visually. When things were apart, and without the motor used to move the X axis, it moved smoothly by hand. So at a slow job, it stutters... which is odd. Even more odd is that I remember seeing this when I was initially tuning. But others thoughts when using clearpaths would be great. They are the SD models. Thanks again to everyone who has provided information. It is greatly appreciated!!
-
What you are calling 'stuttering' MAY be marginal oscillation in the servo. The gain setting may be too high, or maybe the differential setting needs adjusting.
This can happen with any servo drive, and may not be a reflection on the brand - just on the tuning.
Cheers
Roger
-
Will be doing retuning of the x axis. Also refine my velocity and accel. Only things I can think of. Thanks!
-
Hi killrmodz,
I came across this thread and wanted to offer a suggestion that may help with further troubleshooting.
Teknic's ClearPath SD and MC motors come with a motor setup and configuration software called MSP. The MSP software includes many diagnostics tools that can be helpful when troubleshooting a machine’s/system’s performance. For example, MSP includes a digital oscilloscope which can be used to monitor different application variables in real time. For this application, you could use the scope to measure tracking error (i.e. the difference between the commanded position and the actual position of the servo) to help infer whether the issue is caused by the command itself or by the servo's response to the command.
For more guidance on using MSP and the digital scope, here’s a helpful video that walks you through the software: https://youtu.be/-6JbTaa9EPA?si=Q48qSvLpkz4xkqel
Best of luck with the application. If you have any questions, please feel free to contact Teknic at 585-784-7454 or submit a contact request on the Teknic website (https://teknic.com/contact/).
Aimee F. – Teknic Engineer
-
Aimee, thanks for this. Ill have a go at it this week. Falling behind on projects and need to get this resolved!
-
You're welcome - I hope the issue is resolved soon. If you have any questions while troubleshooting, please feel free to reach out. We're happy to help!
-Aimee F.
-
Absolutely!!!
-
Servo is making a clicking sound. Already submitted an email on it.
You're welcome - I hope the issue is resolved soon. If you have any questions while troubleshooting, please feel free to reach out. We're happy to help!
-Aimee F.