CNC 6040 Router Accuracy vs Speed Testing
Вставка
- Опубліковано 5 жов 2024
- After making several workpieces with my CNC 6040 Router, I learned that the faster I let the machine move, the larger the inaccuracies in my workpieces get. So slowing the machine down is necessary if a workpiece requires a specific accuracy. For small programmes the difference between 3 or 4 minutes waiting isn’t that big of a deal. But for larger programmes the difference between 3 and 4 hours is quite substantial.
...
Affiliate links to budget chineese end mills:
Tool 1: ⅛ inch 2 flute: bit.ly/2N2DyEE
Tool 2: 4 mm 3 flute: bit.ly/2N2DPXZ
Tool 3: 5 mm 3 flute: bit.ly/2N2DPXZ
Tool 4: 4 mm 2 flute: bit.ly/2N1KYb2
Tool 5: 4 mm 2 flute: bit.ly/2SxnT6g
Tool 6: 6 mm 2 flute chamfer: bit.ly/2tmihfX
...
To test the relation between Accuracy and the Speed the Gantry of the CNC 6040 Router moves, I decided cut one design 7 different times. Every time with different Feed, Acceleration and Rapids. To compensate for the crooked bed, I fased some plywood to create a flat workspace. Because all the sevens wooden blocks had the same pre-cut holes setting up a new job was pretty easy.
This was the first time that I did repetitive work on the CNC 6040 Router and I got a small collision with a screw head. So before I could proceed with the test I had to calibrate the axis.
The calibration of the axis was done quick and dirty. My stepper driver is of such low quality that I have to be happy with a 0.2-0.3 mm accuracy. So I didn’t bother to warm up the motors, make sure my dial gauge was 100% level and I measured only 25 mm instead of the entire table length. I think that this video gives a much better in-depth explanation of the process. goo.gl/di8MRN
When al the workpieces where finished I it was finally time to measure everything and see what the different CNC values had produced. The final conclusion is that I can save 27 % in cycle time by speeding up my CNC 6040 Router, without losing the 0.2-0.3 mm accuracy.
Affiliate links to budget chineese end mills:
Tool 1: ⅛ inch 2 flute: bit.ly/2N2DyEE
Tool 2: 4 mm 3 flute: bit.ly/2N2DPXZ
Tool 3: 5 mm 3 flute: bit.ly/2N2DPXZ
Tool 4: 4 mm 2 flute: bit.ly/2N1KYb2
Tool 5: 4 mm 2 flute: bit.ly/2SxnT6g
Tool 6: 6 mm 2 flute chamfer: bit.ly/2tmihfX
![[UA] NAVI проти G2 Esports | Blast Premier Fall Final](http://i.ytimg.com/vi/QWlIm4FGESQ/mqdefault.jpg)
Six beginner/low budget end mills for aluminium(Affiliate links):
Test video with Feeds and Speeds: ua-cam.com/video/M-1dYJWh3rw/v-deo.html
Tool 1: ⅛ inch 2 flute: bit.ly/2N2DyEE
Tool 2: 4 mm 3 flute: bit.ly/2N2DPXZ
Tool 3: 5 mm 3 flute: bit.ly/2N2DPXZ
Tool 4: 4 mm 2 flute: bit.ly/2N1KYb2
Tool 5: 4 mm 2 flute: bit.ly/2SxnT6g
Tool 6: 6 mm 2 flute chamfer: bit.ly/2tmihfX
Nice Video Sir, I have some notes for enhancing your machine accuracy :
- Don't use flexible coupling because the rubber can compress and cause positioning error especially when you increase speed, there is special kind of coupling for linear motion.
- Use a ball screw with ballnut if possible for high accuracy (they came in different accuracy levels) consider a preload ballnut for high accuracy but it needs constant lubricant.
Even on heavy and rigid industrial CNC machines a high feed rate affect accuracy not because positioning error but because tool deflection so consider a finishing pass to get accurate results.
Qqqa
Are you sure that your measurement device is perfect parallel to your axes? I assume that it isn't. One indication is that you get a smaller measurement every time. Just draw an image of a non parallel measurement device. What happens?
I want to notice i think you calibrate not properly. I think axis of move your gauge must be parallel to the axis, if you have any angle between axis of move gauge stick and axis then your stick distance is not equal to the axis distance. Its simple you create a triangle one arm is axis and one arm is axis stick gauge. So if you had at gauge equal 25 mm its not equal to moved distance at axis.
That's because your stepper motors skip stepa, change them for a closed loop stepers and you'll see that accuracy will maintain over different velocities
Thank you for posting your Feed, Rapid and Acceleration numbers. I am setting up my new 6090 and your numbers are a good starting point. Though after reading some of the comments, I think I will get the spec sheets for the steppers, drivers and ball screws and then double check everything. I am using Drufel CNC instead of Mach 3.
There is so much flexibility in the X and Y gantry components that I will probably replace the rods with supported rail movements. I wonder how much inertia plays a part in the inaccuracies. But you could measure different areas of the cut with different inertia affected areas to find that out. I will not even try to speculate on your issues but I am sure after all these years you have everything worked out.
I have the 110V 2.2KW spindle version. The way it is wired with a 15 Amp cord and socket on the controller, I can see large loads on the spindle dropping current going to the drivers causing the system to loose steps. My first mod is going to be to run a separate 20 amp circuit from the breaker box to the spindle.
Thanks for the video, I have not tested it as thoroughly as you, but I have found on my 6040 that i started getting errors above 1000mm/min. Good to see you do some research so i can bump up my speed a bit.
lol - drilling the holes at 2:55 - 1, 2, 3, 1 and 4 ? ---> magic!!
Bearing in mind comments. Can you repeat this test since upgrading the table to a steel base plate.
1st problem - indicator stem is angled to y plane. 2nd problem - stepper drivers in this machine are shit. Had same problems - replaced drivers and now I can hold less than 0.05mm for hour long programs.
Replaced with what exactly?
Excellent work. Thank you for that!
the parallelism of the sensor to the axis is also set up at the eye, certainly an error in measurements, the closer the sensor is, the tip of the sensor is measured in a different place on the milling cutter and that it is round, it measures further or closer, when measuring with such accuracy it could be taken into account.
The Problem of Acurracy are the stepper motors losing microsteps while fast moves^^ Buy a closed loop system or real servo motors, to avoid this^^
First time I have ever seen anyone test out the repeatability. Also shown the micrometer test.
Do you remember where you bought your machine from. The exact seller?
Nkw you know the column version of this machine being sold as the 5 axis. Do you know for one if its good, and also if it can use just the entire flat bad in 2d cutting/milling as well.
Thinking to get one. Just need those questions answered and a direction to a trusted seller.
Thx
James
faster produces more cutter deflection, chip loading using end mills for aluminum, use only two flute with high rake high relief specially ground for plastic and an air blow to keep the cutter from eating its chips anything less is just fustrating, same for drill bits regular bits produce chipping
If speed is the problem then it could be related to something mechanical, for instance ball screw accuracy or backlash, or perhaps table vibration at high speed excites a harmonic in some sensitive component. But the first place I would look would be the stepper drivers: are they keeping up outputting the pulsing rate demanded of them at high speeds? You need an O-scope to check whether the pulse waveform is nice and square or rounding or ringing and not keeping pace.
You will never beat newton's third law unless you go with an extreme heavy machine rack.
Following on from "MP5 Works"
I have the same machine (6090 longer version) and have had a long conversation with JFT about the inaccuracy of it.
One moment it's sort of accurate and the next loses and even gains a mm or so.
Tried 3 computers (one of which is dam fast) same results.
The factory suggested solutions have not fixed it.
But i have found a solution....
The Stepper Drivers (JFT) come set to 1600 pulses.
The ball screws are 5mm/rev and if checked by hand (only show a small error after 800mm of travel.)
So the Mach3 setting should be bang on 320 and not 321.xxxx
But as you found it ends up being something between 321 to 322 which means something is not right.
Go set the stepper driver to 400 and Mach3 to 80
You'll find it's dead nuts on accuracy.
I'll test 800/160 tomorrow to see if that can keep the accuracy
But suggest you try testing across the full range of travel, you might get a nasty surprise.
FlyingHaggisRacing can you explain what 400/80 and 800/160 means? How to switch driver stepping. What value for Step and Dir Pulse are you using. What about acceleration? Thanks
I stopped using Feed rate override function entirely. It overrides rapid moves as well and what I found with my Mach3 is that it will start losing steps at anything above 1500mm/min. Rapid moves get impacted by this behaviour as first. I have tested with steppers disconnected from the machine and it does the same. My next step is to check with quicker pc to eliminate that possibility. Will report if I find a solution.
Same here, i set max vel to 1500.
How were you using FR override - to go faster ?
By default I had it set to 2800 as per manufacture recommendation in motor tuning. Then in settings I was using rapid override @1500. However I found a main reason why I was loosing steps. I had a bad habit of using stop cycle button every time program was ending while the machine was moving back to home position (basically just saving me some time moving machine back to where my work piece was for tool change etc). While machine is moving at high speed sudden stops would shift my Y axis by up to .3mm each time depending on the speed. Not using stop cycle button anymore and all seem to be fine so far. Using rapids at 2800mm/min again
hehe nice. but the you have to calculate the repeat error of the ball screw bearings. and get some nice closed loop steppers. :-)
I'd be really keen to see a test on some aluminium
Coming up ;)
Would have been nice to have the same project with all the processes in fusion360 then all the processes in mach3.
For a beginner like me there is too much jumping between stages
I have almost the same material as you a 6090
But I don't find a tutorial for beginners with all the steps.
Thank you already for your video !
Hi Paul, Thanks for your feedback! Can you describe in short what the perfect video for you would be?
The same project but:
The complete construction of the drawing in fusion360 (Model)
And especially the (Cam) Setup / the cutting speed (All stages)
Exporting in Macht3
And if you have time
the setup between mach3 and the cnc
As in your film (the small problem) but all the steps that must be done since the unpacking of the cnc.
But do not cut anything. as a beginner, all stages are important.
Especially since we try to reproduce it
Thank you for your time my cnc is waiting for 4 months
I wanted to make myself happy but it's not my job and I discover everything
hello, thanks for the video, it's very interesting. what parameters do you suggest to enter? did you change them to mach3 in the "motor tuning" section or did you change the cutting speed in the Fusion 360 tool?
nice one, can you actually get 2800 mm/min out of your machine? Mine with the stock config can only get a max of 1492 mm/min on rapid, looks to be the same machine too
Music? 😍
what was your settings in mach3 on constand velocity?
I'd like to comment on three things about this video:
1) Your tests are correct, but your conclusions are not. The machine is capable of far greater accuracy - even my leadscrew-driven 3040 does 50um at over 1000mm/s. Your stepper motor drivers are failing to hold the microsteps necessary at the speeds you're driving the machine at. You should either up the drive current(and keep an eye on the thermals), reduce the number of microsteps, or upgrade to closed-loop servos(most are LPT-compatible anyway). Also, make sure to tighten up the ballscrew mounts. They do come loose on these machines after a bit of rattling.
2) The music. Please, just stop using this god-awful music. It was fine for one video, but not the same track on loop on every single video. There's tons of royalty-free music that isn't cliche UA-cam crap.
3) Don't stretch Fusion when editing - your circles are non-circular in the video. White or black bars work better in this case.
Otherwise, good video and editing, keep it up :)
Thanks for your comment! Point 2 and 3 are noted for the next videos :)
Do you have any information/forum threads about upping the driver current?
The driver ICs are different in some machines - just google their datasheet and follow the schematic. It should be really easy to figure out.
I believe a gecko driver allows you to adjust .. and even on the stock controllers there should be a potentiometer, its tiny and needs to be adjusted.
Which program do you use?
the problem is with material. try some cut in a good wood and get wood cutter
I hope you learned how to use fusion and cad software better 😄 watching you make the block made me mad 😄 I was like why are dtawing it like that.
In school I was taught how to draw stuff for real, I also took architecture desgine in HS for 3 years in which we used just AutoCad.
Anyways just a pet peive. Good video though 🙃🙃
what drawing program do you use? thank you in advance for your reply.
Fusion 360
What software is that please (the CAD software)?
FUSION360
You didnt understand the real problem.
Thanks a lot.
Well this is close to 3d printers.. it needs tuning and calibration.
Please i need your opinion after using it, do you recommend me to buy the 6040 4 axis cnc machine for about $1300 including shipping (yes it is high bcoz of shipping) ? what i am going to do with it is doing mainly aluminum and woods for projects and small quantity of selling products.
Thanks
I think that you always get what you pay for. The same with this machine. so for 1300 you got a great hobby cnc, that has some known issues. If you can work around the shortcomings and can fix things when they break I would buy the 6040. If you want something reliable that you want to depend on for your business, I would buy something else. Good luck!
Thank you a lot.
Well, this breaks the thinking and messing around on thoughts because i thought it is a bit reliable.
But i think i will go for it, as i dont have that big room for pro cnc.
Thanks for your help sir, appreciated.
Best Regards
Gauge R&R ;)
I buy 6040 and it wery shit. I think chinese ashamed to have this model. I follow all videos and it still only make scrap for my garbage man.
you can accurately cut wood at 300mm/min with a 10W proxxon 50Espindle in 1mm depth increments. This one is 2200W and seems to cut at around 1000mm? Bottom line is: on a windy day using two wind generators out of NEMA 17 motors (60W output) you can run a belt driven CNC router and cut these same parts ECO FRIENDLY and for FREE. Learn to love this planet while you can still breathe.
I would like to see When you go back to zero from 25 mm. Then you can see the backlash on the axis..