looks great :)
Printable View
looks great :)
That is definitely a great success. The cut looks very nice. Chamfer also needs a finishing pass. Same like normal side milling. I am almost sure you did not do that? So no real reason of disappointment. 0.05mm or 0.1mm will do the job next time. Also workpiece needs to be tightened to bed very strong.
You are correct, I did not do it. Lesson learned again. Thanks Boyan!Quote:
Originally Posted by Boyan Silyavski
Hi guys!
I have an issue with my machine.
Please look at the pictures.
Outside dimension should be nominally 50.00mm and inside 30.00mm.
Same toolpaths with finishing pass.
I use Fusion360 for cam. Mach3 for controller. Probably messed up settings/offsets somewhere?
https://uploads.tapatalk-cdn.com/201...69e56519eb.jpg https://uploads.tapatalk-cdn.com/201...84e136302a.jpg https://uploads.tapatalk-cdn.com/201...2cc7fdf15e.jpg
What cutter are you using, and how accurate do you believe it is?
I normally use single-flute cutters, and my experience is that they tend to cut as if they are slightly smaller than their marked size. For example, I get pretty accurate results from my nominal 6mm cutter if I specify it as 5.7mm in my CAM software.
I spent a lot of time chasing this kind of inaccuracy and whatever tests I did, the machine seemed to be pretty good. However, when I cut a simple test square and measured it, the size was always wrong. Now, I do a quick test with every new cutter to work out its "real" diameter. The problem is that you cannot just measure single-flute cutters. Maybe they are just cheap cutters and no-one else seems to have this problem, but this is the only way I can get accurate, consistent, results.
Looking again at your measurements, though, I'm not sure that this is the case here. If, for example, the cutter were cutting 0.3mm undersize, the internal dimension would also be wrong (29.7mm). Similarly, a scaling error would also see errors in the internal dimension. Maybe it's something silly like using an offset in the CAM settings for the external cut? Or did you forget the finishing cut? Sorry if this sounds rude, but sometimes the answer is a small error that you have made and you only find it when you show it in public! That is something that you might find if you examine the gcode.
I use 2 flute 6mm cutter from sorotec.de.
Inside measurement is spot on. Thats the reason I can not blame the cutter diameter? Or am I missing something?
I think this kind of problem has been there from the start. But now I recalibrated all my axis with Mitutoyo digital indicator and all is repetable within .005mm (i dont claim the machine to be that accurate) First axis calibration was with digital caliper which accuracy cant be trusted below 0.02mm. So now I see this strange variation.This part X and Y measurements with caliper dont vary more than 0.02mm.
I've just made an edit that you might not have seen. Yes, it does look like an offset problem in the CAM settings for the exterior cut. Can you go back to your F360 operations and double-check?
For Compensation Type I use In Computer.
Attachment 23593
I will try other options today - will try changing tool offsets in Mach3.
Although I have been using my machine for a year now, I am quite clueless about G-code :( If I need something to change I always go back to CAM and repost...
For faster/better workflow I should compensate in Mach3 I guess.
For roughing passes I leave 0,5mm stock which is default by F360.
Im not good at this stuff, but just a thought..
Do you get the same error in the rough cut?
Skickat från min SM-G955F via Tapatalk
I will measure in the evening. I use adaptive cuts for roughing out - little circular cuts with 2,4mm load. But I guess I should be able to measure 0,3mm. Will try.
Very nice build! Is this really your first one??? Can I ask a couple of questions?
Did you use epoxy on the gantry as well or did you get it machined flat?
Did you have any problems getting everything square?
I see you have the z rails on the moving z plate, which is the better method. Is the z plate rigid enough with just the rails bolted to it? I wondered if C channel steel or a box made from ali plate would be better? (the rest of the machine looks super rigid)
(that's 3 questions! Sorry!)
1. Yes, its my first CNC machine build.Quote:
Originally Posted by Richard
2. I used epoxy on the gantry too.
3. Squareing is time consuming. Have to use creativity and long machinist squares.
4. I can measure quite a lot of deflection on my Z. For sure you can design it stronger. Steel is more rigid. It all depends on your requirements and budget. It is nice to have a super strong machine, but it all comes with a cost. Also I like to point out that good machinist can produce superb parts with mediocre machine. There is no black and white, only gray :D
Ha! very true!Quote:
There is no black and white, only gray :D
Your design is very similar to one I started to draw up, with the deep sides to avoid the bendy gantry sides many people have...
I wanted to as long Z travel as possible, so I used steel channel for the Z plate. The fusion360 sim showed that a flat plate would flex a lot.
Your build is very impressive and it looks really nicely finished.
Do you make any mistakes at all???
Would have done anything differently?
So how did you do the epoxy on the gantry? I can see you have rails on the top and underside of the gantry (good design to keep Z close to the gantry!), so the top epoxy would be easy! The bottom????
Hi Toomast,
In support of post #84 (but not an answer !) I also use F360 for CAM and Mach3 for control. I had a similar problem, although the other way around.
One of the last parts I cut out (a simple rectangle) was roughed out with 0.1 mm stock to leave, then a finish pass was applied. I forgot to do the finish pass on one of the parts (and had removed it from the machine) but as the dimension was not critical I left it there. One the second part I made sure it did the rough and the finish.
When I measured them both the first part was spot on the dimension (but should have been 0.1 mm larger) and the second part was 0.1 mm underside (but should have been OK). As they were simple cover plate it did not matter but I wondered what had happened.
I've cut lots of part out and not seen that before. Unfortunately my machine is now a pile of bits so I won't find out any time soon. Just thought I'd post my experiences . . . .
Hi routercnc,
Thanks for shareing. Las night I did some experiments. But now I realize I only did contour cuts. I tried all compensation options but results where the same. Inside dimension spot on and outside 0,3mm extra.
Does Mach3 DRO shows coordinates with compensation or without? Since DRO (both work and machine) was spot on but cut was 0,3mm off... Right now I already doubt my axis calibration. Workshop is a bit colder but i dont think shrinking is to blame here.
If you are using an adaptive toolpath in fusion, are you also following it with a 2d contour ? In fusion when you use 2d adaptive, there will be an option for "radial/axial stock to leave" You can of course set these to 0, but I always follow adaptive toolpaths with a 2d cleanup contour.
Just a thought !
Yes. In 2d adaptive I leave 0,5mm stock. Which I clean up with 2d contour as a finishing pass.Quote:
Originally Posted by needleworks
I think its normal - I do exactly the same with all my cutters, the 2mm SF carbides i use only give true results if cam'med for 1.96mm so a test is needed if dims are critical.Quote:
Originally Posted by Neale
Sorry if someone else mention it and I didn't notice.
Have you try with another program? Design the same object and cut?
Managed to sort out my problem. Almost too embarrassed to tell but.. as it turns out it is not wise to calibrate machine axis with a dial indicator. At least a machine that has not super accurate ballscrews. Also the testpart that showed inside measures correctly and outside wrong, was coincidence I believe. I re-calibrated my axis with a 300mm caliper. Problem solved.
So anyway here is a little cherry on my strawberry cake:
Material: S335
8000rpm, 350mm/min, DOC 3mm, 2-flute TiAIN 6mm endmill.
Some chatter but nothing major... Shitty workholding. Nice blue chips where flying :)
Attachment 23858
Attachment 23859
Attachment 23860