Official Luthiers Forum!

Having a bad habit of not leaving well enough alone, I decided to change things out on my 24x48 XZero Raptor a few months ago. I had been running it using Mach 3, G540, 380 oz Kelling steppers on all three axes. I was able to run without losing steps up to almost 400 ipm/50 ipsps, but the long axis (Y in my case) was the limiting factor with the big 25mm ballscrew. Sometimes it would stall when it was cold, and in general it seemed to be wanting just a little more power.

So, I put Kelling 570 oz/5A steppers on all three axes (2.5mH per phase) using Gecko G213V drivers (got a good deal on them) and a 48 V linear power supply. The results were underwhelming. I could easily double and then some my rapid speeds, but even with the low inductance I had acceleration trouble on the long axis.

I upgraded the long axis to a NEMA 34 906 oz 6.1 A (3.3mH) motor, and things weren't really any better. It would run very fast, but lost steps here and there.

Then I got to reading up on the Gecko drive I was using and found that people who were running 3D tool paths consistently had problems with the G901X step multiplier that is part of the G213V drivers I was using losing steps. I removed the step multipliers and replaced them with opto-isolators and then things changed quickly. After the "downgrading" of the Gecko drives, I was then able to run much faster without losing steps and life was getting pretty good.

Finally, after watching Andy's YouTube video comparison of Mach 3 and UCCNC, I decided to take the plunge and replace Mach3 with a UC100 and UCCNC. It took a little tweaking, but the machine doesn't even sound the same using the new software. It's faster, smoother, even quieter (when cutting air where I can hear the motors) and the parts come out better. Huge improvement, and the machine really ran very nicely before I started the whole process. I don't get any stalling or lost steps now, programs run faster, results are better, and even set extremely conservatively I'm running now 600 ipm rapids with 50 ipsps acceleration. It will run much faster if I want, but there's really no need. And with the stronger steppers, I can cut much more aggressively without problems than before if I choose to.

The latest version of the UCCNC software is a definite improvement over the previous version, but still lacks some of the maturity of Mach. There are some missing G codes, the motor tuning routines aren't as robust--fairly insignificant stuff. But its trajectory planning is much better, and everything runs much smoother for me than it did in Mach. Of course, YMMV.

Anyway, just thought I'd check in and throw some stuff out on the table.

Dave

Hey Dave,

That's great to hear. What PSU are you using now at how many volts?

I'm planning an upgrade to my large machine (30x48) and have purchased some step servos from applied motion but I haven't plopped those in yet. According to the numbers on the spec sheet, I should be able to get to the same numbers that you're at as well. I was originally planning on going Kflop for this new machine and eventually switching the old one over to Kflop but with the latest release of UCCNC, I'm thinking I'm changing my plan to stick with it.

I get that Kflop is probably better but the thought of sifting through all that code to just get it running is causing me to continue to procrastinate so...I'm going UCCNC/UC100 but I'll hang on to the Kflop hardware for a bit in case I get inspired.

I was having a convo with Ger from CNCzone and we were wondering if simply putting a bigger PSU/Drivers would speed things up even with the 380s. I got friend and family price on the step/servos but list for them is like $1500 so it's good to know that there's a much cheaper option out there. Nevertheless, with the volume I'm doing I think I'm going to enjoy the piece of mind that the closed loop will give me.

Best,

_________________
http://www.birkonium.com CNC Products for Luthiers
http://banduramaker.blogspot.com

I'm running 48 volts--I think 51 was what I had calculated as max for the motors I started with--I could go higher now. I think the 380s would probably have run faster at 48 than the 42 I was running with the G540, but probably not all that much. The biggest difference I'm seeing is that the "stalls when cold" issue is gone now. I think the NEMA 34 with its extra heft does a better job with that initial inertia requirement the big screw has. Closed loop would be great, but I'm still too cheap....