DIY CNC Design. criticism wanted.

[joel0407]

Good to see you've made a start.

Oh yeh. I've been started for a little while now. All the gantries are ready to go on. I'm pretty sure I'll be finished the physical build by tomorrow afternoon then it's on to the electronics. When I say physical build finished, I mean the pre-build. I have hand cut all my gantry parts and I'm not that precise with wood stuff. It's all bolts up but not to the tolerances I'm happy with. So it'll be built from MDF then I'll use the machine to cut the actual parts from Marine ply. There maybe some changes from how I build it from the MDF as well. I have already made some small changes.

Happy Days

[ger21]

Holding the motor in my hand got pretty intense but I managed up to get it to stall at 400mm/second with 1500mm/second acceleration. My 16mm ball screws wont handle anything close to that so it's good to know that neither my motors, drivers or computer wont be the limiting factors in my gantry speed.

More than likely, the actual speeds you get when the machine is together will be much different than when holding a motor in your hand. You may find that the slightest misalignment of components can cause stalling at much lower speeds than you expect.

[joel0407]

More than likely, the actual speeds you get when the machine is together will be much different than when holding a motor in your hand. You may find that the slightest misalignment of components can cause stalling at much lower speeds than you expect.

Thanks mate. Like I said it's massive over kill. It'd have to be stalling pretty low for it to be a problem for me. It was quite violent when changing direction with acceleration at 1500mm/second. If I rapid at 100mm/second that's 20 RPS or 1200 RPM which I think I'll be stretching to 1000mm 1605 screws for whip. Even with acceleration at 100mm/second will have it at full pace in 1 second and then it will have travelled near 100mm. That's only 10 seconds for a run from one end to the other. For a home hobby machine I think that's pretty quick.

Happy Days

[ger21]

You'll quickly find that you wished it were faster.... We all do.

[A_Camera]

Thanks mate. Like I said it's massive over kill. It'd have to be stalling pretty low for it to be a problem for me. It was quite violent when changing direction with acceleration at 1500mm/second. If I rapid at 100mm/second that's 20 RPS or 1200 RPM which I think I'll be stretching to 1000mm 1605 screws for whip. Even with acceleration at 100mm/second will have it at full pace in 1 second and then it will have travelled near 100mm. That's only 10 seconds for a run from one end to the other. For a home hobby machine I think that's pretty quick.

Happy Days

Are you sure about those figures? I mean, what sort of super stepper are you using which is capable of 1500mm/s² acceleration? That is HUGE... Maybe you have overlooked something, because even half that is VERY fast. My machine has 700mm/s² acceleration and even if it can do a little more that that as well, it is very far from your values. Admittedly, I never checked my values without any load because that is pretty unimportant, what's important is to be able to keep up the speed and acceleration on all axes under load without losing a single step and to be able to that over and over again with any number of direction changes. With that acceleration you really need to consider rigidity and how to prevent the machine from shaking like crazy, especially when you are making many short moves and direction changes.

On the other hand, 100mm/s2 is a pretty low acceleration. You will fall asleep...

Regarding the speed, my speed is 9000mm/min which is 1800rpm for my 1605 so you should also be able to get that, or at least near that, since the speed I have is not the absolute top speed but the one I can use and find it comfortable on my small machine.

[joel0407]

Are you sure about those figures? I mean, what sort of super stepper are you using which is capable of 1500mm/s² acceleration? That is HUGE... Maybe you have overlooked something, because even half that is VERY fast. My machine has 700mm/s² acceleration and even if it can do a little more that that as well, it is very far from your values. Admittedly, I never checked my values without any load because that is pretty unimportant, what's important is to be able to keep up the speed and acceleration on all axes under load without losing a single step and to be able to that over and over again with any number of direction changes. With that acceleration you really need to consider rigidity and how to prevent the machine from shaking like crazy, especially when you are making many short moves and direction changes.

On the other hand, 100mm/s2 is a pretty low acceleration. You will fall asleep...

Regarding the speed, my speed is 9000mm/min which is 1800rpm for my 1605 so you should also be able to get that, or at least near that, since the speed I have is not the absolute top speed but the one I can use and find it comfortable on my small machine.

Hey mate, With further testing, I've found the acceleration and speed drop off with micro steps. I won't really know how fine I'll want the steps and what speed and all until it's fully up and running. At the moment though, I am not at all worried it will exceed my expectations.

Good to hear you are getting that speed with the 1605 screws. What length are you running that RPM? My longest are 1000mm.

Happy Days

[NIC 77]

Are you sure about those figures? I mean, what sort of super stepper are you using which is capable of 1500mm/s² acceleration? That is HUGE... Maybe you have overlooked something, because even half that is VERY fast. My machine has 700mm/s² acceleration and even if it can do a little more that that as well, it is very far from your values.

1500mm/s^2 = 1.5m/s^2
1.5/9.81 = 0.15 G

Joel, I haven't seen drawings for your design modifications? Or are you proceeding with your initial design? I think machine flex might be your limiting factor in determining acceleration. I don't think you'll be running this at 0.15G, but many people are happy with less, and many people also select parameters far less than what their machine is capable of to get within their comfort zone.

As far as 0.15G being huge in general, not specifically related to Joel, here is some info I have found on the zone:

Jim Dawson is running at around 0.12G on his big moving table CNC using a single Nema 34 1280 ozin at 400 IPM cutting speed with a table of approximately 700 lbs.

Ger21 mentioned his goal is 0.5G for his build. Is that right Ger? Of course he's using 400W servos.

The last build that 1Jumper10 did for a friend, he mentioned it was running happily on 0.13G (50 in/sec^2) but that he detuned it to 30 In/sec^2 (0.077 G) to where it's less anxious to learn on.

linux_fan said "Acceleration is 1500mm/s*s Can go almost as high as you want with those mitsu servo motors, but it hurts even watching gantry at some high acc rates" in his post about his awesome machine here:

http://www.cnczone.com/forums/uncate...aluminium.html

This is linux_fan's video, so this is what 1500 mm/s^2 looks like

[NIC 77]

I can only insert one video per post, so here is another one for comparison.

This DIY CNC that I found a video of is running at 0.5G according to the creator's comments.


Also, I believe that some of the really expensive brand name high speed modern VMCs are capable of around a G. Correct me if I'm wrong.

[A_Camera]

Hey mate, With further testing, I've found the acceleration and speed drop off with micro steps. I won't really know how fine I'll want the steps and what speed and all until it's fully up and running. At the moment though, I am not at all worried it will exceed my expectations.

I think you will want to run at 8x or 10x micro stepping. That should give you smooth running and not losing too much torque. I don't think you would be happy without microstepping.

Good to hear you are getting that speed with the 1605 screws. What length are you running that RPM? My longest are 1000mm.

Happy Days

My screws are 60, 50 and 30 cm long. They are actually oversized because I planned for a larger machine but then I realized that there is no point, I won't need larger one, but by then I had the screws between China and Sweden already... so I just built the machine with too long screws.

[A_Camera]

1500mm/s^2 = 1.5m/s^2
1.5/9.81 = 0.15 G

Joel, I haven't seen drawings for your design modifications? Or are you proceeding with your initial design? I think machine flex might be your limiting factor in determining acceleration. I don't think you'll be running this at 0.15G, but many people are happy with less, and many people also select parameters far less than what their machine is capable of to get within their comfort zone.

As far as 0.15G being huge in general, not specifically related to Joel, here is some info I have found on the zone:

Jim Dawson is running at around 0.12G on his big moving table CNC using a single Nema 34 1280 ozin at 400 IPM cutting speed with a table of approximately 700 lbs.

Ger21 mentioned his goal is 0.5G for his build. Is that right Ger? Of course he's using 400W servos.

The last build that 1Jumper10 did for a friend, he mentioned it was running happily on 0.13G (50 in/sec^2) but that he detuned it to 30 In/sec^2 (0.077 G) to where it's less anxious to learn on.

linux_fan said "Acceleration is 1500mm/s*s Can go almost as high as you want with those mitsu servo motors, but it hurts even watching gantry at some high acc rates" in his post about his awesome machine here:

http://www.cnczone.com/forums/uncate...aluminium.html

This is linux_fan's video, so this is what 1500 mm/s^2 looks like

I never said it is not possible to get that high, or even higher acceleration, but those machines are not driving the screws directly, are more rigid (or lighter) or have some sort of gearing/belt involved. 1500mm/s² is indeed a lot on a small machine with direct driven screws using steppers. Yes, machine flex is a serious issue at high accelerations, especially with many short and fast direction changes. I had to change my original machine when I got the speed and acceleration up because of flexing caused by the rapid direction changes.

[TumbiUmbiTek]

1500mm/s^2 = 1.5m/s^2
1.5/9.81 = 0.15 G

Joel, I haven't seen drawings for your design modifications? Or are you proceeding with your initial design? I think machine flex might be your limiting factor in determining acceleration. I don't think you'll be running this at 0.15G, but many people are happy with less, and many people also select parameters far less than what their machine is capable of to get within their comfort zone.

As far as 0.15G being huge in general, not specifically related to Joel, here is some info I have found on the zone:

Jim Dawson is running at around 0.12G on his big moving table CNC using a single Nema 34 1280 ozin at 400 IPM cutting speed with a table of approximately 700 lbs.

Ger21 mentioned his goal is 0.5G for his build. Is that right Ger? Of course he's using 400W servos.

The last build that 1Jumper10 did for a friend, he mentioned it was running happily on 0.13G (50 in/sec^2) but that he detuned it to 30 In/sec^2 (0.077 G) to where it's less anxious to learn on.

linux_fan said "Acceleration is 1500mm/s*s Can go almost as high as you want with those mitsu servo motors, but it hurts even watching gantry at some high acc rates" in his post about his awesome machine here:

http://www.cnczone.com/forums/uncate...aluminium.html

This is linux_fan's video, so this is what 1500 mm/s^2 looks like

What size is that stepper on z axis? I was thinking it was a second spindle there for a second..


Sent from my LG-H870DS using Tapatalk

[NIC 77]

What size is that stepper on z axis? I was thinking it was a second spindle there for a second..


Sent from my LG-H870DS using Tapatalk

You'll have to ask linux_fan in his thread. It's not my machine.

But I believe he is using servos, and what you're looking at is a second spindle driven by a large servo motor. IDK, you'll have to check out his thread. The link is posted.

[joel0407]

I think you will want to run at 8x or 10x micro stepping. That should give you smooth running and not losing too much torque. I don't think you would be happy without microstepping.



My screws are 60, 50 and 30 cm long. They are actually oversized because I planned for a larger machine but then I realized that there is no point, I won't need larger one, but by then I had the screws between China and Sweden already... so I just built the machine with too long screws.

Yep. Already realize they are pretty rough feeling without a bit of micro stepping.

My longest screw is 1m so I wont reach the RPM you do without whipping.

Happy Days

[joel0407]

I know you guys have said that my motors wont do in the machine what they do on the bench but just out of interest. I got the cable to connect my Leadshine AM882 drivers to my computer today, to tune them with ProTune. Previously they would stall at 400mm/second. After tuning they will now run on the bench at 400mm/second and 1500mm/second acceleration without stalling.

[Biggs427]

That's fast. What's the microstep settings you're using? What's the step/ mm?

I'm doing 16m / min rapid with 10mm pitch with AM882 at 1/4 microstep and 40 volt PS. It stall at 16.5m/sec. I backed off my rapid speed of 25% to 12m/min to have some headroom.

When all is on your machine, find your max feed then back off a certain amount. Don't try to squeeze every tiny oz of speed, you can manage to find. I've learned the hard way that there's a big difference between testing and real life.

[zeeflyboy]

IMO what acceleration/speeds you can reach on the bench with no loads is not massively relevant, and you also have no way of knowing if your motors are actually doing as commanded - all you know is that they are not stalling.

Push speeds/acceleration too high and you will risk losing steps which will cause you some massive headaches. Other than returning to zero and checking the position there is no way to know that a normal stepper has missed steps or not, sometimes it will be obvious when your work gets completely messed up, other times it may be less blatant.

Also, I'm sure it has been mentioned but consider critical speed of the ball screws. It varies depending on length/diameter and bearing arrangement but there are calculators out there to give you a clue. With an optimal bearing set up and medium length screws you're probably looking in the range of 2500 maximum to be safe. With e.g. 5mm pitch that gives a max of 12,500mm/min or roughly half the 400mm/s you mention there.... just another factor to consider.

[joel0407]

That's fast. What's the microstep settings you're using? What's the step/ mm?

I'm doing 16m / min rapid with 10mm pitch with AM882 at 1/4 microstep and 40 volt PS. It stall at 16.5m/sec. I backed off my rapid speed of 25% to 12m/min to have some headroom.

When all is on your machine, find your max feed then back off a certain amount. Don't try to squeeze every tiny oz of speed, you can manage to find. I've learned the hard way that there's a big difference between testing and real life.

16m / min? 16m is 1600mm so that's 1600mm / min. 1 min is 60 second so that's 96000mm / Second. <<<<<< Yes I'm and idiot. (added comment later)

Yeh. I'll be backed off heaps. As I've said before, I'm only aiming for 100mm/second. Surely I wont have too much trouble reaching 1/4 of what they do on the bench.

I'll be interested to try it and see haw far it'll go though.

Happy Days

[joel0407]

My main point was, I was impressed before and I don't think I'll have any problems reaching my target speed. Tuning the AM882 drivers has raise the speed further.

Happy Days

[zeeflyboy]

16m / min? 16m is 1600mm so that's 1600mm / min. 1 min is 60 second so that's 96000mm / Second.

Yeh. I'll be backed off heaps. As I've said before, I'm only aiming for 100mm/second. Surely I wont have too much trouble reaching 1/4 of what they do on the bench.

I'll be interested to try it and see haw far it'll go though.

Happy Days

interesting maths!

16m/min is 16000mm/min which is 266mm/s

100mm/s is definitely reliably achievable on good quality 5mm pitch screws with good steppers and drivers and controller. I have actually set 6000mm/min as my machine's max speed and it's relatively comfortable there.

[joel0407]

interesting maths!

16m/min is 16000mm/min which is 266mm/s

100mm/s is definitely reliably achievable on good quality 5mm pitch screws with good steppers and drivers and controller. I have actually set 6000mm/min as my machine's max speed and it's relatively comfortable there.

Yep. Sorry mate. Terrible maths.