CNC - Temperature Distortion

For discussions related to designing and making ski/snowboard-building equipment, such as presses, core profilers, edge benders, etc.

Moderators: Head Monkey, kelvin, bigKam, skidesmond, chrismp

Richuk
Posts: 1146
Joined: Sun Jan 31, 2010 9:53 am
Location: The Duchy of Grand Fenwick

Post by Richuk »

Richie - this is what I think I know based on what I have read in the past

Ballsrcews - the main practical issue is whipping, The main technical issues relate to those outlined in the decisions made when designing the Z-axis: http://www.cncroutersource.com/leadscrew.html . Resonance, with the additional issue of gravity to consider, leads to the main issue being whipping.

Wear does happen, but I've not noticed many reflect on excessive damage from wood or metal chips. If so, the issue is poor chip control, as ballscrew nuts have been designed with this issue in mind. Chip control -vacuum and or high pressure air (wood and or metal) and flipping the table vertical. Mainly contolled by good placement of the ballscrew from what I have read. Have you considered designing a vertical table? A great space saver and the design of the table appears to be a little simpler.

Missed steps, I thought with a closed loop system the motors know where they are? If you have gone for high-end motors and drivers, the question is what will they reasonably cut. The table design follows those parameters? You might be under selling the choices you have already made if you limit your design to just cutting wood.

With steppers, the approach appears to be x-axis driven by two ballscrews. I assuming this improves the available cutting force, ensuring that the available force can be applied across the table. Decreasing wear within whatever carriages are being used and decreases the potential for twist or racking are other factors, particularly if there are issues with the stiffness of the design or accuracy of the build. There is a tendency to drive stepper motors indirectly, via a pulley and belt. This reduces the transmission of motor resonance, (I think is reduces backlash, when cutting at speed) and allows for gearing. I don't think the issues of resonance applies to servo's - I'am sure Islandrider would have a view.

Racking caused by missed steps is an issue. Missed steps - I'm not sure of all the causes: exceeding design parameters of the electronics, electrical noise, power issues, holding torque issues, poor workmanship on the build and other avoidable problems. Essentially they appear to be caused by a mis-match of design parameters and or poor practical application of the design. There are stepper motor drivers available that provide for stall alert. They monitor back EMF (from memory).

Hope some of this helps man, its based on what I have read, so needs double checking )
User avatar
richie
Posts: 184
Joined: Wed Jun 23, 2010 7:25 pm
Location: New Zealand
Contact:

Post by richie »

Richuk wrote:Richie - this is what I think I know based on what I have read in the past

Ballsrcews - the main practical issue is whipping, The main technical issues relate to those outlined in the decisions made when designing the Z-axis: http://www.cncroutersource.com/leadscrew.html . Resonance, with the additional issue of gravity to consider, leads to the main issue being whipping.

Wear does happen, but I've not noticed many reflect on excessive damage from wood or metal chips. If so, the issue is poor chip control, as ballscrew nuts have been designed with this issue in mind. Chip control -vacuum and or high pressure air (wood and or metal) and flipping the table vertical. Mainly contolled by good placement of the ballscrew from what I have read. Have you considered designing a vertical table? A great space saver and the design of the table appears to be a little simpler.

Missed steps, I thought with a closed loop system the motors know where they are? If you have gone for high-end motors and drivers, the question is what will they reasonably cut. The table design follows those parameters? You might be under selling the choices you have already made if you limit your design to just cutting wood.

With steppers, the approach appears to be x-axis driven by two ballscrews. I assuming this improves the available cutting force, ensuring that the available force can be applied across the table. Decreasing wear within whatever carriages are being used and decreases the potential for twist or racking are other factors, particularly if there are issues with the stiffness of the design or accuracy of the build. There is a tendency to drive stepper motors indirectly, via a pulley and belt. This reduces the transmission of motor resonance, (I think is reduces backlash, when cutting at speed) and allows for gearing. I don't think the issues of resonance applies to servo's - I'am sure Islandrider would have a view.

Racking caused by missed steps is an issue. Missed steps - I'm not sure of all the causes: exceeding design parameters of the electronics, electrical noise, power issues, holding torque issues, poor workmanship on the build and other avoidable problems. Essentially they appear to be caused by a mis-match of design parameters and or poor practical application of the design. There are stepper motor drivers available that provide for stall alert. They monitor back EMF (from memory).

Hope some of this helps man, its based on what I have read, so needs double checking )
Yes there are smarter closed loop stepper systems from Leadshine and others and these should alleviate any concerns really for those wanting a real value for money stepper system I recon they would be awesome.

I ended up with AC servo drive system by chance for than by design, originally I was looking at the closed loop Leadshine steppers but then I came accross a beautifully built second hand dismantled large DIY router and purchased the Granite VSD-XE servo drives and the large AC servo motors and the screws and liner rails/carriages. These had already been coupled up nicely to 25x25 pitch ball screws (1:1 ratio) and were driving a powerful router that was cutting all kinds of materials well, mainly hardwoods and plastics and some ali plate.

So I am expecting my machine will do some alloy cutting work (ski cassettes maybe) at a stretch although as stated its primary purpose is wood/plastic. With relatively low rotational speeds these 1.8m long ball screws should not whip I hope, and although my servo motors spin impressively at say 3000RPM on the bench in real life with these ball screws even on rapids I will not be spinning them at more than 500RPM for a start which gives me a rapid of just over 0.2m/sec which for my purposes is fast enough and provided the screws behave I'll up my speeds with experience. Of course higher accelerations and higher target rapid speeds will be governed by the weight I am throwing around and how much torque I can get out of my servo motors and servo drive current and how well I can tune the servo drives to position and settle quickly with well damped low overshoot and low following error.... lots of parameters and I am doing this by feel to be honest but reasonably confident I have the right gear , I hope!!!!!

My main axis will be built oversquare - a bit like a high reving engine! which means the bearing carriages will be spaced along the axis longer than the width between the 2 rails on the axis, and this should help a lot to stop any racking with a very stiffly built gantry sitting on top of those carriages too. And I will not be trimming any excess weight off so I have a solid damped system to absorb vibrations, even if it is at the expense at the top end of acceleration I don't expect that to be a factor for me. The oversquare is available to me due to the fact I want to build a machine narrow for my workshop but I have conveniently long 2.6metre main axis THK SSR30 linear bearings/rails so my gantry bearing spacing on x rails will be 800mm length x 600mm width give or take a bit.

As you can see I am using a lot of gut feel and am pinning my hopes of a great machine on a high powered servo drive setup and quality linear rails/screws and a solid bench/gantry. To be honest I am sweating more over the trueness of my table and leveling it for my main x axis rails and ballscrew mounts and at this stage rather than expensive heavy machining I am seriously looking at epoxy leveled trenches to mount the rails on which I should be able to get down to low levels of error if I do it right with the right epoxy etc..... more fun with epoxy !!!

Anyway this is good discussion Rich and its all helpful and interesting lets keep it going!

cheers
Rich
MonkeyCAM and SnoCAD - https://github.com/mikemag
Ski binding mounting https://github.com/splitn2/DrillSki

Richard Harcourt | www.splitn2.com | Christchurch New Zealand
rich@splitn2.com | www.facebook.com/splitn2
Richuk
Posts: 1146
Joined: Sun Jan 31, 2010 9:53 am
Location: The Duchy of Grand Fenwick

Post by Richuk »

If the motors and drivers were cutting aluminium on another machine, I don't see why you should think they won't on you design. The rails you have are made of solid stuff, this is above the minimum I have seen used on large format machines cutting deeply into 2 inch aluminium plate.

Take a look at Jonathan's second build - a sufficiently strong machine.

Do you know what cutting force is required for the materials you will be cutting and how that relates to torque? This calculator helped me put things in perspective.

http://www.orientalmotor.com/support/motor-sizing.html

I understand what you mean by oversquare. What is the width of the gantry? As it is important to consider the effect on the connection between the carriage plate and the ends of the gantry. The greater the disparity between the length of the carriage plate and the width of the gantry, the great the strain on the gantry end plate fixings. It becomes a lever - I think? Twizz? What kind of gantry shape are you considering? An L-shape gantry would resist racking well. How thick will your end plates be? I'm assuming Aluminium and Aluminium extrusion?

CNC is about smooth, accurate cutting, not top end speed. Torque, servo's have it on torque right? Toque and speed )

The trueness of your table, really isn't that important. You just need the x-axis to be parallel and level. Again, Jonathan offer some good analysis on the epoxy method.

Weight is certainly your friend. The recommended approach to dealing with any residual vibration is to fill the box section with sand. I looked at this an thought it might be possible to use urethane or RTV silicone instead - applied to just one internal face.
User avatar
richie
Posts: 184
Joined: Wed Jun 23, 2010 7:25 pm
Location: New Zealand
Contact:

Post by richie »

Richuk wrote:If the motors and drivers were cutting aluminium on another machine, I don't see why you should think they won't on you design. The rails you have are made of solid stuff, this is above the minimum I have seen used on large format machines cutting deeply into 2 inch aluminium plate.

Take a look at Jonathan's second build - a sufficiently strong machine.

Do you know what cutting force is required for the materials you will be cutting and how that relates to torque? This calculator helped me put things in perspective.

http://www.orientalmotor.com/support/motor-sizing.html

I understand what you mean by oversquare. What is the width of the gantry? As it is important to consider the effect on the connection between the carriage plate and the ends of the gantry. The greater the disparity between the length of the carriage plate and the width of the gantry, the great the strain on the gantry end plate fixings. It becomes a lever - I think? Twizz? What kind of gantry shape are you considering? An L-shape gantry would resist racking well. How thick will your end plates be? I'm assuming Aluminium and Aluminium extrusion?

CNC is about smooth, accurate cutting, not top end speed. Torque, servo's have it on torque right? Toque and speed )

The trueness of your table, really isn't that important. You just need the x-axis to be parallel and level. Again, Jonathan offer some good analysis on the epoxy method.

Weight is certainly your friend. The recommended approach to dealing with any residual vibration is to fill the box section with sand. I looked at this an thought it might be possible to use urethane or RTV silicone instead - applied to just one internal face.
Yeah my build will be steel for as much as possible of the big stuff and perhaps some ali tool plate in the gantry , it will be as stiff as economically poss. I'm aiming for a design with no "side plates" its hard to explain but when you see it you'll get it. I recon it will do ali nicely once I get the right , feed speeds and spindle tool rpms and right tools and tool paths...... I only want "real speed" for rapids not for cutting, just rapids are ideal to squash up into small amount of time possible and then I will do the cutting at the speeds that work.

So many resources to read, luckily so much of the hard work has been done by others so yeah its good to get all the input I can eh!
MonkeyCAM and SnoCAD - https://github.com/mikemag
Ski binding mounting https://github.com/splitn2/DrillSki

Richard Harcourt | www.splitn2.com | Christchurch New Zealand
rich@splitn2.com | www.facebook.com/splitn2
User avatar
chrismp
Posts: 1443
Joined: Fri Feb 13, 2009 9:00 am
Location: Vienna, Austria

Post by chrismp »

How will you drive your long axis? You say you want 2.6m of travel, but only mentioned 1.8m ballscrews...that's why I'm asking ;)
knightsofnii
Posts: 1148
Joined: Tue Jan 08, 2008 6:02 am
Location: NJ USA
Contact:

Post by knightsofnii »

I've noticed many of the precision flaws of my machine were not the machine itself, some examples:

1) my profiling toolpath kept starting and ending at different points along the x axis. I watched it go for what seemed like a few hours, noticed it was stalling in spots during travel along x. Thought it might be a programmatic error, or a problem with motors. Discussed with cnc supplier. I took the drives apart and discovered the tiny setscrews holding the pinion gear to the motor had backed out. Some blue loctite, viola. So I wouldn't say this is exactly a machine issue, parts do break, more of a plan/build issue I guess? Glad I figured it out before catastrophe.

2) my profile thicknesses are not exact. I've discovered in my G code that as it travels over spots that are supposed to be flat, it's actually varying in height slightly in the code itself. I believe it may be something to do with the way i import the 3d profiled core as a .stl into my cam software. Still investigating.

3) wavy spots when cutting core shape and base shape. I discovered the lines/curves in my cad files are not exactly tangent, and line thicknesses are different all over the place, and can impact where the tool cuts, especially if you're cutting inside or outside the line.

many others but i'll stop there.
Doug
User avatar
richie
Posts: 184
Joined: Wed Jun 23, 2010 7:25 pm
Location: New Zealand
Contact:

Post by richie »

chrismp wrote:How will you drive your long axis? You say you want 2.6m of travel, but only mentioned 1.8m ballscrews...that's why I'm asking ;)
Correct I have about 1.75m of usable driving length when I allow for a bit of unused at each end, just enough to sneak in a snowboard core nicely.

The rails were 2.6m long when I purchased them and I cannot bring myself to cut them down shorter. At this long length it gives me the luxury of making a long but narrow gantry, over square I'll call it.

So say my gantry bearing carriages are spread 900mm apart along the rails length and the gantry is 600mm wide, its way way less likely to have any measurably concerning racking issues.

In the long term if this MK1 machine performs well I may reuse these rails in a different design machine if I get around to it, this machine is being built with what I have so has a few ideas I want to put to the test. I will be running only one screw on that long axis hence the design to control racking torques generated by work on the extreme limits of the gantry.

Hope that makes sense, it will be interesting to see if it works as well as I hope!!!!
MonkeyCAM and SnoCAD - https://github.com/mikemag
Ski binding mounting https://github.com/splitn2/DrillSki

Richard Harcourt | www.splitn2.com | Christchurch New Zealand
rich@splitn2.com | www.facebook.com/splitn2
Richuk
Posts: 1146
Joined: Sun Jan 31, 2010 9:53 am
Location: The Duchy of Grand Fenwick

Post by Richuk »

Have you sketched out the design? It would be great to see.

I wondering why you rejected two balls screws one motor? I would have though this was a great solution if racking is a worry?

Nii, wow. Guru!!!
User avatar
richie
Posts: 184
Joined: Wed Jun 23, 2010 7:25 pm
Location: New Zealand
Contact:

Post by richie »

Richuk wrote:Have you sketched out the design? It would be great to see.

I wondering why you rejected two balls screws one motor? I would have though this was a great solution if racking is a worry?

Nii, wow. Guru!!!
I did consider 2 screws one motor and that can be tried if one screw alone doesn't prove adequate. However I am designing my machine specifically to counteract racking by making it oversquare and so this makes a second screw redundant. If I do run 2 screws off one motor , say using a 25mm T10 timing belt and pulleys to each screw I still have to allow for small amounts of belt error anyway and the extra resonance in the system this will add, so I am preferring for one screw direct drive system and keep it simple.

Yesterday I visited a guy in my home town here in NZ and he has a nice Mach3 CNC mill he has purchased second hand and retrofitted new servos and Granite Devices drives to it. Very clever guy making his own high performance car parts so different project goals but same electronics and software as I'm using so it was a really enlightening discussion.

Yeah I'll sketch up my design and post it when I get time, mostly in the garden this weekend!

cheers
Rich
MonkeyCAM and SnoCAD - https://github.com/mikemag
Ski binding mounting https://github.com/splitn2/DrillSki

Richard Harcourt | www.splitn2.com | Christchurch New Zealand
rich@splitn2.com | www.facebook.com/splitn2
Richuk
Posts: 1146
Joined: Sun Jan 31, 2010 9:53 am
Location: The Duchy of Grand Fenwick

Post by Richuk »

It sounds like you have a really interesting plan in mind and I look forward to seeing it sketched or stitched together. Great to hear you've found someone who can help with the tuning the motors!
Post Reply